[Home] [Databases] [World Law] [Multidatabase Search] [Help] [Feedback]
	England and Wales High Court (Patents Court) Decisions
You are here: BAILII >> Databases >> England and Wales High Court (Patents Court) Decisions >> Chiron Corporation v Evans Medical Ltd & Ors [1997] EWHC 359 (Patent) (03 November 1997) URL: http://www.bailii.org/ew/cases/EWHC/Patents/1997/359.html Cite as: [1997] EWHC 359 (Patent)

[New search] [Printable RTF version] [Help]

IN THE HIGH COURT OF JUSTICE
CHANCERY DIVISION
PATENTS COURT

		Royal Courts of Justice Strand, London, WC2A 2LL
		3 November 1997

Dated: 16/1/98

Mr. Justice Laddie:

A. General introduction

These conjoined actions are to determine the validity and infringement of European Patent 0 162 639 which is concerned with antigenic preparations for use in acellular vaccines against the serious and debilitating disease known as whooping cough. The patent was applied for by the Wellcome Foundation Limited but was assigned to Evans Medical Limited (Evans) which has granted an exclusive license under it to Smithkline Beecham Biologicals SA. The patent is alleged to have been infringed by Chiron Corporation (Chiron). Chiron has both petitioned for revocation and commenced proceedings for a declaration of non-infringement in relation to two vaccines made by it. Evans has counterclaimed for relief for infringement. Although both validity and infringement are in issue, most of the argument before me concerned the former. Before turning to the patent in suit, it will be helpful to set out some of the relevant background.

Whooping cough and the infective agent which causes it

Whooping cough is caused by an infective bacterium called Bordetella pertussis (or B. pertussis). Until the development of a pertussis vaccine in the late 1940's, whooping cough was one of the most frequent and severe diseases of infants. It continues to be so in populations where children are not immunized. When news spread of alleged serious side effects with the original pertussis vaccine, rates of immunization fell and the incidences of whooping cough rose as a consequence. It is estimated that annually 51,000,000 cases and 600,000 deaths still occur worldwide in unvaccinated infants.

Bacteria are very small life forms. Their structure consists of an outer membrane surrounding a gel-like interior. The gel-like substance that forms the bulk of cellular material is called the cytoplasm. It is enclosed by the cytoplasmic membrane. The cytoplasm contains the cellular components which the cell requires to carry out its everyday activities including growth, metabolism, and replication. The cytoplasmic membrane is the selective barrier which separates the cytoplasm from the external environment. It is composed of a layer in which are embedded various proteins. Those proteins control what molecules are able to pass into and out of the cell across the cytoplasmic membrane. Bacteria can be classified into two groups on the basis of the structure of their cell membrane. These structural differences cause the bacteria to stain differently in a procedure called the Gram stain. B. pertussis is a Gram-negative bacterium which means that the outer layer of its outer membrane is composed of a molecule called LPS (lipopolysaccharide - often called "endotoxin"). This substance is important in pathogenic (i.e. disease causing) bacteria because it is highly toxic to humans and deaths from infections of Gram-negative bacteria often are the result of severe reactions to LPS. B. pertussis is also covered by several filamentous protein rod or fibre like structures attached to the cell surface called fimbriae. Fimbriae provide bacterial pathogens with the ability to adhere to the surface of host cells and this is the function they appear to perform in B. pertussis.

Proteins

In the preceding paragraph mention was made of proteins. These are complicated chemicals made by joining together in a string chemical building blocks called amino acids. There are twenty different amino acids. Two of the amino acids which are of relevance to this action are proline (abbreviated to "Pro") and glutamic acid ("Glu"). It is the precise sequence of amino acids which determines the shape and chemical activity of the protein. The sequence of each protein is programmed in the genetic material of the bacterium. There are a very large number of proteins in bacteria. They are found in the cytoplasm and, as mentioned above, embedded in the outer membrane. There are also some proteins which are secreted, i.e. ejected from the bacterium into the surrounding medium. Proteins vary not only in their precise sequence of amino acids but also in their overall size. Some are very large indeed, being made up of thousands of amino acids. Others are much smaller. Different proteins carry out different functions. Some, for example, mediate or catalyse other chemical reactions. These are called enzymes. Others mediate the transport and storage of small molecules and ions, or are involved in the immune system or control the growth and differentiation of cells.

In many cases it is only a part of the protein which is important to its chemical activity in the cell. For example in the case of enzymes, it is the shape and chemistry of a part of the protein molecule, called the active site, which is mainly responsible for the protein's catalytic activity. There are, for present purposes, two consequences of this. First all proteins can be broken up or digested by enzymes (called proteolytic enzymes) which are specific to the destruction of proteins. The process of proteolysis may result, at least at an initial stage, in the creation of large fragments of the protein which is being digested. Some of these fragments may retain all or much of the specific chemical activity of the whole protein from which they came. Secondly, it is possible for, say, a particular enzyme found in related life forms to have the same active site even though there may be differences in other parts of the protein molecule. This appears to be the case, for example, in the case of the enzyme known as adenylate cyclase which plays an important part in this litigation. Adenylate cyclase is the enzyme which catalyses the chemical reaction which turns a molecule called adenosine triphosphate (ATP) into cyclic adenosine monophosphate (c-AMP). This enzyme is found in all bacteria and host cells. However the version of it found in B. pertussis appears to be slightly larger than the version to be found in a very close bacterial cousin known as B. bronchiseptica. This means that these two versions of adenylate cyclase will have the same, or substantially the same, active sites but there will be differences in other parts of the protein molecules. Incidentally B. bronchiseptica is the pathogen which causes a disease in pigs called atrophic rhinitis.

Every molecule has a particular mass, known as its molecular weight. It is possible to calculate the precise molecular weight of a protein if its chemical structure is known by adding together the weights of the individual atoms of which it is made. The molecular weight of large molecules like proteins is measured in units known as kilodaltons or kDa. If the precise chemical structure of a protein is not known, there are ways of assessing within a certain margin of error what its molecular weight is. This is a topic which will have to be addressed in more detail below.

The body's immune response and vaccines

A healthy human inherits a number of sophisticated evolved mechanisms to recognise and destroy invading hostile foreign matter. One of these mechanisms, called the humoral immune response, involves the generation of antibodies. At its most simple, this operates as follows. An invading agent, such as a bacterium, is made up of a number of different components, including proteins. It may also secrete proteins. These are called extracellular proteins. If the host can recognise the bacterium or its secretions as foreign it may then be able to attack, destroy or block the toxic effect of them. A foreign protein almost always has regions on it which are capable of being identified by the host. If this is so, it is recognisable as foreign. When considered in the context of its ability to induce a humoral immune response, a protein is called an antigen. In fact on each antigen there are usually a number of locations which, as a result of their unique shape and chemistry, are identifiable by the host. Each such identifiable location is called an epitope or antigenic determinant. Specialised defence cells in the host, called B lymphocytes, can learn to recognise each epitope and then produce another protein which is complementary to it so that it will bind onto the epitope with a "lock and key" type of fit. The protein produced by these defence cells is called an antibody. Each antibody is specific to the particular epitope it is designed to bond to. Each B lymphocyte is only capable of producing antibodies to a single epitope. Antibodies are collectively known as immunoglobulins, which may be abbreviated as Ig. There are a number of different types of immunoglobulins. Antibodies for a particular foreign epitope can activate the body's immune systems to destroy the antigen carrying that epitope. In some cases the body's defence mechanism can "remember" that it has produced a specific antibody to a specific epitope. If that is the case then when a foreign body containing that epitope tries to enter the body for the second time it will be recognised faster as alien and the body's defence mechanisms will switch into operation faster. Therefore in relation to that particular epitope the body has a primed defence ready and waiting.

It is this ability to prime the body's defence mechanism which lies behind the use of vaccines. Essentially a vaccine consists of a mixture of some or all of the antigens, and therefore the epitopes, of a pathogenic agent such as a bacteria. It is administered in a form which is either not pathogenic at all or in which its pathogenicity is severely compromised. The body learns to recognise the antigens and their epitopes and generates antibodies to them. If the pathogenic agent subsequently infects the body it will be more rapidly met with an effective immune response. In fact some antibodies do not confer immunity on the host. For example a protein may be embedded deep inside a pathogen. Although that protein is capable of giving rise to an antibody response when exposed to B lymphocytes in the host, those antibodies may be ineffective to protect the host. This may be because when there is a repeat infection, the host does not "see" the antigen buried deep inside the pathogen so that no antibodies will be raised. When an antigen elicits an antibody response and that response is effective to confer on the host some degree of protection against further infection, the antigen is said to be protective.

There are a number of ways of making vaccines. They can be made in some cases by simply killing the pathogen, for example with chemicals, and then injecting that dead material into the body. Although killed, many of the proteins in the pathogen are still present and will still generate an immune response. B. pertussis whole cell vaccines made from killed bacteria have been used with considerable success in the past. It was such whole cell whooping cough vaccines which were introduced in the late 1940's and were very effective in suppressing the disease in vaccinated populations. It may be expected that most if not all of the antigens present in the live pathogen, including those which are protective, will be present in such a whole cell vaccine. As an alternative, in some cases it is possible by prolonged culture under artificial growth conditions to produce colonies of pathogens which, although in most respects identical to their wild-type ancestors, have become feeble so that they are no longer seriously pathogenic. A vaccine made from such attenuated organisms will also produce an immune response. In these cases substantially all the antigens of the pathogen are used in making the vaccine. Sometimes vaccines are made which contain selected parts of the totality of the antigens in the pathogen. These are called acellular vaccines. For example it may be possible to make a vaccine simply from the mixed proteins excreted by the pathogen. Alternatively one or more antigens carrying their epitopes or parts of antigens carrying some of their epitopes may be selected out of the totality of antigens in the wild organism and put into a vaccine. The selection may be achieved by isolating the antigens directly from fragmented pathogens or by recombinant DNA means, i.e. by genetic engineering to produce the antigens of choice. Some B. pertussis vaccines contain identified, isolated components of the bacterium. Among the advantages of such acellular vaccines is that it may be possible to exclude some or all of the other materials found in whole cell vaccines which are toxic. Based on the above it may be convenient to split acellular vaccines, and particularly B. pertussis vaccines, into three broad groups. They were categorised as follows in the evidence^[1]:

(a) simple, undefined cell extracts or culture supernatants

(b) semi-defined preparations, free of undesirable components, consisting of a number of co-purified components, some or most of which are known and characterised, others not and

(c) defined preparations consisting of one or more isolated components, fully characterised with regard to antigenic content and the presence or absence of undesirable components.

The proteins and other large molecules relevant to this dispute

Many of the issues in this dispute turn on the behaviour or detection of a number of large molecules found in or secreted from B. pertussis. They are as follows.

1. Filamentous Haemagglutinin (FHA)

Filamentous haemagglutinin is a secreted protein which forms filamentous like structures on the bacteria's cell surface. These help the bacteria to stick to the ciliated epithelium cells of the respiratory tract. In addition, as its name indicates, it can agglutinate red blood cells. This means that it can stick together numerous red blood cells. The latter ability forms the basis of a test which can be used to detect the presence of FHA.

2. Pertussis Toxin (PT)

Pertussis toxin is an extracellular protein (exotoxin). It is involved in the adhesion of B. pertussis to the lining of the trachea and to cells involved in the host's immune response called phagocytes. Systemically, PT alters the hormone activities that are regulated by c-AMP, resulting in an increase in insulin production leading to hypoglycemia and an increase in sensitivity to histamine which results in increased capillary permeability, hypotension and shock. A part of PT also is responsible for disrupting the regulation of adenylate cyclase activity in the host cell. This results in an uncontrolled elevation of c-AMP levels in affected host cells. This in turn results in the widespread death of human lymphocyte cells which occurs during whooping cough infection. These characteristics make PT highly toxic. In some of the technical documentation pertussis toxin is referred to as LPF.

3. Pertactin (p69)

It is the discovery and identification of Pertactin which lies at the heart of the Evans patent. Pertactin is a protein believed to be involved in the adhesion of B. pertussis to a host cell. It is produced in the cell from a larger protein pre-cursor. The form of pertactin derived from B. pertussis has a molecular weight of 60.5 kDa. However in one of the methods used to assess molecular weight, known as the SDS-PAGE gel method, it behaves as if it had a molecular weight of about 69 kDA. It is for this reason that pertactin is referred to in some of the literature as p69.

4. Adenylate Cyclase

B. pertussis not only stimulates host adenylate cyclase via pertussis toxin, but also secretes its own adenylate cyclase. In B. pertussis it was first identified as a 70 kDa protein. However, it is now known that the 70 kDa protein is a fragment of a larger 216 kDa protein. Since 1988 B. pertussis adenylate cyclase has been known as a bifunctional enzyme because it also catalyses hemolysis, that is to say the destruction of red blood cells. For this reason sometimes it is now referred to as adenylate cyclase-hemolysin. Adenylate cyclase is a very well known enzyme. It was identified many decades ago.

5. Lipopolysaccharide (LPS)

Lipopolysaccharide (LPS) is a component of the bacterial outer membrane. It is toxic.

The patent in suit

The patent in this case has had an unusual history. The inventor named is Dr. Pavel Novotny. The patent arises out of work done by one of Dr. Novotny's students, Juan Montaraz (now a professor), as part of the latter's PhD thesis. It claims priority from a priority document filed on 12 May, 1994. That document states that the inventor has discovered that adenylate cyclase is a major protective antigen against B. pertussis and that this discovery should make it possible to make vaccines based on this antigen. The priority document sets out details of certain experimental work which it suggests had been performed and which led to this antigenic adenylate cyclase. Between the filing of the priority document and the grant of the patent itself, it appears that Dr. Novotny or others within Wellcome decided that the protective antigen which had been discovered was not the known adenylate cyclase but was a quite different protein, subsequently identified as pertactin. There is no dispute between the parties that the claims of the patent as finally approved by the European Patent Office are intended to be and are in fact directed to pertactin, vaccines containing pertactin, a method of producing and the use of pertactin. It is this change in emphasis from the priority document to the patent which is central to the attack on priority and played a significant part in this trial. Therefore it will be necessary to consider with care what was disclosed in the priority document. However it is convenient first to start by considering the patent and its claims. The shift in teaching between the priority document and the patent explains why in many respects the description in the latter appears to be confusing. Large tracts which made sense when located in a document directed towards a newly found characteristic of a known protein, adenylate cyclase, do not sit easily in a patent directed to the entirely different protein, pertactin.

The patent starts with a general description of the prior art and states, accurately, that perceived defects in the existing whole cell whooping cough vaccines resulted in research being carried out to find an acellular alternative. The discovery behind the invention is then described as follows:

"It has now been discovered that certain proteinaceous material, associated with adenylate cyclase activity, as hereinafter described, found in the cultures of B. pertussis, is capable of providing protection against challenge by B. pertussis when administered to experimental animals. This discovery that the proteinaceous material usually associated with adenylate cyclase activity is a major protective antigen against B. pertussis permits the preparation of vaccine formulations comprising antigenic preparations which are free from, or contain reduced amounts of, other known B. pertussis components which may be responsible for the toxic side-effects demonstrated by whole cell vaccines.

The term 'proteinaceous material associated with adenylate cyclase activity' (abbreviated to 'ACAP' hereinafter) is used herein to refer to proteinaceous material which is extracted together with adenylate cyclase activity when extraction of the adenylate cyclase activity is performed using an aqueous, acidic (pH3) solution of glycine (0.25 M). The ACAP as defined above may comprise the adenylate cyclase enzyme per se or a binding protein for the enzyme.

Adenylate cyclase activity was assayed by the method of Hewlett, E., and Wolff, J. (J. Bacteriol. (1976) 127 890-898)."

Although this passage, which in large part is derived from the priority document, refers extensively to adenylate cyclase, it must be recalled that the claimed invention is said not to relate to adenylate cyclase but to pertactin. The result is that when this passage says that the discovered ACAP "may comprise the adenylate cyclase enzyme per se or a binding protein for the enzyme" it is, at the very least, confusing since pertactin is neither adenylate cyclase nor a binding protein for it. This confusion is to be found in other parts of the specification. For example in another passage derived wholly or in large part from the priority document, the specification refers to the fact that adenylate cyclase has been previously isolated from B. pertussis but only in a yield of about 20%. It goes on:

"An extraction process is therefore required by which the ACAP can be obtained in high purity and yield, in order to afford sufficient quantities, on a commercial scale, of ACAP for use in the above-mentioned antigenic preparations...We have now discovered that in contrast to the use of detergents, extraction of B. pertussis organisms using regulated, mildly acidic conditions results in the extraction of substantially increased yields (about 40 x better than previously reported techniques) of adenylate cyclase from the outer membrane in a form which is water-soluble." (emphasis added)

This made sense when the inventor thought that he was trying to obtain large quantities of adenylate cyclase. It is harder to make it fit into a specification which is not concerned with the recovery of adenylate cyclase but with the recovery of an entirely different protein. The same problem arises in relation to page 4 of the specification where reference is made to the ability of ACAP to elicit a protective response in a mouse assay known as the Kendrick test (to be explained in greater detail below). It says:

"The supernatant extract obtained has been tested in the Kendrick Test, as described below, and has been found to provide protection in mice against intracerebral challenge with B. pertussis. Control vaccines containing no adenylate cyclase activity were found to provide little or no protection against challenge with B. pertussis, suggesting that ACAP may, in fact, be the most important factor in immunity. Analysis of batches of non-protective whole-cell vaccine has also shown that non-protection tends to be associated with a lack of adenylate cyclase activity, further suggesting that ACAP may be the key antigen necessary for eliciting an immune response against B. pertussis."

Since, as noted above, the patent says that ACAP can consist of pure adenylate cyclase, this passage appears to be teaching the reader that the invention covers vaccines made from that enzyme in the pure state. Similarly Table 1, which demonstrates the alleged efficacy of the product of the invention, describes it as "B. pertussis immunopurified adenylate cyclase".

These ambiguities aside, the specification sets out a series of examples by which it is said that the antigen of the invention may be produced in and separated from cultures of B. pertussis. Essentially it describes a method by which the harvested B. pertussis bacteria are split open by a process known as acid glycine hydrolysis and the crude outer membrane proteins (sometimes referred to as "OMP's") are collected. In other words the outer membrane proteins are separated from the intracellular proteins. As the OMP fraction will include a large number of proteins, the specification describes a series of steps which can be taken to select out the protein of choice. Whether these steps in fact produce the protein of choice, that is pertactin, is another of the major issues in this case which will be considered later in this judgment. The fraction produced after acid glycine hydrolysis is centrifuged and the supernatant is found to give positive results in the Kendrick test. The specification notes that:

"The supernatant extract used in the Kendrick Test may, however, also contain the ACAP in small quantities complexed with other proteins including fragments of LPS, in which case, it may be desirable to purify further the material for use in the vaccine formulations according to the invention."

It goes on to explain that the protein of the invention may be mixed with pharmaceutically acceptable carriers to form vaccines.

Construction of the claims

It is against this background that the claims must be considered. Only three were said to be independently valid, Nos. 1, 2 and 8. The first of these reads as follows:

"A purified Bordetella pertussis antigen which is extracted together with adenylate cyclase activity when extraction of the activity is performed using an aqueous pH3 solution of 0.25M glycine and which is characterized by the following features:

a relative molecular weight of between 67,000 and 73,000 as determined by 12% (w/w) polyacrylamide gel electrophoresis; and

a ratio of proline to glutamic acid of substantially 1:1 as determined by amino acid analysis;

said antigen being substantially free from intracellular B. pertussis material."

The words in italic were added in the EPO, apparently in answer to an added matter attack. Since the claim covers, that is to say at least includes a pure antigen which contains no adenylate cyclase, it will be appreciated that these words again are somewhat confusing.

There are arguments on construction in relation to three parts of this claim. The first two can be considered together. First Chiron argues that, in the context of the patent as a whole, the word "purified" must mean highly pure. It suggests that purity must be of the order of 90+% pertactin with no significant contaminants. Two reasons are put forward for this. First it is said that the claim requires a proline to glutamic acid ratio of substantially 1:1 and it is not possible to work out the amino acid ratio without using extremely pure protein. Secondly the specification describes a number of steps leading to immunopurification with a monoclonal antibody. As will be discussed more fully below, such immunopurification can yield extremely pure protein. In fact it can yield protein far purer than 90+%. Chiron argues that the claim, read in the context of the specification, must not only include such immunopurified protein (a contention not disputed by Evans) but must be limited to proteins of such purity. Against this, Evans argues that "purified" only means that the antigen has undergone a process of purification. It notes that the claim does not say that the antigen is "pure".

The second dispute concerns the words "substantially free from intracellular B. pertussis material". Here Evans suggests that there is no need for absolute freedom from intracellular material. It says that the words should not be construed so as to exclude the presence of amounts of intracellular material which are "insignificant by contrast with prior art whole cell vaccines". This would mean that there could still be very large quantities of intracellular material present. Chiron argues that these words mean that the antigen has so little intracellular material mixed with it that it can for practical purposes be treated as containing none.

I can find nothing in the claims or the patents which support Chiron's construction in relation to "purified". Had the patentee wished to say "pure" he could and should have done so. The fact that high purity may need to be achieved to work out the amino acid ratios and that such purity can be achieved by immunopurification with a monoclonal antibody throws no light on what is meant by the word "purified" in the claim. The proline/glutamic acid ratio is a characteristic of the protein of interest within the purified bulk. Specifying that ratio in the active ingredient says nothing about the degree of purification. All that the claim requires is that the antigen has been purified. It does not require any particular level of purification. Any antigen which has been through a process of purification and otherwise meets the features of claim 1 is covered by it.

On the other hand the claim does specify a degree of freedom from intracellular B. pertussis material. The antigen must be substantially free. The construction advanced by Evans only compares the product within the claim to prior art whole cell vaccines. The latter contain virtually all the intracellular material of B. pertussis. There is nothing in the patent to support Evans' construction on this point. Prima facie the words in the claim should bear their normal meaning which is the one advanced by Chiron. That would also be consistent with the teaching of the specification which is directed to separating the intracellular material from the OMP's of B. pertussis and then selecting out of the latter the antigen of choice.

The remaining point on construction concerns the words "which is extracted together with adenylate cyclase activity when extraction of the activity is performed using an aqueous pH3 solution of 0.25M glycine". Evans argues that this refers to an inherent characteristic of the antigen of claim 1, namely that if it is extracted by that particular method, then it will be accompanied by adenylate cyclase activity. This is a product claim to the antigen (so identified) as such. It may be extracted or made by other methods which do not involve the co-extraction of adenylate cyclase activity. The claim only asserts that adenylate cyclase activity will be present "when" acid glycine hydrolysis extraction is used. Evans also refers to the passage in the specification quoted above which states that the proteinaceous material containing the major protective antigen is "usually associated with adenylate cyclase activity". This clearly indicates that it sometimes is not so associated. The claim must be construed to cover the antigen when not associated so as to be consistent with the specification. Chiron says that if this is correct, then the words have no limiting effect at all. It says that the words must be construed as a limitation to the antigen made by the specified method.

I think Evan's construction is correct. The words in question simply confirm that when the antigen is extracted by the acid glycine process then it will be accompanied by adenylate cyclase activity. In other words it says no more than that the antigen will be found in a particular fraction when made from B. pertussis in a particular way. It does not require the antigen to be made in that way. It follows that the words are redundant, as Evans suggests. They serve no purpose save, perhaps, to preserve a linguistic bridge between the claim and the disclosure in the priority document. I note that since the acid glycine extraction process is likely to recover all or most of the OMPs, this is not even a description which serves to usefully define the antigen of the claim. Further, the claim does not purport to be of a product-by-process type, even were that type of claim allowable under EPO practice.

It follows that claim 1 covers the specific antigen, now called pertactin, substantially free of significant quantities of intracellular material and purified, at least to some extent. As both parties agree, the claim is broad enough to cover essentially 100% pure pertactin.

Claim 2 is in the following terms:

"A vaccine comprising an antigen as claimed in claim 1 in admixture with pharmaceutically acceptable carrier."

The only point which arises in relation to this is how much antigen must be present to fall within this claim. The answer to this question may affect whether some of the prior art anticipates or renders the claim obvious and whether the Chiron vaccine infringes. Evans suggested that since the claim is to a vaccine, it is inherent that there be sufficient pertactin in it to give rise to protection in an average patient into whom it is injected. The reliance on protection has this significance. The mere fact that antibodies may be raised does not mean that the vaccine is protective. As noted above sometimes antibodies may be raised to an antigen but they are ineffective to protect the host. Even if it were shown that the vaccine gave rise to protection, for example in the Kendrick test, this is no proof that it would give rise to protection in man. This is because the behaviour of a vaccine in a mouse test does not prove that the vaccine will behave in the same way in humans. However, if this is what the claim means, it would follow that no prior art would be able to invalidate and no subsequent use would be proved to infringe until full human clinical trials, preferably comparing protective responses to vaccines with and without pertactin, had been carried out. It is most unlikely that the claim could have any such meaning. Furthermore, if that was what it meant, it would mean that the specification itself would be expected to give details of clinical trials carried out by the inventor. No such clinical trials are recorded, nor could they be. This patent was applied for years before clinical trials would be possible. Chiron argues that all that is needed is that there is sufficient antigen present to give rise to an antibody response. I agree.

Finally, claim 8 is as follows:

"A method for the isolation of an antigen as claimed in claim 1 which comprises treating a culture of B. pertussis cells with an aqueous amino acid buffer of pH 2.5-3.5, comprising a hypertonic concentration of said amino acid with respect to the cells, separating the cells from the resulting supernatant and isolating the antigen from the supernatant."

This is a claim to obtaining pertactin from B. pertussis bacteria by use of what is known as the acid hydrolysis process. No points of construction arise in relation to it.

B. Validity

As noted above, there is no dispute that the patent is directed at pertactin whereas the priority document appears, at least superficially, to be directed at the different protein, adenylate cyclase. The difference between these proteins is marked. Not only does the former not exhibit the enzymatic activity which is the hallmark of the latter but the latter does not have a proline to glutamic acid ratio anywhere near 1:1. There are numerous other differences. Based on this, Chiron mounted a detailed attack on priority entitlement.

Priority date - the law

Section 5(1) of the Patent Act, 1977 stipulates that, except as provided for by other provisions of the Act, the priority date of a patent is the date of filing the application. In this case, that would give a priority date of 10 May 1985. However s. 5(2)(a) provides that subject to complying with certain formal requirements:

"if an invention to which the application in suit relates is supported by matter disclosed in [an earlier] application or applications, the priority date of that invention shall instead of being the date of filing the application in suit be the date of filing the [earlier] application in which that matter was disclosed ..."

In this case the earlier application relied on is the priority document filed on 12 May 1984. In a number of recent cases it has been held that the requirement that the invention be "supported by" the priority document means that it must be shown that the latter contains an "enabling disclosure", that is to say a disclosure which would enable a notional addressee of the document to produce the products or work the methods of the invention. Chiron drew my attention to Biogen Inc. v. Medeva plc [1997] RPC 1 at 48 where Lord Hoffmann pointed out that the specification must enable the invention to be performed "to the full extent of the monopoly claimed". Since the invention encompasses substantially pure pertactin and this is a preferred embodiment in the patent^[2], Chiron says that the patent must enable that. I did not understand Evans to disagree. However there are three issues on which there does appear to be a difference of approach between the parties namely, onus, whether both functional and linguistic support was necessary and how easy must it be for notional man skilled in the art to achieve the product or method of the invention by following the disclosure in the priority document.

Priority - onus

There is no dispute between the parties that the onus is on the patentee to justify the earlier priority date pursuant to s. 5(2) of the Act. But there appears to be a difference of opinion as to what this involves in practice. Although the onus is on Evans to prove entitlement to the earlier priority date, it conducted no experiments to show that the teaching of the priority document would enable someone in the art to make the antigen which is the subject of the claims. All the experiments were conducted by Chiron. Evans stance is to attack the experiments carried out by Chiron. Is that enough or is the burden on Evans to put forward evidence supporting its claim to priority?

The issue of onus was canvassed in the Court of Appeal in Biogen [1995] RPC 25 at 105 where Hobhouse LJ said:

"In our judgment, the burden of proof is upon the patentee to establish his entitlement to an earlier priority date when he seeks to rely on one. However if the patentee establishes that prima facie entitlement and if the challenger wishes to displace that prima facie position, the challenger may have to allege further facts. The challenger will then have to discharge an evidential burden of proof and prove the facts he has alleged. In other words, the question of upon whom the burden of proof lies on any particular issue falls to be determined by asking who has to make the relevant allegation."

In virtually all cases the patentee will seek to obtain early priority, thereby cutting down on the prior art. Although the onus is on him, it would be strange if this had the result in most cases of requiring the patentee to carry out experiments simply to prove entitlement failing which the defendant could simply assert that early priority had not been proved. Although the onus is on the patentee, in a normal case the court will proceed on the basis that the technical results set out in the priority document can be achieved as asserted. It is enough if the patentee asserts that the instructions in the priority document work to produce the results claimed for them. A claim to priority is, implicitly, such an assertion of enablement. If that is not effectively challenged then the patentee has done enough.

It follows that a defendant should do more than merely issue a challenge to priority. He must explain the manner in which he says there is failure to enable. This should be set out in his pleadings. He will have to support his assertions by evidence. The court should be able to proceed on the basis that the experimental data in the priority document is not only right but reasonably repeatable. Although the onus remains on the patentee, the evidential burden shifts to the defendant. In some cases he may not be able to discharge it without supporting experiments.

I understand Chiron to accept this. Mr. Kitchin QC, who appears on its behalf, says that although the onus lays on the patentee, potentially, the evidential burden could shift from party to party. Accordingly, in circumstances where the priority document on its face appears to support the patent, it may well be that, absent any other circumstances, that would be enough to establish a prima facie entitlement to priority. Absent some specific contention by the challenger, that prima facie position will be accepted by the court.

However Chiron says that this is not a normal case. Here, if the court assumes that what is said in the priority document is correct, then what it promises is not pertactin but adenylate cyclase. Therefore this is not a case where, prima facie, the priority document supports the invention claimed. Furthermore in this case there has been a clear challenge directed at the accuracy of the experimental details set out in the priority document. Chiron says that Dr. Novotny's own subsequent published work demonstrates that much of the experimental details set out in the priority document are false in the sense that the reported experiments were not carried out on the whooping cough bacterium, B. pertussis, and fractions derived from it at all but were carried out in relation to B. bronchiseptica. It says that that conclusion is confirmed by a consideration of the discovery relating to Novotny's work which has been disclosed in these proceedings. Although Mr. Howe QC, who appears on behalf of Evans and SKB, is not prepared to formally concede that underlying assertion of fact, it became clear during the course of the trial that Chiron's criticisms of the authenticity of the experimental work are entirely justified. Insofar as the priority document says that following the procedures it describes in relation to B. pertussis the inventor in fact obtained the results described, it is false.

I have come to the conclusion that Chiron is right on the question of onus for both of the reasons it gives. In the circumstances of this case, the evidential burden of proving enablement is on Evans.

In any event, no matter who the evidential burden may be on, Chiron has put forward a positive case of non-enablement which it says is supported by experiments carried out on its behalf for the purpose of these proceedings and evidence from highly qualified scientists. That material is alleged to show that following the procedures of the priority document would not lead a worker in the art to pertactin. Rather it would either lead to adenylate cyclase or would leave him in a complete fog as to what antigen, if any, is recovered. Although Evans did not carry out any experiments of its own, it says that Chiron's experiments prove that the teaching of the priority document would lead the man in the art to pertactin even though the author described it as adenylate cyclase. It therefore relies on its opponents' experiments to support priority.

Priority - nature of the test

Chiron says that even if the detailed teaching of the priority document is in fact enabling, that is not sufficient unless it also describes the invention. For this proposition it relies on the following extract from the speech of Lord Jauncey in Asahi Kasei Kogyo KK's Application [1991] RPC 465 at 547:

"... an invention is only supported by prior matter for the purposes of paragraph [5(2)(a)] if the earlier relevant application not only describes the invention but also contains an enabling disclosure thereanent."

Chiron also relies on the words of Aldous J, as he then was, in Schering Biotech's Application [1993] RPC 249 at 252, adopted by Jacob J in Beloit v. Valmet [ 1995] RPC 705 at 732, to the effect that

"'support' means more than that [i.e. than mere textual mention of the relevant features] and requires the description to be the base which can fairly entitle the patentee to a monopoly of the width claimed."

If Chiron is correct that there is a double test for enablement, then it follows that if there is a significant error in the way in which the priority document describes the technology which is the subject of the claim to priority, priority will be lost and, by parity of reasoning, if the same defect is found in the patent, it will fail for insufficiency. Evans challenges this approach. It says that it is in order to describe a material which is not yet perfectly characterised, so long as its useful properties have been identified and a practical route for its preparation has been set out. It says that there is no burden on the applicant for a patent to wait before filing until he has resolved all interesting scientific issues about the nature of his material, even if they are very fundamental and important from a scientific point of view. Although it accepts that the naming of the protective antigen in the priority document as adenylate cyclase has proved to be mistaken, it submits that this is a mere case of disclosing an incorrect theory. It suggests that the court should adopt the approach of Fletcher-Moulton LJ in "Z" Electric Lamp Mfg Co v. Marples, Leach & Co Ltd (1910) 30 RPC 737, at 746:

"The patentee's obligation is not to be omniscient; the patentee's obligation is to put the public in the possession of his invention, and if he does that bona fide in such a way that they know its advantages practically, and they can obtain those advantages practically, the fact that he has formed an erroneous view in theory of that which procures those advantages, or the state of things in which those advantages occur, does not in my opinion militate against him."

As Mr. Howe puts it, a rose smells as sweet, and an antigen is as protective and useful in a vaccine, under any other name. He says that the inventor can call his produce "Gook X" if he likes so long as he discloses a practical method for its preparation.

I do not accept Chiron's argument that there is a double test of enablement. In the passage from his speech in Asahi quoted above, Lord Jauncey was not purporting to set such a test. He was only saying that describing a result as being achievable was not enough to prove it to be achievable in fact. It is the latter which amounts to enablement. Nor was such a test being proposed in Schering or Beloit. As Lord Hoffmann said at page 47 in Asahi, the requirement of an enabling disclosure in a patent application is a matter of substance not form. Therefore, to this extent I accept Evan's argument. The reader will not necessarily be misled by incorrect nomenclature or other errors of description in the priority document. I think Evans is right when it says that it is sufficient if the document sets out the useful properties of the technical advance and identifies a practical route which achieves it. That does not mean that the general description in the priority document is to be ignored and only the examples looked at. The priority document must be read as a whole. Its general description is no less a part of the technical content than the examples it contains. But the "rose by another name" argument can lead one astray. There is a great deal of difference between referring to a known or a new substance by a new and unknown name, such as "Gook X" on the one hand, and referring to something by a name which already is well known for a particular product on the other. Shakespeare wrote, "What's in a name? That which we call a rose, by any other name would smell as sweet." The same approach applies in patent law. The question must always be "what is in" the name or terminology used by the inventor. What does it convey? "Gook X" in Mr. Howe's example is only the patentee's shorthand for "this is the thing I have discovered". It is a name which does not introduce known characteristics or qualities. But if the inventor describes his advance by reference to well known technical expressions or terminology, the usual meaning of those expressions or terminology becomes part of his technical disclosure. To refer to a rose as "Gook X" does not mean that Gook X is not a rose. But to refer to a plant as a rose does convey that it is a living organism within that well known genus and which exhibits the common characteristics of members of the genus.

There is no reason why the notional skilled addressee of the document should be assumed to ignore any part of the description in it. Absent special circumstances, an addressee is likely to approach the document assuming that all its parts, general and specific, are equally accurate. If following the examples in the priority document produces a result which is inconsistent with what the inventor describes as his achievement then the addressee will appreciate that one or other is wrong. If it is apparent that the general teaching is wrong and the examples are correct and the addressee can readily follow them, then the technical content of the latter is enabled. On the other hand if it is apparent that one or more of the examples is incorrect and the addressee can follow the general teaching and such examples, if any, as are clearly accurate without undue difficulty, then what is in the general teaching is enabled. If the reader does not know which is correct, then it is difficult to see how the priority document is enabling.

Priority - the extent of effort required of the notional addressee

Mr. Kitchin says that the approach to be adopted is that set out by Aldous J in Mentor v. Hollister [1991] FSR 557 which was supported and expanded upon by the Court of Appeal, [1993] RPC 7. Both courts emphasised that what amounted to sufficient instruction in a document depended on the facts of the case and the nature of the invention. Aldous J. adopting the words of Buckley LJ in Standard Brands Incorporated's Patent [1981] RPC 499, said that it was still the law that an inventor is not entitled to set the reader of his specification a puzzle and call it a specification. In the Court of Appeal Lloyd LJ drew particular attention to and relied upon the following passages from Valensi v. British Radio Corporation [1973] RPC 337:

"We think that the effect of these cases as a whole is to show that the hypothetical addressee is not a person of exceptional skill and knowledge, that he is not to be expected to exercise any invention nor any prolonged research, inquiry or experiment. He must, however, be prepared to display a reasonable degree of skill and common knowledge of the art in making trials and to correct obvious errors in the specification if a means of correcting them can readily be found.

Further, we are of opinion that it is not only inventive steps that cannot be required of the addressee. While the addressee must be taken as a person with a will to make the instructions work, he is not to be called upon to make a prolonged study of matters which present some initial difficulty: and, in particular, if there are actual errors in the specification - if the apparatus really will not work without departing from what is described - then, unless both the existence of the error and the way to correct it can quickly be discovered by an addressee of the degree of skill and knowledge which we envisage, the description is insufficient."

This passage is of significance to this case because it is not in dispute that the priority document, if it is supposed to lead to the production of pertactin, contains a number of significant errors. Chiron's case is that the priority document fails to give adequate instructions to allow the addressee to make pertactin without excessive trial and error. Indeed it says that following the teaching of that document as well as he could, the notional addressee would either give up in despair or modify the procedures so as to arrive at adenylate cyclase. In any event, it says that Evans' case, even if right on the facts, comes down to no more than an assertion that some skilled addressees might follow the teaching in the priority document and, by making the right choices when the teaching of the document was not capable of being reproduced, arrive at pertactin. Chiron says that even if that is the case, the document is not enabling. Enablement involves leading any reasonable skilled addressee to the desired goal. It is not a lottery.

Mr. Howe disputes this. First he says that, on the evidence, a skilled addressee would obtain pertactin by following the teaching of the priority document. Chiron's evidence to the contrary is not credible. Secondly he says that you have to take one notional skilled person and you have to work out, on the balance of probabilities, what he would have done. If it is more likely that he would go down the enabling route than the other, there is enablement. Even if his client is wrong on the first point it should succeed on the second.

The question of whether the evidence unequivocally supports Chiron's or Evans' case on the teaching of the priority document is a matter which I will consider below. But in my view Mr. Howe is wrong on his second point. His suggestion that it is a question of the balance of probabilities can have the effect of hiding defects in the priority document. It can result in an artificial all or none answer to the question of what the priority document teaches. The reality may be that the document is so obscure that different results can be achieved by trying to follow its teaching. It is not enough if the instructions are such that a number of equally qualified notional addressees can arrive at completely different end points, some within the scope of the claimed invention and some not. If reasonable addressees can come to different conclusions there is a conundrum as to which is right. That is not enablement. This view appears to be consistent with the approach of the Technical Board of Appeal of the EPO in Unilever case T 226/85 which was referred to with approval both by Aldous J and the Court of Appeal in Mentor v. Hollister:

"Even though a reasonable amount of trial and error is permissible when it comes to the sufficiency of disclosure in an unexplored field or - as it is in this case - where there are many technical difficulties, there must be available adequate instructions in the specification or on the basis of common general knowledge which would lead the skilled person necessarily and directly towards success through the evaluation of initial failures or through an acceptable statistical expectation rate in the case of random experiments." (emphasis added).

The priority document - the general teaching

The priority document bears the general title "Antigenic Preparations and the Isolation of Such Preparations". Its stated concern is to produce antigenic preparations for use in acellular vaccines against B. pertussis^[3]. Within that general objective, Dr. Novotny sets out with clarity a number of contributions to the art which he believes he has made. The first is expressed as follows:

"We have now discovered that adenylate cyclase, an enzyme found in the cultures of B. pertussis, is capable of providing protection against challenge by B. pertussis in experimental animals. This discovery that adenylate cyclase is a major protective antigen against B. pertussis permits the preparation of vaccine formulations comprising antigenic preparations which are free from, or contain reduced amounts of, other B. pertussis components which are responsible for the toxic side-effects demonstrated by whole cell vaccines."^[4]

Secondly he says that that discovery can be put to good use:

"According to one feature of the present invention we provide a vaccine formulation for protection against B. pertussis, which includes an antigenic preparation comprising adenylate cyclase, optionally toxoided e.g. using formalin, glutaraldehyde or B-propiolacto n, derived from B. pertussis and having a relative molecular weight of 67,000 to 73,000, substantially free from intracellular material, together with an adjuvant and/or carrier for the said antigenic preparation."^[5]

Toxoiding is a process by which certain cellular components are rendered less toxic by chemical treatment. The document then proceeds to discuss how adenylate cyclase has been prepared previously by isolation from cultures of B. pertussis. It refers to a paper by Hewlett and Wolff which, according to the evidence before me, was extremely well known to those in this art at the time and which would have been referred to by any skilled addressee of the priority document. The priority document notes that only about 20% of the total adenylate cyclase enzymatic activity, representing 0.5% of the total adenylate cyclase in B. pertussis, is recovered by the existing isolation techniques. It goes on to note:

"An extraction process is therefore required by which the adenylate cyclase can be obtained in high purity and yield, in order to afford sufficient quantities, on a commercial scale, of adenylate cyclase for use in the above mentioned antigenic preparations."^[6]

In response to this need, Novotny states:

"We have now discovered that ... extraction of B. pertussis organisms using mild acidic conditions results in the extraction of substantially increased yields (about 40 x better than previously reported techniques) of adenylate cyclase from the outer membrane in a form which is water-soluble."

As a consequence Novotny asserts that a further feature of his invention is the provision of :

"a method, for the isolation of an antigenic preparation containing adenylate cyclase from B. pertussis which comprises treating a culture of B. pertussis cells with a buffer of pH 2.5-3.5 preferably comprising a hypertonic concentration of glycine or other amino acid as a buffer component, separating the cells from the resulting supernatant and isolating an antigenic preparation containing adenylate cyclase from the supernatant."

There is no doubt that all of this amounts to an assertion by Dr. Novotny that he has found a new and valuable antigenic property of a well known protein, namely adenylate cyclase, and an improved method of isolating that protein. This is a strong and clear technical message which the notional skilled addressee would not ignore. The statement in the patent specification, referred to above, that the desired antigen is usually associated with adenylate cyclase activity, carrying with it the implied statement that it is not always so associated, is not to be found in the priority document. Indeed the implied statement is quite contrary to the express teaching of the latter document. This wording was inserted for the first time in the patent application in May 1985 and, as already explained, is relied on by Evans to show that claim 1 covers pertactin which is not even extracted in conjunction with adenylate cyclase.

However the general statements are not the only technical information in the priority document. It goes on to set out some examples which are said to illustrate Dr. Novotny's invention. Evans' case is that working through the examples sensibly will lead to the antigen which Dr. Novotny had discovered and which is of use in making acellular vaccines, namely pertactin. The priority document is therefore enabling and gives the necessary support to the invention claimed in the patent. To understand those examples it is necessary to have a basic understanding of some more technology.

Apparent molecular weight

As mentioned already, although they are made up from the same 20 amino acid building blocks, proteins vary greatly in size and composition. They can therefore vary in their molecular weight. A number of ways exist of calculating or assessing the molecular weight of a protein. One of these, referred to in the priority document and the patent specification, is SDS polyacrylamide-gel electrophoresis, or SDS-PAGE.

Electrophoresis is a method which can be used to separate mixtures or proteins or nucleic acids according to the size and charge of the components of the mixture. When applied to proteins, the technique relies in part on the fact that proteins usually have a net positive or negative charge. When an electric field is applied to a solution containing protein molecules each will migrate at a rate dependent upon its net charge, size and shape. Therefore, different proteins will move at different rates, and electrophoresis can thus be used to separate mixtures of proteins. This phenomenon can be used to give an indication of molecular weight as follows. In SDS-PAGE, proteins are first dissolved in a solution of sodium dodecyl sulphate (SDS), a negatively charged detergent. This causes the protein molecules to unfold (denature) and frees them from links with other molecules. Negatively charged SDS ions bind with the denatured proteins to form complexes that have a total net negative charge that is roughly proportional to the mass of the protein. This means that the migration of the proteins is determined by their size and not by the net effect of the charge of the protein. In SDS-PAGE, the proteins are made to migrate down a polyacrylamide gel (PAGE).

The polyacrylamide-gel is usually prepared immediately before electrophoresis, and its pore size can be chosen so as to retard the migration of the protein of interest. The proteins in solution with SDS are loaded in wells at the top of the gel and an electric current is applied. The proteins migrate toward the positive electrode, with smaller proteins moving much more readily through pores in the gel than larger ones. This results in a series of discrete protein bands, arranged in order of molecular weight, the smaller proteins travelling further through the gel. The location of proteins on the gel are then detected by staining them with a dye, such as Coomassie Blue, or by other means. If a number of standard protein markers of different and known actual molecular weights are also run on the gel at the same time, they also will be separated out by the electrophoresis. By comparing the position of a target protein with the positions of the markers it is possible to arrive at an apparent molecular weight. The technique produces a result like this:

Fig. 1:

[Diagram or picture not reproduced in HTML version - see original .rtf file to view diagram or picture]

The depth of the blue stain gives some qualitative indication of the total amount of protein at each location on the gel (deeper and thicker blue = more protein). However because the stain which is used stains all proteins without distinguishing between them, it is not possible to determine whether any particular stained patch consists of one protein or a number of proteins of the same or substantially the same apparent molecular weight. It necessarily follows that if there is more than one protein present at the same location, this method will not be able to give an indication of the relative proportions of them.

The apparent molecular weight arrived at by SDS-PAGE is not precise. The evidence before me was that there can be an error in apparent molecular weight of a few percent either way. Furthermore it must be borne in mind that what is being measured by SDS-PAGE is not the actual weight of the protein but its mobility under defined conditions. The result is that the apparent molecular weight of the protein may differ by a significant margin from its actual molecular weight. As noted already, pertactin's apparent molecular weight as measured by SDS-PAGE is about 15% higher than its actual molecular weight.

Ion-exchange chromatography

Proteins can be separated on the basis of their net electric charge by ion exchange chromatography. An ion exchanger is a solid that has chemically bound charged groups to which ions are electrostatically bound; it can exchange these ions for ions in aqueous solution. The principle of ion exchange chromatography is that charged molecules adsorb to ion exchangers reversibly so that molecules can be bound and then washed out ("eluted") by changing the ionic environment. Separation on ion exchangers is usually accomplished in two stages:-

(1) the substances to be separated are bound to the exchanger, using conditions that give stable and tight binding;

(2) the column is then washed with buffers (a liquid having an acidity or alkalinity - "pH" -which is resistant to change) of different pH or different ionic strength, and the components of the buffer compete with the bound material for the binding sites.

If a protein has a net positive charge at pH 7, it will usually bind to beads of resin containing carboxylate groups, whereas a negatively charged protein will not. A protein with a net positive charge bound to such beads can then be removed by increasing the concentration of positively charged ions (e.g. Sodium ions - Na⁺) in the eluting buffer. Such ions compete with positively charged groups on the protein for binding to the column. Proteins that have a low density of net positive charge will tend to emerge first, followed by those having a higher charge density. Therefore this technique can separate proteins which differ in their positive or negative charge density. Other factors such as affinity for the supporting matrix and the pore and mesh size of the matrix can also influence the behaviour of proteins on ion-exchange columns. Negatively charged proteins can be separated by chromatography on positively charged diethylaminoethyl-cellulose (DEAE-cellulose) columns. Positively charged proteins can be separated on negatively charged carboxymethyl-cellulose (CM-cellulose) columns.

Isoelectric focussing

Isoelectric focusing is a technique that allows separation of molecules purely on the basis of their electrical charge characteristics. Proteins contain both positively and negatively charged groups and therefore the net charge of a protein depends on the pH of its environment. At low pH, proteins are more likely to be positively charged, whereas at a high pH proteins are likely to be negatively charged. Because the net charge of the protein depends on pH, a pH gradient can be used to separate proteins based upon their isoelectric point (pI), the pH at which the net charge of a protein or molecule is zero.

The pH gradient of a gel is first formed by electrophoresing on that gel a mixture of small multi-charged polymers having a range of pI values. This mixture produces a pH gradient between the electrodes attached to opposite ends of the gel. The solution containing the proteins to be separated is added to the gel and an electric field is applied. This causes the proteins to migrate to the position of their isoelectric points. At the point that the pH is equal to the pI of a protein, the net charge of the protein becomes zero and it no longer moves through the gradient. Because the protein will migrate towards the point at which it its net charge is zero, it does not matter where on the gel the protein sample is put initially. It will eventually end up at the point on the gel which corresponds with its pI. The technique can be said to focus the protein at that point.

In this case two types of isoelectric focussing (IEF) are referred to. The first is analytical IEF. In this, very small quantities of the sample which is being analysed are put on the gel. The application of the electric field to the gel can result in a very clear separation of proteins even with very similar pI's. This precision of separation is sometimes referred to as its "resolving power". IEF used analytically has high resolving power. It can separate proteins whose pI's differ from each other by as little as 0.2 pH. I understand that analytical IEF is still used in this art.

IEF can also be used preparatively. In this case it is being used to separate larger quantities of proteins from one another than would be the case with analytical IEF. A larger sample of the mixed proteins is put on the gel and, as with analytical IEF, they migrate to their correct pI's. However the bands of proteins will spread out on either side of their respective pI's because there is so much protein present. Therefore the resolving power will decrease when the gel is run in preparative mode. It should be noted that the pI's of the proteins do not change, it is just that the quantities of protein are so large that adjacent proteins will tend to diffuse to some extent into each other's band on the gel. Preparative IEF is not used commercially now.

Techniques using monoclonal antibodies.

One of the most important advances in this field was the development of techniques for making what are known as monoclonal antibodies (MAB's) and their use.

At the beginning of this judgement, reference was made to the way in which a B lymphocyte cell can make antibodies to a foreign epitope. In fact it appears that each B lymphocyte cell can only make a single antibody although it may produce large quantities of it. In the early 1970's Kohler and Millstein worked out a technique by means of which they could "immortalise" B lymphocyte cells. If such an immortal cell was already producing a single antibody and was separated from other cells, it would grow into a colony of identical cells, all of which would produce exactly the same single antibody. The colonies of cloned cells would therefore produce significant quantities of a single antibody which could then be harvested. An antibody produced in that way is called a monoclonal antibody. One method by means of which B lymphocytes could be immortalised was to fuse them with a particular type of cancer cells. The hybrid product of such a fusion is called a hybridoma. Kohler and Millstein invented and described this technique.

In outline the method of making monoclonal antibodies is as follows. A population of B lymphocytes is exposed to a protein. The epitopes on the protein will stimulate some of the B lymphocytes to produce antibodies. The population of lymphocytes will then be subjected to a special treatment in the presence of cancer cells. A number of them will fuse with the cancer cells to produce hybridomas. Some of those hybridomas will be producing antibodies. The cells are then separated from each other and cultured. A number of colonies will grow up which contain nothing but one type of cell (i.e. perfect clones). Some of those monoclonal colonies will thus contain numerous cells each of which is producing the same antibody - i.e. each such colony is producing monoclonal antibodies. As one of Chiron's witnesses, Professor Higgins, put it, raising monoclonals demands skill, luck and time. At the priority date of the patent, it might take many months to produce any suitable hybridomas. The element of luck can be explained as follows. Assume that the lymphocytes are challenged with three proteins, A, B and C, rather than one. This is likely to be what will happen in practice because most protein preparations contain mixtures of proteins. Also assume that each of the proteins has four epitopes e.g. A¹, A², A³ and A⁴ and so on, so that the lymphocytes are being challenged with 12 epitopes. Assume that each of those epitopes induces the production of antibodies from some of the B lymphocytes. The B lymphocytes are now subject to a treatment which randomly will turn some of them into hybridomas. The result may be that 5 hybridomas will be produced which can grow into clone colonies. However there is no way of predicting in advance which hybridomas will be created. It may be that all 5 hybridomas produce antibodies to epitope A³. Or it may be that two produce those antibodies while the other three produce antibodies to epitopes on protein B and so on.

At the molecular level, monoclonal antibodies may be likened to homing pigeons. They are produced in relation to a particular epitope and they will always return to, and take up residence on, that epitope. If one has a technique for finding the monoclonal antibody, for example if it is made with a radioactive marker, it is possible to follow the protein to which it binds.

There is one other feature of antibodies which played a part in the issues and evidence in this case. Consider the case of an antibody which binds to an epitope on a protein which has chemical activity, e.g. it is an enzyme. When the antibody binds to the epitope it may disrupt the chemical activity of the protein. For example the epitope may be so close to the active site of the enzyme that it will prevent it carrying out its catalytic activities. When the antibody attaches to the enzyme, the latter's activity is turned off. If the antibody is subsequently removed, the enzymatic activity may return.

(a) Immunoaffinity purification

Monoclonal antibodies can be used to purify out a single protein from a mixture. For example, antibodies specific for the protein can be coupled to a matrix consisting of inert polymer beads. This can take the form of a column, called an immunoaffinity column. If a mixture containing the protein of interest and other proteins are applied to column, the protein of interest binds to the specific antibodies while the other molecules in the mixture pass through. The protein of interest can then be eluted from the column. The evidence given before me was that, properly run, an immunoaffinity column can produce essentially pure protein^[7]. However there is one feature of this technique which deserves special mention. If the column is loaded with a particular monoclonal antibody, it will select out of a mixture of proteins only those which are specific to the antibody. However that tells you nothing about the protein. If an epitope on an unknown protein X induced the production of the antibody by the B lymphocyte now immortalised in the hybridoma, it is inevitable that anti-X antibody will only select out protein X. But that does not tell you what protein X is.

This may be of significance to one of the issues in this case and can be explained further by way of analogy. Assume that it is desired to find a specific person in the audience at a full Wembley Stadium. Each person has his own specific bundle of characteristics. A certain way of identifying the target would be to finger print or DNA test the audience, assuming of course that you already know what the correct finger print or DNA profile of the target is. On the other hand if someone is randomly pulled out of the crowd and finger-printed or DNA profiled, it is inevitable that he will be the only member of the crowd who matches that finger print or profile. But that does not tell you that he is the target individual. If, however, there is other material which enables you to be certain that he is the target, e.g. because of a combination of other known defining characteristics (e.g. weight, height, hair colour, scars, accent etc), then taking the fingerprint or DNA profile will be useful to help find him a second time round. So too with immunoaffinity purification. If you know that your monoclonal antibody is specific to an epitope on a particular protein, the column can be used to select out that protein in high purity from a mixture. If you do not know the identity of the protein with the epitope to which the monoclonal antibody binds, all that the immunoaffinity purification column will allow you to do is to produce significant quantities of the protein which carries the epitope used for making the monoclonal antibody in the first place. This does not mean that immunoaffinity purification is valueless when an anonymous monoclonal antibody is used. The ability to prepare significant quantities of an unknown protein may itself be a very important step in further analysing the protein for the purpose of characterising it. For example, once purified, its amino acid composition can be worked out and it can be seen whether it has a specific chemical or biological activity.

(b) Western blotting

Following electrophoresis using a SDS-PAGE, proteins may be further studied using an immunoassay technique known as Western blotting. This blotting procedure involves transferring the proteins in the gel onto a nitrocellulose filter, either by electrophoresis or by diffusion. For example the gels illustrated in Fig. 1 above may be pressed against some "blotting paper". Very small quantities of the proteins located in each of the blue stains shown in that illustration will then be transferred to the paper. This results in a band or pattern of bands on the paper which corresponds exactly to that which was in the gel. After this a specific protein may be identified by its interaction with an antibody. To do this, the filter is incubated in a solution containing an antibody that is specific for the protein of interest. The interaction is visualised by the use of an appropriate label. For example the antibody-antigen complex can be detected by rinsing the paper with a second antibody which is specific for the first. The second antibody may be conjugated with an enzyme, which generates a coloured product upon developing, or has a radioactive label which produces a dark band when exposed on a x-ray film. By this means it is possible to identify which particular mark on the SDS-PAGE contained the protein for which the antibody is specific. The enzymatic version of this assay is called ELISA

The Kendrick test

The Kendrick test is an assay which was used to assay the potency of sample antigenic preparations compared to that of standard whole cell vaccines from B. pertussis. The test samples and standard vaccines are made up in a range of at least three dilutions and a small amount of each dilution is used to immunise groups of mice (between 10-20 mice per group). This is done by injecting the samples into the peritoneum of the mice. If a particular sample induces an immune response, to a greater or lesser extent the mice will be protected against developing an infection if subsequently exposed to B. pertussis. To test the efficacy of the samples in generating such protection, approximately fourteen days after the first injection, the mice are challenged with an intra-cerebral injection of B. pertussis bacterial suspension. At the end of the next fourteen day period, the number of mice surviving in each group is plotted against the dosage of antigen sample or standard vaccine. A straight line dose response should be obtained for standards and test samples. The performance of the test samples is gauged by reference to the standards. If fewer mice survive in the samples than in the equivalent whole-cell vaccine mice, the test vaccine is less protective.

In nature B. pertussis is not a serious pathogen in mice. Furthermore when it is pathogenic in humans, it attacks the respiratory system. It is not injected into the human brain. It will be appreciated that this means that the way in which the Kendricks test was conducted bore little apparent relationship to the natural modus operandi of B. pertussis. Nevertheless from the 1940's and still in the mid 1980's it was regarded as the golden test for B. pertussis vaccine efficacy. It is no longer used.

The priority document examples

Against this technical background, I turn to consider the examples which are said to illustrate Dr. Novotny's invention. Although 5 examples are set out, they are really steps in a single experimental procedure which claims to lead the reader from a culture of B. pertussis to a protective antigen suitable for incorporation into a whooping cough vaccine.

Example 1

B. pertussis bacteria have to be split open so as to release the target protein. To this end the bacteria are grown and then harvested, ³/₄ the way through their "exponential phase" (the growth of this culture is described in Example 4^[8]). This is the phase in which the bacteria are still growing strongly. They are then broken open by acid glycine hydrolysis. The OMP fraction is collected. This is a crude preparation which will contain many different proteins.

Example 2(a)

The crude OMP fraction from Example 1 is sent through an ion-exchange chromatography step which, as explained above, is capable of separating proteins according to their different charge densities. After the fraction has been passed through the ion exchange bed, the latter is washed with salt solutions of different strength, namely 0.1, 0.2, 0.3 and 1.0 Molar. The specification then asserts:

"The 69000 m.w. protein was present in the material unretarded by the column, as shown by SDS-PAGE, but was also present in the retarded material eluted by 0.2 M NaCl"

In other words, some of the target protein went straight through the gel while some came out only in the fraction washed with 0.2 Molar salt. The presence of the protein in the latter fraction was shown by subjecting that fraction to SDS-PAGE. If this is right, then a Coomasie Blue stain should appear in about the 69 kDa location on the SDS-PAGE gel of this 0.2 M fraction.

Example 2(b)

The target protein is said to be separable from other proteins to be found in the B. pertussis OMPs by isoelectric focusing. The experimental details of this procedure are set out at pages 9 and 10 of the priority document^[9]. The target protein is said to have a pI of 7.6 - 7.2 under preparative conditions^[10].

Example 3

The target protein is separated from the other proteins which are present in the material which runs through the chromatography column in example 2(a) by use of an immunosorbent column. In other words immunoaffinity purification was carried out on this material using a monoclonal antibody specific for the target protein^[11]. No description of how to make such a monoclonal antibody is given in the specification, nor was one generally available to a man in the art from any existing public source.

Example 5

The immunopurified protein was used in the Kendrick test. It gave pronounced protection to inoculated mice. The comments about the absence of a description of how to make and the lack of public availability of the monoclonal antibody for use in immunopurification applies as much to this example as it does to example 3.

For Evans to succeed on its claim to enablement, it must discharge the onus on it to show that following these examples, the notional skilled addressee of the priority date would, in May, 1984, have obtained pertactin, even though he was told by Dr. Novotny that he would obtain adenylate cyclase. He would also need to realise that what he obtained was what the inventor was targetting. If it was apparent that the obtained material was not adenylate cyclase, the addressee would need to appreciate that the references to adenylate cyclase in the document should be ignored as erroneous. With the exception of the Kendrick test of example 5, Chiron carried out experiments to show that Evans could show no such things. Those experiments, which were repeated at Evans' request are said to show the following:

(a) Example 2(a) does not produce the results promised. To show where the pertactin was in fact, Chiron used antibody to it which is now available but was not available in 1984. The use of this antibody demonstrates that the pertactin was located in the flow through (as the priority document suggested) but it was not in fraction eluted by 0.2M salt buffer.. In fact it was located in the fraction eluted by the 0.1M salt buffer. Furthermore the 0.2M eluate did not appear to contain any 69 kDa protein at all. Adenylate cyclase activity was also not to be found in the 0.2M eluate.

(b) Example 2(b) does not produce the results promised. What happens if this procedure is carried out is represented graphically and by SDS-PAGE gel as follows:


 


Fig 2:

[Diagram or picture not reproduced in HTML version - see original .rtf file to view diagram or picture]
 


The gel was cut into 38 adjoining strips to form 38 fractions with gradually increasing pH's. Some of these (from 12 to 38) are represented in this illustration by the vertical lines against which the legend "pH" is written. The priority document says that the target protein has a pI of 7.6 to 7.2. This is marked accordingly in the figure. Fractions 25 - 27 are in this pH range. It will be seen from the SDS-PAGE gels that none of these fractions have an measurable quantity of protein in the 67-73 kDA range, although they have a number of stained bands which show that a number of smaller proteins are present. On the other hand fractions of a much lower pH, particularly fractions 18 - 20, do give bands on SDS-PAGE indicating that one or more proteins in the 67 - 73 kDa range are present within them. Numerous other bands are also shown in these fractions. These fractions have pH's between 5.6 and 6.00. In fact, although it is not visible in Fig. 2, a stained band in this apparent weight range can be detected by SDS-PAGE between pH 5.33 and 6.18. The Western Blot at the bottom of Fig. 2 was produced by using anti-pertactin antibodies (again, not available at the priority date). It confirms that the pertactin is present, particularly in fractions 18 to 19 and that it is not to be found in fractions 25 to 27 - i.e. in the fractions which, according to the priority document, should have had them as a result of IEF. Testing for adenylate cyclase activity showed that very low levels of such activity was to be found in fractions 23 and 24 only. Once again these are fractions which do not have the pH taught in the priority document.

It follows that if fractions 18 -20 are chosen for further analysis because they contain one or more proteins of about 69 kDa, they do not exhibit adenylate cyclase activity nor the pI promised in the patent. If fractions 23 and 24 are chosen because they exhibit adenylate cyclase activity, they do not contain a 69 kDa protein or have the pI as promised in the patent. Finally, if fractions 25 - 27 are chosen because they have the correct pI, they did not contain any 69 kDa protein nor do they exhibit any adenylate cyclase activity.

The same analysis was carried out in relation to the 0.1M elute. This was chosen rather than the 0.2M elute, as taught by the priority document, because (as noted above) the latter did not contain any pertactin. The results of carrying out the same analysis on the 0.1 elute are as follows:


 
Fig. 3:

[Diagram or picture not reproduced in HTML version - see original .rtf file to view diagram or picture]


 
Once again it will be seen that the fractions which the priority document suggests will contain the target protein in fact have none of the pertactin, and no SDS-PAGE band in the 67 - 73 kDa range. The only fractions containing such bands are nos. 15 to 20. In fact both SDS-PAGE and ELISA (using antibodies not readily available at the priority date) show that the pI of the pertactin is centered on pH 5.94. As for adenylate cyclase activity, this was found, at a very low level, in disparate fractions only one of which was within the 7.2 - 7.6 range set out in the priority document.

(c) Example 3 in the priority document was not carried out because it is for immuno purification of the target protein. This involves using a monoclonal antibody which, according to Chiron, was not available at the priority date.

(d) Example 5 is for the Kendrick test. This also was not carried out. Although the priority document claims that "immunopurified adenylate cyclase" gives pronounced positive results in the test^[12], it is not in dispute that it has been found in subsequent work that pure pertactin in fact gives negative results in the test. That later work shows that if the pertactin is contaminated with very small amounts of PT a synergistic effect is produced and this mixture will produce a positive result in the test.

Evans responds to this material in two ways. First it says that the experiments were not carried out properly so as to repeat the teaching in the priority document. Secondly it argues that the experiments clearly establish that anyone taking reasonable steps to find the antigenic target protein would readily do so. He would be able effectively to put his hand on it, and obtain it in the degree of purity needed for vaccine purposes, even without a raising monoclonal antibody. It says that Chiron's approach to purportedly following the steps of the specification is based on a number of fallacies. It also says that Chiron has erred by treating indications from the examples as to what fractions the material was found in when the inventor carried out the experiment as if they were instructions to be blindly followed, rather than indications to be given appropriate weight with regard to the variabilities of the techniques involved.

The first of these responses can be dealt with easily. The main points advanced by Evans were put forward by Dr. Guiso, who is a highly regarded and clearly very able researcher whose professional life has centered on whooping cough research. She raised a number of objections to Chiron's experiments. For example she asserted that Chiron had failed to harvest the culture of B. pertussis during the exponential phase of growth. She suggested that it might have been harvested at the end of that phase. If that was so many of the bacteria would be dying. This would result in proteolytic enzymes being released into the culture medium. Those enzymes could attack the pertactin. If this was happening, it would "affect the amount of p69 obtained"^[13]. Even if this criticism was made out, I do not see what relevance it has to the issues in this case. It is not suggested that failure to harvest the B. pertussis at the correct time would affect the electrical, weight and other characteristics of such pertactin as is in fact produced. For example whatever the quantity of pertactin produced, it is not suggested that these growth conditions would affect its pI. There is no dispute that pertactin was produced by the B. pertussis used by Chiron in its experiments. The criticism that the quantity of it produced might be affected by the culture conditions is therefore irrelevant. However it is, I regret to say, symptomatic of Evans' attitude to the experiments conducted by Chiron. Evans' witnesses appear to have proceeded on the basis that it was their duty to pursue all possible criticisms of Chiron's work, no matter how big or small each criticism might be and no matter how peripheral it may be to the issues.

In any event, I have no doubt that Dr. Guiso's criticisms are unfounded. The evidence on this issue took up a considerable time. As far as Dr. Guiso is concerned, I came to the conclusion that she was prepared to teach but not to listen. She dug her heels in whenever a technical point was raised which did not suit the position she adopted. She did not appear to be objective in her evidence. In particular in relation to the issue of when the B. pertussis was harvested, I found her evidence did no credit to her obvious and considerable scientific abilities. It was apparent that Chiron had ensured that harvesting occurred at the time specified by Dr. Novotny. No other reasonable conclusion was possible. On this issue Chiron's position was supported by Dr. Maria Kontakou whose evidence I found clear and convincing.

The more important argument is the second one. Would the notional addressee have arrived at the wrong result from following the priority document, as Chiron says, or would he have ignored such signposts as were misleading so as to arrive at pertactin? Before dealing with this, it is appropriate to dispose of one issue. It is apparent that the results set out in the priority document were not in fact obtained by Dr. Novotny and his team. Professor Higgins expressed the view that it was not in accordance with good practice to put fake results in a scientific document like this. Furthermore he thought that Dr. Novotny's underlying work, even had it been properly reported, was of very poor quality. These criticisms appear to me to be of peripheral relevance. If an inventor through clever foresight or lucky guess work describes something which works and how to do it, his disclosure is enabling. It is nihil ad rem that he never carried out the experiments themselves or faked the results. The more complex the area of technology, the less likely it is that the inventor will be able to predict the results of experiments he never carried out or that he will strike lucky, but what is important is what the document teaches, not how the contents got there. Nevertheless the fact that Dr. Novotny's experiments were not conducted on pertactin derived from B. pertussis does provide a possible explanation for why his results are different to Chiron's. It may well be that carrying out the examples in the priority document on B. bronchiseptica and on the particular pertactin produced by that bacterium, as Dr. Novotny did, would always give the results there set out. As will be explained hereafter, there is no doubt that pertactin from B. bronchiseptica is different to pertactin from B. pertussis. It is very likely that the results obtained by Chiron in repeating the examples in the priority document in relation to B. pertussis would have been obtained by any other competent worker in the field in the mid-1980's or, indeed, for many years thereafter. I have no reason to doubt that Dr. Novotny would have obtained such results had he carried out his work on B. pertussis. The issue is how those results would have been interpreted.

Chiron's argument is simple. A reader of the priority document is told that the target protein is adenylate cyclase. That protein had at the priority date one characteristic which was known to be unique to it, namely its enzymatic activity. An important part of any process to isolate the target protein would therefore involve tracking that activity. In following the examples of the priority document it would soon be found that the results claimed were not achieved. This would put the reader in a quandary. He might think that his experimental technique was wrong and, if so, he would repeat it. But eventually he would realise that the results as reported were not achievable. Faced with that he would be likely to track adenylate cyclase activity which would lead him away from pertactin. Alternatively he would bear in mind that the inventor was not claiming to have found a new protein but only a new characteristic of a known protein and a way of obtaining higher yields of it. He would know of Hewlett and Wolff's paper on the preparation of adenylate cyclase from B. pertussis, not only because it was one of the most important and well known papers in the field but because the priority document refers to it as a starting point. The witnesses agreed that the notional skilled addressee would refer to it. That teaches a method of preparing adenylate cyclase from B. pertussis. It states that adenylate cyclase from B. pertussis has a calculated molecular weight of 70.6 kDa and an apparent molecular weight calculated by SDS-PAGE of 69 kDA^[14]. That would confirm, if any confirmation was necessary, that it is the same protein as is being described by Novotny. Hewlett and Wolff would therefore be followed as an alternative route to that set out in the priority document. Again it would lead to adenylate cyclase not pertactin. Finally, if, notwithstanding all the discouragement of inaccurate instructions in the priority document, the skilled addressee was minded to try to see whether the protein he wanted was the one with an apparent molecular weight of about 69 kDa found by SDS-PAGE which might be noticed in the wrong IEF fractions (e.g. the protein in the band at about that weight in fractions 18 to 20 in Fig. 2), he would purify it as much as possible. He might try to make a monoclonal antibody to it, which would take many months of work and would then use that monoclonal to purify the protein from samples of B. pertussis. If he did that he would find, of course, that the monoclonal antibody did not bind adenylate cyclase but that it would bind another, unknown, protein. He would use his immuno purification column carefully to ensure that the protein he recovered was not contaminated with other proteins which might give false results. That pure protein would still not have the right pI but if he decided to go further to try and find out the characteristics of this unknown protein he would be likely to try it out in the Kendrick test, because the objective of the priority document is to make a whooping cough vaccine. The pure anonymous protein would fail that test because, as noted above, pure pertactin is not effective in the test, only impure pertactin is. He would be left not knowing where to go.

Evans' says that the skilled addressee's approach would be very different. He would note that the priority document is directed at a new antigen which is of potential value in whooping cough vaccines. The enzymatic activity of the target protein is irrelevant to this purpose. He would therefore pay no or little regard to monitoring for adenylate cyclase activity. Even if he looked for adenylate cyclase activity, he would not be alarmed by its absence because adenylate cyclase is a fairly sensitive enzyme and it is easy for it to lose activity. He would analyse all the samples produced in accordance with the experimental technique set out by Dr. Novotny, not only those which he is told will contain the target. He would follow the 69 kDa protein which he would find in the wrong place. That would lead him to making a monoclonal antibody which would then be used to make a purified protein. However it would not be too pure. In particular it would be likely to contain PT as a contaminant and this would mean that it would give a positive result in the Kendrick test - the gold test for a vaccine at that time.

These competing arguments were supported by extensive passages in the parties' respective experts' reports. They gave rise to lengthy and detailed cross-examination. It is not possible to do full justice to that evidence here. However my conclusions on the main disputes are as set out below. In coming to them I have had to consider not only the witnesses' evidence but the voluminous technical material put to them. In coming to the conclusions that I do, I have found the evidence from some witnesses more compelling than that from others. I have already said that I found Dr. Guiso lacking in objectivity. The same could be said, although to a lesser extent of Professor Shaw, who was called by Evans, and Dr. Lowe, who was called by Chiron. The other major witness was Professor Higgins. He was most impressive. His views appeared to me to be based on clear and compelling logic.

I cannot accept the basic premise of Evans' argument, namely that the notional skilled worker would have decided, on the basis of the description in the priority document, that the crucial marker for the target protein is that it has an apparent molecular weight of 69 kDa and that its adenylate cyclase activity is of little significance. There is nothing on the face of the document to support such an approach. On the contrary, the one thing the priority document is firm on is that the protein of interest is adenylate cyclase. As far as apparent molecular weight is concerned, although the examples refer to something of 69 kDa, this must be read in the context of the whole document which describes the protein as having a relative molecular weight of 67 to 73 kDa^[15] or "generally ... about 69,000"^[16]. Dr. Guiso accepted under cross examination that this indicated that the molecular weight of the adenylate cyclase might not be precise^[17]. Further, this teaching is to be considered in the light of the fact that apparent molecular weight measurements are not precise and that a man in the art would expect bacteria to contain many proteins of this sort of size. By comparison, adenylate cyclase activity is a precise and unique marker. Professor Higgins was unshaken in his evidence that any good scientist would keep assaying for adenylate cyclase activity to make sure that he was tracking the right protein. Dr. Guiso disputed this. However under cross examination she was obliged to accept that in the art adenylate cyclase had invariably been identified and tracked by its enzymatic activity and its molecular weight was measured afterwards. This was demonstrated by the published work of a large number of separate research teams from 1970 onwards. For example it was what Hewlett and Wolff did in, inter alia, their seminal paper referred to in the priority document^[18]. It was also what was done by Messrs Confer and Eaton in another paper^[19] agreed as seminal in this field and which Dr. Guiso described as the most important publication before 1984^[20]. In fact, Dr. Guiso eventually accepted that she would have tracked activity:

"Q. The way the enzyme has been tracked by all the workers that we have been looking at, both before and after 1984, is to track activity? A. Yes.

Q. They have promised that it is better than Hewlett and, accordingly, is it not the first thing that you check, namely the activity at each stage of the purification protocol?

A. Probably I would have checked the activity, but I would not have paid attention to the activity, because for me, and really for me in 1984, the most important result is the product activity.

Q. You tracked activity in your paper that we looked at, the Ladant paper at bundle K2, page 532.

A. Yes. I told you. I would track activity, but it is not the most important thing as the product activity because it is different from the other vaccine, first; and, secondly, if I am losing activity, I will not be really afraid in 1984 because Confer and Eaton told you already that you can lose activity -- invasive or enzymatic activity."^[21]

Dr. Guiso's evidence that she would have tracked activity and then ignored it, I find difficult to accept. I suspect that this is one of the passages in the evidence where she simply dug her heels in and lost sight of objectivity. In any event, I do not accept that her stated approach would have been the approach of the notional skilled addressee.

What was also apparent from reading the papers in the art was that by 1984 it was known that B. pertussis adenylate cyclase had been reported as having a variety of molecular weights and that its extraction from the bacterium could well result in partial proteolytic digestion, thereby producing active adenylate cyclase proteins but of different sizes. This additional material would have reinforced the view which would have been held by a man in the art in any event that molecular weight was not an exclusive marker. These conclusions are supported by many parts of the evidence given before me including the following:

Per Professor Higgins:

"If you were told the protein had enzymatic activity, then I have no doubt that that is what you would follow. The crucial defining characteristic of the protein you are studying is the enzymatic activity. So that would be your first and easiest, certainly the easiest screen, and, in my view, it would be the first screen as well."^[22]

"Given that there are so many proteins of around these sorts of molecular weights in all bacterial cells, it just indicates that molecular weight itself is not a single defining characteristic. It is one helpful characteristic certainly, but not a single-defining characteristic."^[23]

Per Dr. Lowe:

"Q. Do you agree that the SDS-PAGE molecular weight of a protein is a very important characterizing feature of the protein?

A. It is a characterizing feature. It is not very precise. As you know, with a 12% gel you may be plus or minus 5% on that particular measurement of molecular weight. Then, of course, there are many other proteins in that molecular weight range. The only defining characteristic of this antigen, as from the priority document, is that it has adenylate cyclase activity. That is a clearly defined enzyme with an enzyme commission number. You know exactly where you stand with that."^[24]

Furthermore I was not persuaded by Evans' suggestion that the worker would ignore adenylate cyclase activity because it was susceptible to degradation. Of course if there was no adenylate cyclase to be found, then it might be that the protein was still present but its active site was damaged, removed or blocked. This would mean that it would not exhibit enzymatic activity. But even so, it was apparent on the evidence that if the skilled addressee followed through the procedures set out in the priority document to find the target protein, he would at the end purify it and using it to make monoclonal antibodies. Those would then be back checked against B. pertussis OMP containing adenylate cyclase activity to ensure that the monoclonal attached to the enzyme. This would be demonstrated either by it specifically blocking the enzyme activity (e.g. if the antibody attaches itself near the enzyme's active site) or by precipitating it out of solution. In either case certain confirmation that the purified protein is or is not adenylate cyclase would be provided.

Furthermore, even if the absence of adenylate cyclase activity would not cause the skilled addressee to stop, that does not mean that when it is present he would ignore it. If adenylate cyclase activity is detected he would be likely to assume that he is tracking the target protein because that is what the priority document tells him. If this is correct then if he gets to the stage illustrated in Fig 2 above (i.e. assuming he is not put off by the fact that there is nothing in the 0.2 M eluate), he is likely to speculate that what he wants should be in fractions 23 and 24, but the fact that this has the wrong pI and contains no protein of the right apparent molecular weight would, at the very least, concern him. In these circumstances, if he picks any route forward he is as likely as not to try these fractions. Whatever he gets out of them, if anything, will not be pertactin.

The matter does not end there however. Even if one assumes that the addressee ignores adenylate cyclase activity, would he be led directly or confidently to pertactin? On the evidence, the answer must be "no". There are two issues arising in relation to this. The first is that, even ignoring enzymatic activity, the teaching in the priority document is so confusing that it is impossible to determine what should be looked for. As noted above, Evans says that the addressee would concentrate on the molecular weight data virtually exclusively. I do not accept that. The point was put clearly by Dr. Lowe:

"Q. ---- would you not say that that is the obvious candidate for the protein that you were looking for?

A. No, I do not think I would at all. I agree it is a candidate, but then, you know, I would ask myself, why should I take any notice of the molecular weight? All the other information is incorrect. Why should you suddenly switch into another gear and say, "This is clearly the figure"? I do not think you would at all in those circumstances. We know that the pI is wrong. We know that the activity is incorrect. It comes off the wrong place on the DEAE column. Why should I suddenly say that they have all that wrong but they may have the molecular weight correct? I would also argue that if, indeed, the molecular weight was the critical parameter, as you are suggesting, I would suggest, first, with the SDS-PAGE, you would probably lose a lower percentage so that you expand that region a little bit more. That is the first thing I would suggest. Secondly, the purification protocol would include a gel filtration step that is based on molecular size because you know 69k is an important area. Why does it not include that? This whole purification protocols means to me charge, charge, charge. That is because of its ion-exchange chromatography, and there are three isoelectricfocusing steps. All it is focusing on is the charge. I would suggest to you that what you would actually do is you would look at the pI value as the critical parameter. There are three examples in the patent of pI. You know, it is no coincidence that is in there. It focuses the attention on it."^[25]

Although at the end of this passage, Dr. Lowe is talking about the patent specification, which contains very similar information, the sentiment that it was impossible to know what was wrong and what was right applies at least as forcefully to the priority document. That was a view which came out clearly in the evidence of both Dr. Lowe and Professor Higgins. In my view it is correct.

The second point concerns the use of monoclonal antibodies. The priority document says that the target protein can be purified by using "a monoclonal immunoglobulin [i.e. antibody] specific for the 69Kd protein". It was not in dispute that if such a monoclonal antibody was available or could be made, then pertactin could be found although this would still involve ignoring all the other examples in the priority document. It is also apparent that Dr. Novotny had such an antibody. It was given the code name BB05. There are no instructions in the priority document as to how to make this tool. Evans' suggested that what a man in the art would do, faced with the problems of following the priority document, would be to pursue the 69 kDa protein, use it to make an anti-69 kDa monoclonal antibody and then use that to purify out pertactin. Chiron says this would not work. To understand why it is important to distinguish between what Dr. Novotny did and what an addressee would do.

Dr. Novotny developed material produced in the course of Juan Montaraz's PhD work. Montaraz happened to make monoclonal antibodies which bound to an unknown protein to be found in both B. pertussis and B. bronchiseptica. These antibodies were identified as BB05 and BB07. Once these were available it would be possible to find much more of the antigen to which they bind. This antigen happened to be pertactin. To go to the Wembley Stadium example I gave earlier in this judgement, Dr. Novotny and Juan Montaraz fingerprinted a target pulled more or less randomly out of a pool of proteins. Once they had the fingerprint, they could find the same protein again. In fact the monoclonal antibodies react with pertactin from both B. pertussis and B. bronchiseptica. Thus by making either BB05 or BB07 from B. bronchiseptica they could then use it as a tool for finding pertactin in B. pertussis. Furthermore, the pertactin from the two sources is different in apparent molecular weight (although both include the same epitope to which BB05 and BB07 bind). The version from B. bronchiseptica is 68 kDa. If a worker in this field has this information, it is possible to test whether any monoclonal antibody he has reacts both with a 69 kDa protein from B. pertussis and a 68 kDa protein from B. bronchiseptica. If it does, it is extremely likely to be a monoclonal antibody to the same protein that Dr. Novotny was working with, i.e. pertactin. It may not be specific to the same epitopes bound by BB05 and BB07 but it is binding to an epitope on the same two proteins. He can therefore be fairly confident that he has a monoclonal antibody to the right protein. However the priority document does not refer to BB05 or BB07, the fact that they were derived from B. bronchiseptica or that they are specific to two similar but different proteins of identified apparent molecular weight derived from the two related bacteria. This information is to be found however in a paper published under the names of Montaraz, Novotny and Ivanyi after the priority date. This document is pleaded as prior art to the patent if Evans cannot hold on to its earlier priority date.

Evans' case is that, even though the priority document does not tell you how to select the correct monoclonal antibody which is specific to pertactin and that neither BB05 nor BB07 were available from any public source, it would be possible to make a suitable monoclonal antibody and thereby obtain impure enough pertactin to carry out the Kendrick test and confirm that you had the correct antigenic protein referred to in the priority document. All that was necessary was to go through the examples in the document and find the 69 kDa protein on an SDS-PAGE, make a monoclonal antibody from that and use it to immunopurify B. pertussis extracts but not so thoroughly as to leave essentially pure protein. The slightly impure protein would give a positive Kendrick test. That is pertactin.

I do not accept that anyone would go down this path. On the evidence, even if an addressee went down the route of picking out a non-adenylate cyclase 69 kDa protein as a possible candidate, he would conclude that he could not be certain that the protein he selected was the right one. Without a pre-existing defining feature or set of features which together are defining, finding a monoclonal antibody is of no value. Here the only pre-existent defining feature was that the protein was adenylate cyclase. Finding a monoclonal to pertactin would fail to satisfy that feature. On the other hand if the combination of features set out in the priority document were adequate to define the right protein, again a monoclonal to pertactin would fail to isolate a protein which met that combination of features. This was explained by Professor Higgins:

"A. ---- without knowing anything about bronchiseptica. I could raise antibodies, I would suspect, against a protein of 69 kD. I suspect I could do that. The question would be: is it the same protein as Montaraz has antibodies against? And that is a very difficult question to answer. I do not know how you know that unless you look at other characteristics of the protein other than molecular weight. I think in this paper we are not given any other characteristics. There will be multiple proteins in pertussis as in bronchiseptica of around that sort of molecular weight size. So you have to know you have an antibody against the same one that Montaraz had and that is the problem.

Q. One approach, which of course may depend on how many multiple ones you pull out, is that you could look for the one with the same molecular weight and you could then immunopurify them if you had more than one and test them for their protective effect in animals.

A. You could test .... What you are suggesting is that we screen purely on the basis of molecular weights. We find antibodies cross-reacting with a protein of around 69 kD, grow up those, and purify them. We find the ones that work in immunopurification, which in itself is not trivial (you might hit lucky but you might not), purify a lot of the protein and then use it in animal tests. Is that what you are suggesting?

Q. Yes.

A. That would be a lot of work. You could probably succeed but you still would not .... Even if it was protective, even if it was, you would not know it was the same protein that Montaraz and people were referring to without other characterizing features. Molecular weight is not a distinguishing feature of a protein. It is one characteristic, but it does not define a protein.

Q. I follow. So you might find the useful antigen if it works in the antibodies ----

A. You might find a useful antigen.

Q. ---- but you would not know whether or not it was the same one? That is your ----

A. For example, the 70 kD fragment of adenylate cyclase I believe is active in the Kendrick test. So that is one that I might obtain antibodies against. Used in the Kendrick test, it would show positive results, but it is not the protein that has been referred to here and I have no doubt there will be plenty of others. As I say, molecular weight is not a defining feature, it is just one characterizing feature."^[26]

and:

"A. But if you will forgive me, what that is referring to is the use of this antigen in vaccine formulations. It is not referring to the use of the Kendrick test to test whether whatever antigen you pull out with your antibody is the right antigen. This is that is a very different issue. We are using the Kendrick test in your hypothesis to determine whether the antibody we have raised by the circuitous route is against the right protein, as spelt out in the patent. How do we know whether it is the right protein? We are told that it should be adenylate cyclase, and it is not. We are told that it should have a certain pI, and it is not. We are told that it should have a certain molecular weight, which, OK, we have gone along that route, you have told me it has gone that route, and we are told the purified thing is active in the Kendrick test, which it is not, so you have the wrong antibody. You cannot know you have the right antibody.

Q. You are told that when purified by the procedure set out in the document, it is protect[ive] in the Kendrick test?

A. Yes, and that gives you BB05. If you gave me BB05, maybe, but I do not have BB05, so we are talking about .... You are talking about how you would get the antibody, if you did not have BB05. If you make an antibody against a 69k protein, and it reacts with a 69 kD protein, you have to know that that is the same 69k protein as defined in the patent. How do you [confirm] that? You check the pI and it is wrong. You check whether it precipitates adenylate cyclase activity and it does not. You check whether when you immunopurify it using that the antigen it immunopurifies is active in the Kendrick test, which it will not be if it is against P69, and, therefore, you will discard it."^[27]

and

"A. ... Tell me this. If I have raised a monoclonal which cross-reacts with a 69k protein, how am I going to know that monoclonal reacts with the same protein that BB05 interacts with? OK, it is the same molecular weight but it is the wrong pI, it comes off in the wrong place on the DEAE column, it does not have adenylate cyclase activity and it does not work in the Kendrick test. It is the wrong thing."^[28]

I am not persuaded that any notional addressee at the relevant time would have tried to make monoclonal antibodies. He would not have known what to make such antibodies to. To have engaged on such a potentially valueless course would not have been the natural way forward. The disincentives to going down this path would only have been increased by the very great difficulty which surrounded the making of monoclonal antibodies at the relevant time. The evidence before me was that to have tried to develop a monoclonal antibody from scratch at that time might well have taken a year's work. It is most unlikely that anyone trying to follow the priority document would have embarked on such a project when it was apparent from the outset that it was likely to produce results of little value.

In fact Professor Shaw accepted^[29] that a skilled addressee might go down the route of making a monoclonal to adenylate cyclase. That, of course, would have taken him away from pertactin. The whole of Professor Shaw's evidence on this subject emphasises that no worker could be certain which course, if any, would be the right one to follow. It also follows that the addressee would not know, if he followed one of these courses, that he had chosen the right one so that what he had ended up with, if anything, was what Dr. Novotny was targetting.

In the end I think Professor Shaw came to much the same conclusion as to the inability of a man in the art to follow with confidence the teaching of the priority document. He expressed it in less direct language but the meaning was the same. He said that some of the teaching was a "puzzle" and that the reader would be "furious" that he had not been supplied with the appropriate monoclonal antibody. He also said:

"Q. Be that as it may, if you do run them all out, I think you are agreeing with me that there is nothing which satisfies the description in the priority document of the appropriate pI, molecular weight and activity.

A. There is no single fraction that satisfies all of those. That is right.

Q. What I put to you is that if you were performing this priority document, the one thing that you would be looking for is residual activity.

A. At this stage, if I had gotten this far with the protocol and seen the results we have in front of us, putting aside the monoclonal data, and forget that for the moment, I would have stopped thinking about enzyme activity at this stage because I already knew, when I began this latest episode, that I had practically no measurable activity there. I had slogged through this, and now I have a bunch of fractions that have something of that 69k. I would say, "Thank God for small favours."

Q. But you have no idea if that 68 to 73, whatever you like to call it there precisely, that molecular weight .... You have no idea what it is. It is not associated with activity, nor is it at the right pI.

A. That is correct."^[30] (emphasis added)

and:

"A. My answer to that is precisely the same answer I gave in my report. They would not find it where they had expected to find it. They find a protein of the right description, but not the activity in the 0.1 molar salt, nor indeed would they find the activity in the flow-through. But they would find a protein of the requisite molecular weight in the flow- through and in the 0.1 molar salt.

Q. But not with activity?

A. But not with activity.

Q. I put to you that, because you are told here to look for adenylate cyclase or an adenylate cyclase-bound protein with activity, you would be concerned if you had lost the activity?

A. Yes, I would be concerned, but I would not give up."^[31] (emphasis added)

and, perhaps most tellingly:

"A. Well, to use your own words, the person reading this says, "It is adenylate cyclase or a binding protein, and the patent's teachings appear to have failed on two points." My own eyes tell me that, in one respect, there is something there. I am looking after an antigen, which is going to be protective. I do not believe the patent tells me that there must be enzyme activity in the final product. I am not on firm ground. I might be on soft sand, but I have not sunk at this point. I think that I would not go back to the beginning at that stage simply having gotten this far and having fractions, at least three fractions, of having material which could be easily purified. If I wanted to use that material, I would not put [pull?] it at this stage.

Q. You know, do you not that, adenylate cyclase comes in different molecular weights?

A. I do.

Q. We have been through the exercise of looking at the variability of the molecular weight of adenylate cyclase. Yes?

A. Yes.

Q. Looking at this document in bundle M, the fact that you do not see something of the precise molecular weight is likely to be the least of your worries.

A. Sorry. I found something of the correct molecular weight.

Q. But not where it tells you it will be.

A. I have it that abundantly clear that the average skilled protein scientist would do the best he could under the circumstances and be delighted to find a protein of the right molecular weight in a different fraction. Having gone this far, he would not scrub the whole thing but say, "Let us go on and finish it off and see what we get. Let us immunize." I will not go into the details of what could be done. At this stage, one would not scrub the whole procedure, I think."^[32] (emphasis added).

The fact that a determined scientist with no obvious way forward might soldier on in the hope that something might turn up does not qualify this as an enabling disclosure. On the contrary, the evidence demonstrates that a skilled addressee would not know how to find the target protein if it is not adenylate cyclase. Depending on his obstinacy and willingness to follow up inaccurate clues he might find pertactin after months of trial and experiment. He would never be certain that it was the protein referred to in the priority document. He is more likely not to find pertactin. In any event this document gets nowhere near leading a skilled person necessarily and directly to pertactin. I think that most addressees would give up at an early stage and, like Professor Higgins, view this as bad science.

In the circumstances, Evans has failed to make out its claim to priority. It is not necessary for me to deal in any further detail with the arguments raised in relation to this topic. I will only add that Professor Shaw criticised IEF in his report as being in some way an unreliable test which no one would rely. Under cross examination Professor Shaw's objections disappeared. The accuracy and ease of use of IEF is described in two documents at Tab 22 of Volume V of the papers in this action. In particular the document authored by Dr. Righetti claims that IEF requires only a limited number of chemicals, is completed within a few hours, is less sensitive than most other techniques to the skill (or lack of it) of the operator and allows excellent resolution of proteins, with the protein bands being very sharply defined due to the focusing effect. Professor Shaw could not deny Righetti's assessment. This was but one of the areas which did not reflect well on Professor Shaw's objectivity in giving evidence. Furthermore, even if there was something in Professor Shaw's written criticism of IEF, it would not have helped the patentee here. Dr. Novotny used that test to define one of the characteristics of his protein. If the test gives unreliable results, it should not have been put in the priority document. The patentee has to bear the consequences of having put before the reader technical information which can confuse rather than clarify.

The consequences of loss of priority - intervening publication

If, as I have found, priority is lost, a number of intervening publications come into play as prior art on which an attack of lack of novelty or obviousness is based. As I understand it, by the end of the trial Chiron only relied on two groups of such publications namely (i) an article by Montaraz et al published in March 1985, Infection and Immunity, 1985, vol 47, pp 744-751^[33] ("the Montaraz March paper") and (ii) various materials disclosed by Dr Novotny in and for the purpose of a talk given by him at the Fourth International Symposium on Pertussis, held in Geneva in September 1984. Although Chiron relied on both of these, it did not suggest that the second added anything of significance to the first and the arguments centered on the Montaraz March paper.

Intervening publication - the relevant law:

The Montaraz March paper was written by Juan Montaraz, Dr. Novotny and Professor Ivanyi. The Defendants have pleaded section 6 of the 1977 Act as excluding the publications from being part of the state of the art even if is lost. This argument was considered in detail by Jacob J in Beloit v. Valmet [1995] RPC 705, and rejected. The point was not considered on appeal. Evans accepts that on this interpretation, section 6 will not help it. It does not feel it appropriate to revisit the issue before me. However it has indicated that it reserves the right to argue against the construction of section 6 reached in Beloit if this case goes further. It follows that this is an issue with which I am not concerned. Furthermore, in relation to these publications it is not suggested by Mr. Howe that the complications arising out of the decision in the House of Lords in Merrell Dow v. Norton [1996] RPC 76, which are considered below, have any bearing.

Evans argues that the disclosure in the Montaraz March paper is not significantly different to that in the priority document. Therefore, if the priority document is not enabling, it must follow that it cannot render the patent invalid either for anticipation or obviousness. Chiron disputes this both on the facts and as a matter of law. It says that the disclosure in the Montaraz March paper is crucially more informative than the priority document and, in any event, the test for obviousness is not the same as the test for enablement. On the latter point I accept Chiron's submission. For a document to be enabling, it must direct the reader to the invention clearly. Obvious errors may be corrected but that does not mean that this is another version of the test of obviousness. Take a case where a document does not describe or cover a particular technical step, where the author clearly had never thought of it, where carrying out the instructions in the document does not achieve it and where it would be regarded as a further development of what is disclosed. Such a disclosure, if embodied in a priority document, is unlikely to give enabling support to a patent claim directed to the step. However if the disclosure is incorporated in a piece of prior art it may render the step obvious. For example a reader might think that there is sufficient prospect of success to warrant trying it out. The tests of enablement and obviousness differ.

Intervening publication - the facts

There are a number of respects in which the disclosure in the Montaraz March paper differs from that in the priority document. I accept Chiron's argument that these differences are technically significant. Montaraz describes, accurately, the work which had been done in relation to B. bronchiseptica. It is not suggested that the disclosures in this document could not be followed by a skilled worker. Montaraz identifies the existence of a 68 kDA OMP in that organism which appears to have the ability to immunize mice against further challenge with B. bronchiseptica. It is therefore a protective antigen. The paper raises the possibility that this protein might be a possible candidate for a vaccine against atrophic rhinitis in pigs. It describes how to make BB05 and BB07 and how to use them to purify the target protein. Furthermore it discloses the "binding profile" of both monoclonal antibodies. This shows that they react with a similar protein in B. pertussis extracts. They show such cross binding not only with a B. pertussis but with a protein of similar weight also to be found in another close relative, B. parapertussis. It does not make the mistake of calling the target protein adenylate cyclase. It also shows the use of acid glycine hydrolysis as a means of producing the extracts, just like example 1 of the patent does, and that these extracts may be passed through an anionic exchange column, just like example 2(a) of the patent, before they are passed through the affinity chromatography column containing the monoclonal antibody.

On the basis of this, Chiron says that it is obvious that the equivalent protein in B. pertussis is likely to be a candidate for vaccine preparation against the disease caused by that bacterium, namely whooping cough. Further, the cross binding in relation to the three types of Bordetella bacteria would provide the means by which a worker could be sure that he had found the correct protein, pertactin. Since Dr. Guiso and Professor Shaw had expressed the view that the priority document with all its errors would have led a skilled worker to pertactin, they were hardly in a position to deny that the Montaraz March paper, which contained no such errors but did contain other important technical information would lead a worker there. In any event under cross examination Dr. Guiso accepted that in the light of this publication it would be obvious how to proceed to make and identify a particular monoclonal specific to the protective antigen now called pertactin^[34]. Once that was available there is no dispute but that a worker could extract pertactin from B. pertussis. The disclosure in the Montaraz March paper would make it obvious that this antigen would be likely to be protective in relation to B. pertussis.

Very similar evidence was given by Professor Shaw^[35] and by Professor Higgins both in his report^[36] and under cross examination^[37]. Dr. Lowe was also cross examined on this although it was not a subject he had considered in his report. His evidence is somewhat confused^[38] not least because he did not have a close familiarity with the Montaraz March paper and appeared to have some difficulty in understanding what was being put to him. I did not find his evidence on this topic to be of any value. The argument that the disclosures in the Montaraz March paper would enable a worker in the field to make the correct monoclonal antibody to pertactin from B. pertussis appears to be confirmed by work subsequently done by Brennan^[39]. The correct monoclonal was found by screening it against the pattern of binding to the different proteins from B. pertussis, B. bronchiseptica and B. parapertussis^[40]. This is the approach Professor Higgins says it would have been reasonable for a worker to have adopted on the basis of the disclosure in the Montaraz March paper. He was not challenged on this issue.

I have come to the conclusion that the Montaraz March paper does not anticipate the patent in suit. All of its teaching is directed primarily at B. bronchiseptica. However it does show that the protective antigen in that bacterium does have a close cousin in B. pertussis. In my view it renders obvious taking the equivalent steps including acid glycine hydrolysis to produce, purify and use that antigen, i.e. pertactin, from B. pertussis . It follows that all the claims for which independent validity is asserted are invalid.

In the light of this finding, it is not strictly necessary to consider some of the other attacks on validity or the arguments, such as they were, on non-infringement. However I had the benefit of a full examination of these issues during the trial. They raise issues of fact which are not entirely straightforward. Since it is to be assumed that this case will go further, I think it is right that I should deal, as briefly as the technology permits, with these issues.

Insufficiency

Chiron attacks the patent on the ground that it does not disclose the invention clearly and completely enough for it to be performed by a person skilled in the art (s. 72(1)(c)). In large part this is based on the same arguments which succeeded in relation to enablement and the priority document. The examples in the patent specification include those in the priority document, described in essentially the same way, together with two examples of analytical IEF^[41]. Notwithstanding these similarities, there are some differences between the priority document and the patent specification which are argued to affect the outcome of this issue. It is not in dispute that as a result of the decision of the House of Lords in Biogen v. Medeva [1997] RPC 1 at 48, the specification must enable the invention to be performed to the full extent of the monopoly claimed.

At page 3 line 52^[42], the patent says that the antigen, i.e. pertactin, can be obtained in high purity. It can therefore be argued that, to the extent that pertactin is not per se adenylate cyclase, a skilled addressee would not be too concerned by the absence of adenylate cyclase activity if he tries to carry out the examples in the specification. Therefore this misleading signpost which was very prominent in the priority document is missing. Evans says that this avoids a finding of insufficiency.

In my view this argument fails. Whatever the characteristics of the antigen in Claim 1 may be, the patent specification is full of references to the presence of adenylate cyclase activity. Indeed, unlike the priority document, it appears to suggest that such activity is tracked in the preparation of the target antigen. Adenylate cyclase activity is said to be associated with the antigen if acid glycine hydrolysis is used and it such hydrolysis which is the starting point for the examples in the specification. The fact that adenylate cyclase activity is a useful flag which identifies that the right fraction is being retained is emphasised in a number of passages. Some of these have been set out at the beginning of this judgment, including the following:

"Control vaccines containing no adenylate cyclase activity were found to provide little or no protection against challenge with B. pertussis, suggesting that ACAP may, in fact, be the most important factor in immunity. Analysis of batches of non-protective whole-cell vaccine has also shown that non-protection tends to be associated with a lack of adenylate cyclase activity, further suggesting that ACAP may be the key antigen necessary for eliciting an immune response against B. pertussis."^[43]

Analysis for adenylate cyclase activity is also referred to specifically in examples 2(b) and (c). It follows that the specification is no less misleading and inaccurate than the priority document in describing how to obtain pertactin. Mr. Kitchin goes further than this. He says that the additional information in the specification compounds the confusion. This is for two reasons.

First, the priority document says that the target protein has a pI of 7.2 to 7.6 measured under preparative conditions. Example 2(b) does not qualify or add to that assertion. But the patent specification refines and modifies the equivalent data. In the general description it says that the target protein as detected by preparative IEF:

"generally has ... an isoelectric point of 7.0 - 7.4 under preparative conditions"^[44]

and

"In more detail ... as two bands, one having an isoelectric point (pI) of, about 7.0, the other (diffuse) band having an isoelectric point of 7.2-7.4. Adenylate cyclase activity was associated almost entirely with the neutral band (pI-7.0) but monoclonal antibodies to ACAP bound both bands strongly."^[45]

These general statements are based on example 2(b) which is the same example as recorded in the priority document but somehow now is recorded as having produced slightly different results. In particular it states that the target protein:

"... was detectable as two bands, one of pI 7,0, and the other (diffuse) band of pI 7.2-7.4. Adenylate cyclase activity was associated almost entirely with the central band (pI 7.0) but monoclonal antibodies to [the target] bound both bands strongly."

Examination of Figs. 2 and 3 above demonstrate that this more precise data would again lead the worker away from the fractions containing pertactin.

Secondly, the patent now contains two more examples, 2(c)(i) and (ii) which are analytical IEF procedures. It teaches that the results from example 2(c)(i) were the same as for example 2(b). For 2(c)(ii) it specifies the pI's at which the target protein was to be found and states where the associated adenylate cyclase activity was observed. Chiron says that if example 2(b) had not worked, as it did not, then a worker would not have bothered to go on to perform either part of 2(c). Even if that submission is incorrect, repetition of example 2(c)(i) was not the same as for 2(b) as promised or even similar and pertactin was not focused in the place where the patent says its should have been. The repetition of example 2(c)(ii) gave results which were visibly quite different from what the patent promised, and again pertactin was not where it was said that it would be.

I accept these additional criticisms of the patent. Insofar as the invention is directed to pertactin, it is insufficient.

Insufficiency - a further point

There appears to be another argument of insufficiency which was not advanced by Chiron although it falls within the scope of its pleadings. It seems to me to arise inevitably out of the material put before the court. If so, the court should not turn a blind eye to it.

In accordance with the Biogen decision, if the claims include a number of discrete methods or products, the patentee must enable the invention to be performed in respect of each of them. Chiron has pleaded that he has failed to do this^[46]. It is necessary to look again at what the claims cover. To do this, the claims must be construed in the light of the rest of the specification.

Claim 1 is for a purified B pertussis antigen which is extracted together with adenylate cyclase activity when extraction of the activity is performed using acid glycine hydrolysis and which has certain other characteristics. This claim reads back directly to the part of the specification which describes the inventor's claimed technical breakthrough. This passage has been set out at the beginning of this judgment but, to simplify understanding of the point under consideration, it is repeated here:

"It has now been discovered that certain proteinaceous material, associated with adenylate cyclase activity, as hereinafter described, found in the cultures of B. pertussis , is capable of providing protection against challenge by B. pertussis when administered to experimental animals. This discovery that the proteinaceous material usually associated with adenylate cyclase activity is a major protective antigen against B. pertussis permits the preparation of vaccine formulations comprising antigenic preparations which are free from, or contain reduced amounts of, other known B. pertussis components which may be responsible for the toxic side-effects demonstrated by whole cell vaccines.

The term 'proteinaceous material associated with adenylate cyclase activity (abbreviated to 'ACAP' hereinafter) is used herein to refer to proteinaceous material which is extracted together with adenylate cyclase activity when extraction of the adenylate cyclase activity is performed using an aqueous, acidic (pH3) solution of glycine (0.25 M). The ACAP as defined above may comprise the adenylate cyclase enzyme per se or a binding protein for the enzyme." (emphasis added).

The italicised sentence made sense when it was in a priority document directed to the purification and use of adenylate cyclase but it makes no sense if the patent is directed not at adenylate cyclase at all but at a different protein, pertactin, only. However the patent is full of references to adenylate cyclase. Not only does it say that the ACAP can be pure adenylate cyclase but it also repeats from the priority document that what is needed is a method of obtaining higher yields of adenylate cyclase than was obtained in the prior art and that the yield of adenylate cyclase should be at high purity^[47], that the new method of extraction yields such higher yields of adenylate cyclase^[48], that vaccines without adenylate cyclase activity do not seem to work^[49], that such activity was to be found in the correct samples as demonstrated by the examples and that it was immunopurified adenylate cyclase which gave the best results in the Kendrick test^[50]. If the claims do not encompass adenylate cyclase then the specification is, at the least, grossly misleading. However there is nothing in the claims which, on their face, exclude adenylate cyclase. On the contrary the claims appear to encompass adenylate cyclase and, although it may not be strictly relevant, this was the understanding of all the major technical witnesses; Professor Higgins^[51], Dr. Lowe^[52], Professor Shaw^[53] and Dr. Guiso^[54].

If it was not the intention of the patentee to cover the purification and use of adenylate cyclase he should have excluded all the material in the specification which said that that enzyme was at the heart of the invention. One may speculate that the failure to take that course was so that the patentee could maintain the position that this patent was for the same invention as described in the priority document. Whatever the explanation, it appears to me that the claim must be read as covering, inter alia, adenylate cyclase - that is to say it covers both adenylate cyclase and pertactin. However it is not in doubt that there is no known adenylate cyclase which meets one of the other characterising features of Claim 1, namely a ratio of proline to glutamic acid of substantially 1:1. No such protein can be extracted from B. pertussis nor does the patent suggest that any such protein could be made in any other way. A claim which covers a protein which does not exist and where a man in the art would not know, and there is no teaching of, how to make it is also insufficient.

This particular problem might be curable by an amendment excluding adenylate cyclase from the claims. However that would highlight the difference between inventions described respectively in the priority document and the patent. That is a difference which the language in the patent appears designed to obscure.

Anticipation and obviousness if early priority is retained

Whether or not the patent is entitled to its earlier priority date, Chiron says that it is invalid for anticipation and obviousness. This provoked a sharp difference between the parties as to the scope and effect of Merrell Dow v. Norton [1996] RPC 76 in the House of Lords. Before turning to that, I will deal with the factual issues. Chiron's case is put in a number of ways. First it is said that it would have been obvious to a person working in the whooping cough vaccine field to make outer membrane protein extracts from B. pertussis as potential sources of protective antigens. Acid glycine hydrolysis and use of DEAE were standard and obvious ways of making and purification. Such extracts would in fact contain pertactin and would fall within the scope of claim 1 of the patent in suit. This renders the other claims obvious. Secondly it is said that the claims are anticipated or rendered obvious by the use in Japan of a whooping cough vaccine made by Takeda Yakuhin Kogyo KK ("Takeda"). Thirdly it is said that the claims are anticipated or rendered obvious by the publication of Takeda's UK patent application 2,083,358^[55]. It was not suggested that that in any of these cases at the relevant date a skilled worker would know that his protein extract contained the specific protein now called pertactin. It is this latter point which gives rise to the Merrell Dow arguments.

(1) OMP's and Acid Glycine Hydrolysis generally

Chiron's argument proceeds as follows. At the priority date, if an alternative to a whole cell vaccine was wanted, as was the case by 1984, the obvious place to start was to try to make a vaccine from the B. pertussis outer membrane proteins, i.e. the proteins which would be seen and responded to first by the host's immune system. This was not really disputed. Professor Ivanyi volunteered that there was a working hypothesis, that a protective antigen is going to be an outer membrane protein^[56]. Dr. Guiso thought that the majority of interest was on two particular proteins, PT and FHA, but she accepted that some laboratories were interested in outer membrane proteins generally^[57] and she accepted that OMP's were thought of as good vaccine candidates^[58]. Dr. Chubb, a witness called by Evans, agreed that the outer membrane protein seemed the most obvious part to be the active component in a vaccine^[59]. Similar evidence was given by Professor Higgins. Chiron also called as one of its witnesses a Dr. Redhead, who at the relevant time was a microbiologist employed by the National Institute for Biological Standards and Control in the Division of Bacteriology, an organisation set up in the 1970's to control the quality, efficacy and safety of all materials used in human medicine in the UK. He says^[60] that before 1984 most research focused on a few specific antigens, including those known to be present in whole cell pertussis vaccines (PT, FHA, agglutinogens and adenylate cyclase), together with a general search for unknown cell surface located antigens, in particular, outer membrane proteins. He explains that the interest in antigens at that location was based on the appreciation that they were likely to be involved in the early stages of host colonisation. The cell surface antigens would be the first pertussis antigens to be seen by the host immune system. He says that it seemed reasonable in those circumstances to investigate the possibility that targeting such antigens for inclusion in a vaccine may provide a means to block pertussis infection at a very early stage. Even the patent confirms that work in the field had concentrated on isolating and purifying the surface antigens of B. pertussis and characterising their ability to induce immune responses^[61]

Acid glycine hydrolysis was a known method of removing all the cell surface antigens. That this was an obvious method of preparation to use was also not really disputed. In any event it was confirmed to be the one of the methods which would have been known to all in the art at the relevant date by Dr. Guiso^[62]. Dr. Guiso also agreed that those in the art in the time would have wanted to remove or reduce the amount of LPS in such an OMP extract because LPS was reputed to be associated with unpleasant side effects. Among the well known way of doing this in 1984 was the use of a DEAE column. This evidence was consistent with the evidence of Professor Higgins and Dr. Redhead.

On the basis of this it was obvious at the priority date to make an OMP extract from B. pertussis and to "purify" it by removal of some or all LPS. Such an extract inevitably would contain pertactin and would fall within claim 1. However it was not shown or suggested that any worker would have analysed the content of such an extract so as to elucidate its pertactin content. On the evidence, he would have known or suspected that his extract contained many proteins and protein fragments. He would hope or expect it to raise an antibody response. He would not know that it contained the particular 69 kDa protein now called pertactin.

The Takeda vaccine

This is a vaccine which had been placed on the market in 1981 and used to vaccinate very large numbers of patients. It is still on the market. In its current form it contains in the region of 5% pertactin. Pertactin is highly antigenic. 5% of it in the Takeda vaccine represents a significant fraction. Evans says that this vaccine infringes the patent in suit. Chiron agrees. However Evans challenges the assertion that the current version of the vaccine is the same as the pre-May 1984 version in regard to its pertactin content and, in any event, it says that it is irrelevant because no one knew that it contained pertactin.

In relation to the issue of the precise composition of the pre-May 1984 vaccine, a large quantity of evidence was put before the court. No purpose would be served by going through it all. The Takeda vaccine, because of its method of manufacture, almost inevitably contained pertactin. In fact Dr. Guiso agreed that probably the Takeda vaccine had pertactin in it^[63]. Chiron argues that the pertactin content of the vaccine before 1984 was substantially the same as it is now. Evans suggests the opposite. It is probable that the pertactin content has varied from time to time. However, on the evidence, it appears that the pertactin content of the early vaccines did not differ vastly from the content now and I so find. The vaccine sold and used from 1981 to 1984 would have fallen within the patent claims just as the more recent batches do. Once again, it is not in dispute that no one at the time knew the detailed contents of the Takeda vaccine. They did not know before May 1984 that it contained pertactin.

The finished product does not disclose how it was made. No one in the art is taught how to imitate it. If this product invalidates the patent in suit it is on the basis of anticipation only. That issue is dependent on the impact of Merrell Dow.

The Takeda patent application

This was published in March 1982. Essentially it describes a method of making a B. pertussis product for development into a whooping cough vaccine which consists of taking the culture supernatant of a B. pertussis strain, concentrating it and then removing endotoxin (i.e. LPS) from it. It will be recalled that pertactin is to be found both in the OMPs and as an excreted protein in the supernatant over growing B. pertussis cultures. Chiron's case is that this application both anticipates and renders obvious the patent in suit. In support of the argument of anticipation it purported to repeat example 1 of the Takeda application to show that the resultant product contained a sizeable quantity of pertactin. There is no doubt that the experiment as originally run for the notice of experiments and as repeated at Evans' request produced an end product with a pertactin content similar to that in the current Takeda vaccine.

Evans says that the experiment is valueless because it fails to carry out with precision the detailed instructions in example 1 of the application. In particular it points to the fact that the strain of B. pertussis used in example 1 of the application is called the Tohama strain, whereas Chiron used a different strain, that the growth medium used in the experiments was different to that specified, that the sucrose gradient used for separating out the proteins in the supernatant was different and so on. There was no dispute that these detailed differences exist. However Mr. Kitchin pointed out that all of the variations were well within the general teaching of the application and that none would have a marked effect on the quantity of pertactin produced. Based on this he put his argument four ways. First he says that his client has, to all practical intents and purposes, carried out a repetition of example 1. It proves anticipation. Secondly he says that the experiment carried out is good evidence of what would have been achieved if all the minutiae of example 1 had been followed. In other words it is evidence upon which the court can decide, on a balance of probabilities, what carrying out example 1 would have produced. Thirdly he says that in any event the experiment was well within the general teaching of the patent and as such it proved that the general teaching anticipated. Finally he says that the changes made to the protocol of example 1 were trivial workshop modifications. He relies on Professor Higgins' assessment that the experiment appeared to him to be a bona fide and good attempt to carry out the teaching of the Takeda application. The Professor expressed the view that the experiments had been done to the best possible standard^[64] and were a very genuine best possible attempt effectively to put into practice the directions of the patent application^[65]. The Takeda application therefore renders the patent invalid for obviousness at least. In addition to these points, Mr. Kitchin argues that the general teaching of the application, would lead one in the art to make a vaccine which, in fact, would contain pertactin. So, even in the absence of the example, the document renders the patent invalid for obviousness.

There was much evidence on the significance or otherwise of the departures made by Chiron in their experiments from the precise protocol of example 1. To understand these and the argument based on them, it is necessary to have in mind the teaching of the Takeda application.

The application sets out its claimed contribution to the art in wide terms. It says that it has been found that a pertussis toxoid (i.e. a pertussis extract which has been rendered less toxic by chemical treatment) can be obtained by

"... removing endotoxin [i.e. LPS] from a culture supernatant or a concentrate thereof and flocculating pertussis exotoxin [i.e. PT] in the resultant fluid by permitting formaldehyde to act upon the fluid in the substantial absence of basic amino acid"^[66]

The process as described here consists of two steps. First the complete supernatant, with all its contents, is treated to remove a particular ingredient namely LPS. The remnant is then subject to a chemical treatment (with formaldehyde) to render the PT in it less poisonous. The applicant then goes on to describe how this is achieved. It says that the supernatant is from "a Bordetella pertussis phase 1 strain or a concentrate thereof". In other words any B. pertussis phase 1 strain will do. The application then provides that:

"The cultivation of the B. pertussis phase 1 strain can be carried out in a manner known per se. Thus, for example, the strain is cultivated in a liquid medium (Cohen-Wheeler medium, Stainer & Scholte medium, etc.) at about 35 to 37^oC for about 5 to 7 days."

What is apparent from this is that the precise details of the way in which the B. pertussis is grown is not critical. Any suitable known growth medium will do and two particular growth media are identified as convenient candidates. After describing the collection of the supernatant, the application goes on to describe how the LPS is removed:

"In accordance with this invention, the above-mentioned supernatant or concentrate is treated to remove its endotoxin. This removal of the endotoxin can be accomplished by any of such procedures as sucrose density gradient centrifugation, potassium tartrate density gradient centrifugation, cesium chloride density gradient centrifugation, gel filtration, etc. A particularly advantageous procedure comprises centrifuging the above-mentioned supernatant or concentrate on a sucrose density gradient of about 0 to 60 W/W % at R max. about 62,000 to 122,000G for about 10 to 24 hours."

This passage refers to a well known technique whereby molecules can be separated from each other on the basis of size using the technique of ultracentrifugation. The gravitational forces (or "G forces") generated by a centrifuge will tend to make suspended molecule sediment out. Larger molecules will sediment out at lower G forces, and therefore at lower rotational speeds, than smaller ones. A refinement of the basic centrifugation technique is the use of a density gradient to separate out the components in a solution. This involves the use of an increasingly dense solution of an inert, highly soluble material such as sucrose. The sample is layered as a narrow band on top of the sucrose gradient, which acts to stabilize the sedimenting components against mixture by convection. Under such conditions, on centrifugation the different components sediment at different rates, separating into different fractions which may be collected individually. The rate at which each component sediments depends on its size and shape and is normally expressed as its sedimentation coefficient or "s" value. Centrifuges can produce forces of up to 500,000 G thereby separating even relatively small macromolecules on the basis of their size. It will be appreciated therefore that the application gives general instructions to separate out the LPS on the basis of size and suggests density gradient centrifugation techniques using different inert gradient materials. One of the variety of recommended candidates is sucrose density gradient centrifugation. The range of 0 to 60 W/W % in fact covers the whole range of possible density gradients which are possible when sucrose is used. The output from density gradient centrifugation is split into fractions. The fractions differ from each other in the size of the molecules they contain.

After removal of the fractions containing LPS, the "most essential feature"^[67] is employed, namely the toxoiding of the LPS-deficient remnant with formaldehyde. The resultant product is called a pertussis toxoid. The application then goes on to give one example of the manufacture of such a toxoid. As is normal, the example is stated to be illustrative but not limitative of the invention. The example includes a detailed description of making a batch of pertussis toxoid. Because it is an example of what was done, it gives precise details of all the ingredients and the processing parameters used. In the example a 5 W/V% to 30 W/V% sucrose density gradient centrifugation is used. The fractions recovered from this are tested and those which contain significant quantities of LPS are discarded. This is in accordance with the general teaching of the application. However the example also includes another refinement. The inventor does not use all the LPS-deficient fractions. He selects only those which are said to be high HA reactive. HA is a reference to heamagglutination - the ability to make red blood cells stick to each other. This is an ability possessed by FHA and PT. The patent example therefore directs that high HA fractions which are also low in endotoxin are harvested.

It is this example which Chiron purported to repeat in their experiments. The results obtained on the repetition observed by Evans are displayed graphically as follows:


 


Fig. 4:

[Diagram or picture not reproduced in HTML version - see original .rtf file to view diagram or picture]
 


Chiron excluded from retention all those fractions having significant endotoxin content (fractions above 28). No criticism was levelled at this. From what was left it retained only those fractions which had significant HA activity (i.e. fractions 15 to 28). In the example in the Takeda application, fractions having HA activity of "not less than 20 titers per ml. preferably not less than 500 titers per ml." were retained. If one looks at Fig. 4, HA activity is marked as the graph made by joining together the plots indicated with square boxes. If very high HA activity is chosen as the criterion for retained fractions, the number of retained fractions will be reduced. An example of this is marked on the graph by means of dark turquoise lines. Chiron chose an HA cutoff point somewhere between those specified in the Takeda example. This was the subject of criticism by Evans. It says that higher levels of HA activity should have been chosen. This argument was based on another prior art article produced by the same inventor as in the Takeda application, a Mr. Sato. I will consider the Sato paper later.

In any event, these criticisms aside, it will be seen from Fig. 4 that if high HA, low endotoxin fractions are pooled, the resultant mix of proteins will contain significant quantities of pertactin. The total quantity of that antigen is represented by the red-shaded area under the pertactin graph line. Incidentally it will be seen that if all endotoxin-lean fractions are pooled, as is taught by the general disclosure in the patent, then fractions 1 to 28 will be mixed together. More pertactin will be in the resultant product. One other feature of this experiment is illustrated by the SDS-PAGE blot below the graphs. It emphasises that what are being retained are fractions containing a number of proteins. No attempt was made in the Takeda application to make pure single protein ingredients and then add them together. The Takeda application is describing the manufacture of a semi-defined vaccine. The SDS-PAGE blot carried out by Chiron confirms what would have been apparent to anyone in the art, namely that the Takeda product contains a number of different proteins.

This latter point may be of significance. During his opening submissions, Mr. Howe argued that the stated aim of the Takeda application was actually to collect specific antigens and try and get rid of everything else altogether, whether antigen or not. His argument was that Takeda promised that if you emptied out his vaccine "bag" only two things would fall out, namely PT and FHA^[68]. I have no doubt that this submission is wrong. The Takeda application is not directed at making a vaccine of pure PT and FHA. It is trying to get rid of was LPS, not to purify the remaining fractions. This is a matter I will return to in relation to the Merrell Dow point.

As far as Chiron's experiments are concerned, once again it appeared to me that Evans' experts adopted the position of raising all possible objections, whether or not they made any significant difference to the results. This stance can be illustrated by the objections taken to the choice of medium in which the B. pertussis strain was to be grown. In example 1 of the Takeda application Cohen-Wheeler liquid medium was used as the nutrient medium. In Chiron's repeat Stainer & Scholte medium was employed. As noted above, the general teaching of the application states explicitly that either of these named media, as well as others, can be used. Since Chiron had not used the precise medium specified in example 1, Evans says that the result does not prove anticipation. However there was no real suggestion that changing from one of the mediums actually mentioned by Takeda to another would have any or any significant impact on the percentage of pertactin in the culture supernatant. Similar objections were taken to, inter alia, the strain of B. pertussis used, the precise gradient used in the glucose gradient centrifugation step and the precise level of HA activity chosen as the limit for the retained fractions.

In the end I have come to the conclusion that none of the objections raised by Evans is of substance. Chiron's experiments were not precise repetitions of example 1 of the application. Therefore Mr. Kitchin's first argument fails. However I accept his second argument. One of the issues before the court is whether the carrying out of a specific example described in the prior art falls within the scope of the patent in suit. Normally the easiest way of demonstrating that it does is for the attacker to carry out that example precisely to see what results are obtained. If he makes alterations to the experimental protocol the court may well conclude that the alterations made a significant difference to the results obtained. If that is so, then the attacker will not have proved what happens when the described example is carried out. But the fact that this is normally the easiest way of proving what the example produces does not mean that it is the only way. For example a piece of prior art may include a reference to carrying out an experiment in a conical flask made by a particular manufacturer. By the time of the repetition that manufacturer may not exist. Carrying out the experiment in a flask from an alternative manufacturer would normally not be expected to have any impact on the course of the reaction. The experiment still proves, on a balance of probabilities, what would be the result of carrying out the example in the prior document. To approach this issue in any other way would be to subject the repetition of prior art examples to a tyranny of precision which has little to do with the issue before the court or the science to which the prior art relates. I have come to the conclusion, on the basis of the experiments carried out by Chiron and to the degree of certainty needed for civil proceedings, that an exact repetition of example 1 in the Takeda application would have resulted in a product containing in the region of 3 - 5% pertactin. In so holding, I have concluded that none of the Evans objections, including in particular the objections to the precise sucrose gradient used, would have any material affect on the results obtained. I also have come to the conclusion that the criticism directed to the level of HA activity chosen as a cut off for fraction selection is misconceived. The point chosen by Chiron was well within the teaching of the Takeda example.

I also accept Mr. Kitchin's third and fourth arguments. The experiments were a fair working of the general disclosure of the application and, in addition, were an obvious way of putting that disclosure into practice. This means that, subject to the Merrell Dow point, the Takeda application anticipates and renders obvious the claims 1 and 2 of the patent in suit. Furthermore I accept the more general argument advanced by Mr. Kitchin. The Takeda application would lead a man in the art to make a co-purified vaccine which would contain, in minor but significant quantities, pertactin even if he did not follow the precise teaching of the example.

The Sato paper

This also is pleaded as a piece of prior art upon which an argument of obviousness was based. It reports the manufacture of an endotoxin-depleted pertussis vaccine. It says that vaccines made in accordance with the described preparation had been on the market in Japan since 1981. This is a reference to the Takeda vaccine. In fact evidence was given that the Sato method was adopted not only by Takeda but also (and subject to some modifications) by other companies. Dr. Guiso gave evidence^[69] that the Takeda vaccine and the general features of how it was made were well known in the art in the early 1980's. She confirmed that the initial step in the preparation described in the Sato paper, namely precipitation using ammonium sulphate, would bring down all proteins in the supernatant from the B. pertussis culture. She also explained that the lipopolysaccharide, LPS, was thought to be responsible for the undesirable side affects allegedly associated with whole cell pertussis vaccines. She said that it was for that reason that scientists tried to make acellular vaccines containing only protein devoid of this endotoxin^[70]. She accepted that what the Sato paper describes and what was known about the Takeda vaccine was that the vaccine was produced by co-purification of the FHA and PT, rather than by making those ingredients separately, and that someone operating the Sato disclosure would be taking all the proteins in the selected fractions into the vaccine^[71]. It appears that this document, like the Takeda application would lead a man in the art to produce a co-purified pertussis vaccine which would contain a significant amount of pertactin. It is also clear that, at the time, no one identified pertactin as being present.

Because Sato used a simple process of co-purification with the retention of fractions of the total supernatant, it was not producing a vaccine consisting of only pure FHA and PT. In this respect the Sato document states explicitly what is implicit in the Takeda application. Dr. Redhead said:

"The initial material that was produced was claimed to contain more or less just PT and FHA, although testing at the time with SDS-PAGE did [not]^[72] suggest that other components might be present. It was only over the years, as the test methods have improved, it was possible to define what these other components might be."^[73]

In fact the Sato paper discloses that its vaccine was not made of pure FHA and PT. Other components were likely to be present. Thus it states:

"The pertussis component vaccine contains two main protective antigens, filamentous HA (F-HA) and leucocytosis-promoting factor (LPF-HA) ..."^[74] (emphasis added)

"The crude extract of F-HA and LPF-HA still contains a large quantity of substances, such as endotoxin, that cause side-effects. ... Fractions containing HA but no endotoxins ... are pooled and used as raw material for the component vaccine. Polyacrylamide gel electrophoresis of this raw material showed two main protein bands, a slow moving band corresponding to F-HA and a faster one to LPF-HA"^[75] (emphasis added)

The understanding at the time of what was in the Sato/Takeda vaccine is also reflected in a paper by Oda, Izumiya and Sato published in July 1983^[76]. Referring to the Sato vaccine it states:

"... the acellular vaccine contains F-HA, LPF-HA, and small amount of the agglutinogens as the predominant immunogens."^[77] (emphasis added)

As with the Takeda application, the aim of the Sato paper was not to collect two specific antigens and try and get rid of everything else altogether. Subject to Merrell Dow, it renders the patent in suit invalid for obviousness.

Merrell Dow v. Norton

Mr. Howe argues that in the light of Merrell Dow to invalidate a patent it is not enough to demonstrate that the invention is inherent in the prior art. He says that the invention is only part of the state of the art and invalid "if the information which has been disclosed enables the public to know the invention under a description sufficient to work the invention". He is prepared to concede that that something far less than a chemical structure would suffice to make a substance known so as to destroy the novelty of a later claim to that substance. But he says that the present case involves no recognition of the existence of the substance, pertactin, in the art at all and therefore falls within category of cases where there may be inherency but no invalidity. He argues that, as happened in Merrell Dow itself, a "few passing words" may be enough to amount to invalidating disclosure, even if they do not characterise what is being disclosed but this was not such a case. I put to him a hypothetical example in which an encyclopaedia, readily available before the priority date, contained an entry describing with precision how to make a whooping cough vaccine. If followed, the instructions would always produce a vaccine containing significant amounts of pertactin. However the presence of that particular antigen was not known or obvious. He agrees that a vaccine so made would infringe, but he says that it is a necessary consequence of his interpretation of Merrell Dow that, as long as nobody knew that it contained this particular antigen, the patent is not bad for obviousness or anticipation. Mr. Howe agrees that his argument is unpalatable but he says that is what the House of Lords has decided is the law. Furthermore, he says that if he is right in relation to anticipation, the same must follow in relation to obviousness.

Mr. Kitchin put an entirely different construction on Merrell Dow. He said that the House of Lords confirmed that one could not obtain a valid patent for characterising or analysing a known product. What you cannot do is get a patent for an old product merely because you have discovered a little bit more about its chemical make-up and structure. He says that all that has been done, in terms of disclosure by this patent, is to say, "A part of what was previously known is provided by this component" -- perhaps. The component produces a response, but precisely what that contribution is and what it is doing is, even now, very much an open question. But in terms of the information relevant to the invention, there is no more now than there was before.

In Merrell Dow the patentee had obtained a patent (the first patent) for an anti-histamine drug called terfenadine. It was subsequently discovered that terfenadine was turned into a metabolite in the liver of the patient. It was the metabolite which produced the beneficial anti-histamine effect. The patentee therefore applied for a patent (the second patent) to the newly discovered metabolite. If that patent was valid it would have had the effect of preventing others who were making terfenadine in accordance with the teaching of the first patent from selling it because that would amount to knowingly supplying consumers with the means for putting the invention in the second patent into effect contrary to s. 60(2) of the 1977 Act. The decision of the House of Lords dealt with two attacks on the validity of the second patent. The first was an allegation of anticipation by prior use as follows. Before the priority date of the second patent certain clinical trials of terfenadine had been conducted. It was said that the patients who had participated in those trials had taken the terfenadine and had produced the metabolite in their livers. The second allegation was that the patent was invalid because of anticipation by prior disclosure. It was said that following the instructions in the first patent would inevitably produce terfenadine for consumption by patients who would inevitably create the metabolite in their livers. In all cases neither the patients nor anyone else knew or had reason to suspect that a metabolite existed or was the effective agent. The decision of the House of Lords was that the prior user attack failed but the second attack succeeded.

Mr. Howe puts particular emphasis on Lord Hoffmann's reasoning in relation to prior use. Evans' case is that even if following the prior art would inevitably produce a vaccine which falls within the claims of the patent in suit, it does not invalidate the latter because no one before the relevant priority date knew that such a vaccine contained pertactin. However anyone following the prior art now would infringe, whether or not he knew that the vaccine he made contained pertactin. As I have noted above, Mr. Howe readily accepts that this argument seems unpalatable but he points out that Lord Hoffmann warned against coming to an "intuitive response" or from being too swayed by "the extraordinary consequences" of or "one's initial incredulity" at a finding of validity. Adopting Lord Hoffmann's words, he says that Evans' arguments are serious and "cannot be dismissed simply on the grounds that they produce results which seem contrary to common sense"^[78]. However, at times it appeared to me that Mr. Howe's argument turned Lord Hoffmann's words of caution on their head. It seemed that any reading of Merrell Dow or the Patents Act which gave rise to a conclusion consistent with common sense was suspect.

It seems to me that the easiest way to analyse Merrell Dow is to consider first that part of Lord Hoffmann's speech which deals with anticipation by prior disclosure. The question to be asked is whether the prior art specification "conveyed sufficient information to enable the skilled reader to work the invention"^[79]. In answering that it must be remembered that

"Section 2(2) does not purport to confine the state of the art about products to knowledge of their chemical composition. It is the invention which must be new and which must therefore not be part of the state of the art. It is therefore part of the state of the art if the information which has been disclosed enables the public to know the product under a description sufficient to work the invention."^[80]

Lord Hoffmann went on to point out that where the invention is a new product, in most cases knowledge of the product's chemical composition will be necessary to enable the public to work the invention. But that is not always so. Having cited with approval the EPO BAYER/Diastereomers Decision T12/81 he said:

"In other words, if the recipe which inevitably produces the substance is part of the state of the art, so is the substance as made by that recipe. CPC/Flavour Concentrates Decision T303/86 [1989] 2 E.P.O.R. 95 was a case about actual recipes for cooking. The application was to patent a process for making flavour concentrates from vegetable or animal substances by extraction with fat solvents under pressure in the presence of water. The claim specified certain parameters for the ratio between the vapour pressure of the water in the meat or vegetables and the vapour pressure of the free water. Opposition was based upon two cookbook recipes for pressure-frying chickens and making stews which in nontechnical terms disclosed processes having the same effect. The, Technical Board of Appeal said (at page 98):

"It is sufficient to destroy the novelty of the claimed process that this process and the known process are identical with respect to starting material and reaction conditions since processes identical in these features must inevitably yield identical products."

Furthermore, it did not matter that the cook did not realise that he was not only frying a chicken, but also making a "flavour concentrate" in the surplus oil. It was enough, as the Board said, that "some flavour of the fried chicken is extracted into the oil during the frying process even if this is not the desired result of that process."

... [the Board's decision] proceeded on the basis that for the purpose of being part of the state of the art, a process for making flavour concentrates was sufficiently described by a recipe for cooking food which did not expressly refer to the flavour concentrates but would inevitably have the effect of making them"^[81]

In my view this is a passage which is crucial to the understanding of this part of the law. It is not necessary to describe the product or process in the same terms as used in the subsequent patent. Indeed it is not even necessary for the reader to realise that the process was being used or product was being made. This is because whether or not the prior art enables the reader to work the claimed invention:

"is an objective fact independent of what he knows or thinks about what he is doing."^[82]

It appears to me that none of this departs from or modifies the well accepted principles of anticipation by prior disclosure. First one must identify what the alleged invention is, that is to say what is covered by the claims in the patent, and then one must decide whether or not that invention, or any part of it, would be made inevitably by following the instructions in the prior art. If it would be, then it does not matter whether the skilled reader of the prior art would realise that he was working within the area claimed in the subsequent patent. The same approach applies just as much to the issue of obviousness. If the prior art renders it obvious to do or make something which falls within the subsequent claims, they are invalid. This view of the law makes obvious sense. Were Mr. Howe's approach correct, any product could be repatented by discovering and patenting previously unknown inevitable byproducts or contaminants within it. I do not accept that Merrell Dow suggests that a sensible view of the law is the wrong view. Further this approach is consistent with that adopted by the Technical Board of Appeal in BAYER/Diastereomers Decision, cited with approval by Lord Hoffmann, namely that the purpose of Art. 54(1) of the EPC is to prevent the state of the art being patented again.

But is this understanding of the law of anticipation affected by what was said in Merrell Dow about prior use? It will be remembered that terfenadine had been ingested by a number of patients who did not know that they were as a consequence producing the metabolites. This was held not to invalidate the second patent. The very narrow basis upon which the argument of prior use was advanced appears from the following passage in Lord Hoffmann's speech:

"It is important to notice that anticipation by use relies solely upon the fact that the volunteers in the clinical trials took terfenadine and therefore made the acid metabolite. There is no suggestion in the Agreed Statement of Facts and Issues that the volunteers were also at liberty to analyse the terfenadine to discover its composition. If it was open to them to have done so, they would have been in the same position as if they had read the terfenadine specification and the arguments for anticipation by use would have been the same as for anticipation by disclosure. If anticipation by use is to be treated as a separate argument, it must be assumed that the volunteers were given terfenadine capsules by employees of Merrell Dow for the sole purpose of swallowing them. They took them without knowing their composition and produced within themselves a substance which was not then readily capable of being identified but is now known to have been the acid metabolite."^[83] (emphasis added)

Thus the prior use relied solely on the fact that the pills of terfenadine, which did not in themselves contain the metabolite, were ingested. In no real sense were the persons carrying out the alleged prior use, i.e. the patients, in possession of any information which would allow them to make the metabolite. This is emphasised towards the end of Merrell Dow:

"It may be helpful at this point to highlight the similarities and the distinctions between the case for anticipation by use, which I have rejected, and the case for anticipation by disclosure, which I have accepted. In both cases no one was aware that the acid metabolite was being made. In the case of anticipation by use, however, the acts relied upon conveyed no information which would have enabled anyone to work the invention, i.e. to make the acid metabolite. The anticipation in this form relies solely upon the fact that the acid metabolite was made... It disavows any reliance upon extraneous information, such as the formula for making terfenadine and the instructions to take it for its antihistamine effect. Anticipation by disclosure, on the other hand, relies upon the communication to the public of information which enables it to do an act having the inevitable consequence of making the acid metabolite. The terfenadine specification teaches that the ingestion of terfenadine will produce a chemical reaction in the body and for the purposes of working the invention in this form, this is a sufficient description of the making of the acid metabolite. Under the description the acid metabolite was part of the state of the art."^[84] (emphasis added)

Therefore not only were the users, i.e. the patients, ignorant of the creation of the metabolite but the case was not put forward on the basis that they had been told to take the pills for the antihistamine effect. Under the pre-1977 law this would have been called secret prior use. As Lord Hoffmann explains such uninformative use is no longer invalidating. But this does not affect the well established principle that prior art directions or information which will inevitably result in the use of a patented process or creation of the patented product invalidates by anticipation.

With this in mind, it is possible to assess the impact of the prior art in this case. First however one must be clear as to what the invention in issue is. Claim 1 includes a wide range of partially purified protein mixtures containing pertactin. For the reasons set out above, I have come to the conclusion that a number of pieces of the prior art teach or render it obvious to manufacture such mixtures. The claim is therefore invalid. Similar considerations apply to the other claims said to have independent validity. The only hesitation I have arises in relation to the prior art Takeda vaccine. It could be said that since the pertactin content of that product was unknown there is no difference between this case and Merrell Dow. I have come to the conclusion that even that argument is incorrect. In Merrell Dow the patients were never in possession of a product containing the metabolite. The metabolite only came into existence inside their livers. But here the Takeda vaccine itself is a product within the claims. It was open to anyone to analyse it. It is as invalidating as the other prior art.

On this issue it is only necessary to add one point. I have mentioned already Mr. Howe's concession that to the effect that Merrell Dow would allow a "few passing words" to amount to invalidating disclosure, even if they do not characterise what is being disclosed. It is clear that I do not accept that Merrell Dow is limited to this concession. However, even if it is, in my view the Sato document expressly and the Takeda specification implicitly disclose that their vaccines contained other unspecified antigens. They are semi-defined mixtures, as described at the beginning of this judgment. That disclosure would be enough to fall within Mr. Howe's concession.

Added matter

Chiron also ran an added matter attack. In the light of the findings made above, it is not necessary to deal with this issue.

C. Infringement

Chiron's proposed vaccines contain about the same or just slightly more pertactin than the Takeda vaccine. Chiron's arguments of non-infringement is primarily directed to putting Evans in a squeeze on validity. If, as Evans suggests, the pertactin must be present in sufficient quantitities so as to give rise to protective efficacy, then there is no evidence to support a finding of infringement. There is nothing to show that the pertactin present in the Chiron vaccines has any protective effect at all in humans. I have already rejected this view of the scope of Claim 1. What is required is that there be sufficient pertactin present to generate an antibody response. Pertactin is highly antigenic. It is not in dispute that the quantities present in the Chiron vaccines would generate an antibody response. If claim 1 was valid, it would be infringed.

D. Miscellaneous Matters

The Primer

There is one final matter which deserves mention. About a week before the trial commenced, Chiron served a technical primer on Evans. Evans had not been told that this document was in preparation and it objected to me seeing it, at least before Evans and its co-parties had had time to comment upon it. Unfortunately I was not aware of that objection until after I had read the document. On the other hand Evans did not produce its own primer. It was happy for the court to start this action without any general and simple introduction to the technology. I suppose that the only technical introduction, if that is the right word, which was to be found in Evans' documents was contained in the Dr. Guiso's main witness statement. This is an immensely complicated document which touches on many issues which had nothing or little to do with the issues in the action. It is also drafted in a form which does not appear designed to simplify the issues or make them more readily comprehensible to a novice. Mr. Howe indicated that many of the paragraphs in this were intended to explain and convey similar information to that which Chiron had set out in its primer, albeit more fully. Evans' objections to Chiron's primer came down to a complaint that it was served late and without warning and that it was put forward as a stand alone document rather than as a part of the evidence of one or other witness. I directed Evans to consider the primer and to notify me and Chiron of the passages to which it objected, so that the areas of technology where the parties differed could be resolved by oral testimony. In the end, the criticisms of the primer were minor and a modified version with which both sides agreed was used for the rest of the trial.

I must say that I think that this kind of dispute is quite unnecessary. The technical issues which arise in litigation over patents like the one in suit in these proceedings are complicated. The financial consequences of a finding of validity or invalidity are likely to be sufficiently significant to justify a proper analysis of the issues raised by the parties. In some crucial areas the differences both in detail and overall flavour between the technical evidence filed by the expert witnesses was so pronounced that the parties had no choice but to challenge their opponents evidence. To this end it was necessary to try to put the court in a position where it understood the technology involved. No criticism can be levelled at the parties for having taken the trouble to tackle the issues in detail. I can not see how the disputes in this action could be resolved fairly without a proper understanding and examination of them.

However this gives added importance to the parties' and their legal advisers' duty to do whatever they can to limit the areas of dispute and to help the court understand the science involved. A properly drafted primer helps to achieve these objectives. In 20 years experience of the use of primers in patent actions, I can remember no case in which they failed to assist the court. Evans was content that the court should start this action "cold", without being eased into the technology. That was not a helpful stance to adopt and, had it been matched by Chiron, inevitably would have resulted in the trial taking longer with obvious cost implications. Here as in other cases, I found the primer to be useful. The production of such documents in trials involving complex technology should be encouraged. On the other hand it would be much better if the parties attempt to agree the contents of a primer before it is submitted to the court. In the circumstances it would have been helpful if Chiron had sent their draft of the primer to Evans well enough in advance of the trial to give the latter adequate time to consider it and suggest amendments and this is the course which should be adopted in future.

Note 1   Statement of Dr. Redhead. Bundle E tab 7 para. 5.    [Back]

Note 2   Bundle A page 53 line 45.    [Back]

Note 3   Bundle A page 7, first para.    [Back]

Note 4   Bundle A page 9, top para.    [Back]

Note 5   Bundle A page 9, middle para.    [Back]

Note 6   Bundle A page 10, middle para.    [Back]

Note 7   Professor Higgins, Transcript Day 10 p. 1279.    [Back]

Note 8   Bundle A p. 17 bottom para.    [Back]

Note 9   Bundle A pp. 15 - 16.    [Back]

Note 10   Bundle A p. 9 bottom para.    [Back]

Note 11   Bundle A p. 16 bottom para.    [Back]

Note 12   Bundle A page 19.    [Back]

Note 13   Bundle I p. 76, para 64.12.    [Back]

Note 14   Bundle K 1 p. 401 at 403.     [Back]

Note 15   Bundle A page 9 middle para.    [Back]

Note 16   Bundle A page 9 bottom para.    [Back]

Note 17   Transcript Day 3 p. 297.    [Back]

Note 18   Bundle K1 p. 401.    [Back]

Note 19   Bundle K1 p. 148    [Back]

Note 20   Transcript Day 2 p. 234.    [Back]

Note 21   Transcript Day 3. p. 307.    [Back]

Note 22   Transcript Day 8 p. 992.    [Back]

Note 23   Transcript Day 9 p. 1162.    [Back]

Note 24   Transcript Day 13 p. 1664.    [Back]

Note 25   Transcript Day 13 p. 1689.    [Back]

Note 26   Transcript Day 9 p. 1143.    [Back]

Note 27   Transcript Day 10 p. 1294.    [Back]

Note 28   Transcript Day 10 p. 1279.    [Back]

Note 29   Transcript Day 6 p. 742 et seq.    [Back]

Note 30   Transcript Day 6 p. 723.    [Back]

Note 31   Transcript Day 6 p. 712.    [Back]

Note 32   Transcript Day 6 p. 727.     [Back]

Note 33   Bundle D Tab 9.    [Back]

Note 34   Transcript Day 3 pp. 372 to 375.    [Back]

Note 35   Transcript Day 7 pp 813 to 823.    [Back]

Note 36   Bundle F, tab 1 paras. 78 to 83.    [Back]

Note 37   Tanscript Day 9 pp 1136 to 1144.    [Back]

Note 38   Transcript Day 13 pp 1652 to 1657.    [Back]

Note 39   Bundle K1 p. 68.    [Back]

Note 40   A particularly clear explanation of how this is done can be found in the cross examination and answers of Dr. Guiso: Transcript Day 3 pp 364 to 368.    [Back]

Note 41   Bundle A p. 56 lines 20 to 24.    [Back]

Note 42   Bundle A p. 53.    [Back]

Note 43   Bundle A p. 54 lines 25 to 30.    [Back]

Note 44   Bundle A page 53 line 27.    [Back]

Note 45   Bundle A page 53 lines 35 -37    [Back]

Note 46   Particulars of Objection, para 9(d).    [Back]

Note 47   Bundle A p. 53 lines 47 et seq.    [Back]

Note 48   Bundle A p. 54 lines 5 - 10.    [Back]

Note 49   Bundle A p. 54 lines 24 - 30.    [Back]

Note 50   Bundle A p. 58 table 1.    [Back]

Note 51   Bundle F, Tab 1 paras 50 - 53.    [Back]

Note 52   Bundle F, Tab 2 paras 18 - 23.    [Back]

Note 53   Transcript Day 6 p. 680 lines 15 to 22 and p. 682 line 8 to 683 line 18.    [Back]

Note 54   Transcript Day 3 pp 315 - 317.    [Back]

Note 55   Dundle D tab 6.    [Back]

Note 56   Transcript Day 2 p. 126.    [Back]

Note 57    Transcript Day 2 p. 196.    [Back]

Note 58    Transcript Day 2 p. 200.    [Back]

Note 59    Transcript Day 7 p. 891.    [Back]

Note 60   Bundle E, Tab 7 p. 27-8 para 7, 8.    [Back]

Note 61   Bundle A p. 52 lines 24 et seq.    [Back]

Note 62   Transcript Day 2 p. 204.     [Back]

Note 63   Transcript Day 4 p. 457.    [Back]

Note 64   Transcript Day 8 p. 1030.    [Back]

Note 65   Transcript Day 8 p. 1018.    [Back]

Note 66   Bundle D p. 48 lines 23-5.    [Back]

Note 67   Bundle D p. 48 line 46.    [Back]

Note 68   Transcript Day 1 p. 55.    [Back]

Note 69   Transcript Day 2 p. 188 et seq.    [Back]

Note 70   Transcript Day 2 p. 192.    [Back]

Note 71   Transcript Day 2 p. 194-5.    [Back]

Note 72   The word “not” appears in the transcript. It is clear that it was not what was intended by the witness.    [Back]

Note 73   Transcript Day 11 p. 1389.    [Back]

Note 74   Bundle D p. 58 left column.    [Back]

Note 75   Bundle D p. 58 right column.    [Back]

Note 76   Bundle K4 tab 9.    [Back]

Note 77   ibid p. 1302.    [Back]

Note 78   [1996] RPC 79 at p. 84 line 30.     [Back]

Note 79   [1996] RPC 79 at p. 87 line 36.    [Back]

Note 80   [1996] RPC 79 at p.89 line 16.    [Back]

Note 81   [1996] RPC 79 at p.90 line 8.    [Back]

Note 82   [1996] RPC 79 at p.90 line 51.    [Back]

Note 83   [1996] RPC 79 at p. 85 line 12.    [Back]

Note 84   [1996] RPC 79 at p. 91 line 15.    [Back]