England and Wales Court of Appeal (Civil Division) Decisions

IN THE SUPREME COURT OF JUDICATURE
COURT OF APPEAL (CIVIL DIVISION)
ON APPEAL FROM MR. JUSTICE JACOB
CHANCERY DIVISION PATENTS COURT

		Royal Courts of Justice Strand, London, WC2A 2LL
		17th July 2003

Vice-Chancellor :

Introduction
1. The claimant ("Celltech") is a company incorporated in England. On 11th July 1991 it applied, pursuant to the Patent Cooperation Treaty, to the US Patent and Trade Mark Office ("USPTO") for a patent in relation to humanised antibodies. In the course of the prosecution of that application Celltech had to deal with a number of objections as to patentability raised by USPTO. It did so partly by amending the draft claims and partly by advancing arguments in opposition to the objection. On 4th September 1996 USPTO indicated that it would allow the claims in their then form and the patent ("the Adair Patent") was duly granted to Celltech on 12th January 1999.
2. In the meantime, on 19th January 1998, Celltech granted a licence under the prospective Adair Patent to the defendant ("MedImmune"), a company incorporated in the State of Delaware, United States, to develop, make, use and sell antibodies. The consideration for the licence was, in part, a royalty of 2% of net receipts, as defined, from all those antibodies which, but for the licence, would infringe a valid claim in the Adair Patent.
3. MedImmune has developed a humanised antibody called palivizumab for use in the treatment of respiratory syncitial virus. It sells such antibodies in large quantities in the United States and elsewhere under the trade name SYNAGIS. It claims that the manufacture and sale of SYNAGIS would not, in the absence of the licence, infringe a valid claim in the Adair Patent. In consequence it denies any liability to pay a royalty to Celltech by reference to its receipts from sales of SYNAGIS.
4. Celltech does not agree and issued these proceedings in October 2000. It is not disputed that the issue is to be determined in accordance with the Patent Law of the United States and that our approach to the application of that law should be in accordance with the guidance afforded by the decision of this court in MCC Proceeds Inc. v Bishopsgate Investment Trust plc (4th November 1998). Celltech relies on a principle of US Patent law called the doctrine of equivalents by which, as alleged in paragraph 20(7) of its particulars of claim,

"if it is established that a defendant's product contains elements which are identical or equivalent to each claimed element of the patented invention it will be held to fall within the scope of the claim."

It admits for the purposes of these proceedings that one element of SYNAGIS is not identical to the corresponding element in the Adair Patent but, it contends, it is an equivalent within the doctrine.

5. In its defence, served on 27th December 2000, MedImmune denies that SYNAGIS infringes the Adair Patent whether under the doctrine of equivalents or otherwise. In addition it contends that the doctrine of equivalents is not applicable in this case. The basis for that contention is the doctrine of prosecution history estoppel which also forms part of US Patent Law. It has two branches. The first, as alleged in paragraph 20 a) of the defence and counterclaim ("Amendment Estoppel") provides that:

"where a patentee has limited or amended the claims of a patent during prosecution to satisfy the requirements of the Patent Act and the amendments narrow the patent's scope the patentee is precluded from obtaining under the doctrine of equivalents coverage of the subject matter that had been relinquished or distinguished during the prosecution of the patent application."

The second branch ("Argument Estoppel") is to the effect that unmistakeable assertions made by the applicant to the USPTO about the scope or meaning of the claim or an element in the claim may also operate to preclude the applicant from asserting equivalence between an element or limitation of a claim and a substitute for that element or limitation. In its original particulars in support of its reliance on prosecution history estoppel MedImmune referred to and relied on the decision, given on 29th November 2000, of the US Court of Appeals for the Federal Circuit (the court having exclusive jurisdiction over all patent cases) in Festo Corporation v Shoketsu Kinzoku Kogyo Kabushiki Co. 187 F.3d 1381 ("Festo 1").

6. On 9th March 2001, on the application of MedImmune, Pumfrey J directed the trial of the following preliminary issues:

"(1) whether the claimant is precluded from relying on the US doctrine of equivalents in these proceedings by virtue of the US doctrine of prosecution history estoppel,

(2) whether as a result the claimant's claim should be dismissed."

The hearing of those preliminary issues was refixed by Laddie J on 20th July 2001 in the light of the fact that the US Supreme Court had decided to entertain an appeal from the decision of the Court of Appeals for the Federal Circuit in Festo 1. The appeal was successful for the reasons given by the Supreme Court in their judgments given on 28th May 2002, see Festo Corporation v Shoketsu Kinzoku Kogyo Kabushiki Co. 122 S.Ct.1831 (2002) ("Festo 2").

7. The preliminary issues were heard by Jacob J on 1st and 2nd October 2002. He concluded that Celltech was precluded by Argument Estoppel, but not by Amendment Estoppel, from relying on the alleged infringement by equivalents and dismissed the claim. Both sides now appeal to this court with the permission of the judge. Celltech contends that the judge was wrong to hold that there was an Argument Estoppel and, in consequence, to dismiss the claim. MedImmune submits that the judge was wrong to have found that there was no Amendment Estoppel. Both of them argue that the judge's conclusions are mutually inconsistent. In those circumstances it will be convenient to consider Amendment Estoppel before Argument Estoppel. But before I consider either it is necessary to explain, so far as necessary, (1) the background science, (2) the prosecution of the Adair Patent, (3) the alleged infringement, and (4) the relevant principles of the US Patent Law.
Background Science
8. As the judge recorded, the expert evidence was not, for present purposes, controversial. He acknowledged that his exposition of the background science largely came from the skeleton argument of counsel for MedImmune with his own embellishments. For my part I acknowledge with gratitude that what follows in paragraphs 9 to 15 is taken with minor additions, subtractions and alterations from paragraphs 21 to 27 of the judgment of Jacob J.
9. Antibodies, which are useful for a variety of therapeutic and research purposes, are produced by cells of the human immune system called B lymphocytes. They are protein molecules. A protein is a long chain of amino acid residues called a polypeptide. The proteins of living organisms use 20 specific amino acids. Antibodies are proteins which bind to other proteins foreign to the human body called antigens. A typical antigen would be a virus. Antigen/antibody binding depends on a close and specific physical fit between particular areas on each. The antibody "locks on" to part of the antigen, thereby disabling its invasive abilities. Antigens may be many and varied. Each particular kind will need a particular kind of antibody to disable it. When the body is invaded the B lymphocytes set about producing antibodies specific to the invading antigen. This is why it takes time to deal with an infection such as the common cold, and why, when antibodies have built up, it is cured.
10. Antibodies have a generally "Y" shape, though the shape is more elaborate than that. The "Y" is made from two heavy (relatively long) and light (relatively short) chains of amino acid residues. A heavy chain is about 440, and a light chain about 220 amino acids long. The Y consists of two identical heavy chains joined together for about 220 amino acids which form the trunk of the Y. The two tines of the forks (of about 220 amino acids) each have attached to them a light chain. The trunks of the Y consist of constant regions, that is to say they are the same in different antibodies. The branches are variable, i.e. different in different antibodies. In broad terms, the tips of the forks of the "Y", where the variable regions of the heavy and light chains meet, are the parts which make the antibody "fit" with its complementary antigen. There are three loops at each tip. It is their shape and composition which make the antibody "fit" the target antigen. The regions for fitting into the antigen are called "CDRs" (complementarity determining regions). I reproduce as appendix A a helpful schematic diagram taken from the first witness statement of Dr Martin, Celltech's expert witness, showing the general structure of an antibody.
11. Producing human antibodies is difficult. As a result, researchers have taken to producing antibodies in mice or other rodents. But mice antibodies themselves provoke an immune reaction in humans (the "HAMA" response). Researchers have therefore worked for some time on using antibodies which are "humanised", being part mouse and part human. The goal is to arrive at an antibody which is the right shape for the target antigen but is human enough not to provoke a HAMA response. In such antibodies, the mouse elements are called "donor" and the human elements are called "acceptor".
12. There is a standard, conventional means of referring to individual specific amino acids in the protein structures of this kind of antibody. It is called the Kabat numbering system. It is used to designate a given amino acid position in an antibody by a number. In the present case position 23 is the subject of most of the argument.
13. Part of Celltech's argument in this case revolves around what are called "conservative substitutions". As previously stated, there are 20 amino acids available to make up the chain. Amino acids differ from one another in a variety of ways and to different degrees. To replace one amino acid with another which is very similar is described as making a "conservative substitution" of one for the other. Such a substitution at a non-critical part of the molecule is unlikely to make a significant difference to its function. There may be a local slight difference in shape but it does not matter. In critical areas, such as CDRs, even a conservative substitution may make a substantial difference.
14. Celltech summarise their invention in simple terms in the skeleton argument they used at the hearing before Jacob J:

"one takes the human antibody and grafts into it the CDRs and certain other non-CDR residues which are necessary to achieve binding. Humanised antibodies of the kind described in the patent are therefore designed to combine the best attributes of murine and human antibodies: the important binding elements from the mouse monoclonal antibody are used to target the antigen of interest and, because the remainder of the antibody contains human sequences, the human body is less likely to mount an immune response to it.

Celltech identified a specific set of positions at which it was necessary to ensure that the residue was the relevant murine residue (the so-called "donor" residue) to obtain the desired binding. This set was described in the US Adair patent [1] and claimed."

15. Thus grafting a CDR obtained from a mouse onto an otherwise human antibody is not enough because the desired binding is not achieved. Specified donor (mouse) amino acids or residues are needed at specified positions. These are near the CDRs. I attach as appendix B another helpful diagram from the witness statement of Dr Martin showing where position 23 is. I should emphasise that the process of grafting does not involve any physical transposition or joinder; rather it refers to the creation of a protein with the requisite sequence of amino-acids by the processes of genetic engineering referred to in greater detail in the decision of this court in Chiron Corporation v Murex Diagnostics [1996] FSR 153.
The prosecution of the Adair Patent
16. As I have already indicated the application for the Adair Patent was filed with USPTO on 11th July 1991. The relevant claim set out in the application (emphasis added) was:

"1. A CDR-grafted antibody heavy chain having a variable region domain comprising acceptor framework and donor antigen binding regions wherein the framework comprises donor residues at at least one of positions 6, 23 and/or 24, 48 and/or 49, 71 and/or 73, 75 and/or 76 and/or 78 and 88 and/or 91."

The references to "residues" are references to amino-acids and references to a donor or an acceptor residue are references to the amino-acid to be found in the protein sequence of the donor or mouse or the acceptor or human sequence as the case may be. Thus position 23 in claim 1 may be occupied by any one of the 20 amino-acids if but only if one of the other 6 positions (treating alternatives as a single position) is occupied by the amino-acid in that position in the corresponding donor sequence.

17. The prior art at that date included an article by Lutz Riechmann and others published in Nature Vol 332 on 24th March 1988 entitled "Reshaping human antibodies for therapy" ("Riechmann"). This, amongst other matters, formed the basis of objections to patentability advanced by USPTO. If Celltech's application was to succeed it had to overcome these objections either by amendment or by argument sufficient to satisfy USPTO that its objection was not well founded.
18. The relevance of Riechmann, as explained by Dr Martin, is not that it taught that the relevant positions (6, 23, 24 and 49) should be any particular amino-acid. Rather it arose from the fact that he showed sequences in which positions 6 and 49 are the same in both the donor and acceptor sequences, thereby anticipating claim 1 in relation to those positions. But positions 23 and 24 are not only different from but not conservative substitutions the one for the other.
19. The examiner's objection was based on §102(b) of the Patent Trademark and Copyright Law of the US which requires that the claimed invention shall not have been "...described in a printed publication in this or a foreign country....more than one year prior to the date of the application...". Celltech proposed an amendment in which each of positions 23, 24 and 49 should be donor amino-acid residues. In explaining such amendment Celltech stated to the examiner in its Response dated 19th January 1993:

"In part A of this rejection, claim..1..... [was] rejected as anticipated by Riechmann et al. The Examiner stated that claim 1...[was] interpreted to mean that the framework has donor residues in at least one of any of positions 6, 23, 24, 48, 49, 71, 73, 75, 76, 78, 88 or 91.....and thus the teachings of Riechmann et al anticipate the invention as claimed.

The Examiner contends that the original claims lacked novelty over Riechmann et al. Claim 1...[has] been cancelled without prejudice and submitted as new claims that more distinctly point out certain aspects of the present invention.

In present claims....it is specified that residues 23 and 24....should be donor residues. However, as can be seen...in Riechmann et al in the recombinant antibody shown there residues 23 and 24 are acceptor residues."

20. This and other arguments satisfied the Examiner who, on 4th September 1996, accepted as suitable for grant the claim ultimately appearing as claim 1 in the Adair Patent. In the Adair Patent, as granted, the background to the invention and the work carried out by Celltech is set out in columns 1 to 3. The outcome is explained in column 3 lines 54 to 61 to be that:

"This has enabled us to establish a protocol for obtaining satisfactory CDR-grafted products which may be applied very widely irrespective of the level of homology between the donor immunoglobin and acceptor framework."

21. The invention is summarised in columns 3 to 8. The protocol, contained in columns 8 and 9, states:

"It is first of all necessary to sequence the DNA coding for the heavy and light chain variable regions of the donor antibody, to determine their amino acid sequences. It is also necessary to choose appropriate acceptor heavy and light chain variable regions, of known amino acid sequence. The CDR-grafted chain is then designed starting from the basis of the acceptor sequence. It will be appreciated that in some cases the donor and acceptor amino acid residues may be identical at a particular position and thus no change of acceptor framework residue is required."

22. In column 10 line 17 position 23 is referred to as "a key residue" in the context of the residues identified by Celltech as having an effect on net antigen binding. In column 26 it is stated that:

"These and other results lead us to the conclusion that of the 11 mouse framework residues used....it is important to retain mouse residues at all of positions 6, 23, 24, 48 and 49..."

23. Claim 1 in the Adair Patent as granted claims, omitting irrelevant elements and adding emphasis is for:

"An antibody molecule....wherein, according to the Kabat numbering system, in said composite heavy chain: said CDRs comprise donor residues at least at residues 31 to 35, 50 to 58, and 95 to 202; and amino acid residues 6, 23, 24 and 49 at least are donor residues."

24. In summary the relevant changes made in consequence of the Examiner's objections based on Riechmann are to the nature of the amino-acid residue at position 23. Whereas in the original claim 1 it was not essential that the residue at position 23 was a donor residue, for there were the other six alternatives, under the amended claim 1 it is. The effect of the alteration is to restrict the monopoly claimed to those antibodies which do have a donor residue at position 23, as well as at positions 6, 24 and 49.
The alleged infringement
25. In SYNAGIS the amino-acid at position 23 is threonine. In the donor sequence it is serine. In the opinion of Dr Martin the difference between the two is so slight that the one may be properly regarded as a conservative substitution for the other. It is common ground that SYNAGIS has all the other elements set out in claim 1 of the Adair Patent. It follows that SYNAGIS does not literally infringe the Adair Patent because the amino-acid at position 23 is not serine. But if Celltech is entitled to rely on the doctrine of equivalents and the evidence of Dr Martin is accepted at the trial then SYNAGIS infringes claim 1 notwithstanding that the amino-acid at position 23 is threonine rather than serine.
26. But, in this connection, it should also be noted that if SYNAGIS infringes claim 1 of the Adair Patent as granted it would also have infringed claim 1 as set out in the original application. This is because if the doctrine of equivalents applies and threonine is a conservative substitution for serine the presence of threonine at position 23 would have satisfied the relevant element in both the original and the amended claim 1. It is this consideration which forms a large part of the reasoning of Jacob J and the case for Celltech.
US Patent Law
27. There is no dispute as to the nature of the doctrine of equivalents. It was established by the decision of the Supreme Court in Graver Tank & Mfg. Co. v Linde Air Products Co. 339 US 605 (1950). It applies to cases in which a product or process which does not literally infringe the express terms of a patent claim may nevertheless be found to infringe if there is equivalence between the elements of the accused product or process and the claimed elements of the patented invention. The doctrine was reaffirmed by the same court in Warner-Jenkinson Co. Inc. v Hilton Davis Chemical Co. 520 US 17 (1997), notwithstanding legislative revisions made by Congress to the Patents Act in 1952.
28. In the latter case, at page 29 the Court emphasised that:

"Each element contained in a patent claim is deemed material to defining the scope of the patented invention, and thus the doctrine of equivalents must be applied to individual elements of the claim, not to the invention as a whole. It is important to ensure that the application of the doctrine, even as to an individual element is not allowed such broad play as to effectively eliminate that element in its entirety. So long as the doctrine of equivalents does not encroach beyond the limits just described or beyond related limits to be discussed....we are confident that the doctrine will not vitiate the central functions of the patent claims themselves."

The Court also accepted that the doctrine of equivalents may be excluded by prosecution history estoppel.

29. The principles of prosecution history estoppel have not been so clear in recent years. I will refer first to that branch of it called Amendment Estoppel. In Festo 1 the Court of Appeals for the Federal Circuit concluded that there was "a brightline rule" to the effect that any amendment to a claim relating to its patentability, whether voluntary or otherwise, creates such an estoppel as to exclude the doctrine of equivalents in relation to the amended element. No doubt it was the principle so formulated in Festo 1 (decided on 29th November 2000) which led MedImmune on 4th January 2001 to seek the trial of the preliminary issues ordered by Pumfrey J on 9th March 2001. But the issue was reconsidered by the Supreme Court who reached a different conclusion in Festo 2.
30. The Supreme Court described the principle in these terms:

"When the patentee responds to the rejection by narrowing his claims, this prosecution history estops him from later arguing that the subject matter covered by the original broader claim was nothing more than an equivalent. Competitors may rely on the estoppel to ensure that their own devices will not be found to infringe by equivalence."

The appeal was based on two points, first, that the principle had been held to apply whatever the reason for the amendment and, second, that if and when the estoppel arises it operates as a complete bar to any claim based on an equivalent to the element in the amended claim. The Supreme Court rejected the first ground. Thus the principle is applicable to any relevant amendment whatever its cause.

31. The Supreme Court did not agree that a prosecution history estoppel gave rise to a complete bar. In paragraphs 6 to 11 Justice Kennedy set out the Court's reasons for arriving at that conclusion. Those paragraphs are too long to cite in full but I accept the submission of counsel for Celltech that they establish the following propositions:

(1) the principle applies in respect of the amended element only,

(2) the extent of the estoppel depends on the extent of the subject matter surrendered by the amendment,

(3) the purpose of the principle is to hold the patentee to representations he has made and inferences which may fairly be drawn from the amendment,

(4) the principle is to be applied flexibly rather than rigidly,

(5) subject to (6) it is to be assumed that by the amendment the patentee has surrendered all the territory between the original claim and the amended claim,

(6) such presumption may be rebutted, for example in cases where the rationale for the amendment bears no more than a tangential relation to the equivalent in question.

32. In relation to the fifth and sixth propositions, at page 1842, Justice Kennedy said:

"There are some cases, however, where the amendment cannot reasonably be viewed as surrendering a particular equivalent. The equivalent may have been unforeseeable at the time of the application; the rationale underlying the amendment may bear no more than a tangential relation to the equivalent in question; or there may be some other reason suggesting that the patentee could not reasonably be expected to have described the insubstantial substitute in question. In those cases the patentee can overcome the presumption that prosecution history estoppel bars a finding of equivalence."

33. He concluded by way of summary of those propositions in these terms:

"This presumption is not, then, just the complete bar by another name. Rather, it reflects the fact that the interpretation of the patent must begin with its literal claims, and the prosecution history is relevant to construing those claims. When the patentee has chosen to narrow a claim, courts may presume the amended text was composed with awareness of this rule and that the territory surrendered is not an equivalent of the territory claimed. In those instances, however, the patentee still might rebut the presumption that estoppel bars a claim of equivalence. The patentee must show that at the time of the amendment one skilled in the art could not reasonably be expected to have drafted a claim that would have literally encompassed the alleged equivalent".

34. Before leaving Festo 2 it is also necessary to note three other relevant principles on which counsel for MedImmune relied and which I did not understand counsel for Celltech to dispute. First the court must consider the literal effect of the amendment in narrowing the scope of the claim. Cf Festo 1 at page 587 and Aclara Biosciences v Caliper Technologies Corpn. 125 F.Supp.2d 291 (N.D.Cal.2000) at page 398/9. Second, regard must be had only to what is in the public file. An objective test is then to be applied to that material. Thirdly, the court is not concerned with whether the amendment went beyond what was strictly necessary to overcome the objection, see Festo 1 at p.568 and Warner-Jenkinson at p.33 note 7.
35. I can deal with Argument Estoppel more shortly. In Texas Instruments Inc. v US International Trade Commission 988 F.2d 1165 (Fed.Cir.1993) the Court of Appeals for the Federal Circuit noted that as a general proposition prosecution history estoppel is based upon showing an amendment to a claim. It added (p.1174):

"Amendment of a claim in light of a prior art reference, however, is not the sine qua non to establish prosecution history estoppel. Unmistakable assertions made by the applicant to the Patent and Trademark Office (PTO) in support of patentability, whether or not required to secure allowance of the claim, also may operate to preclude the patentee from asserting equivalency between a limitation of the claim and a substituted structure or process step".

As Jacob J noted the test is "unmistakeable assertion".

Celltech's application for an adjournment
36. Following the successful appeal to the Supreme Court in Festo 2 the matter was remitted to the Court of Appeals for the Federal Circuit. As Jacob J noted in paragraph 17 of his judgment the latter court invited further argument in relation to the rebuttal of the presumption arising from an amendment. It highlighted issues of whether foreseeability, tangentialness, or the reasonable expectations of those skilled in the art were issues of law or fact and what factors are encompassed by those criteria. At the end of the first day of the hearing before us we were informed that such further argument had occurred on 3rd February 2003 and judgment had been reserved.
37. We were also told that on 16th September 2002 Celltech had instituted further proceedings against MedImmune in this jurisdiction in relation to sales of SYNAGIS in Germany. They alleged that but for the licence they would have constituted infringements of the European Patent corresponding to the Adair Patent and thus similarly generated a right to a royalty as prescribed by the Licence. Celltech relies on the construction of the European Patent and a German doctrine of equivalents. In its defence MedImmune relies on limitations to that doctrine rather than any German principle of prosecution history estoppel.
38. This information led us to question whether we should determine this appeal before judgment had been given in what for convenience I will call Festo 3 or merely take account of the fact that judgment had not been given so that to that extent the US Patent Law might be unclear. Counsel for Celltech invited us to set aside the order of Jacob J so that this action might proceed in tandem with the action concerning the European Patent, to adjourn the appeal to await the outcome of Festo 3 and subsequently dismiss this action if we do find some prosecution history estoppel. We dismissed this application and indicated that we would give our reasons later.
39. For my part I consider that the application was wrong in principle and too late in practice. MedImmune has a judgment in its favour. At the time the application was made we had not heard any argument from its counsel. MedImmune was not prepared to agree to the order of Jacob J being set aside. In the absence of such consent I do not consider that this court could properly set aside the order having heard argument from one side only even though its effect might be restored later. Further, the relevant facts had been known to Celltech for many months. If the hearing of the appeal was to be adjourned the application should have been made long before the hearing of the appeal started.
Amendment Estoppel
40. I turn then to consider MedImmune's cross-appeal. It seems to me to be logical to deal with this aspect of the case first because the nature and effect of the assertions relied on by Jacob J depend in part on the nature and effect of the amendment. The case for MedImmune is set out in paragraph 21 e) of its amended defence and counterclaim. MedImmune contends that there is an Amendment Estoppel because in response to the prior art objection based on Riechmann Celltech amended its claim to require that residue 23 must be a donor residue so that

"having narrowed the claim to avoid prior art disclosing an antibody with an acceptor residue at position 23 the claimant cannot rebut the presumption that it is estopped from covering that which was surrendered, ie an antibody with an acceptor residue at position 23. Thus the claimant is estopped from relying on the doctrine of equivalents to cover an antibody (such as SYNAGIS) with the acceptor residue at position 23 rather than the donor residue."

41. This contention was rejected by Jacob J. In paragraphs 33 and 34 of his judgment he said:

"33. Celltech say that the equivalent in question, namely a conservative substitution, is simply irrelevant to the amendment made. I think they are right on this point. The amendment cut down the number of possibilities claimed. But it did not in any way cut down whether or not the claim covered equivalents of donors. Consider the case of a product which, at position 23, had an equivalent of a donor residue but did not have a donor residue at positions 6 or 24. And suppose it otherwise fell within the unamended claim. The argument for infringement of that claim would have had to be on the doctrine of equivalents. It would be exactly the same as it is now. The amendment is irrelevant.

34. One can look at the same point in another way. Festo requires one to examine what is "relinquished" or "surrendered" by the amendment. What was surrendered here is a number of possibilities. Focussing on position 23, 6 possibilities were given up. Prior to amendment an argument lay in relation to all seven possibilities that a conservative substitution would infringe by virtue of the doctrine of equivalents. All that has happened is that 6 of the possibilities have been surrendered. The surrender has no connection, no nexus, with what is alleged to be an equivalent."

42. The judge considered that his conclusion was warranted by analogy with Aclara Biosciences v Caliper Technologies Corpn. 125 F.Supp.2d 391 (N.D.Cal.2000). He rejected the counter-argument of MedImmune which he described as being to the effect that once a claim had been amended the Festo rules apply to all aspects of the claim, even those untouched by and unrelated to the amendment. Similarly Jacob J did not consider that the case on which MedImmune relied, Glaxo Wellcome v Impax Laboratories 2002 US Dist Lexis 15966, supported its argument.
43. Counsel for MedImmune submits that the judge was wrong. He disclaims the argument attributed to him by the judge as ever having been his case. He does not suggest that any amendment will preclude all equivalents. His case is simply that the amendment narrowed the claim because before amendment position 23 might be any of the 20 amino acids provided that one of the other six positions was occupied by donor residue but after amendment position 23 had to be occupied by donor residue. Put another way, the exclusion of the other six alternatives means that a compulsory element is introduced, namely that position 23 must be donor residue. Thus, he contends, the territory surrendered is the ability for position 23 to be occupied by any one of the other 19 amino acids, in particular the acceptor residue. In those circumstances the presumption applies and cannot be rebutted.
44. Counsel for Celltech supported the judge's reasoning and conclusion. He submitted that there had been no narrowing amendment with regard to position 23 because donor residue in that position would have infringed the unamended claim just as it infringes the amended claim. He points out that there was no amendment sufficient to alter the meaning of donor residue in relation to position 23 so that conservative substitutions before amendment are not altered or surrendered by the amendment. For these reasons he contends that there could be no presumption of any surrender, but if, contrary to that submission there is, then the presumption can be rebutted. The basis for the rebuttal on which he relies arises from the fact that to avoid Riechmann it was only necessary to make amendments with regard to positions 6 and 49. Consequently, he asserts the additional amendment regarding position 23 was tangential within the principle formulated in Festo 2.
45. I prefer the submissions for MedImmune. First, it is necessary to consider the amendment made in respect of the particular element. That element is position 23. It seems to me to be plain that the amendment was both made in relation to that element and one which narrowed the claim. Before amendment position 23 might be occupied by any one of the 20 amino acids or residues if at least one of the other 6 positions was occupied by a donor residue. After amendment position 23 must be occupied by a donor residue or amino-acid even if all the other previously alternative positions were similarly occupied. The removal by amendment of the other six options had the effect of introducing a compulsory element. Thus the amendment both related to the relevant element and narrowed the monopoly claimed by removing all those categories of antibody which had donor residues at all or one or more of the six alternative positions but not at position 23.
46. In my view the contrary argument based on the fact that SYNAGIS would have infringed both the amended and the unamended claim does not affect this conclusion. The proposition is true but it does not satisfy the Festo principle. The relevant part of the Festo principle requires the court to consider the unamended claim in its widest ambit and then to ascertain whether it has been narrowed by the amendment. The comparison relied on by Celltech does not do that because, in effect, it concentrates on one only of the seven alternatives. The question is not whether SYNAGIS infringed either or both the amended and unamended claims; rather it is whether the amendment was related to the relevant element and narrowed the claim.
47. Once the literal comparison is made it can be seen that the territory surrendered includes the occupation of position 23 by any amino acid other than the donor residue in all those cases where a donor residue is present at one or more of the other alternative positions. As the amendment also required a donor residue at each of those alternatives its effect was to exclude all possibility of any amino acid at position 23 other than the donor residue. It must follow that as a conservative substitution could only be one of the remaining 19 amino acids each of them is to be regarded as surrendered territory.
48. The suggestion that the rationale for the amendment is tangential to the equivalent or conservative substitution is in my view unsustainable. There is no doubt that the amendment was made to overcome the objections raised by the examiner based on Riechmann. The fact that the amendment may have gone further than was strictly necessary for that purpose is irrelevant as contrary to the third proposition referred to in para 34 above. Similarly the public is entitled to expect a patentee who wishes to encompass conservative substitutions to position 23, elsewhere in the Adair Patent described as a key residue, in amended claim 1 to say so; it could be done in virtually the terms I have used.
49. I am unable to agree with the analysis of Jacob J in paragraphs 33 and 34 of his judgment which I have already quoted. I do not think that the question he apparently posed for himself in paragraph 33, namely whether the conservative substitution of threonine for serine was relevant to the amendment was the right question. Similarly in paragraph 34 he gives no recognition to the effect of the deletion of the alternatives in creating a mandatory requirement for a donor residue at position 23. In paragraph 35 the judge drew an analogy with Aclara Biosciences v Caliper Technologies Corpn. 125 F.Supp.2d 391 (N.D.Cal.2000) and concluded

"The position is the same here – the amendment did not surrender any subject matter related to an equivalent of a donor residue."

I am unable to agree. In surrendering the territory occupied by all amino acids except the donor residue the amendment necessarily surrendered any equivalent of a donor residue.

50. In opening his case on Celltech's appeal counsel made a number of submissions as to what the skilled man would have understood as to the need for particular amino acids in particular positions given the need to obtain the best compromise between maximum adhesion to the antigen and avoidance of the HAMA immune response. I do not consider that any of these considerations touch on the questions posed by the Festo principles, confined as they are to a literal comparison by a man skilled in the art of the claims before and after amendment, the territory presumed to be surrendered in consequence and whether and if so how that presumption may be rebutted.
51. For all these reasons I would allow the cross-appeal. The consequence is, if the other members of the court agree, that the order of Jacob J must stand irrespective of the fate of Celltech's appeal. It would be possible therefore not to consider the arguments based on Argument Estoppel. Nevertheless I think that it would be right to do so. Not only may this case go further, but both parties relied on alleged inconsistencies between the judge's conclusions arising from his rejection of any Amendment Estoppel but acceptance of an Argument Estoppel. Accordingly it is appropriate to test my conclusion on the former by reference to the arguments for the parties on the latter.
Argument Estoppel
52. Jacob J correctly concluded that for such an estoppel to arise there must have been an unmistakeable assertion. In paragraphs 40 and 41 of his judgment he quoted two passages from Responses made by Celltech in the course of the prosecution of the Adair Patent. They are

"looking at the heavy chain Table, it can be seen that in all the successfully "superhumanised" antibodies produced by the applicants, in the heavy chain residues 23, 24, 31 to 25 49 to 58 and 95 to 102 are all donor residues."

"It can thus be seen that Riechmann does not disclose a procedure which leads to an antibody as defined in the present claims. As far as the heavy chain is concerned, Reichmann does not even mention Kabat residues 23 24 and 49, let alone change them to the rat [i.e. donor] residues."

In paragraph 42 the judge referred to other documents which he considered showed that Celltech was insisting that not only must position 23 be occupied by a donor residue but also that it should not be an acceptor residue.

53. The judge then recorded the submissions for Celltech that as nothing was said to USPTO about conservative substitutions there could be no estoppel precluding Celltech from relying on infringement by equivalents. He noted the submission that Riechmann said nothing about position 23. He recorded the evidence of Dr Martin to the effect he had read the prosecution file as leading to the conclusion that Celltech was surrendering a non-donor residue at position 23 but not a conservative substitution.
54. The judge expressed his conclusion in paragraph 45 of his judgment in the following terms:

"I reject all these submissions. [Prosecution History] estoppel does not require one to consider the cited prior art and work out what might have been said to distinguish it. One looks at what was said. It may be just enough to avoid it or it may go further, either by accident or on purpose. Here it may well be that Reichmann could have been distinguished in a different way - but what the patentee said is that position 23 had to be a rat (i.e. donor) residue and not an acceptor residue. He cannot now claim that an acceptor residue is within his monopoly even if it is a conservative substitution which would make no difference. As to Dr Martin's evidence, it misses the point - in particular that the patentee not only surrendered a non-donor residue at position 23 but also pointed out that Reichmann was different because he had an acceptor residue at that position. I would add this that in any event I am not impressed by the evidence - it states Dr Martin's conclusion having read the file, but does not give his reasons for that conclusion. He does not deal with what it actually says."

55. Counsel for Celltech criticises this conclusion on a number of grounds. First he submits that the conclusion is one which the skilled man would realise leads to absurd consequences because it would mean that in cases where the donor residue was the same as the acceptor residue it is impossible to exclude the latter without excluding the former. With respect, it seems to me that the absurdity lies in the implication, notwithstanding the teaching in the Adair Patent quoted in paragraph 21 above, that the skilled man would consider that this requirement led to the exclusion of the amino acid which was the same in both the donor and acceptor sequences rather than the other 19 which are not.
56. Then it is suggested that the passages the judge relied on do not support his conclusion. The objection of counsel for Celltech is to the implication the judge made that the residue at position 23 should not be the acceptor residue. In substance this is the same as the first point. There is no such problem once it is recognised that the requirement that position 23 should not be occupied by the acceptor residue arises only where it is not the same as the donor residue.
57. Counsel contended that the judge was not entitled to reject the evidence of Dr Martin in the way he did or at all. I do not agree. It is clear that the Adair Patent said nothing about conservative substitutions. Equally where it referred to donor residues it meant donor residues and not donor residues and conservative substitutions for donor residues. I agree with the judge that in the passage relied on Dr Martin was either speculating on whether Celltech had given away more than it need have done to avoid the objection based in Riechmann or was expressing views contrary to the meaning of the Adair Patent. In any event the true meaning and effect of the statements made by Celltech to USPTO was a matter for the judge to be determined by the application of an objective test to the words used when read through the eyes of the skilled man.
58. The final submission made by counsel for Celltech was that the judge's conclusion on Argument Estoppel was inconsistent with his conclusion on Amendment Estoppel. I agree, but as I consider that the judge was wrong on the second it follows that I agree with him on the first.
59. For all these reasons I would dismiss the appeal.
Conclusion
60. In summary I conclude that the judge was wrong to find that Celltech was not precluded from relying on the doctrine of equivalents because of an Amendment Estoppel, but was right to find that Celltech was so precluded by an Argument Estoppel. It follows that I would allow the cross-appeal and dismiss the appeal but leave undisturbed paragraph 1 of the order of Jacob J by which he dismissed the action.
Appendix A Structural Domains of an IgG Antibody

 
Appendix B
Structural Location of Residues in Claim 1 of the U.S. Adair Patent


Arden LJ

61. I am indebted to the Vice-Chancellor for his comprehensive exposition of the matters in issue in this appeal, which I gratefully adopt. I will use the same abbreviations. I have also had the advantage of reading the judgment of Longmore LJ in draft.
62. I have, however, come to a different conclusion. It is convenient to deal with the cross-appeal first.
MedImmune's cross-appeal (prosecution history estoppel)
63. The patent in suit concerns a biotechnological invention using recombinant DNA technology and monoclonal antibody technology. The appeal requires the application of the federal law of the United States of America. We have effectively approached the questions of foreign law as questions of law, and not as fact. Both the background science and federal law are unfamiliar. However, the appeal and cross-appeal turn on the more familiar territory of legal analysis.
64. As the judge indicated at the start of his judgment, federal law adopts a literal approach to the construction of patents. A claim is interpreted according to the exact meanings of the words used as those meanings would appear to a person skilled in the art. The doctrine of equivalents allows claims to be interpreted more broadly. "If two devices do the same work in substantially the same way and accomplish substantially the same result, they are the same even though they differ in name, form or shape.": Graver Tank & Mfg. Co. v Linde Air Products, Co. 70 S.Ct. 854 Sup. Ct. (1950) page 856.
65. In Warner Jenkinson v Hilton Davis 117 S.Ct. 1040, US 1997, the Supreme Court held that equivalents should be determined on an element by element basis, i.e. "… the doctrine of equivalents must be applied to individual elements of a patent claim not to the invention as a whole." An element is a distinctive feature or limitation in the claim. Each of the positions specified in unamended claim 1 is a separate element of that claim.
66. The doctrine of prosecution history (or file wrapper) estoppel applies where a claim is amended in the course of the application for a patent. The Vice-Chancellor has described the estoppel in detail and for my purposes the critical propositions established in Festo 2 and summarised in paragraph 31 of his judgment are:-

"(2) The extent of the estoppel depends on the extent of the subject matter surrendered by the amendment

…

(5) … it is to be assumed that by the amendment the patentee has surrendered all the territory between the original claim and the amended claim …"

67. As the Supreme Court said in Festo 2:-

"A patentee's decision to narrow his claims through amendment may be presumed to be a general disclaimer of the territory between the original claim and the amended claim." (page 1842 per Justice Kennedy).

68. Prosecution history estoppel applies only to the territory surrendered by the amendment. This point is made several times in Festo 2. For instance, at pages 1841 to 1842, Justice Kennedy, giving the judgment of the Supreme Court said:-

"In Warner-Jenkinson we struck the appropriate balance [between patentees and the public] by placing the burden on the patentee to show that an amendment was not for purposes of patentability:

'Where no explanation is established, however, the court should presume that the patent application had a substantial reason related to patentability for including the limiting element added by amendment. In those circumstances, prosecution history estoppel would bar the application of the doctrine of equivalents as to that element' Id, at 33, 117 S.Ct. 1040.

When the patentee is unable to explain the reason for amendment, estoppel not only applies but also 'bar[s] the application of the doctrine of equivalents as to that element'. (Ibid)"

69. In Festo 2, the patentee added two limitations which were disclosed for the first time in the application for an amendment (page 1836). The Supreme Court held that the critical issue was what territory those amendments had surrendered (page 1842). It, therefore, remitted the matter back to the CAFC or the District Court to determine what territory the amendments had surrendered.
70. The critical question on this cross-appeal is also what territory was surrendered by Celltech's amendment. Festo 2 does not explicitly give guidance on how the territory question is to be determined in any case. In my judgment, what the court must do is examine the meaning of the terms used in the unamended claim using a literal approach. The critical wording (to be found in paragraph 16 of the Vice-Chancellor's judgment) is crafted round various "positions" which are identified by number. Taking account of the mandated combinations, the unamended claim identifies seven possibilities, one of which must be donor residue. (The seven possibilities include a number of combinations of individual elements or limitations, as in "23 and/or 24"). The "element" is not the position but the requirement that the residue at a specified position be donor residue. No mention is made in the unamended claim of any amino acids other than "donor residues". No claim is made that any position be acceptor residue.
71. Next, the court must, in my judgment, examine the narrowing amendment. The narrowing amendment limited the statement in the claim that "the framework comprises donor residues at at least one of the positions 6, 23 and/or 24, 48 and/or 49, 71 and/or 73, 75 and/or 76 and/or 78 and 88 and/or 91" and reduced it to the following:- "amino acid residues 6, 23, 24 and 49 at least are donor residues". In the amended claim, there are now four elements, all of which must be donor residues.
72. The nature of the narrowing amendment is to delete the elements constituted by donor residue at positions 48, 71, 73, 75, 76, 78, 88 and 91 and to introduce a new limitation that donor residues be present at each of, not one of, positions 6, 23, 24 and 49. Thus to assert infringement of the amended claim by a donor residue in the infringing product at position 24, the patentee would also have to show that there were donor residues at positions 6, 23 and 49. On a literal interpretation, before amendment, a product infringes if there was donor residue in any one of the seven possible cases and after amendment, a product infringes if there is donor residue in any one of four specified positions (being within the seven positions previously specified), provided that in each case there was also donor residue in each of the three other of those four positions. It is important to read the amendment this way for two reasons. First, federal law adopts the literal approach. Second, the patent was not cancelled and reissued. It was simply the subject of narrowing amendment, and this is the only basis on which the amendment could take effect as such. The net effect, where the residue in the alleged infringing product is the equivalent of donor in only one of the specified positions, is that prosecution history estoppel does not apply to that element.
73. The role of position 23 in the amended claim is different from its role in the unamended claim because donor residue at this position now becomes a compulsory as opposed to an optional requirement. However, the element constituted by the requirement for donor residue at position 23 itself remains unchanged. An untechnical example may help to demonstrate that last point. Suppose that there are seven green bottles on a wall in a coconut shy. When a contestant starts to play, he has a choice as to which bottle to hit. But, when he has knocked down six, he has no option but to try to hit the last one. That last bottle is the same as the one that stood there at the beginning of the contest. There has been no change to it. It has remained constant. It has always been a target. The fact that it is now obligatory to hit that one has not changed it at all. Analogies, however, must not be taken too far. I have used a single colour for the bottles in my example because all the elements in issue in this case must have one particular quality, namely be donor residue.
74. Mr Antony Watson QC, for the respondent, submits the territory surrendered by the narrowing amendment was "the ability to assert literal infringement against an antibody that does not have a donor at position 23." (Transcript page 182/9 – 11). However, when in Festo the Supreme Court refers to surrendered territory it is clearly referring to the difference, on a literal interpretation of the patent, between the elements in the original claim and the elements of the claim as narrowed, and not to a legal approximation of that difference. One of the elements of the unamended claim was a requirement for donor residue at position 23. The narrowing amendment reduces the original claim by requiring that that element must co-exist with donor residue at positions 6, 24 and 49. That was a new feature of the claim. But, as I have explained, the element constituted by donor residue at position 23 undergoes no change.
75. Moreover, the result of Mr Watson's analysis is that Celltech cannot claim infringement on the basis of the doctrine of equivalents for a non-donor residue at position 23: that would be to seek to recover the territory which the patentee is presumed to have disclaimed or renounced by the narrowing amendment. Here, however, Celltech relies on the doctrine of equivalents in relation to donor residue at position 23. Donor residue at position 23 is not within the territory surrendered. This court must ignore the fact that the residue in SYNAGIS at position 23 happens to be an acceptor (viz non-donor) residue because Celltech does not put its claim to infringement either on the basis of literal infringement by a non-donor residue at position 23 or on the basis of an equivalent for a non-donor residue at position 23. It is making its claim to infringement on the basis of an equivalent to a donor residue at position 23. This court cannot determine that claim because it is an issue for determination at trial. I make no assumption that Celltech will succeed on that at trial. It may very well fail to do so.
76. For the reasons given in this judgment, I do not accept Mr Watson's analysis that "the mandatory integer was added" (Transcript page 175/25).
77. I appreciate that Mr Watson's argument is good if the starting point is the element constituted by donor residue at a position other than 23, say 6, i.e. if one of the other possibilities in the unamended claim is taken. In relation to that possibility, the requirement for donor residue at position 23 is part of the narrowing amendment. However, Festo was not a case where there were, as here, alternative possibilities (the seven possibilities, as Mr Watson puts it). In Festo there was a single claim into which two limitations were introduced (page 1836). There is nothing in Festo to require the court to apply the effect of a narrowing amendment in relation to some possibilities to every other possibility. It is enough, as I see it, that Celltech can show that the element constituted by donor residue at position 23 is not part of a narrowing amendment for the purpose of one of those possibilities, namely the possibility in the original claim based on the requirement for donor residue at position 23 and/or 24. That requirement is not part of the territory surrendered for the purposes of that possibility: ergo, prosecution history estoppel does not apply.
78. To apply Festo to alternative possibilities is to reduce the protection available to patentees. The importance of that protection is recognised by the Supreme Court in Festo in its rejection of the "bright line rule" in favour of the flexible bar. This case, moreover, extends the doctrine of prosecution history estoppel: we have not been shown any case where the estoppel was applied to an element which is not itself changed by the narrowing amendment and which accordingly appears in the amended claim in the same form as it did in the original claim. Mycogen Plant Science v Monsanto Company (2001) 252 Fed Rep. 2^nd 1306 was said to be authority for the rejection of this distinction but that case was a case where the elements did not remain the same and the case (so far as relied on) merely decides that prosecution history estoppel can apply to a cancellation and reissue of a patent just as it does to a narrowing amendment on the grounds that to hold otherwise would be to elevate "form over substance". But that principle cannot be applied to interpretation of a claim. That has to be interpreted by the application of the doctrine of literal interpretation. To extend Festo in the way required in this case disturbs the balance between the patentee and the competitor which was so carefully struck by the Supreme Court in Festo (pages 1837 to 1839 and 1840 to 1842). Furthermore, the competitor reading the file would see that the requirement for donor residue in position 23 was in both the unamended claim and the amended claim. The rationale for prosecution history estoppel – namely that the patentee should not be able to claim back by the doctrine of equivalents an element he represented to the public to have been surrendered by the narrowing amendment – simply has no application to the facts of this case.
79. Accordingly, the limitation that position 23 should be donor is not part of the territory surrendered. Accordingly, it is not to be assumed that the equivalent of a donor at position 23 is surrendered.
80. It is important to look at the question of infringement as the judge did (Judgment, paragraph 33). Infringement is the acid test of the effect of the amendment. It helps to define the territory surrendered.
81. Support for the judge's conclusion is to be found in Aclara Biosciences v Caliper Technologies 125 F.Supp 2d 391 (ND Cal 200). Some support for the judge's analysis may also be found in AMP Incorporated v Hellerman Limited [1962] RPC 75, HL. In this case an optional feature of a crimping machine, namely a "stop", was incorporated as an essential feature in an amendment made after acceptance of the specification. The question arose whether the amendment was a disclaimer for the purposes of section 31(1) of the Patents Act 1949 or a new claim. The House of Lords held that the amendment was a disclaimer. After analysing various authorities, Lord Denning said this:-

"In view of those cases and others, I am clearly of the opinion that a 'disclaimer' takes place whenever the patentee reduces the ambit of his monopoly: for he thereby reduces his previous claim in its fullest scope and limits it to a narrower scope. This renunciation need not, however, be done in express terms. It is sufficient if it is done implicitly: and this is what I think happened in this case. The patent here is not for a single new device. It is for a combination of old devices. But the combination of these, so it is said, is new and it produces new and useful results. If so, it is patentable and anyone who takes all the essential features of the combination is an infringer. The original specification, however, is framed so widely that a "stop" is not an essential feature of the combination. It includes both crimping tools without stops and crimping tools with stops. If the patent, as originally claimed, were good, it would give the patentee a monopoly in this way: if anyone produced a similar crimping tool which contained all the essential elements of the combination, he would be an infringer. Even though it does not include a 'stop', it would be an infringement; for a 'stop' is not an essential feature of the combination. Now the patentee desires to amend by making the 'stop' an essential feature. By doing so he reduces the ambit of his monopoly for the simple reason that, once he makes the 'stop' an essential feature, he can only complain of crimping tools with stops. He cannot complain of crimping tools without stops. If, after the amendment, anyone should make a crimping tool without a stop, he would not be an infringer, because he would not have taken all the essential features. In the result therefore the patentee by his amendment disclaims all crimping tools without stops and limits his claim to crimping tools with stops. That is, I think, a 'disclaimer' within the Act. Putting it in specialist language, once you add another essential feature to the combination, you produce a subcombination; and amendment so as to limit the scope of the specification to a subcombination which was within the original claims is disclaimer. See Baker Perkins Ltd's application [1958] RPC 267 at 276 by Lloyd-Jacob J, approving the judgment of the Assistant Controller." (page 71).

82. Accordingly, the effect of incorporating an optional feature as an essential feature in the amended claim was to produce a sub-combination of elements, not a new claim. The essential question in the present case is: what was the effect of the narrowing amendment? In my judgment, by making the donor residue at position 23 essential Celltech reduced the unamended claim just as the patentee had done in AMP Incorporated: the analysis of Lord Denning set out above provides support for my analysis of the narrowing amendment and the territory surrendered in this case. I note also that Lord Denning tested the situation by reference to infringement.
83. Accordingly, in my judgment, the judge was correct to conclude that it was open to Celltech to claim that an equivalent of donor residue at position 23 infringed its patent. The requirement for donor residue at position 23 was not within the territory surrendered. As I read his judgment the judge decided this issue on the basis of the territory surrendered, not tangentiality. That explains his use of the term "irrelevant to the amendment" in paragraph 33 of his judgment.
84. In this case, we have looked only at intrinsic evidence on the patent file. However, I would not exclude the possibility in other circumstances of extrinsic evidence, such as evidence of the prior art.
85. Accordingly, I would dismiss the cross-appeal.
Celltech's adjournment application
86. I agree with the Vice-Chancellor that the Celltech (adjournment) application described in his judgment should be rejected. In this case it is not a question of the rebuttal of the presumption but of the identification of the territory surrendered. Festo 3 does not indicate that this is a question of fact.
Celltech's appeal (argument estoppel)
87. I start from the position that the narrowing amendment did not relate to position 23 on the basis that this is the correct interpretation of the narrowing amendment and the basis on which a competitor should have read the file. I would, however, accept as the judge did that there is no logical reason why a claim in argument estoppel should not succeed even if one based on prosecution history does not. What has to be shown is that there was an "unmistakeable assertion" that for the future position 23 could only be a donor residue. Celltech certainly represented to the USPTO that position 23 should be a donor residue. But the judge read the file as meaning - and this is the sheet anchor of his conclusion on argument estoppel – that position 23 could never be anything else. That, as Mr David Kitchin QC for the appellant, in his argument demonstrated, is inconsistent with the teaching of the patent itself. In particular, the specification stated that "in some cases the donor and acceptor amino acid residues may be identical at a particular position and thus no change of acceptor framework residue is required". In my judgment, in the light of that teaching there could not be an "unmistakeable" or "unequivocal" (to use the term used by the judge) assertion that position 23 could never be anything other than donor residue. Mr Watson sought to meet Mr Kitchin's point by submitting that the word "donor" in the file, and in the patent itself, must be read with the qualification in the teaching of the patent referred to above. However, the patent was not drafted in this way. In addition, the judge's conclusion that the effect of Celltech's representations to the USPTO was that the residue at position 23 should be donor residue cannot be so read. His conclusion is unqualified and depends for its legitimacy on what he saw as the unlimited nature of the representation made by Celltech.
88. In addition, nothing was said in the file about conservative substitutions. This factor is not determinative but has some significance in the circumstances of this case. As I see it, as the patentee was not amending position 23, the question of equivalents did not arise for consideration. The court must proceed on the basis that both the patentee and the patent office knew that the doctrine of equivalents would apply to the unamended elements of claim 1.
89. The three representations which the judge identified are very brief. In the first, the terms of the unamended claim and prior art are simply set out. The second did not amount to a representation about the amendment. The third is an analysis of Riechmann (the prior art). I do not consider that in any event any of these passages is sufficiently unambiguous to make an unmistakeable assertion that the residue at position 23 in Celltech's amended claim could not be acceptor, any more than it makes any representation that that residue could be a residue which was neither acceptor nor human.
90. In any event, the judge's conclusion on argument estoppel makes no commercial sense. Why should Celltech have given up the equivalent of an unamended element of its claim? In those circumstances, a competitor is not entitled to conclude that the representations meant more than the exact words used.
91. In the circumstances, I need not deal with the position of Dr Martin.
92. Accordingly, I would allow the appeal.

Longmore LJ

93. Like Arden LJ I gratefully adopt the Vice-Chancellor's exposition of both the science and the United States law relevant to this appeal.
94. Claims 1 and 2 of the patent before it was amended read as follows:-

"1. A CDR-grafted antibody heavy chain having a variable region domain comprising acceptor framework and donor antigen binding regions wherein the framework comprises donor residues at at least one of positions 6, 23, and/or 24, 48 and/or 49, 71 and/or 73, 75 and/or 76 and/or 78 and 88 and/or 91.

2 A CDR-grafted heavy chain according to Claim 1 comprising donor residues at positions 23, 24, 49, 71, 73 and 78, or at positions 23, 24 and 49."

95. After amendment the relevant part of the claim as granted set out the claim and added:-

"wherein, according to the Kabat numbering system, in said composite heavy chain: said CDRs comprise donor residues at least at residues 31 to 35, 50 to 58, and 95 to 202; and amino acid residues 6, 23, 24 and 49 at least are donor residues."

96. Mr Kitchin QC submitted that, although the claim had been amended, the amendment was not a relevant amendment for the purpose of prosecution history estoppel because there was no amendment to the status of Kabat 23 as a donor residue. Both before and after the amendment, position 23 could be an acceptor residue if the amino acid constituting such acceptor residue was a conservative substitution within the United States doctrine of equivalents. The judge accepted this argument stating that the amendment cut down the possibilities claimed. But it did not in any way cut down whether or not the claim covered equivalents of donor residues.
97. While acknowledging my own inexperience in this field, I cannot agree with the judge's conclusion.
98. His conclusion is, in my judgment, inconsistent with the approach of the United States Supreme Court in what may be called Festo 2 which the judge set out fully in paragraph 16 of his judgment but the last three paragraphs of which bear repetition:-

"Just as Warner-Jenkinson held that the patentee bears the burden of proving that an amendment was not made for a reason that would give rise to estoppel, we hold here that the patentee should bear the burden of showing that the amendment does not surrender the particular equivalent in question.

……….

The patentee, as the author of the claim language, may be expected to draft claims encompassing readily known equivalents. A patentee's decision to narrow his claims through amendment may be presumed to be a general disclaimer of the territory between the original claim and the amended claim. Exhibit Supply, 315 U.S., at 136-137, 62 S.Ct. 513 ("By the amendment [the patentee] recognized and emphasized the difference between the two phrases and proclaimed his abandonment of all that is embraced in that difference"). There are some cases, however, where the amendment cannot reasonably be viewed as surrendering a particular equivalent. The equivalent may have been unforeseeable at the time of the application; the rationale underlying the amendment may bear no more than a tangential relation to the equivalent in question; or there may be some other reason suggesting that the patentee could not reasonably be expected to have described the insubstantial substitute in question. In those cases the patentee can overcome the presumption that prosecution history estoppel bars a finding of equivalence.

This presumption is not, then, just the complete bar by another name. Rather, it reflects the fact that the interpretation of the patent must begin with its literal claims, and the prosecution history is relevant to construing those claims. When the patentee has chosen to narrow a claim, courts may presume the amended text was composed with awareness of this rule and that the territory surrendered is not an equivalent of the territory claimed. In those instances, however, the patentee still might rebut the presumption that estoppel bars a claim of equivalence. The patentee must show that at the time of the amendment one skilled in the art could not reasonably be expected to have drafted a claim that would have literally encompassed the alleged equivalent."

99. This indicates that there is a presumption that the territory between the original and the amended claim has been disclaimed but that there will be some cases where the amendment cannot reasonably be viewed as surrendering a particular equivalent. One has, therefore, to ascertain the difference between the original and amended claims and then ask whether the amendment can reasonably be viewed as surrendering a particular equivalent. In my view the surrendered territory, in relation to position 23, is the ability to have that position as either a donor residue or an acceptor residue. Before amendment, position 23 was one of a number of optional places for a donor residue; after amendment it is compulsory that it be a donor residue. Once there is no option but to have 23 as a donor residue, it is impossible, by reason of estoppel, to recover the ground by saying that after all one has the option to make 23 an acceptor residue by virtue of the doctrine of equivalents. It is, therefore, not right to say that the equivalent in question is "simply irrelevant" to the amendment made. It is inconsistent with it.
100. Having identified the surrendered territory one must then ask whether the amendment can reasonably be viewed as surrendering a particular equivalent. The amino-acid alleged by Celltech to be equivalent for the purposes of infringement is the amino-acid threonine. But they cannot rely on threonine being an equivalent because it is an acceptor residue and they have disclaimed any entitlement to an acceptor residue at position 23. After amendment, only a donor residue at position 23 can infringe. That to my mind is not merely a reasonable view of the amendment but, in fact, the only reasonable view. None of the particular reasons for displacing the presumption set out in Festo 2 have any application nor did the judge decide that they did.
101. These particular reasons, however, presuppose that it may be necessary to ascertain the rationale of the amendment. One underlying rationale of the amendment was to deal with the prior art of Riechmann. As the judge himself said, Riechmann had shown residue 23 to be an acceptor residue whereas residue 23 was, by the amendment, required to be a donor residue. Once that rationale has been discerned, it becomes clear that the asserted right to make a conservative substitution is indeed relevant to the amendment made and the judge himself effectively so said in the last sentence of paragraph 39 of his judgment.
102. It follows that I agree with the Vice-Chancellor and would allow the cross-appeal. It also follows that I must reluctantly differ from Arden LJ. In paragraph 13 she helpfully gives what she calls an untechnical example to help demonstrate the point; she supposes seven green bottles on a wall in a coconut shy. The last bottle remains at the end and is the same bottle waiting to be hit as it was at first. But if a requirement of the competition is that any one of the bottles must be green but the others can be a different colour and that it is a green bottle that must be hit, the last bottle may be a white bottle and thus outside the competition altogether. Her example does not, therefore, help to persuade me that position 23 is not part of the surrendered territory, since the ability to have an acceptor residue at that position has been abandoned.
103. I also agree with the Vice-Chancellor in his disposition of Celltech's application for an adjournment and of their own appeal.

Order: Cross appeal allowed; The appeal dismissed with costs of which £60,000 ro be paid on interim basis.

(Order does not form part of the approved judgment)