T 0063/94 (Determination of antigens/AKZO NOBEL N.V.) 18-01-1995

Reasons for the Decision

1. The appeals are admissible.

Late-filed evidence

2. Of the material sought to be relied on by the parties only at a very late stage, the Board was prepared to admit into the proceedings only document (8) published just before the priority date of the patent in suit, because this contains a helpful summary of two site assays with polyclonal antibodies. As to the other material sought to be relied on, the Board regarded this as not filed in due time, and exercised its discretion under Article 114(2) EPC to disregard it, as it appeared no more relevant than other material already on file, and in some cases it was not even clear when it was published or whether it was prior art or not.

The main request: inventive step (Article 56 EPC)

3. The only point at issue in respect of this request is the inventive step of Claims 1 to 7 filed on 14 February 1990. For this purpose it is only necessary to discuss the independent method Claims 1 and 5.

4. The Board sees in document (5) the closest prior art. This document is a review of radioassays and other related labelled reagent techniques. In particular, the article discusses solid-phase "sandwich" or "two-site" assay methods in which use is made of two separate radioisotopically-labelled antibodies that recognise two different binding sites on the same antigen molecule (see, in particular, page 223 and Figure 2). The use of some other non-isotopic labels such as enzyme or fluorescent labels, is also discussed (see page 235 onwards). In its final part (see pages 238 to 239), the article outlines two major obstacles to be overcome, these being the reduction of "background" signals and the improvement of the purity of the labelled reagents, i.e. - in general - labelled antibody. In respect to this latter point, the article states as follows: "Resolution ... is in sight in consequence of the exciting new developments pioneered by Milstein and others involving in-vitro fusion of mouse myeloma and antibody producing spleen cells. These techniques, now under intensive examination in my own laboratory hold out the promise of in-vitro monoclonal antibody production in unlimited amount, giving us for the first time the opportunity to tailor molecules possessing exactly the structural-recognition capacity which we require for use in two-site assay systems." In the conclusion, the view is expressed that the use of monoclonal antibodies will generate antibody-based analytical methods with improved speed, sensitivity and precision (see, page 239, last paragraph of the discussion).

5. In view of the cited prior art document the problem to be solved is to be seen in overcoming one or more of the above disadvantages of the prior art "two-site" or "sandwich" assay methods.

As a solution thereto the patent in suit proposes a method for antigen determination and means therefor (test kits) wherein use is made of at least two different sorts of monoclonal antibodies directed against the same antigen, one of which is labelled. The said method can be carried out in soluble, suspension or solid phase. The solution proposed in independent method Claims 1 and 5 is based fundamentally on the same principle. The major difference between the Claim 1 and Claim 5, is that Claim 5 is further limited to one sort of monoclonal antibody being labelled with particular labels, namely with red blood cells (haemagglutination), polystyrene latex spheres (latex agglutination), finely- suspended dyestuff particles or metal sol particles. Moreover, the method according to Claim 1 excludes by way of a disclaimer "sandwich assays" in which two monoclonal antibodies are used, one of which is insolubilized by bonding to a solid surface and the other is labelled.

On the basis of the whole body of evidence in the present case, including the results presented in the patent specification (see, for example, Table 1) and put forward during the prosecution of the case (see, for example, the Table reported in the minutes of oral proceedings before the Opposition Division on 5 October 1993), the Board is satisfied that the method of Claim 1 or Claim 5 solves the underlying technical problem.

6. The development by G. Köhler and C. Milstein [see document (4)] of the methods of in vitro production of monoclonal antibodies constituted a major breakthrough in the field of immunochemistry, as it permitted the synthesis of unlimited amounts of single species of antibody directed against a single antigenic site. It is also generally accepted that the applicability of monoclonal antibodies to immunoassays was immediately apparent to the skilled person.

7. The relevant question to be asked in the present case is whether the skilled person, faced with the underlying technical problem, would have in a straightforward manner replaced the polyclonal antibodies by monoclonal antibodies in a "two-site" or "sandwich" assay with the reasonable expectation of any improvement.

8. With respect to the above question, the divergent positions of the parties were essentially as follows.

According to the Appellants, the skilled person had definite reasons to proceed to the replacement in question in view of the many known advantages of monoclonal antibodies, such as their availability in large amounts, their specificity for single antigenic determinants and the constant affinity therefor [see document (7)]. In their submissions, there was no prejudice in the art against such a replacement; on the contrary, there was an incentive to use monoclonal antibodies [see also documents (3) and (5)]. In the view of the Appellants, the person skilled in the art would have empirically tried for a combination of monoclonal antibodies which produced as good an assay as possible. In their submission, the method suggested in the patent in suit to obtain a good combination was just this (see the patent specification, page 4, lines 48 to 53).

The Respondent maintained that, although the advantages of monoclonal antibodies such as their reproducibility and their high specificity were recognised in the art, the skilled person was also aware of the fact that they had a lower affinity than polyclonal antibodies. This lower affinity was due partly to the fact that monoclonal antibodies were prepared from spleen cells whereas affinity maturation of antibodies takes place in the bone marrow, and partly to the difficulties in selecting with respect to affinity. The skilled person in order to improve "sandwich" or "two-site" assays had several options open such as, for example, the search for better polyclonal antibodies or the substitution of only the first polyclonal antibody by a monoclonal antibody. However, he or she would not have readily replaced both antibodies by monoclonal antibodies in such assays, being aware of the lower sensitivity of the latter.

9.1. The Board observes that Claims 1 and 5 at issue are quite generally worded and cover methods for antigen determination in which at least any two different monoclonal antibodies directed against the same antigen are used, irrespective of their specificity and/or affinity characteristics. As stated also in the specification (see page 4, lines 48 to 53), "not every arbitrary combination of two or more antibodies will satisfy reasonable requirements regarding sensitivity and specificity, and it is even possible that such a combination will not be at all effective. Establishing the best combination depends exclusively on empiricism." (emphasis added). Thus, the claimed innovation lies merely in the proposal of using a combination of at least two monoclonal antibodies in "sandwich" or "two- site" assays, the burden being then left to the skilled reader to find empirically the appropriate combination of antibody partners. As a matter of fact, the present patent specification gives virtually no contribution to lessening this burden of work.

9.2. The skilled person would have readily derived the suggestion of replacing both polyclonal antibodies by monoclonal antibodies in "sandwich" or "two-site" assays from document (5) (see, in particular, page 239). The reproducibility of monoclonal antibodies, their availability in large amounts as well as their specificity and the constant affinity for single antigenic determinants would have constituted an incentive in this direction. The skilled person would have seen in monoclonal antibodies the suitable substitutes for products (polyclonal antibodies) which had disadvantages such as the non-reproducibility and low specificity. Based thereupon, the skilled person would have expected the replacement of not only one but both polyclonal antibodies by monoclonal antibodies to bring about an improvement of the assays. The skilled person would, of course, have been aware of the fact that the identification of monoclonal antibodies with affinity and specificity characteristics required for the assay of an analyte could be in some cases arduous and time-consuming. However, these technical difficulties, which underlie also the present patent specification, would not have constituted for the skilled person a prejudice against the replacement because, he or she knew that the finding of a suitable antibody, whether by conventional or hybridoma-based techniques, varied considerably from one analyte to another and was merely a matter of time, perseverance, and luck, and that the benefits of finding a suitable monoclonal antibody which could be reproducibly produced justified in general the efforts required.

9.3. In the Board's view, the arguments based on the affinity characteristics of monoclonal antibodies in comparison with polyclonal antibodies which have been put forward by the Respondent (see point 8, third paragraph, above) are irrelevant in the present case. In fact, the underlying technical problem here is not specifically the improvement of the sensitivity of "sandwich" or "two-site" assays, but the overcoming of one or more of the disadvantages of these assays, i.e. - in other words - the generic improvement of any aspect of such assays. The solution proposed, as reflected by the general wording of the claims, embraces any combination of at least two different monoclonal antibodies against the same antigen, including combinations which are no better than would be expected from adding the affinities of the two separate antibodies. Such solution is fully in line with the suggestion in document (5) to replace polyclonal antibodies by monoclonal antibodies.

9.4. In any case, as the skilled person was certainly aware of the fact that sensitivity of an assay determined the lower limit of detection, it was desirable to ensure the highest possible sensitivity and this was merely a matter of finding empirically at least two appropriate antibody partners.

9.5. On behalf of the Respondent it was also argued that the skilled person would not contemplate operating within the scope of the claims because there was no certainty that a successful assay against any particular antigen could be produced. This would be a strong argument, if the claims had been directed to an assay against a particular antigen, and the description had shown that this was difficult but provided a reproducible way of overcoming these difficulties. Here, however, the claims cover assays against any antigen, and the suggested way of carrying out the claimed method is embarking on a process of trial and error using only the information to be found in the prior art on polyclonal assays and the making of monoclonal antibodies. If this prior art knowledge is not sufficient to make an assay against a particular antigen, the Board can find nothing in the specification that would enable the skilled man to do so. The specific embodiments described identify (e.g. by deposit) neither the monoclonal antibodies used, nor the hybridomas used to make them, in a way that would enable the specific embodiments to be reproduced. The specific embodiments thus convey only the information that pairs of unidentified monoclonal antibodies do exist with which much better results can be achieved than by using a single sort of these unidentified antibodies. This information does not reduce the amount of work anybody else would have to do to come up with an assay against an antigen they wish to test for. The Board understood from the co-inventor who gave evidence, that in Example 1 the five antibodies referred to as A, B, C, D and E had in fact been selected after testing the antibodies from more than a hundred hybridomas producing monoclonal antibody against HCG, so that the amount of work in producing a suitable assay against any one antigen is very substantial. Yet in the patent, the Respondent treats this amount of work as reasonable for making something falling within his claims. On this basis it seems fair to the Board to assume that a skilled person seeking to solve the problem above posed would also have thought this amount of work reasonable, and would, on the basis of the indications existing in the prior art, inevitably have arrived at something falling within the scope of Claims 1 and 5, when adapting prior art polyclonal assays at least for some antigens.

9.6. The Board, therefore, concludes that the solution proposed in general terms in both Claim 1 and Claim 5 at issue is not different from what a person of average skill would have derived in a straightforward manner from the prior art. In this respect, it is further observed that both the manner in which the antigen- antibody reaction complex is formed (in solution, in suspension, in solid phase; cf. Claims 1 and 5) and the specific nature of the label recited (Claim 5) are features of ancillary importance which in the present case do not contribute to inventive step, being for the skilled person merely routine aspects of putting the assay method into practice [in this respect, see, for example, the review of the state of the art made in the introductory part of document (8)]. In respect of the labels, it is observed that the patent specification itself acknowledges at page 1, lines 27 to 30 that these labels have been employed to label [polyclonal] antibodies. Whereas reliance on the requirement of such labels may serve to render a claim novel, it does not introduce any new aspect into the consideration of inventive step.

10. In conclusion, the Board considers that Claims 1 and 5 at issue lack an inventive step and, therefore, the main request which contains them cannot be allowed.

The auxiliary requests: inventive step (Article 56 EPC)

11. As regards auxiliary requests I to V, as the claims involved are essentially more restricted than the claims of the main request, the Board sees no objections to them under Articles 54 and 123 EPC. However each of them contains at least one claim that lacks inventive step.

12. Thus, Claim 1 of auxiliary request I is essentially identical to Claim 5 of the main request, and thus lacks an inventive step for the reasons given above in relation to that claim.

13. Claim 1 of auxiliary request II is identical to Claim 1 of the main request, and lacks an inventive step for the reasons given above in relation to that claim.

14. Claim 1 of auxiliary request III is essentially the same as Claim 5 of the main request, subject to the further limitation that the label is limited to being finely- suspended dyestuff particles or metal sol particles. As explained in point 9.6 above, this is known for assays using polyclonal antibodies, so that the arguments for lack of inventive step remain as for Claim 5 of the main request.

15. Claim 1 of auxiliary request IV differs from Claim 1 of the main request essentially in that the manner in which the immunological determination is made is specified, this being agglutination. However, as already stated (see point 9.6, above), the Board is of the opinion that this feature cannot contribute to inventive step because for the skilled person it is merely a question of routine to decide the manner in which the antibodies are reacted with the antigen. In fact, the principles of antibody-antigen reactions are fundamentally the same in suspension, soluble or solid phase. Moreover, no particular improvement and/or advantage linked to agglutination has been put forward by the Respondent. Thus, in the Board's view, also Claim 1 of auxiliary request IV does not involve an inventive step.

16. Claim 5 of auxiliary request V is essentially identical to Claim 5 of the main request, and thus lacks an inventive step for the reasons given above in relation to that claim.

17. As each of the auxiliary requests contain a claim for which an inventive step cannot be recognized, none of these auxiliary requests is allowable.