T 0805/06 (Membrane protein/HIRANO TOSHIO) 13-12-2007

Main request

Novelty (Article 54(1)(2)(3) and (4) EPC; Article 54(4) and (5) EPC 1973)

1. The respondent maintains that the subject-matter of claims 1 and 7 lacks novelty in view of document D1, disclosing the "HM1.24 antigen" of molecular weight of approximately 30 kDa (see document D1a, page 6, end of first paragraph) and antibody HM1.24 recognising this antigen (ibidem, page 3, end of first paragraph).

2. For a prior publication to take away the novelty of a claim, according to the jurisprudence of the boards of appeal, the subject-matter of the claim must be clearly and unambiguously disclosed in the prior publication, and also in a manner which enabled the skilled person to carry it out. The question thus arises whether or not document D1 provided sufficient technical information for the skilled person to arrive, without exercise of undue burden, at the "HM1.24 antigen" and monoclonal antibody HM1.24 recognising this antigen.

Enabling or non-enabling character of document D1 for the monoclonal antibody of claim 7

3. Many years after Köhler and Milstein's 1975 publication of the results of their pioneer work concerning the production of monoclonal antibodies by using the hybridoma technique, the production and screening of hybridomas secreting a monoclonal antibody with specific, desired features consisted basically of a sequence of widely known routine technical steps where all that was normally called for was perseverance. However, as transpires from the analysis below (see points 4 and 5 infra), the selection of the immunogen used to raise the antibodies as well as the screening method for selecting the hybridoma secreting a monoclonal antibody with a desired specificity still remained critical steps.

4. As regards the immunogen, the authors of document D1 (see document D1a, page 2, last full paragraph), used KPC32 and MHB cells. These were in-house cell lines established in the laboratory of the authors of document D1 (see document D4, page 1922, r-h column, under "Material and Methods"). Therefore, serious doubts arise as to public availability of the immunogen (i.e., the KPC32 and MHB cells) to the skilled person wishing to reproduce the teaching of document D1. The skilled person using cells other than KPC32 and MHB cells (possibly not bearing the "HM1.24 antigen" on their surface) would of course possibly miss the hybridoma/antibody ("HM1.24") looked for.

5. As for the screening method, this is a fundamental step allowing the selection of a hybridoma secreting a monoclonal antibody characterised by a well defined spectrum of binding activities from a great number of hybridomas. The authors of document D1 (see the passage bridging pages 2 and 3 of document D1a) indeed collected 248 hybridomas (termed "antibodies" in document D1a) and applied "FCM and an enzymatic antibody method" to various cell lines in order to arrive at the 4 kinds of antibodies (HM series) shown in Table 1 of document D1, of which antibody HM1.24 exhibited the highest specificity to unfixed plasma cells surface antigens (see page 3, lines 7-9 of document D1a).

However, the skilled person was not taught which specific cell(s) among the many possible cells (see ibidem: "peripheral blood cells, bone marrow cells, lymph nodes of normal individuals and hematopoietic tumor patients") have actually been selected for performing the screening. Moreover, the cells listed in Table 2 of document D1, taken from patients ("T.K.", "H.M.", etc) suffering from lymphoid malignancies, were also not available to the skilled person wishing to reproduce the teaching of document D1. Therefore, in the absence of the above critical information as to how the screening method had to be carried out, the skilled person was prevented from arriving at the four antibodies ("HM series") referred to in Table 1 of document D1, let alone isolating a monoclonal antibody designated HM1.24.

6. The respondent argues that the skilled person was in a position to perform the screening against the publicly available cell line RPMI 8226 in order to select an antibody having the binding profile shown in Fig. 2 of document D1 and/or having the capacity of precipitating a "broader band of MW30kD". However, while the RPMI 8226 cell line was publicly available from the Japanese Cancer Research Resources Bank (see document D4, page 1922, lines 13-14 from the heading "Materials and Methods"), in the board's view, the skilled person would have to establish a binding curve (7 different concentrations of **(125)I-labeled free antibody in abscissa vs. the bound form) for each of the 248 hybridomas/antibodies referred to on page 2, line 3 (from the bottom) of document D1a, which borders on undue burden. This conclusion extends to the possibility of using the antibody's capacity of precipitating a "broader band of MW30kD", which would require that an autoradiography of immunoprecipitates (as shown in Fig. 5, lane 1 of document D4) be performed for each of the 248 hybridomas/antibodies.

7. Furthermore, binding to RPMI 8226 cells and/or precipitating a "broader band of MW30kD" were necessary but not sufficent conditions for arriving at monoclonal antibody HM1.24. This is shown by post-published document D4, cited by the respondent for questioning the novelty. On page 1924 of this document (see under "Results") it is stated that the screening for obtaining monoclonal antibody HM1.24 had to be made not only against RPMI 8226 cells but also against KPC-32, Raji, CEM and THP-1 cells. After this first screening, three hybridomas were selected for their specificity for plasma cells. Of the three, two antibodies (implicitly binding inter alia to RPMI 8226 cells) were discarded and only moab HM1.24 was retained for its best behaviour in cytometric analysis and its cytolytic activity against RPMI cells. The skilled person, however, could not derive all this critical information from document D1.

8. Finally, the board observes that the skilled person could not make use of hybridoma FERM BP-5233 producing monoclonal antibody HM 1.24 because this hybridoma had been deposited on 14 September 1995 (see document D3, page 6, lines 13-17), i.e., after the publication date of document D1 (1992) and after the priority date (15 October 1993) of the patent in suit.

9. In conclusion, the board considers document D1 as a non-enabling disclosure for the skilled person wishing to obtain monoclonal antibody HM1.24, which the respondent argues falls under the definition of present claim 7.

Enabling or non-enabling character of document D1 for the polypeptide of claim 1

10. As regards the antigen recognised by monoclonal antibody HM1.24, which the respondent argues falls under the definition of present claim 1, the only technical information which can be taken from document D1a is that "the antigens were immunoprecipitated by a lactoperoxidase method as a broader band of HM1.24 of MW30KD from dissolved RPMI8226 cells labeled with **(125)I" (see document D1a, page 6, end of first paragraph). More details about this way to proceed are given in document D4, page 1924, r-h column, under "Immunoprecipitation": the "broader band of HM1.24 of MW30KD" was the result of the following operations: (i) RPMI 8226 cells were surface-labeled with **(125)I (which bound to any antigen on the cells' surface); (ii) the cells were solubilised; (iii) monoclonal antibody HM1.24 was added (whereby only the antigen recognized by this antibody could precipitate from the clear solution); (iv) a SDS-PAGE and an autoradiography were performed on the (re-dissolved) immunoprecipitate yielding an autoradiography as shown in Fig. 5, lane 1 of document D4.

11. The above steps (i) to (iv) show that the only route open to the skilled person towards the "broader band of HM1.24 of MW30KD" occurred via the obligatory monoclonal antibody HM1.24 (cf. step (iii)). This is because simply lysing RPMI 8226 cells (with or without **(125)I-labeling), electrophoresing (SDS-PAGE) and taking from the gel a band around 30 kDa would have resulted in a preparation contaminated with a great number of proteins having a molecular weight around 30 kDa, but which failed to bind to monoclonal antibody HM1.24 and hence to undergo immunoprecipitation.

12. But since document D1 did not enable the skilled person to arrive at monoclonal antibody HM1.24 (see point 9 supra), the board must conclude that this document was also non-enabling for the skilled person wishing to isolate the "HM1.24 antigen" of molecular weight of approximately 30 kDa, which the respondent argues falls under the frame of present claim 1.

Document D5

13. In a different line of argument, the respondent maintains that the skilled person following the instructions in document D5, representing prior art pursuant to Article 54(3) EPC, would inevitably generate a hybridoma expressing a monoclonal antibody that recognised the BST2 antigen, given that the immunogen was the same in both document D5 and in the patent in suit.

14. The board agrees that document D5 discloses a method for the generation of monoclonal antibodies using the synovial cell line SynSV6-14 as an immunogen, the latter being identical to that used in the patent in suit to generate antibodies recognising the BST2 antigen (compare page 7, line 10 of document D5 with page 7, line 45 of the patent). It is also conceded that immunisation with the cell line SynSV6-14 as an immunogen would generate a great number of hybridomas (cf. the 248 hybridomas referred to in document D1) possibly including a hybridoma expressing a monoclonal antibody that recognises the BST2 antigen.

15. However, the idea that said collection of hybridomas/antibodies itself would anticipate an invention relating to a well defined hybridoma/antibody which may be contained therein somewhere, cannot be sustained. Rather, this well defined hybridoma/antibody would only become part of the state of the art if its existence had recognisably been made publicly available (see decision T 301/87, OJ 1990, 335, point 5.8 concerning a DNA contained somewhere in a cDNA gene bank).

16. As regards the question whether or not document D5 provides instructions as to how a hybridoma expressing a monoclonal antibody that recognises the BST2 antigen has to be isolated from said collection of hybridomas/antibodies, it is stated on page 8, lines 5 of this document that the clones were screened against RASV5-5 and NFSV-1 cells to yield hybridomas RF3 and SG2 (ibidem, page 7, line 49) secreting antibodies RF3 and SG2. The latter antibodies differ from the antibody referred to in present claim 7 because they recognise a protein of 318 amino acids (ibidem, page 10, line 22) which is clearly different from the 180 amino acid-long BST2 antigen of present claim 1 (see SEQ ID No 1). Accordingly, the above question about the teaching by document D5 of means for isolating the hybridoma/antibody against the BST2 antigen, has to be answered in the negative. Hence document D5 does not render publicly available this well defined hybridoma/antibody.

17. In view of the foregoing, the subject-matter of granted claims 1 to 7 is new in the sense of Article 54(1) EPC 1973.

Article 56 EPC

18. The question to be answered is whether or not the claimed subject-matter (including the polypeptide having the amino acid sequence defined by SEQ ID NO. 1 according to claim 1 and the monoclonal antibody according to claim 7) is obvious in the light of the prior art.

19. As emphasised under points 8 and 12 supra, following the instructions in document D1 did not provide by itself an enabling disclosure for obtaining the claimed subject-matter. Stated otherwise, the lack of sufficient information in document D1 prevented the skilled person starting from document D1 from arriving at the claimed subject-matter.

20. Nevertheless, it is the respondent's view that starting from document D1 as closest prior art, the skilled person could derive from this document supplemented by the common general knowledge an alternative and obvious way to the claimed monoclonal antibody/BST2 antigen, which route consisted in using the publicly available RPMI 8286 cells as an immunogen, and selecting a hybridoma secreting an antibody binding to a 30 kDa protein. The antibody, in the respondent's opinion, could be used for purifying the antigen.

21. The board firstly notes that the approach taken by the authors of document D1 was that of identifying a protein associated with the terminal differentiation of B cells, whereas the claimed subject-matter purports to identify and isolate a protein associated with pre-B cell proliferation. Doubts thus arise whether the skilled person looking for a protein associated with pre-B cell proliferation would have turned to document D1, dealing with a protein associated with the terminal differentiation of B cells.

22. However, even assuming in the respondent's favour that the skilled person would have turned to document D1, the board is not convinced by the respondent's line of argument that an alternative obvious route to the claimed subject-matter existed (i.e., using the publicly available RPMI 8286 cells as an immunogen, and selecting a hybridoma secreting an antibody binding to a 30 kDa protein). This is because selecting hybridomas/antibodies on the basis of their capacity of binding to a band around 30 kDa in a gel afforded no certainty that the "right" hybridoma/antibody had been selected, owing to the great number of proteins having a molecular weight around 30 kDa (see point 11 supra). Moreover, once a protein of 30 KD had been purified via this antibody, no amino acid sequence information or biological test were available to confirm that the "right" protein (and hence antibody) had been picked up.

23. Therefore, the claimed subject-matter (including the polypeptide having the amino acid sequence defined by SEQ ID NO. 1 according to claim 1 and the monoclonal antibody according to claim 7) is not obvious from document D1 and/or any other prior art document before the board.

Sufficiency of disclosure (Article 83 EPC)

24. Sufficiency of disclosure was also cited as a ground of opposition. In the decision of the first instance, it was not assessed in respect of the subject-matter now claimed on appeal since the appellant filed new claims before the opposition division. Taking into account the length of the proceedings, the board decides to make use of the provisions of Article 111(1) EPC to evaluate by itself whether or not the claimed invention is sufficiently disclosed.

25. The board firstly observes that the hybridoma secreting antibody RS38 had been deposited according to the Budapest Treaty (see paragraph [0079] of the patent in suit). The respondent never argued that the patent did not provide sufficient information for the skilled person to obtain antibody RS38 or to isolate the intact protein of claim 1.

26. Rather, the respondent argues that the patent gives no guidance as to how to modify/truncate the amino acid sequence defined by SEQ ID NO. 1 and still achieve the function of supporting pre-B cell growth.

27. Methods for obtaining modified or truncated proteins by chemical and/or genetic engineering methods were part of the common general knowledge of the skilled person at the priority date of the patent in suit (see e.g., paragraphs [0051] to [0056] of the patent). A technique for monitoring/measuring pre-B cell growth is disclosed in paragraphs [0097] and [0098] of the patent. There is thus no reason to believe that the selection of modified/truncated forms of the amino acid sequence defined by SEQ ID NO. 1 still achieving the function of supporting pre-B cell growth could be achieved only with an undue burden of experimentation. No evidence to the contrary has been submitted by the respondent.

28. As for the respondent's argument that the patent provides no guidance as to how to select monoclonal antibodies according to claim 7 different from the exemplified antibody (RS38), once monoclonal antibody RS38 and the intact or modified/truncated protein were in the skilled person's hands, it would have been common practice to use these molecules as immunogens and means for screening hybridoma panels for isolating further hydridomas secreting antibodies binding to these molecules. This situation is quite different and not comparable to that encountered by the skilled person attempting to reproduce the teaching of document D1 without having either the antibody or the protein at hand (see points 8, 12 and 22 supra).

29. The board concludes that no case of insufficiency of disclosure has been made out.