T 1911/17 (BNP(1-32) epitope specific antibodies/BIORAD) 11-05-2021

Admissibility of the appeals

1. Both appeals comply with the requirements of Articles 106 to 108 and Rule 99 EPC and are admissible.

Admittance of document D33 (Article 12(4) RPBA 2007)

2. Document D33 was filed with appellant II's (opponent) statement of grounds of appeal. Appellants I (proprietors) requested that the document be disregarded.

3. Article 12(4) RPBA 2007 stipulates that everything presented by the parties with the statement of grounds of appeal or the reply is to be taken into account by the board if and to the extent it relates to the case under appeal and meets the requirements of Article 12(2) RPBA 2007. Yet Article 12(4) RPBA 2007 further provides that, of those materials, the board may hold inadmissible those "facts, evidence or requests which could have been presented or were not admitted in the first instance proceedings".

4. The opponent argued that document D33 was filed in response to an argument not made by the proprietors until the oral proceedings before the opposition division. Therefore, filing document D33 with the statement of grounds of appeal was the earliest point in time when the document could have been filed. The proprietors argued that the argument had been known to the opponent since the proprietors' reply to the notice of opposition and should thus have been filed earlier.

5. In the context of inventive step, the discussion of the data in the patent on BNP(1-32) detection by the antibody 24C5 and the discussion of the opponent's data disclosed in document D11, it was noted in passing in the proprietors' reply to the notice of opposition that "these discrepancies could also be due to a difference in the 24C5 antibody commercialized by HyTest in 2007 (described in D1) and used by the Patentee at that time, and the 24C5 antibody used 8 years later by the Opponent to perform the experiments of D11" (see page 12, third paragraph). In response, the opponent did not comment on this allegation, relying instead on the experimental data in pre-published documents D1, D3 to D5 (allegedly confirmed by documents D6, D9, and D11) to disprove the data in the patent in suit, and commented on the data presented by the proprietors in document D28. In the further course of the written proceedings, neither the opposition division nor the proprietors addressed the argument that the antibody 24C5 might have been different. It was only at the hearing before the opposition division that the argument resurfaced in the context of the novelty of the subject-matter of the main request over the disclosure in document D1. The proprietors' statement is recorded in the minutes as follows: "[...] it must be that the 24C5 antibody available in 2007 and the one available in 2010, are indeed different antibodies".

6. Given these circumstances the board considers that there was no reason why the opponent should have filed document D33 in the opposition proceedings. Consequently, the board decided to take document D33 into account.

Main request, auxiliary request 1 - claim 7; auxiliary request 2 - claim 2

Clarity (Article 84 EPC)

7. Claim 7 of the main request and claim 2 of each of auxiliary requests 1 and 2 derive from a combination of independent claim 9 as granted with (i) the feature of dependent claim 10 as granted ("wherein said ligand is unable to bind a sequence of amino acids of the BNP(1-32) and/or proBNP(1-108) sequence which does not comprise the FGRKMDR epitope in its entirety") and (ii) the feature "and wherein the residues F11, K14 and R17 are essential for binding of said ligand to the FGRKMDR" from the description page 10, lines 19 to 20 of the application as filed.

In view of decision G 3/14 of the Enlarged Board of Appeal, this type of amendment is one where it is appropriate to examine compliance with Article 84 EPC (see also Case Law of the Boards of Appeal, 9th edition, 2019; II.A.1.4 and IV.C.5.2.2).

8. The opponent argued that the claims lacked clarity because the meaning of the ligand-characterising features (i) "unable to bind a sequence of amino acids of the BNP(1-32) and/or proBNP(1-108) sequence which does not comprise the FGRKMDR epitope in its entirety", and (ii) "wherein the residues F11, K14 and R11 are essential for binding of said ligand to the FGRKMDR epitope" was inconsistent.

9. On the one hand, the feature "said ligand is unable to bind a sequence of amino acids of the BNP(1-32) and/or proBNP(1-108) sequence which does not comprise the FGRKMDR epitope in its entirety" is explicit in requiring the epitope to be present in full in order to be bound by the claimed ligand. According to this feature, each of the amino acids is thus essential for the claimed ligand to bind to the epitope FGRKMDR.

10. There was no dispute about this interpretation among the parties, who also referred to paragraphs [0032] and [0045] of the description to support it.

11. The opponent argued that the feature "and wherein the residues F11, K14 and R17 are essential for binding of said ligand to the FGRKMDR epitope" would be understood to mean that only certain amino acids of the epitope FGRKMDR, namely at positions F11, K14 and R17, were essential for binding, i.e. not all amino acids of the epitope. In contrast, the proprietors submitted that the feature provided a further limitation of the binding properties of the ligand, namely that it bound to three specific amino acids within the epitope.

12. Hence, one feature teaches that all seven amino acids of the epitope are essential for binding whereas the other feature teaches that only three of them are. This inconsistency between the meaning of the two features results in a lack of clarity about the binding properties of the ligand.

13. Consequently, claim 7 of the main request and auxiliary request 1 and claim 2 of auxiliary request 2 do not comply with the requirements of Article 84 EPC.

Auxiliary requests 3 and 4 - claim 2

14. Claim 2 of both auxiliary requests 3 and 4 relates to a "monoclonal antibody produced by the hybridoma according to claim 1" (see section IX. above). It is undisputed that this is the antibody referred to as 20G7 in the patent. Inventive step is assessed below with regard to that antibody.

Inventive step (Article 56 EPC)

Closest prior art

15. The board agrees with both parties that document D1 represents a suitable springboard for assessing inventive step.

16. Document D1 discloses that peptides derived from brain natriuretic peptide (BNP) precursor (proBNP), BNP and the N-terminal fragment of proBNP (NT-proBNP) are used as biomarkers of heart failure. The anti-BNP(1-32) and anti-proBNP(1-108) antibody 24C5, generated by immunisation with the BNP(11-22) fragment, is successfully used in assays for the immunodetection of BNP from human plasma samples. The relevant detection limits were found to be 0.4 ng/L for BNP (see abstract and results).

Difference

17. The subject-matter of claim 2 differs from the anti-BNP antibody 24C5 of document D1 in that it relates to an anti-BNP antibody named "20G7", i.e. the one produced by the hybridoma according to claim 1.

18. Both parties accepted that the two antibodies shared the feature whereby they both bind to BNP(1-32) and proBNP(1-108), and that they differed on account of their amino acid sequences, i.e. their structure.

19. There was dispute about whether a further difference resided in the epitope to which the two antibodies bind, i.e. whether or not the antibodies bind to the same epitope.

20. The proprietors did not dispute the opponent's submission, based on documents D9, D11 and D32, that, like antibody 20G7, antibody 24C5 was specific for the epitope FGRKMDR. However, they submitted that this was not known to the person skilled in the art before the priority date of the patent.

21. The epitope to which an antibody binds is an intrinsic consequence of an antibody's structure. Furthermore, it has not been argued that the epitope-specificity resulted in an effect in addition to the binding of the antibodies to BNP and proBNP. Hence, in this situation, knowing the epitope to which antibody 24C5 binds is not crucial for arriving at a proper formulation of the technical problem.

22. Consequently, the difference between the closest prior-art antibody 24C5 and the antibody 20G7 referred to in the claim is their structure.

Technical effect of the difference

23. The opponent submitted that, in view of the proprietors' position that antibody 20G7 performed better than antibody 24C5, the burden to proof this was with the proprietors.

24. The "legal burden of proof" is on the party who relies on a legal consequence arising from an alleged positive fact. Accordingly, the "legal burden of proof" is determined by the legal cases which each party presents. The "legal burden of proof" (unlike the "evidential burden of proof") does not shift (see decision T 301/95, OJ 1997, 519, point 6.2.3 of the Reasons). Whether the burden is discharged or not is assessed by the board in accordance with the appropriate standard of proof on the basis of all the evidence before it. If the party bearing the "legal burden of proof" fails to demonstrate to the required degree the fact(s) on which its legal case rests, the board has to decide against that party. There is no shift of the legal burden of proof in the appeal proceedings. Although an appellant must argue on appeal why the contested decision was wrong, this does not result in a shift of the legal burden of proof on the substance (see decisions T 1210/05, point 2.3 of the Reasons and T 1608/13, point 3.1 of the Reasons).

The legal burden of proof lies with the opponent to establish that the claimed invention lacks an inventive step. The opponent must therefore set forth the state of the art which makes the claimed invention obvious to the person skilled in the art. If, in support of an inventive step, a patent proprietor alleges that the claimed invention has advantageous properties or effects, then the legal burden of proof for the alleged improvement over the prior art rests upon them (see decision T 97/00, point 3.1.6 of the Reasons).

25. The proprietors submitted that the improved properties of antibody 20G7 compared to antibody 24C5 were demonstrated by examples 7, 8, 10 and 13 to 15 of the patent.

26. Of those examples, only examples 8 and 10 are relevant for determining the effect of the difference because they compare both antibody 20G7 and antibody 24C5 directly.

26.1 The examples test the binding to BNP(1-32) and proBNP(1-108), respectively, of three different antibodies, including the antibodies 20G7 and 24C5, in a sandwich-ELISA format using rabbit polyclonal antibody L21016 and hinge 76 antibody, respectively, as capture antibodies.

As shown in Table 3 relating to example 8, antibody 20G7 linearly detects BNP(1-32) at concentrations ranging from 20 to 10 000 pg/ml (see also Figure 5) whereas antibody 24C5 does not.

As shown in Table 5 relating to example 10, antibody 20G7 linearly detects proBNP(1-108) at concentrations ranging from 20 to 10 000 pg/ml (see also Figure 6) whereas antibody 24C5 does not.

Comments in the patent on the results of examples 8 and 10 are that the antibody 24C5 "behave[s] quite differently from the 20G7 antibody", that "20G7 is much more suitable than the 24C5" antibody in the assay formats used, and that antibody 24C5 is "not very effective or not at all effective in detecting BNP(1-32) and proBNP(1-108), even at high concentrations of the analyte".

27. According to the proprietors, the results in document D28 confirmed that the antibody 20G7 displayed better reactivity towards BNP(1-32) than the antibody 24C5 when no capture antibody was used. However, since the experimental set-up in document D28 is not a sandwich-ELISA, it has not been further considered by the board (see point33. below).

28. The opponent submitted that the results of examples 8 and 10 were not suitable to demonstrate the superiority of antibody 20G7 and pointed inter alia to documents D1, D11 and D32. In this context, document D33 provided uncontested evidence that the antibody used in the experiments disclosed in documents D11 and D32 was the same as that used in the patent.

29. Document D1 discloses that "[t]he antibody combination 50E126-32[capture]-24C511-22[detection] manifested the highest detection limit in 1-step sandwich IFA with both synthetic and endogenous antigens ...".

Inter alia, document D11 discloses data (see Figure 4) from a sandwich immunoassay with BNP and proBNP using antibody 50E1 as the capture antibody and 24C5 as the detection antibody. The results show that the 24C5 antibodies "recognize with a very good performance both BNP and proBNP (detection limit is below 0.5 pg/mL)".

Document D32 compares the detection characteristics of antibodies 20G7 and 24C5 towards BNP(1-32) and proBNP(1-108) in a sandwich-ELISA assay using antibody 50E1 as the capture antibody. It discloses that "[m]ab24C5 and mAb20G07 behave quite similarly".

Hence, all three documents demonstrate that antibody 24C5 is very effective at detecting BNP(1-32) and proBNP(1-108). Moreover, document D32 shows similar detection properties for antibody 24C5 and antibody 20G7.

30. Consequently, there seems to be a contradiction between the detection characteristics of antibodies 20G7 and 24C5 disclosed in the patent and those disclosed in documents D1, D11 and D32.

31. The sandwich-ELISA assays in the patent and in documents D1, D11 and D32 were carried out with different capture antibodies. The specificity of the capture antibody determines which parts of the captured molecule are accessible for the detection antibody. Hence, the differences in the detection characteristics could be attributed to the different assay conditions.

32. As submitted by the opponent with reference to the proprietors' submissions in the opposition proceedings, the proprietors have accepted that assay conditions may influence an antibody's detection characteristics:

"The Opponent also contradicted the statement of the Patentee according to which 'the two commercial antibodies 24C5 and 26E2 are not very effective or not at all effective in detecting BNP(1-32), even at high concentrations of the analyte'.

We wish to point out to the Opponent that this statement was made in the context of a specific immunoassay using the L21016 rabbit polyclonal antibody ...

As well-known from the skilled person, different results can be obtained in immunoassays using detection and capture antibodies according to the capture antibody used. This aspect is by the way confirmed in D1 which indicates that a specific antibody combination (50E1 and 24C5 antibodies) enabled obtaining a good detection of BNP (page 868, left column, §2).

The results presented in Example 5 of the patent demonstrates that, when using a different capture antibody than the 50E1 antibody such as the BioRad capture antibody L21016 antibody, the 24C5 and 26E2 antibodies are not very effective in detecting BNP(1-32). There is thus no contradiction between the results of D1 and D6 with the 24C5 and 50E1 antibodies and the results of the Patent with the 24C5 and L21016 antibodies. They are only results obtained using different experimental conditions."(Emphasis in the original; see page 12, last point of the opponent's statement of grounds of appeal together with paragraph 3.4 of the proprietors' submission of 22 December 2015.)

33. On the basis of the evidence on file, the board cannot conclude that antibody 20G7 is generally superior to antibody 24C5 in detecting BNP(1-32) and proBNP(1-108).

34. Consequently, the objective technical problem is formulated as providing alternative antibodies capable of binding to both BNP(1-32) and proBNP(1-108).

Obviousness

35. In the board's view, a person skilled in the art starting from the closest prior-art antibody 24C5 would have been able to provide alternative antibodies capable of binding to BNP(1-32) and proBNP(1-108) using routine methods. Under the case law of the boards of appeal in these circumstances, the claimed subject-matter, i.e. antibodies, is obvious (see e.g. decisions T 735/00, point 26 of the Reasons; T 187/04, point 11 of the Reasons; T 511/14, points 3 and 5 of the Reasons; T 605/14, points 23, 24 and 26 of the Reasons). Therefore, the subject-matter of claim 2 does not involve an inventive step (Article 56 EPC).