T 0235/22 (Anti-CD47 dosage regimen/LELAND STANFORD JUNIOR UNIVERSITY) 06-06-2024

Absence of a duly summoned party (Rule 115(2) EPC and Article 15(3) RPBA)

1. Respondent I (opponent 1) had been duly summoned to the oral proceedings, but had indicated with its letter dated 11 April 2024 that it would not be attending.

Main request

Added subject-matter (Article 123(2) EPC) - claim 1

2. When deciding whether claimed subject-matter extends beyond the content of the application as filed (Article 123(2) EPC), it has to be established what a skilled person would derive directly and unambiguously, using common general knowledge, and seen objectively and relative to the date of filing, from the whole of the documents as filed (G 3/89, OJ 1993, 117; G 11/91, OJ 1993, 125; G 2/10, OJ 2012, 376; referring to this test as the "gold standard").

3. Claim 1 differs from claim 1 as filed by the following additional features (see also decision under appeal, point 22.4):

(i) "treating cancer"

(ii) "in a primate subject"

(iii) "as a primer agent the anti-CD47 agent", i.e. identical to the agent in step (b)

(iv) "sub-therapeutic dose" in step (a)

(v) the sub-therapeutic dose "primes the subject for administration of a therapeutically effective dose of the anti-CD47 agent"

(vi) the sub-therapeutic dose "significantly reduces toxicity due to loss of erythrocytes"

(vii) the anti-CD47 agent "reduces the binding of CD47 to SIRPalpha"

(viii) the anti-CD47 agent "is an anti-CD47 antibody"

(ix) "step (b) is performed in a range from 3 days to 21 days after beginning step (a)"

4. In the present case it is necessary to determine whether the features of claim 1 of the main request can be found in combination in the application as filed. As there is no verbatim disclosure of that combination, it has to be determined whether that combination of features can be derived directly and unambiguously, taking into account the skilled person's common general knowledge, from the application as filed. One approach suggested by the appellant was to start with the disclosure of claim 1 as filed.

5. In doing so, it is necessary to analyse whether the additional features of claim 1 of the main request (see point 3. above) are disclosed in the application as filed and also whether the skilled person would derive the combination of those additional features with the subject-matter of claim 1 as filed in a direct and unambiguous manner from the application as filed. According to the case law of the boards of appeal the content of an application must not be treated as a reservoir from which features pertaining to separate embodiments of the application can be taken and combined in order to artificially create a particular embodiment. Instead, a pointer to a particular combination is required. This can, for example, be an explicit indication that a particular embodiment is preferred or an implicit pointer through the preferential use of an embodiment in the examples (see Case Law of the Boards of Appeal of the EPO, 10th edition 2022, II.E.1.6.1).

6. In the following paragraphs, the disclosure in the application as filed for each of the limiting features (i) to (ix) will be discussed in turn, together with an analysis of how this disclosure would be perceived by the skilled reader when consulting the application as a whole.

6.1 Feature (i) "treating cancer" represents one of the two specified therapeutic applications of the invention, the other being treating "infection with an intracellular pathogen"(see paragraph [0005] in the "Summary of the Invention"). Paragraph [0050] further discloses that "symptoms, illnesses, and/or diseases that can be treated with an anti-CD47 agent include, but are not limited to cancer and infection". In claim 17 as filed, which is dependent on claim 1 as filed, "the subject has cancer or an intracellular pathogen infection". The skilled person would therefore infer that "cancer" and "infection" are the two main disease groups to be treated.

6.1.1 The skilled person would, however, also consult the examples. In doing so, they would realise that none of the examples relates to infection. Example 3 is performed on diseased animals carrying xenografts of various cancer cell lines. Mice carrying xenografts of a human bladder cancer cell line were treated with human anti-CD47 antibody hu5F9-G4 and shown to remain tumour-free in contrast with mice treated with buffered saline (see Figure 11). Further experiments established the effect on metastasis using engrafted human metastatic prostate tumour specimens (see Figure 11B) and a xenograft model of human breast cancer (see Figure 12). All other examples were carried out in healthy animals (mice or monkeys) to test different dosage regimens and to evaluate safety and toxicity of the anti-CD47 antibody (see Examples 1, 2, 4 and 5).

6.1.2 The skilled person would therefore conclude that the application as filed showed a preference for the treatment of cancer by performing the respective experiments while not testing anti-CD47 agents for other diseases, e.g. infection.

6.1.3 Respondent II argued in this respect that Example 3 also referred to infection in its conclusion (see paragraph [00132]) and did not therefore establish any preference for cancer. The board does not accept this argument, because the cited passage merely concludes that, as regards disease, the tested humanised antibody had the same general properties as non-humanised antibodies (i.e. as used in Example 1) and could therefore also be effective for other diseases. This, however, does not render irrelevant the fact that only cancer treatments using dedicated cancer models for human bladder, prostate and and breast cancer were tested in the application, thus indicating a preference to the skilled person.

6.1.4 Respondent II further argued that, due to the wording of claim 1 as filed, which required "treating a subject with a therapeutic dose of an anti-CD47 agent", and the specification of claim 17 "wherein the subject has cancer", no actual treatment of cancer was disclosed but only the treatment of a subject having cancer in which the effects of anaemia were mitigated. The board does not agree, because treating a subject with a therapeutic dose of an agent is a therapeutic treatment of the underlying disease in the subject with that agent. The fact that the invention relates to a specific dosage regimen that mitigates side effects of said treatment does not change this.

6.2 Feature (ii) "in a primate subject" is explicitly mentioned in paragraph [0060] ("particularly relevant when treating primates") with links to features (v) ("methods that use a primer agent") and (vi) ("primates are sensitive to RBC [red blood cell] count and prone to develop anemia"). Moreover, Examples 2, 4 and 5 were performed in non-human primates, namely rhesus or cynomolgus monkeys.

6.2.1 Respondent II argued by reference to paragraph [0053] that, despite the particular relevance mentioned, humans were a more preferred subset in the application. The board does not agree with this reasoning, because the disclosure of a more preferred subset, e.g. humans, does not invalidate the explicit indication that primates, as a target group, are of particular interest. Therefore, the skilled person would have considered the claimed treatment in this subject group as directly and unambiguously derivable.

6.2.2 Respondent II furthermore argued that, as the following paragraph [0061] was limited to ESAs as primer agents, the particular relevance of primates was also limited to this embodiment. The board does not agree, because the next paragraph [0062] discloses that "the primer agent comprises a sub-therapeutic dose of an anti-CD47 agent", thus making clear that the subject group primates was particularly relevant also for the use of anti-CD47 agents.

6.2.3 The skilled person would therefore have concluded that primate subjects were preferred.

6.3 Feature (iii) "as a primer agent the anti-CD47 agent", i.e. the identical agent to that in step (b), is disclosed in claim 3 as filed, which is directly dependent on claim 1 as filed. It is disclosed as one of three options and combinations thereof: "an erythropoiesis-stimulating agent (ESA), a priming dose of the anti-CD47 agent, a priming dose of a second anti-CD47 agent, and combinations thereof".

6.3.1 The appellant argued that using the anti-CD47 agent of step b) as the primer agent was a preferred embodiment, as could be seen from the examples in which escalating doses of anti-CD47 antibodies were tested (see, e.g., Examples 2, 4 and 5).

6.3.2 Respondent II argued that EPO, dexamethasone and benadryl were also used as primer agents in the examples (see Example 5, paragraph [00152] and Table 3). EPO was further used as a primer agent in Example 2 (see paragraph 00125]). Moreover, Table 6 in Example 5 showed that effects on red blood cell morphologies were less pronounced when EPO was used as a primer agent. The skilled person would not therefore recognise a preference for using the same anti-CD47 agent in both steps.

6.3.3 Furthermore, the decision under appeal (see point 22.4.3), referring to paragraph [00152], Table 3, paragraphs [0159] to [0167], and Tables 4 to 6, also considered the use of alternative primer agents (dexamethasone and benadryl or EPO) in the examples as evidence that the skilled person would not have regarded the use of the same anti-CD47 agent as a preferred embodiment.

6.3.4 Taking into account the application as a whole, including the remaining examples (see paragraph [0168] ff.), the board comes to a different conclusion, for the following reasons.

6.3.5 Example 1 using mice finds that "repeated administration of anti-CD47 antibodies does not exacerbate the initial anemia", thereby providing "the basis of the subsequent experiments in non-human primates" (see paragraph [00120]).

6.3.6 Example 2 confirms this finding in non-human primates with or without EPO: "Escalating concentrations of anti-CD47 antibodies do not exacerbate anemia" (see paragraph [00125]) and establishes that a "[s]ingle loading dose of anti-CD47 antibodies enables higher maintenance doses" (see paragraph [00127]).

6.3.7 This then leads to the only therapeutic experiment, Example 3, in which mice with different xenografts receive anti-CD47 antibody, but without a priming dose.

6.3.8 Example 4 using cynomolgus monkeys confirms that "the priming/maintenance dosing strategy allows Hu5F9-G4 to be clinically well-tolerated, even at doses as high as 300 mg/kg" (see paragraphs [00133] and [00134] as well as Figures 13 to 15).

6.3.9 Example 5 first performs in vitro and in vivo tests of humanised anti-CD47 antibody in monkeys. It then tests a single or repeated low dose (1 or 3 mg/kg) of humanised anti-CD47 antibody either alone or with dexamethasone/benadryl in parallel or EPO as a primer agent (minus 5 days) (see Table 3). It concludes that "administration of Hu5F9-G4 as a single dose at 1 mg/kg or once weekly doses for 4 weeks at a dose of 3 mg/kg (alone or with pre-treatment with EPO or in combination with administration of Dexamethasone/Benadryl) was well-tolerated in male cynomolgus monkeys" (see paragraph [00158]).

6.3.10 Respondent II argued that the use of dexamethasone and benadryl or EPO in these experiments would have indicated to the skilled person that those agents were equally preferred agents. The board disagrees. The doses of anti-CD47 antibody used in this study are sub-therapeutic and are also too low to cause anaemia. The skilled person would therefore have considered this experiment to be only an initial test but not a test for the priming/maintenance regimen of the invention. Moreover, dexamethasone and benadryl are commonly known as being neither an ESA nor an anti-CD47 agent, i.e. they do not fall within the two groups of primer agents listed in claim 3 as filed.

6.3.11 In the subsequent experiments of Example 5, monkeys receive escalating doses of anti-CD47 antibody with or without EPO (see paragraphs [00159] ff.). In both parts of the study this leads to good maintenance of the red blood cell count (RBC) even with therapeutic doses of up to 100 mg/kg (see Tables 4 and 5).

6.3.12 The fact that some red blood cell morphologies changed in this study (see Table 6), as pointed out by respondent II, is not seen as an indication that EPO was equally preferred as a primer agent. First of all, EPO is present together with the priming/maintenance dose of the same anti-CD47 antibody, i.e. it represents only an optional additional primer agent, which - according to the claim wording ("comprising") - can be present in the method referred to in the claim. Secondly, the study conducted in the absence of EPO was followed up for a longer period (for 37 days compared to 5 days - see Table 6) and is therefore not directly comparable over the whole period. Thirdly, the red blood cell count (RBC), as the most relevant measure, was even slightly better in the absence of EPO (see Table 5).

6.3.13 The following paragraph [00168] states that "initial administration of Hu5F9-G4 at lower doses enables subsequent administration of higher doses that are tolerated in cynomolgus monkeys" and, based on this, sets out "to evaluate the potential toxicity and toxicokinetics of Hu5F9-G4 when administered as a priming dose at a low dose level followed by multiple maintenance doses at higher dose levels" in order "to model the potential clinical dosing schedule using a priming/maintenance-dosing regimen". This paragraph thus summarises the results of the preceding experiments and presents the aim of using the priming/maintenance regimen with the same anti-CD47 agent (anti-CD47 antibody) in a clinical setting.

6.3.14 In conclusion, the experiments consistently develop a priming/maintenance regimen using the (same) anti-CD47 antibody, which is envisaged for clinical use. The skilled person would thus have considered the use of the same anti-CD47 agent in the form of an anti-CD47 antibody to be a preferred embodiment.

6.4 Feature (iv) "sub-therapeutic dose" is disclosed in paragraphs [0057] and [0062] (also named "priming dose"). It is linked to the use of "the same anti-CD47 agent" throughout the application (see point 6.3 above) and in the examples (see, e.g., Examples 2, 4 and 5), and can thus be seen as generally disclosed in this context.

6.5 Feature (v) "primes the subject for administration of a therapeutically effective dose of the anti-CD47 agent" is a definition pertaining to the terms "primer" and "anti-CD47 agent" (see paragraphs [0033], [0057] and [0059]). It is thus generally applicable.

6.6 Feature (vi) "significantly reduces toxicity due to loss of erythrocytes" reflects the object of the invention by identifying the context for the priming regimen (see paragraphs [0004], [0008], [0059]) and is thus also generally applicable.

6.7 Feature (vii) "reduces the binding of CD47 to SIRPa" is a definition of the term "anti-CD47 agent" (see paragraphs [0033], [0057] and [0059]) and is thus generally applicable.

6.8 Feature (viii) "an anti-CD47 antibody" is disclosed in dependent claim 19 as filed, which is dependent on claim 1 via claim 18. It is also disclosed in paragraphs [0039] and [0007] of the application as filed. Anti-CD47 antibodies are the only anti-CD47 agents which are used in the examples (MIAP410 and MIAP740 in mice, Hu5F9-G4 in primates, see points 6.3.5 to 6.3.11 above). The skilled person would therefore conclude that an anti-CD47 antibody is the preferred embodiment of an anti-CD47 agent mentioned in claim 1 as filed.

6.9 Feature (ix) "in a range from 3 days to 21 days" and the combination thereof with the features in claim 1 as filed is disclosed in paragraph [0005].

6.9.1 Respondent II argued that other more limited ranges were disclosed in the application as filed (see, e.g., claim 5, which is directly dependent on claim 1, and the list of ranges in paragraph [0075]) and were at least equally preferred. The board does not agree, because the range of "from about 3 days to about 21 days" is the only range disclosed at the beginning of the "Summary of the Invention" (see end of paragraph [0005]). It would therefore have been considered by the skilled person as the generally applicable dosage range to be combined with the other features.

7. In conclusion, the board finds that each of the features inserted into claim 1 as filed represents either a general definition (features (iv) to (vii) and (ix)) or a preferred embodiment of the invention (features (i) to (iii), (viii)). The skilled person would therefore have derived the combination of these features with the subject-matter of claim 1 as filed in a direct and unambiguous manner.

8. The subject-matter of claim 1 does not infringe the requirements of Article 123(2) EPC.

Added subject-matter (Article 123(2) EPC) - claims 2 to 8

9. Respondent II considered that claims 2 to 8 infringed Article 123(2) EPC for the same reasons as claim 1.

10. The board has not been presented with any additional reasons to support the view that the subject-matter of claims 2 to 8 extends beyond the content of the application as filed.

11. The board concludes that claims 2 to 8 do not infringe Article 123(2) EPC.

Remittal to the opposition division (Article 111(1) EPC and Article 11 RPBA)

12. There are special reasons in the present proceedings for remitting the case to the opposition division for further prosecution. Firstly, both parties requested remittal if the board concluded that the claims did not infringe Article 123(2) EPC. Additionally, the decision under appeal did not address the other grounds of opposition, particularly novelty and inventive step. The board found it inappropriate to initiate this examination for the first time during the appeal proceedings. This might have required a continuation of the proceedings in written form, possibly followed by further oral proceedings, if deemed appropriate or if requested.