T 1700/19 (Tenofovir alafenamide hemifumarate / GILEAD) 14-03-2023

Reasons for the Decision

1. Procedural violation

1.1 During the proceedings before the opposition division, the parties debated the issue of novelty over D12, which claims priority from D13, and the question of whether TAF-HF is sufficiently disclosed in D12/D13, taking common general knowledge into account.

According to appellant 2, the appealed decision fails to contain a reasoning on crucial points of dispute regarding novelty over D12/D13, specifically on the common general knowledge reflected in D54 and its application in D55 and D56, and on appellant 2's evidence D66 and D67 and arguments regarding the issue of unintentional seeding. The decision would thus be insufficiently reasoned in the sense of Rule 111(2) EPC, and the reimbursement of the appeal fee would be equitable within the meaning of Rule 103(l)(a) EPC.

1.2 The Board does not agree with appellant 2 that the appealed decision is insufficiently reasoned in the sense of Rule 111(2) EPC.

In paragraph 7.3 of the appealed decision, after summarising the arguments of the parties, including those of appellant 2 based on the textbook D54 and the experimental evidence D55 and D56, the opposition division firstly addresses the content of D12 and D13, and concludes that D13 does not contain an enabling disclosure of TAF-HF. The opposition division then indicates (see page 10) that the "experimental data filed by opponent 2" (which implicitly refers to D55 and D56) "cannot be considered to form part of the common general knowledge. Having to resort to an experiment from a text book" (which is understood to refer to D54) "and to follow the experimental data of an active pharmaceutical ingredient (API) completely different from the one as presently claimed cannot lead the person skilled in the art to expect that tenofovir alafenamide hemifumarate would be obtained". Having found that the experiments D55 and D56 could not reflect common general knowledge, it was not essential for the decision to take position on the respondent's additional argument against D55-D56 regarding unintentional seeding, and thus on the appellant's counter-arguments based on D66 and D67.

The Board concludes that the opposition division did not merely repeat appellant 2's submissions in this regard, but duly considered them and gave a reasoning on this point of dispute. A deficiency in the above reasoning would at most constitute an error of judgement on substantive issues, but not a substantial procedural violation.

1.2.1 A further condition for the appeal fee to be reimbursed under Rule 103(1)(a) EPC is that the Board deems the appeal to be allowable. This condition is not met either in the present case, because the Board can uphold the appealed decision, as explained below (see 2.).

1.3 Accordingly, a reimbursement of appellant 2's appeal fee is not justified under Rule 103(1)(a) EPC.

2. Main request

2.1 Novelty over D1

Example 4 of D1 describes the preparation of the fumarate salt of GS-7340, i.e. TAF-MF, by dissolving TAF and fumaric acid in refluxing acetonitrile (ACN) followed by cooling to 5°C. D1 does not explicitly mention any hemifumarate.

According to appellants 1 and 3, the hemifumarate is implicitly disclosed in D1, because example 4 inevitably results in the production of TAF-MF containing detectable quantities of TAF-HF, in light of paragraph [0066] of the patent.

The Board does not share this view. Paragraph [0066] of the patent indicates that, when "suspended in these solvents, the monofumarate form of tenofovir alafenamide [...] partially converts to the hemifumarate form in ACN, ethyl acetate, MTBE, and acetone, as well as at ambient temperatures". Apart from this statement, which does not specifically relate to the conditions used in example 4 of D1, the appellants did not provide evidence showing that TAF-HF is inevitably obtained in D1. In contrast, the evidence cited by the respondent (see the respondent's reply dated 7 January 2020, paragraphs 6.3-6.8, especially D53, and D44-D47) indicates that suspending TAF-MF in ACN, even for extended periods of time, does not necessarily lead to the formation of TAF-HF.

Accordingly, D1 does not anticipate the subject-matter of the main request.

2.2 Novelty over D12

2.2.1 The patent has a date of filing of 15 August 2012, and claims priority from D11, filed on 16 August 2011.

The appellants raised an objection of lack of novelty over D12, filed and published in 2013, but claiming an earliest priority date of 3 February 2012 (from D13).

D12 would be part of the prior art under Article 54(3) EPC, and prejudicial to novelty, only if a subject-matter is:

- disclosed in D12, and entitled to priority from e.g. D13, and

- covered by the claims of the main request, but not entitled to priority from D11.

Following decision G 1/15, for any such subject-matter of claim 1 or D12 to be entitled to priority from D11 or D13, it is necessary and sufficient that it is disclosed for the first time, directly, or at least implicitly, unambiguously and in an enabling manner in the priority document.

2.2.2 Claim 1 of the main request relates to TAF-HF, without any explicit further limitation as to its form. All parties agree that claim 1 covers TAF-HF both as a salt, i.e. with proton transfer between TAF and fumaric acid, and in co-crystal form, which involves no proton transfer (see appellant 1's grounds of appeal, paragraphs 5.7-5.9; appellant 2's grounds of appeal, paragraph 2.4; appellant 3's grounds of appeal, page 6; respondent's reply, paragraph 5.16). Considering that a continuum exists between co-crystals and salts depending on the extent of proton transfer in the solid state (see D27, abstract), the Board accepts this non-limiting interpretation of the wording "hemifumarate" in claim 1.

Document D11, from which the patent claims priority, does not contain the expression "hemifumarate". However, D11 mentions a composition comprising a pharmaceutically acceptable coformer and TAF in a ratio of about 0.5 (see page 2; lines 4-7; claims 1-3 of D11). The pharmaceutically acceptable coformer may in particular be fumaric acid (see page 4, lines 10-11, and claim 8 of D11). Thus these passages of D11 disclose the combination of TAF with fumaric acid at a ratio of about 0.5, in other words TAF-HF, without any limitation as to its form. The invention defined in claim 1 of the main request is not broader than the disclosure of D11 in this respect, but corresponds to it exactly. The Board cannot share the opinion of appellant 1 than the word "coformer" in D11 should be understood as referring exclusively to the formation of co-crystals. Document D15 (see page 360, penultimate paragraph on the left) does not define coformers in such terms. On the contrary, D11 explicitly indicates that the coformer may be any pharmaceutically acceptable compound that is capable of forming a complex, "e.g. a co-crystal complex or a salt", with TAF (see page 4 lines 3-6). The presence of further passages in D11 specifically disclosing TAF-HF in co-crystal form (see e.g. page 1, lines 23-28) does not change the fact that D11 contains a general disclosure of TAF-HF without limitation as to its form.

2.2.3 D12 discloses TAF-HF (see paragraphs [0057] and [0059]) and its preparation (see examples 6-8).

Its earliest priority D13 also mentions TAF-HF (as GS-7340 hemifumarate, see pages 2 and 19) but contains no example of its preparation.

2.2.4 The parties present opposing views as to whether the mere mention of TAF-HF in D13 is an enabling disclosure of this compound, considering the relevant common general knowledge. It is also debatable whether D11 provides an enabling disclosure for the preparation of TAF-HF, as a salt and/or as a co-crystal.

However, in the Board's opinion, if it is accepted that the mention of TAF-HF in D13 is an enabling disclosure of this compound, then it must follow that the equivalent mention in D11 of the combination of fumaric acid and TAF at a ratio of about 0.5, as co-crystal or salt, is an enabling disclosure of TAF-HF. Thus it is not possible for D12 to enjoy priority from D13 for TAF-HF without claim 1 enjoying priority from D11 for the same subject-matter.

In other words, it is not necessary to take position on the validity of the priority claims with regard to enablement. Either D12 enjoys a valid right to priority from D13 in respect of TAF-HF, in which case, for analogous reasons, present claim 1 of the main request also validly claims priority from D11 in respect of the same subject-matter. Or D12 does not enjoy a valid right to priority from D13 in respect of TAF-HF. In both cases, D12 is not prior art under Article 54(3) EPC.

2.2.5 Accordingly, D12 does not prejudice the novelty of the claimed subject-matter.

2.3 Sufficiency of disclosure

2.3.1 The appellants do not contest that example 1 of the patent describes the preparation of a mixture of TAF-HF along with GS-7339-MF and TAF-MF. However, they contend that the patent does not give the necessary information to the skilled person as to how to separate TAF-HF from the obtained solids, so that the claimed subject-matter is insufficiently disclosed.

2.3.2 The Board does not share this opinion. Claim 1 of the main request pertains to TAF-HF as such and does not mandate any degree of purity. Consequently, the Board shares the respondent's opinion that the criteria of sufficiency of disclosure are met already for the reason that example 1 discloses the preparation of TAF-HF.

Contrary to the appellant's view, the fact that TAF-HF is destined to be used as a pharmaceutical does not change this conclusion. Claim 1 does not mandate that TAF-HF be in a form suitable for pharmaceutical use. And in any case, the appellants have not established that the solid mixture of example 1 comprising TAF-MF, TAF-HF and GS-7339-MF would be unsuitable for such a use.

2.3.3 In addition, example 2 of the patent also enables the preparation of TAF-HF, in a form the purity of which the appellants have not contested, by seeding "with tenofovir alafenamide hemifumarate formed in Example 1" (see paragraph [0056]). This further example of the patent would thus enable the skilled person to obtain isolated TAF-HF, supposing arguendo that such a degree of purity would be required.

The appellants countered that the wording of this seeding step in example 2 was ambiguous and possibly referred to purified TAF-HF of example 1, and that the skilled person would understand that the use of pure seed crystals is required.

The Board is not convinced by these arguments. Firstly, there is no reason to consider the reference, in example 2 of the patent, to the TAF-HF formed in example 1 as implying a further isolation or purification step, considering that no such additional step is mentioned in the patent. The Board thus considers that the solid directly resulting from example 1 is meant. Furthermore, there is no debate that the solid formed in example 1 contains some amount of the TAF-HF form which is to be obtained in example 2, and there is no indication in the textbooks D32 (see page 197, §5.2.2) or D2 (see page 1896) that any level of purity is mandatory in the seed crystals for them to effectively promote crystallisation. Thus, the common general knowledge does not teach that a pure seed crystal must be used, and the appellants have not raised serious doubts as to the effective use of the solid of example 1 as seed crystal in example 2.

2.3.4 Accordingly, the criteria of sufficiency of disclosure are met.

2.4 Inventive step

2.4.1 The parties agree on the choice of D1 as closest prior art. D1 discloses TAF-MF (which comprises a 1:1 stoichiometric ratio of TAF to fumaric acid).

The subject-matter of claim 1 of the main request differs in that TAF is in the form of a hemifumarate instead of a monofumarate (i.e. it comprises a 2:1 stoichiometric ratio of TAF to fumaric acid).

2.4.2 According to the respondent, TAF-HF exhibits the following improvements over TAF-MF:

(a) Purging of diastereomeric impurity

Paragraph [0064] and table 2 of the patent, as well as D48 and D49, show the respective preparations and isolations of TAF-HF and TAF-MF under similar conditions, involving dissolving fumaric acid (respectively 0.9 and 0.5 equivalents) and a mixture of TAF (GS-7340) with about 10% of its diastereoisomer (GS-7339) at 70°C in acetonitrile, then cooling down, seeding respectively with TAF-HF (GS-7340-03) and TAF-MF (GS-7340-02), cooling further and isolating the solid product. As summarised in the respondent's reply (see paragraphs 7.7-7.11), the amount of isomer left in the final product is lower in the case of TAF-HF (TAF:GS-7339 ratio 99.35:0.65) than in the case of TAF-MF (92.4:7.6).

Appellants 1 and 2 object that this property is not an intrinsic property of TAF-HF, but results from differences in the process parameters, in particular the amounts of fumaric acid employed resulting in different acidities. However, in the Board's opinion, these amounts of fumaric acid used in the process mirror the differentiating feature, namely the stoichiometric ratio of TAF to fumaric acid in the claimed complex of 2:1 instead of 1:1. It can thus be concluded that, starting from the same starting material (TAF with 10% diastereomer), the formation of TAF-HF by combining TAF with the corresponding 0.5 equivalent fumaric acid and seeding with TAF-HF leads to higher TAF purities than the formation, under comparable conditions, of TAF-MF by combining TAF correspondingly with about 1 equivalent fumaric acid and seeding with TAF-MF. This effect thus credibly arises from the differentiating feature over the prior art.

(b) Increased chemical stability

The patent (see paragraph [0065] and table 3) and the post-published document D51 (see the translation D50, table 1 page 14) credibly show that TAF-HF is chemically more stable than TAF-MF, as measured by the lower rate of decrease of the active ingredient TAF when subjecting TAF-HF to several stress conditions. The effect observed in table 3 of the patent is not caused by the higher initial purity of TAF-HF (0.05% degradation products in TAF-HF vs 0.69% in TAF-MF at t=0), considering that in D50 the effect is also observed in the opposite situation (i.e. starting from TAF-HF with 97.40% purity vs TAF-MF with 97.74% purity). Accordingly, the difference in chemical stability can be attributed to the distinguishing feature of the invention compared with the closest state of the art.

The appellants submit that TAF-HF exhibits a higher relative increase in degradation products under certain storage conditions (namely 40°C/75% RH, cap closed or open) than TAF-MF. However, the relevant technical effect here is the stability of TAF-HF, i.e. the amount of the active ingredient TAF-HF remaining in the product over time, and not the relative increase in degradation products. The appellants' argument does not call into question the fact that TAF-HF has improved stability.

2.4.3 Considering that an inventive step can be acknowledged taking into account the technical effects discussed above, it is not necessary to assess whether the claimed hemifumarate additionally achieves higher thermodynamic or thermal stability.

2.4.4 The objective technical problem to be solved can accordingly be formulated as the provision of an improved form of TAF, having improved chemical stability and an improved ability to purge TAF of its diastereomeric impurity GS-7339.

2.4.5 Obviousness

It was known at the priority date that, in the case of tenofovir disoproxil, the co-crystal with fumaric acid in a 2:1 molar ratio is more stable and less hygroscopic than the 1:1 fumarate salt (see D15, passage bridging pages 368-369; D3, page 3, lines 10-15). The appellants contend that this historical development would have provided the skilled person with a reasonable expectation that, in analogy to tenofovir disoproxil, TAF-HF would exhibit improved stability over TAF-MF.

The Board does not share this opinion. The properties of the hemifumarate in the case of TAF could not be extrapolated from those of tenofovir disoproxil. The prior art does not point to any such predictability. On the contrary, in light of the statements regarding the unique physicochemical properties of API solid forms, including salts and co-crystals in D15 (see page 359, right column) and D28 (see page 317, left column), no such generalisation is possible. In the present case, the prior art gives no reason to assume that the modification involved, namely the use of a 2:1 molar ratio with fumaric acid instead of 1:1, would lead to the same effect in the structurally different compounds TAF and tenofovir disoproxil. Accordingly, the claimed subject-matter is not obvious starting from D1.

2.4.6 For the same reason, the objection of lack of inventive step starting alternatively from D3 is unconvincing. This objection supposes that the skilled person, starting from the tenofovir disoproxil hemifumarate of D3, would change the API to TAF and expect that similar advantageous physicochemical properties be still obtained. The claimed subject-matter is therefore not obvious starting from D3 either.

2.4.7 In conclusion, the main request satisfies the requirements of inventive step.