T 2168/11 (Alzheimer's disease beta amyloid peptide mouse model/ELAN ELI LILLY) 24-06-2015

Summary of Facts and Submissions

I. European patent no. 0 730 643 is based on European patent application no. 94 931 891.9, published as International patent application WO 95/11968. The patent was opposed by three opponents on the grounds as set forth in Articles 100(a),(b) and (c) EPC. The opposition division considered the Main Request and an Auxiliary Request to contravene Article 54 EPC and, accordingly, the patent was revoked.

II. The patentees lodged an appeal against this first decision of the opposition division. In the decision T 1847/06 of 16 December 2008, this board in a different composition considered a second Auxiliary Request to fulfil the requirements of Articles 123(2),(3), 84 and 54(3),(4) EPC (1973). The case was remitted to the opposition division for further prosecution on the basis of the second Auxiliary Request.

III. In an interlocutory decision dated 26 July 2011, the opposition division considered the second Auxiliary Request filed in appeal (which was made patentees' new Main Request) to be entitled to the claimed priority dates and to meet the requirements of Articles 83 and 56 EPC. Thus, the opposition division maintained the patent in amended form on the basis of this request and a description adapted thereto.

IV. An appeal was lodged by opponent 01 (appellant) against this interlocutory decision. The appellant filed new documentary evidence (documents D41-D45) and requested that the decision under appeal be set aside and that the patent be revoked. As auxiliary request, oral proceedings were requested.

V. In reply to the appellant's Grounds of Appeal, the patentees (respondents) requested, as a Main Request, that the appeal be dismissed. As auxiliary requests, the respondents requested that the decision under appeal be set aside and that the patent be maintained on the basis of one of the newly filed first, second or third Auxiliary Requests. As an auxiliary request, oral proceedings were also requested.

VI. Opponents 02 and 03, parties as of right to the present appeal proceedings, did not file any submissions in this appeal proceedings.

VII. In a communication pursuant to Article 15(1) of the Rules of Procedure of the Boards of Appeal (RPBA) annexed to Summons to oral proceedings, the parties were informed of the board's preliminary, non-binding opinion on the issues of the case. In particular, the board considered the Main Request not to fulfil the requirements of Article 56 EPC. The first, second and third Auxiliary Requests were preliminarily considered to be admissible. However, the board noted that no submissions addressing these requests were on file.

VIII. No substantive submissions were filed by the respondents in reply to the board's communication. The respondents informed the board that they would not be represented at the scheduled oral proceedings and they withdrew their request for oral proceedings.

IX. The appellant replied to the board's communication, filed a new document (D46) and, inter alia, addressed the admissibility of respondents' Auxiliary Requests.

X. The oral proceedings were cancelled by the board.

XI. Claim 1 of the Main Request (claims upheld by the opposition division) read as follows:

"1. Use of a transgenic nonhuman animal or stem cell to screen an agent for activity in preventing, inhibiting or reversing Alzheimer's disease, wherein said transgenic nonhuman animal or stem cell:

comprises a diploid genome comprising a transgene encoding a heterologous APP polypeptide comprising the Swedish mutation wherein the amino acid residues at positions corresponding to positions 595 and 596 in human APP**(695) are asparagine and leucine, respectively; and

expresses said APP polypeptide,

wherein said agent is administered to said transgenic animal at a dosage of from 1 ng/kg to 10mg/kg, preferably from 10 myg/kg to 1 mg/kg."

Claims 2 to 5 were directed to preferred embodiments of claim 1.

XII. All Auxiliary Requests were identical to the Main Request, except for amendments introduced into claim 1 which read as follows:

a) "1. ... [as claim 1 of the Main Request] ...

expresses said APP polypeptide; and

produces detectable quantities of ATF-ßAPP,

wherein said agent is administered ...

[as claim 1 of the Main Request] ..."

(first Auxiliary Request)

b) "1. ... [as claim 1 of the Main Request] ...

expresses said APP polypeptide; and

produces quantities of ATF-ßAPP which are at least two-fold higher than the quantities of ATF-ßAPP, produced from wild type human ATF-ßAPP in an equivalent transgenic model,

wherein said agent is administered ...

[as claim 1 of the Main Request] ..."

(second Auxiliary Request)

c) "1. ... [as claim 1 of the Main Request] ...

expresses said APP polypeptide; and

produces detectable quantities of ATF-ßAPP,

wherein said agent is administered ...

[as claim 1 of the Main Request] ...

preferably from 10 myg/kg to 1 mg/kg, and

wherein the effect of the agent on ATF-ßAPP

production is measured."

(third Auxiliary Request).

XIII. The following documents are cited in this decision:

D1: WO 91/19810 (publication date: 26 December 1991);

D2: M. Mullan et al., Nature Genetics, August 1992, Vol. 1, pages 345-347;

D4: M. Citron et al., Nature, 17 December 1992, Vol. 360, pages 672-674;

D9: WO 94/10569 (publication date: 11 May 1994);

D23: D. Wirak et al., Science, 13 March 1992, Vol. 255, pages 1443-1445;

D24: S. Kawabata et al., Nature, 5 March 1992, Vol. 356, page 23;

D25: J. Marx, Science, 6 March 1992, Vol. 255, pages 1200-1202;

D29: D. Quon et al., Nature, 18 July 1991, Vol. 352, pages 239-241.

XIV. Appellant's arguments, insofar as they are relevant to the present decision, may be summarized as follows:

Main Request

Inventive step (Article 100(a) EPC; Article 56 EPC)

The closest prior art

Document D1 disclosed transgenic mice (the "Cordell's" mice) expressing wild-type ß-amyloid precursor protein (ß-APP), in particular the full-length ß-APP**(751) isoform. In Examples 8-11 (Table 1), the expression vector (ß-APP**(751) with neuronal-specific enolase (NSE) promoter), disclosed in Examples 1 and 5, was used to generate transgenic mice. The brains of the NSE-A751 transgenic mice were histologically analyzed in Example 12 and deposits characteristic for Alzheimer's disease (AD) were identified in these mice but not in control wild-type mice. The generation of transgenic mice with the full-length ß-APP**(695) isoform was also mentioned in Examples 2, 5 and 8. Thus, the transgenic mice disclosed in document D1 produced ß-amyloid protein (ß-AP) and showed characteristic amyloid plaques known to be related to AD. As stated in document D1, these transgenic mice, in particular those harbouring the full-length ß-APP**(751) and ß-APP**(695), were suitable AD models for use in screening of drugs for AD treatment and/or prevention.

Objective technical problem and proposed solution

Starting from document D1, the objective technical problem was the provision of a further animal AD model for use in drug screening. According to claim 1, this technical problem was solved by the provision of a non-human transgenic animal carrying a human APP with the Swedish mutation and its use for drug screening.

Obviousness

As stated in document D4, only a very limited number of mutant APP were known at the priority date of the patent. While any of these mutants could have been used to produce a transgenic animal AD model for use in drug screening, a mutant with the Swedish mutation disclosed in documents D2 and D4 was the most promising and preferred, since this mutation was the first mutation known to produce increased ß-AP levels (6-8 fold more ß-AP than cells expressing normal ß-APP), establishing a direct link between this mutation genotype and the clinicopathological phenotype. Document D4 disclosed the use of cells transfected by ß-APP with the Swedish mutation for drug screening. It was an entirely logical development to implement the in vitro experiments of documents D4 into an in vivo setting, i.e. the production of transgenic animals which were at the time usually used for drug screening. Thus, a skilled person would have combined the disclosures of documents D1 and D4 at least with a try-and-see attitude. It only required the replacement of a wild-type ß-APP by an ß-APP with the Swedish mutation in an attempt to improve the "Cordell's" mice AD model provided by document D1.

The broad dosage range in claim 1 was trivial. Such dosages were routinely used in drug screening (even in neurodegenerative diseases like AD), as acknowledged in the patent and in prior art on file. According to the case law, the mere aggregation of obvious and trivial features did not make a claim inventive.

Reasonable expectation of success

Neither the "Wirak's" nor the "Kawabata's" mice referred to in documents D23 and D24, respectively, expressed the wild-type full-length ß-APP used to produce the "Cordell's" mice disclosed in documents D1 and D29. None of documents D23 and D24 discredited the "Cordell's" mice which, as stated in document D25, were still considered a useful AD model at the priority date of the patent.

The transgenic "Wirak's" mice expressed a short (4kD) ß-amyloid fragment (derived from full-length ß-APP) and formed amyloid-like fibrils that were similar in appearance to those found in brains of AD patients. Later, it was noticed that the mouse line ("C57BL/6 mice") used in these studies was inherently prone to formation of amyloid-like structures, raising questions on the extent to which the transgene contributed to the observed phenotype. However, it could not be deduced from document D23 that the transgenic approach failed as such.

Document D24 was a retraction of earlier studies concerning the transgenic "Kawabata's" mice. These mice expressed a C-terminal fragment of ß-APP (which included the 4KD ß-amyloid fragment used in the "Wirak's" mice). In these earlier studies, the overexpression of the C-terminal APP fragment produced extracellular amyloid plaques similar to those found in brains of AD patients. Though the transgenic status of the mice was confirmed, the earlier reported histopathologically findings could not be reproduced in document D24.

In fact, document D25 summarized the development of the three available transgenic AD models in March 1992. The retractions of the "Wirak's" and "Kawabata's" mice models, both expressing short ß-APP fragments derived from the wild-type full-length ß-APP protein, were seen as disappointing. However, the suitability of the third transgenic animal AD model, namely the "Cordell's" mice disclosed in document D1, the only one expressing the full-length ß-APP (ß-APP**(751) and ß-APP**(695) isoforms), was not challenged and was described as the sole AD model "holding up" at that time.

Admissibility of the Auxiliary Requests

In preparation for the first oral proceedings before the opposition division, the patentees filed requests similar to the actual first and second Auxiliary Requests. These requests were however withdrawn and replaced at these oral proceedings. In its first decision, the opposition division revoked the patent, because it considered the then pending Main and Auxiliary Requests not to be novel. In the subsequent appeal proceedings, the board decided that Auxiliary Request II (now the Main Request), containing a dosage feature, was novel and remitted the case back to the opposition division for further prosecution (T 1847/06, supra).

Although the first opposition and appeal proceedings were focused on novelty, extensive submissions on the issue of inventive step were filed by all parties. The same took place before the second oral proceedings before the opposition division. In fact, an objection for lack of inventive step based on a combination of documents D1 with D2 or D4 was already raised in the notice of opposition. By failing to file the present first and second Auxiliary Requests as a precautionary measure (respectively by withdrawing similar requests then on file), the patentees deprived the opposition division from taking a decision on these requests. Moreover, at the second oral proceedings before the opposition division, the patentees were aware that in the prosecution of a parallel patent application, the competent board had decided that subject-matter corresponding to the present Main Request was not inventive in the light of the same documents (D1, D2, D4) (cf. T 1127/06 of 13 August 2008). Thus, the patentees had a fair chance to file the first and second Auxiliary Requests at that point in time.

No explanations were provided by the patentees why these Auxiliary Requests were not filed or further pursued during proceedings before the opposition division. Admitting these Auxiliary Requests raised completely new issues, such as entitlement to priority, and was thus contrary to procedural economy.

XV. Respondents' arguments, insofar as they are relevant to the present decision, may be summarized as follows:

Main Request

Inventive step (Article 100(a) EPC; Article 56 EPC)

The closest prior art

Document D1 disclosed transgenic mice expressing the wild-type ß-APP, the "Cordell's" mice, and their use for drug testing. A comparison between the amount of ß-amyloid plaques (deposits) formed in brains of transgenic mice and control animals determined the drug effectiveness. However, it was not clear from document D1 whether the alleged plaques occurred with sufficient size, frequency and reproducibility to form the end point of a drug screening assay. Transgenic "Cordell's" mice expressing wild-type ß-APP would not have been considered by a skilled person a working AD model. There were doubts based on the many different characteristics of the deposits in the "Cordell's" mice relative to the plaques in AD patients, the difficulties in scoring these deposits and the possibility that the deposits in the "Cordell's" mice could be full-length ß-APP as well as or instead of the cleavage ß-AP product. The inconclusive data and the lack of a convincing AD pathology displayed by the "Cordell's" mice did not render them a robust model animal of the AD pathology.

Objective technical problem and proposed solution

Starting from document D1, the technical problem to be solved was the provision of a robust animal AD model for screening of therapeutic agents. The patent provided such model by selecting transgenic animals expressing ß-APP with the Swedish mutation in the ß-APP transgene and producing detectable levels of the amino-terminal fragment of ß-APP (ATF-ßAPP), a cleavage product used as a marker to identify suitable animal AD models at an earlier stage in the life-cycle than would be required for animals to develop the AD pathology.

Obviousness

At the priority date, there was a concern that laboratory (transgenic) animals were inherently incapable of developing AD because they did not so in nature. Furthermore, scepticism was in the field due to the retraction of two purported animal AD models (infra). Indeed, more straightforward options were available to a skilled person, such as to improve existing in vitro models or ex vivo screening methods using cells from AD patients.

Even if a transgenic approach was considered, multiple options were available to a skilled person. Document D1 disclosed several regulatory sequences (neural-specific promoters) and a number of possible ß-APP constructs (A42, A99, A695), but only one (ß-APP**(751)) was tested. A skilled person could have tried any of these, or other, ß-APP isotypes, isoforms or fragments to see if a robust AD pathology was achieved. Further, a number of other APP mutations were also known, of which any could have been selected. There was no objective reason why a skilled person would have specifically selected a ß-APP with the Swedish mutation in a transgenic approach. It was only with the benefit of hindsight of the patent that this combination was apparent.

The Swedish mutation had been identified in a human family (document D2) and reported to be associated with increased processing of ß-APP in human cell culture (document D4). Indeed, document D4 reported that human 293 cells transformed by a cDNA encoding human ß-APP with the Swedish mutation produced higher ß-AP levels than cells expressing wild-type ß-APP. However, this document did not provide an incentive to create a transgenic animal model expressing ß-APP with the Swedish mutation. On the contrary, the conclusion in document D4 actually pointed away from such a model, since it was stated that the transfected cells described therein, or endogenous cells taken from a AD patient with the Swedish mutation, could be useful for screening and identifying therapeutic compounds. This was in line with document D2 which acknowledged that humans having a ß-APP with the Swedish mutation had a relatively late AD onset and thus, it was unlikely that transgenic animals with this mutation would develop a significant AD pathology. Document D2 taught the skilled person away from a transgenic animal comprising ß-APP with the Swedish mutation as a useful AD model.

Moreover, even if document D4 stated that human 293 cells expressing ß-APP with the Swedish mutation produced higher ß-AP levels than cells expressing the wild-type ß-APP, it was known in the art that, in experiments conducted with wild-type ß-APP, the effects demonstrated in vitro frequently were not replicated in vivo. The results disclosed in document D4 could not be used to predict the effect in vivo. The less so since it was not known whether ß-APP processing would occur in a non-human model and would result in detectable levels of ATF-ßAPP. There was a lack of understanding regarding the cleavage efficiency of human Swedish APP by non-human enzymes and the metabolic stability of ATF-ßAPP generated by such cleavage. Therefore, taking D1 as the closest prior art, document D4 did not provide an incentive for a skilled person to produce a transgenic animal (mouse) expressing ß-APP with the Swedish mutation for screening therapeutic agents.

Reasonable expectation of success

Several efforts to produce an effective animal AD model were made in the years immediately before the priority date. Wirak et al. reported transgenic mice expressing and accumulating human ß-AP in their brains and forming amyloid-like fibrils similar to those seen in brains of AD patients. Kawabata et al. reported transgenic mice overexpressing a C-terminal fragment of human ß-APP and forming amyloid plaques and neurofibrillary tangles in their brains which was associated with neuronal degeneration similar to that seen in the brains of AD patients. However, in document D23, Wirek et al. reported that the published data did not support the conclusion that the expression of human ß-APP transgene caused the formation of amyloid deposits in the brain of the transgenic mice. In document D24, Kawaba et al. acknowledged that the AD pathology identified in the transgenic mice was not reproducible. As stated in document D25, only one transgenic animal AD model was left by these retractions, namely the "Cordell's" mice disclosed in documents D1 and D29, which, however, did not display a robust AD pathology. Indeed, document D25 simply stated that the animal AD model of document D1 was "holding up", without further emphasis or stronger sentiment, thereby showing a lack of confidence in the transgenic mouse approach.

In fact, the transgenic mice disclosed in documents D1 and D29 were not shown to have the characteristic amyloid plaques. Document D29 only stated that there were deposits resembling ß-amyloid structures typically seen in AD patients' brain. A mere resemblance did not establish a robust AD model. The less so since the antibody used to detect the deposits was cross-reactive binding to both the pathogenic ß-AP (generated by ß-secretase cleavage of ß-APP) and the full-length ß-APP. The generation of the pathogenic ß-AP was a prerequisite for an AD model, but it was not even clear whether the transgenic mice disclosed in documents D1 and D29 met this basic requirement. As concluded in document D29, further studies were needed on the quality and quantity of the detected deposits as well as on the display of other AD pathological features, it was thereby shown that the authors of document D29 were not sure whether the disclosed transgenic mice could develop into a robust AD model.

In view of the confusion in the state of the art and the reported failures of transgenic animal models comprising the wild-type human ß-APP to produce significant AD pathology, a skilled person could not have had a reasonable expectation of success that a transgenic animal expressing a human ß-APP with the Swedish mutation instead of the wild-type ß-APP would have provided a robust AD model or that its use in screening assays to identify therapeutic agents for AD treatment and/or prevention would be meaningful.

According to the case law, a skilled person was cautious, with a conservative attitude, never entering unpredictable areas or technology and performing only routine work within the framework of normal practice of filling gaps in knowledge by applying already existing knowledge. In the present case, the skilled person faced a situation where all previous attempts had failed and there was no guidance as to which direction could generate a successful AD model. It was required to perform scientific research and the outcome was unknown. Thus, inventive step was given.

Admissibility of the Auxiliary Requests

No submissions with regard to this issue were made. In particular the respondents did not react to appellant's submissions objecting to the admissibility of the Auxiliary Requests.

XVI. The appellant requested that the decision under appeal be set aside and that the patent be revoked.

XVII. The respondents requested, as their Main Request, that the appeal be dismissed and the patent be maintained on the basis of the set of claims upheld by the opposition division. As auxiliary request, the respondents requested the maintenance of the patent based on any of their first, second and third Auxiliary Requests.

XVIII. No requests were on file from any of the parties as of right.