We’re pleased to bring our readers an in-depth post looking at the grant of a patent application related to the anti-cancer drug Nivolumab, written by Dr. Venkata Subramanian Raman and Ms Sai Ratna Manjari. Dr. Venkata Subramanian Raman is a registered patent agent and proprietor of Patentszone, and Ms. Sai Ratna Manjari is a registered patent agent providing IPR consulting service and also associated with Patentszone. (Please note that given the nature of examination in the current post, it is longer than our usual posts).
Nivolumab and 5C4- Big brother or Twin brothers
Dr. Venkata Subramanian Raman and Ms. Sai Ratna Manjari
In an interesting grant recently, an application (PDF) related to anti-cancer drug Nivolumab (Opdivo, Opdyta in India) survived 4 pre-grant oppositions before finally being allowed by the patent office. Bristol Myers Squibb and Ono Pharmaceuticals were the applicants, and the Indian Pharmaceutical Alliance (IPA), Pankaj Kumar Singh, Restech Pharmaceuticals, and Dr Reddy’s Laboratories were the opponents.
The patent (5057/CHENP/2007) relates to a human monoclonal antibody raised against Programmed Death-1 (PD-1) protein, which was isolated from transgenic mice. Among multitude of anti-PD1 antibodies isolated, applicant arrived at 5C4 antibody based on its affinity and lower non-specificity compared to other antibodies. Among other arguments, opponents claimed that this 5C4 was already disclosed as Nivolumab, whereas the applicants claimed this is as an improved version of Nivolumab. This post endeavors to bring to the fore all the interesting arguments (Opponent/Applicant) during the prosecution and how it was eventually granted.
Background
The patent in this case, 5057/CHENP/2007, belongs to a family of patents (US 8779105), which entered India as a national phase application through PCT route PCT/JP06/309606. This patent covers anti-PD 1 human antibody, 5C4. Antibodies are macromolecules, which are glycoproteins produced by our body in response to invading pathogens to neutralize their effect. From 2008, at least 48 therapeutic monoclonal antibodies (mAbs) have been approved by the US FDA and 18 new approved between 2018 to 2019; the revenue for antibodies is expected to hit US $ 300 billion in 2025 (see here). Nivolumab was the first recognised antibody against PD-1, approved in 2014 for the treatment of advanced melanoma, advanced renal cell carcinoma, non-small lung cancer, and metastatic melanoma.
Nivolumab was first invented by Japanese scientists at Medarex, which was licensed to Ono pharmaceutical in 2005. Later, Bristol-Myers Squibb (BMS) acquired Medarex in 2009 for US $2.4 billion. The first filed patent covering Nivolumab dates back to 2006, (EP 1896582 A1) which was granted in 2013 to BMS and Ono Pharmaceuticals, while the corresponding US patent, US 8779105 was granted in 2014.
The sales of Nivolumab went from US $ 2.1 billion in 2015 to $7.6 billion in 2018. Another antibody, pembrolizumab by Merck, was second best in sales for the year 2015 at US $566 million targeting the same protein. Incidentally, BMS and Ono Pharma signed a global patent license agreement with Merck in 2017, wherein BMS and Ono would get 6.5% royalties on sales of pembrolizumab, for settling patent infringement of PD-1 antibody.
Considering the complicated transaction history of the case, we have attached a flowchart to help readers navigate the case history (Fig .1 below).
The Initial set of 105 claims, were amended to only 34 claims based on the objections raised in the FER, with claims related to a method of treatment removed from the initial set. During the opposition proceedings the claims were further amended to 13 claims and finally this application was granted with 8 claims covering the product (antibody) and a composition comprising the product and a pharmaceutically acceptable carrier. The 4 sets of oppositions were filed at various points ranging from 7 years to 10 years after the date of publication. The Applicant had raised concerns about the conduct of the opponents to delay grant of the patent. But the Patents Act clearly allows opponents to file oppositions under Section 25(1) any time before the grant of a patent.
Figure 1: Simplified transaction history of 5057/CHENP/2007
Objections and Arguments
Section 25(1) (b) Novelty: The recurring argument made by the opponents is that the current case reclaimed Nivolumab. Specifically, that WO/2004/04771, filed on 02/07/2003, disclosed human PD-1 antibody for the treatment in cancer, and the Japanese equivalent of this WIPO patent, JP 4409430, sought a patent term extension covering the active ingredient, Nivolumab. Based on the priority date this anticipates the current case. Interestingly, the applicant agreed but added that the prior art did not cover unique sequences of the antibody disclosed as 5C4 in this case, asserting that a drug product can be covered by different patents with different priority dates.
This of course raises the question of whether it is a suspect under Section 3(d), which requires new forms of known substances to show enhanced efficacy.
In another significant argument of the opposition it was pointed out that the applicant had relied on the examples of WO 2006/121168 (5057CHENP/2007) to prove anticancer activity for a European patent, EP 1537878 (family of WO 2004/04771), during the prosecution. An infringement suit had been filed against Merck by the applicants of this case for exploiting EP 1537878 (family of WO 2004/04771) in their patent for pembrolizumab. The opponents suggested that the applicant might not have used these prior art as references unless there was overlap in its scope and coverage. In essence, the oppositions eluded that the antibody claimed in 5057/CHENP/2007 was already documented in the prior art as Nivolumab. However, the applicant denied these stating that none of the prior art had the sequence of complementarity determining regions (CDRs) of the antibody, as claimed in the current case. Considering all the arguments made by the opponents, the applicant had accepted that prior art claimed Nivolumab, whereas the current case claimed 5C4 antibody, with the only difference between Nivolumab and 5C4 is the sequence of the CDRs domain (Fig. 2). (see here for readings on this).
It is worth noting that the opponent referred to EP 1537878, which discloses Nivolumab, and patents covering recombinant mouse models for inherently anticipating 5C4 antibody. A claimed invention anticipates a prior art even if the feature of the invention is found missing, and the prior art was considered to inherently disclose the missing feature. And the applicant dismissed this saying that inherency cannot be established based on “probabilities” or “possibilities” and reasoned that it is hard to arrive at 5C4 with superior properties without any undue experimentation. This is along the lines of what was held by the IPAB in Enercon v. Aloys Wobben (2013).
The Controller’s decision in the current order was in favor of the applicant since none of the prior art carry specific CDRs sequence similar to the current case and deemed the case as new and novel.
Figure 2: Structure of Antibody
Inventive step: Section 25(1)(e) Multiple oppositions argued that it was ‘obvious’ to identify antibodies against a target with high specificity by using a transgenic mouse, based on the combination of teaching references, such as references of PD-1 as a target for modulating immune response (WO 2001/014557), human antibody against human PD-1 (WO 2004/05685, WO 2004/04771), hamster monoclonal antibody raised against mouse PD-1 (EP 1445264 B1 (corresponding to WO 2003/042402)), generating human antibody, conceptualized by Medarex LLC, from a transgenic mouse (WO 2001/009187).
The applicant argued that mere existence of other antibodies cannot be used as an obviousness argument for human anti-PD1 and the concept of somatic hypermutation, a natural selection process for introducing close to 15 amino acid mutations per antibody, rendered “unexpected superior properties” thereby characterizing the inventiveness of this case. (see here for more on this point).
The applicant compared superior binding property of 5C4 to other related antibodies, disclosed in WO 2004/056875 (PD1-28, PD1-33, and PD 1-35), and concluded that to arrive specifically at the 6 CDRs sequence of 5C4 antibody is difficult despite the existence of human or non-human antibody.
Highlighting the distinction between human interventions and natural selection, the opponents retorted that “selection of human germline to produce antibody has been carried out by the biological system of the mouse” with minimal human intervention (Fig. 3), which seem to fall under the purview of someone skilled in the art.
Data scrutiny by the opponent draws attention to the expert affidavit of the applicant, which compares specificity of 5C4 (Figure 14 of the specification) over a prior art antibodies (WO 2004/056875) under two different experimental conditions. Here data was compared to show superior specificity of 5C4, but the data presented were in two different scales to be comparable. The opponent was trying to point out that if the same scales were used for comparison, the 5C4 antibody would have shown only comparable specificity with prior antibodies, rather than the superior specificity claimed by the applicant. The expert affidavit of an opponent further adds that error bars were missing in Figure 14 of the specification, which makes this inconclusive and also scientifically inappropriate since the data from two different experimental conditions cannot be compared. In addition, the cytokine release which reflects level of antibody in the blood, in response to different antibodies were compared with 5C4, and all the antibodies exhibited comparable cytokine levels with no superior property of one antibody over the other. In essence, the opponent put forth the argument that 5C4 only had comparable properties like other antibodies and not superior.
Interestingly, the applicant had not replied to this and the Controller in the order accepted the defense by the applicant for arriving at specific 6 CDRs sequence of 5C4 is non-obvious and inventive.
It is also interesting to note that the applicant used unique sequences of CDR to get over novelty and inventive steps. It is a fact that the applicant’s case has unique sequences in 5C4, but a skilled artisan at the time of filing of this application had prior knowledge about immunizing transgenic mice with antigenic human protein to generate high affinity antibodies. This may be the reason for deleting method claims involved in isolating the antibody and including only product and composition claims for protection.
Figure 3: Process associated with isolating antibody as in 5057/CHENP/2007
Section 25(1)(f) Non-patentable inventions: Section 3(c) of the Patents act keeps check of inventions that are mere discovery of any living or non-living thing. To overcome Section 3(c), the applicant has to prove that subject matter is not naturally occurring and involves human intervention in isolating them, which has to be novel and inventive. All the oppositions delved on the role of human intervention, since the antibody isolated was obtained from a transgenic mouse, which performs its own selection.
The applicant argued that 6 CDRs of 5C4 antibody is different from 6 CDRS of germline sequences, and the process involved more of human ingenuity than mere isolation from the transgenic mouse and proved they are non-naturally occurring sequences based on alignment with the germline sequences. (see here for more).
Further, it was mentioned that PD-1 is a naturally occurring protein and nature prevents production of antibodies against self-antigens. This raises the question of what promoted sequence change in the CDR region:- is it due to somatic mutation (i.e., natural process) or by human intervention? Both the applicant and the opponents relied on previous granted and refused cases in arguing their case to the Controller. The decision rendered this objection mute since the 6 CDRs sequence were not naturally available, and opponents also failed to prove otherwise.
Under Section 3(e), if either substances or processes for producing substances are mere admixture resulting in aggregation of properties, they are not patentable. The composition claim of this case came under scrutiny, but the applicant muted this objection suggesting the combination of novel and inventive antibody with another substance is also novel and cannot be viewed under the purview of Section 3(e). The Controller’s decision was in the favor of the applicant since the opponent failed to prove the non-patentability under section 3(e).
Finally, under Section 3(d), the opponents challenged the claimed antibody as a new form of known substance and asked for therapeutic efficacy of the antibody over the known antibodies, and the applicant’s rebuttal pointed out that none of the prior art carry sequence similar to that of CDRs of 5C4 antibody, and Section 3(d) is not applicable to novel compounds.
With regard to sufficiency of the disclosure, the opponents raised objections on the know-how for the number of mutations in the variable or the CDR region, and lack of information about the source of hybridoma. The opponents reasoned that if the process of obtaining antibodies is through the formation of hybridoma, any skilled person would not be able to comprehend this information without the source and knowledge of the hybridoma. The applicant insisted that the patent was directed towards antibody, and its sequences are clearly disclosed and no deposition was required. The applicant went on to clarify who is a skilled person in the art, and emphasized that the patent sufficiently discloses information for someone skilled in the art, and sufficiency of disclosure must be ascertained from the viewpoint of person possessing average skill, who can follow the instructions of the examples provided, and not directed to any general public. Further it was pointed out that the application sufficiently discloses the sequence information of the variable region and it must have been obvious to a person skilled in the art that the other regions beyond CDRs are constant.
Conclusion
This case bordered between human intervention and the natural process under the blurred lenses of Section 3(c). But is there a threshold for a natural process beyond which one can term it as human intervention? Transgenic mice are a result of human ingenuity, but these mice have become a natural source for obtaining antibodies from the time of creation. Assuming the steps involved in isolating antibodies from transgenic mice are known in the art, would antibodies produced from transgenic mice constitute inventiveness? The 2013 Indian biotech guidelines (page 11) give direction on non-patentable products isolated directly from nature, but when it comes to the process of isolating them it reroutes to Section 2(1)(j). More clarity in categorizing natural against human intervened process would put the inventive argument to rest. Also patentability criteria for a product by process involves providing evidence for modification in the process that results in products that are distinct in properties over the products known in the prior art. Such technical evidence may vary from case to case and this case brought those issues to the fore.