Biologic drugs are large molecules, such as therapeutic proteins, DNA vaccines, monoclonal antibodies, and fusion proteins, that are typically derived from living cells and used in the treatment, diagnosis, or prevention of disease. Most biologics are produced by genetically engineering living cells to express the therapeutic proteins rather than through traditional chemical synthesis. As proteins are sensitive to the conditions in which they are produced, purified, and stored, biologics are more difficult to chemically characterize and manufacture than small molecule drugs, such that even minor differences in manufacturing processes or cell lines can alter or destroy the desired proteins. Consequently, methods of manufacturing biologics are specialized and critically important to the reliable production of a given biologic.

Patent protection for biologics therefore includes inventive processes for reliably, safely, and consistently manufacturing biologic molecules, in addition to traditional patents such as composition-of-matter patents, formulation patents, and method of treatment patents.  While method of treatment and formulation claims are often attacked on obviousness grounds in view of composition-of-matter-related prior art, method of manufacturing claims in the field of biologics are infrequently challenged given difficulties in identifying invalidating prior art.

On February 15, 2018, in a rare victory for a biosimilar applicant, the Patent Trial & Appeal Board (“PTAB” or “the Board”) cancelled claims 1–17 and 19–24 of Amgen’s U.S. Pat. Number 8,952,138 (“the ’138 patent”), directed to a method of refolding a protein expressed in a non-mammalian expression system in IPR2016-01542.  Only claim 18, which further required that the incubation step of the method be performed under non-aerobic conditions, was not held unpatentable.

During the inter partes review (“IPR”) challenge, Apotex and Amgen also litigated the issues of invalidity and infringement of the ’138 patent in the United States District Court for the Southern District of Florida arising from Apotex’s abbreviated Biologics License Applications seeking permission from the Food and Drug Administration (“FDA”) to market biosimilar versions of pegfilgrastim (Neulasta®) and filgrastim (Neupogen®).[1]  The district court found that Apotex had failed to meet its burden of establishing by clear and convincing evidence that the ’138 patent is invalid for obviousness, but it also found that Amgen had failed to prove that Apotex’s proposed commercial marketing of the two products, pursuant to Apotex’s applications, would infringe the ’138 patent, either literally or under the doctrine of equivalents, as reported here.  On appeal, the Federal Circuit affirmed the district court’s finding of non-infringement, as reported here.[2]  While Amgen argued that the district court had found the claims non-obvious, the Board explained that the standards are different between the two proceedings, and the district court’s decision was not binding on the Board.

Applying the broadest reasonable interpretation standard, the Board construed various claim terms including “protein,” “final thiol-pair ratio,” “redox buffer strength,” “refold mixture,” and “complex protein.”  In resolving a dispute over the construction of “complex protein” the Board grappled with the following competing definitions in the specification: 1) “complex protein, i.e., a protein that (a) is larger than 20,000 MW, or comprises greater than 250 amino acid residues, and (b) comprises two or more disulfide bonds in its native form”; and 2) “complex proteins (e.g., proteins comprising 2-23 disulfide bonds or greater than 250 amino acid residues, or having a MW of greater than 20,000 daltons).”  The Board concluded that the specification set forth the definition of “complex protein” multiple times by using “i.e.” (id est, or “that is”) in contrast to the phrase introduced by “e.g.” (exempli gratia, or “for example”), which the Board concluded “does not indicate a definition.”

The ’138 patent explains that the expression of recombinant proteins in prior art prokaryotic systems was problematic in that the expressed proteins have limited solubility precipitates called inclusion bodies, which are improperly folded proteins.  It states that such complex molecules could not be refolded at high concentrations, i.e., concentrations of 2.0 g/L and higher, with any meaningful degree of efficiency on a small scale, and notably not on an industrial scale.  The ’138 patent’s solution to this problem was a method involving the following steps: (a) contacting the protein with a refold buffer comprising a redox component comprising a final thiol-pair ratio having a range of 0.001 to 100 and a redox buffer strength of 2 mM or greater and one or more of: (i) a denaturant; (ii) an aggregation suppressor; and (iii) a protein stabilizer; to form a refold mixture; (b) incubating the refold mixture; and (c) isolating the protein from the refold mixture.

In assessing motivation to combine, the Board was not persuaded by Amgen’s arguments that a person of ordinary skill in the art would not have combined the references because the secondary reference taught a “chemical approach” to achieve protein refolding at high concentrations, while the primary reference used a “physical approach” by diluting the protein molecules.  The Board found that the primary reference taught using a customizable refolding buffer with a redox buffer option (a chemical approach) and that the secondary reference taught experimenting with concentration ranges to find workable ranges (including dilution).

In assessing reasonable expectation of success, the Board was also not persuaded by Amgen’s arguments that host-cell contaminants would lead one of ordinary skill in the art not to have an expectation of success, as model proteins are not predictive of or applicable to recombinant proteins expressed in mammalian expression systems.  The Board found that the authors of the relied upon exhibit did not come to this conclusion, that Amgen relied on the single worst example in the reference in conflict with the Abstract disclosures of the relied upon exhibit, and extrinsic references introduced by Amgen’s expert insufficiently supported his unpredictability arguments.  Characterizing the determination as “a close call,” the Board concluded that a person of ordinary skill in the art would have looked to solve the problem of refolding proteins at higher concentrations, and would have known that the chemical approach of the secondary reference could apply to the dilution refolding methods of the primary reference.

In attacking the prior art references, Amgen argued that the references did not describe the thiol-pair ratio (“TPR”) and the redox buffer strength (“RBS”) equations disclosed in the ’183 patent’s specification, which Apotex’s expert used to calculate that the prior art taught the same ratios and concentrations of oxidant and reductant as required by the claims.  While characterizing it as an “interesting argument,” the Board rejected Amgen’s argument because “to hold otherwise would eviscerate long-standing legal precedent and simply allow for the patenting of inventions whose only contribution was to quantify into a previously unwritten equation relationships that were discernible to one of ordinary skill in the art from the prior art.” The Board explained that a fact finder must necessarily use these equations to determine whether the prior art taught the ratios and concentrations that are required by the claims.  Accordingly, as the claimed TPR and RBS were found in the prior art, the Board concluded that the combination of the prior art references rendered independent claim 1 obvious.

The Board also rejected Amgen’s argument that refolding of complex proteins as required by dependent claims 9-11 were “extremely difficult” and “challenging,” finding that while refolding proteins is difficult and challenging, the person of ordinary skill in the art is highly skilled and would have recognized that the methods of the prior art references were applicable to those types of proteins.

Finally, in rejecting Amgen’s attempt to antedate a journal publication from 2009 relied on by Apotex’s expert with evidence of earlier-created internal Amgen presentations indicating that a protein code-named “AMG 745” allegedly falls within the scope of the claims, the Board found that the code names of proteins provided no real identification of the type of protein and were insufficient to prove earlier invention.  The Board held that purported antedating documents relied upon to teach a specific type of protein must give a more credible identification of what the protein is in order to be persuasive.

Amgen, in another rarity for a pharmaceutical patent owner, did not assert any objective indicia of non-obviousness.

Thus, in a rare victory for a challenger of method of manufacturing claims by a biosimilar applicant, Apotex successfully convinced the Board that all but one claim of the ’183 patent were unpatentable as obvious over a combination of prior art publications.  Interestingly, by successfully cancelling claims which it was adjudged not to infringe, Apotex may have needlessly cleared the way for competing biosimilar manufacturers.  It remains to be seen if this victory encourages increased challenges of method of manufacturing patents by biosimilar applicants, or if it is chalked up to a rare case involving relatively strong prior art.


[1] Amgen Inc. et al. v. Apotex Inc. et al., No. 0:15-CV-61631-JIC/BSS (S.D.Fla.).

[2] Amgen Inc. v. Apotex Inc., No. 2017-1010, 2017 WL 5256264 (Fed. Cir. Nov. 13, 2017).