Can a prior art reference that does not contain a teaching sufficient enough to allow it to be used in an obviousness combination nevertheless be used as a background reference for evidence of motivation to combine?  Moreover, does it matter if an adjudicating body had concluded earlier in a proceeding that the reference was "inapplicable" to present invention?  Such a situation occurred in IPR2014-00784 (Torrent Pharmaceuticals Ltd. v. Novartis AG), in which the PTAB stated in an institution decision that a particular prior art's "stated reasons for using mannitol in liquid pharmaceutical compositions are inapplicable to its potential use in connection with solid pharmaceutical compositions."  Id., Paper 11, at 13 (emphasis added).  The Board did institute trial on other grounds, and ultimately invalidated the challenged claims as obvious.  On appeal, Novartis, the patent owner, argued that it had relied on this statement made by the PTAB in the institution decision when it prepared its Patent Owner Response and otherwise fashioned its trial strategy.  In other words, Novartis claimed it would have conducted the trial differently had it known that the Board would eventually use this reference in its obviousness calculus.  Last week, however, in Novartis AG v. Torrent Pharmaceuticals Ltd., the Federal Circuit was unpersuaded, and affirmed the Board's invalidation decision.  Specifically, the Court reasoned that the Board had not changed theories midstream, and as such, the due process requirements of the Administrative Procedure Act (APA) were satisfied.  The Federal Circuit also affirmed the challenges to the PTAB's obviousness determination.

The technology at issue in U.S. Patent 8,324,283 (the '283 patent) is solid pharmaceutical compositions of a sphingosine-1 phosphate (S1P) receptor agonist and a sugar alcohol, such as is found in the formulation of Gilenya^® marketed for the treatment of certain forms of multiple sclerosis.  According to the patent, S1P receptor agonists are normally difficult to formulate as solid compositions.  However, the inventors found it was easier to do so by using a sugar alcohol excipient like mannitol.  The particular S1P receptor agonists used in Gilenya^® is FTY720, or fingolimod.  The two independent claims of the '283 patent are:

1.  A solid pharmaceutical composition suitable for oral administration, comprising:
    (a) a S1P receptor agonist which is selected from 2-amino-2-[4-(3-benzyloxyphenoxy)-2-chlorophenyl]propyl-1,3-propane-diol or 2-amino-2-[4-(3-benzyloxyphenylthio)-2-chlorophenyl]propyl-1,3-propane- -diol, 2-amino-2-[4-(3-benzyloxyphenylthio)-2-chlorophenyl]-2-ethyl-1,3-pr- opane-diol, and its phosphates or a pharmaceutically acceptable salt thereof; and
    (b) a sugar alcohol.

19.  A solid pharmaceutical composition suitable for oral administration, comprising mannitol and 2-amino-2-[2-(4-octylphenyl)ethyl]propane-1,3-diol or a pharmaceutically acceptable salt thereof.

Torrent filed its IPR petition on May 27, 2014, alleging that all claims were obvious in view of the combination of U.S. Patent 6,004,565 (Chiba) and Pharmaceuticals: The Science of Dosage Form Design (Aulton).  Torrent also alleged two additional grounds that part of the claims were anticipated by U.S. Patent 6,277,888 (Sakai), and all claims were rendered obvious by the combination of Sakai and Chiba.  Chiba disclosed the use of immunosuppressive compounds, including fingolimod.  Aulton taught the use of tablets and capsules to administer drugs orally, and explained that mannitol was "expensive," but "commonly used."  Finally, Sakai described liquid formulations of fingolimod, and explained that the use of sugar alcohols could result in liquid solutions that were less irritating.  In its institution decision, the Board made the statement quoted above about Sakai -- that the stated reasons for using mannitol were inapplicable to solid formulations.  After institution, Apotex and Mylan also filed a petition that was joined to the Torrent proceeding.

In its Final Written Decision, the PTAB found that Chiba and Aulton strongly suggested the claimed invention, but also cited to Sakai as "directly instruct[ing] that the two ingredients should be combined."  Therefore, according to the Board, there was sufficient reason to combine Chiba and Aulton.  Part of this decision rested on a paper authored by Novartis' expert Dr. Stephen Byrn, in which he said "solution studies can be very helpful" in understanding solid state drug degradation.  At the same time, the Board also found that additional references provided motivation to use mannitol as a diluent in tableting.  Finally, the Board rejected both of Novartis' arguments related to objective indicia of non-obviousness.  First, even though the low dose stability of fingolimod when combined with mannitol may have been unexpected, the Board found that the independent claims were "not commensurate in scope" with this fact because they were not limited to a particular dose range.  Second, the Board found that the nexus for long-felt need, industry praise, and commercial success was missing because solid oral multiple sclerosis treatments were already known in the art.

During the oral argument, the judges focused on Novartis' due process issue, and asked probing questions of both parties.  It was clear the Court was concerned that Novartis might not have been afforded proper notice and a chance to heard, as required by APA.  In other words, it wanted to ensure that the Board had not changed theories "midstream" by first rejecting Sakai, but ultimately relying on it to invalidate the claims.  After reviewing the entire record, however, the Federal Circuit concluded that the APA had not been violated.

First, the Federal Circuit found that the use of Sakai as a background reference "relevant to one of skill in the art in deciding which excipients to use in formulating a solid oral dosage form of fingolimod" was not contrary to its rejection as an anticipatory or primary obviousness reference.  It was Chiba and Aulton that strongly suggested the combination, according to the Court, and Sakai was only cited to reinforce this finding.  The second reason was more suspect.  The Federal Circuit claimed that Novartis could not allege "surprise" because it had debated the merits of Sakai throughout the proceedings.  According to the Court, Novartis addressed Sakai in its Patent Owner Response.  However, Novartis had apparently only done this because Torrent had cited Sakai as a reason to combine Chiba and Aulton.  Moreover, Novartis' response appeared to be nothing more than a parroting of the Board's conclusion that Sakai was not relevant to solid formulations.  As Novartis alleged, it would probably have more adequately addressed Sakai if it had known that the Board would rely upon it so heavily.  Nevertheless, Novartis' position was weakened by its expert's report, which addressed Sakai in some detail.  In addition, Novartis was accused of questioning Torrent's expert "at length" on the issue.  Perhaps if Novartis was a little more silent on the issue, the outcome would have been different.  The Court finally bolstered its conclusion by noting that the Board had additionally cited to several other references, even though none of these appeared to be as relevant as Sakai to its ultimate conclusion.

Novartis also challenged the Board's obviousness decision as allegedly not taking into account the teaching of the references as a whole.  Specifically, Novartis argued that the Board overlooked the known disadvantages of using mannitol, such as its difficulty to manufacture, its impurities, and its expense.  However, the Federal Circuit found that the Board at least addressed the cost of mannitol, but concluded that it was nevertheless "commonly used."  Moreover, the Court found that the Board's consideration of the issue was commensurate with the argument presented by Novartis in the Patent Owner Response.  As such, the Court could not fault the Board for its limited treatment of the arguments.  The Federal Circuit otherwise found that substantial evidence existed to support the Board's finding on the motion to combine Chiba and Aulton.  Finally, the Court could not fault the Board's conclusions with regard to the objective indicia of non-obviousness.  As such, it affirmed the PTAB's decision.

Novartis AG v. Torrent Pharmaceuticals Ltd. (Fed. Cir. 2017)
Panel: Circuit Judges Taranto, Chem, and Stoll
Opinion by Circuit Judge Chen