Will Teva’s PGR against Corcept’s ‘214 patent covering Korlym be instituted?

Zachary Silbersher

The saga over Corcept Therapeutics’s ($CORT) patent battles against prospective generics for Korlym® is approaching another stage.  We previously discussed Teva’s ($TEVA) petition for post-grant review (PGR) of the ‘214 patent.  Corcept has filed its preliminary response to Teva’s PGR, and the PTAB is scheduled to decide whether or not to institute the PGR by about November 23, 2019.  The ‘214 patent is potentially the strongest patent Corcept is currently wielding against generics, since it arguably does read upon Korlym®’s label.  Now that the papers are in, what are the odds Teva’s PGR is instituted? 

The ‘214 patent is directed to a dosing regime for mifepristone when co-administered with a CYP3A inhibitor, such as ketoconazole.  The patent specifically covers a method of reducing a once-daily dosage of mifepristone from 900 mg to 1200 mg down to 600 mg when the drug is co-administered with a CYP3A inhibitor, including ketoconazole.

For decades, both mifepristone and ketoconazole have been recognized treatments for patients suffering from Cushings syndrome and related conditions.  Cushings syndrome is essentially characterized by hypercortisolism, or excess cortisol levels in the body.  Ketoconazole and mifepristone can each treat excess cortisol levels, but it in different ways.  Ketoconazole is a steroidogenesis inhibitor, which can inhibit the synthesis of cortisol.  Mifepristone is a glucocorticoid antagonist (GRA) that binds and blocks cortisol receptors, and thus, reduces the effects of cortisol on the body.

Likewise, when the FDA approved Corcept’s application to distribute Korlym® for Cushings syndrome in 2012, the agency was aware that patients are likely to also be taking ketoconazole.  In addition to being a steroidogenesis inhibitor, ketoconazole is also a CYP3A inhibitor.  That means, it could potentially create toxicity issues when ketoconazole is co-administered with mifepristone.  Mifepristone is a CYP3A substrate, and is thus metabolized by CYP3A enzymes, which means that co-administering ketoconazole with mifepristone could increase mifepristone levels. 

Because of this concern, the FDA approved Corcept’s application for Korlym®, but simultaneously instructed the company to conduct studies to evaluate the drug-drug interactions (DDI) between ketoconazole and mifepristone.  These recommendations were memorialized in an FDA pharmacology review of the Korlym® NDA, which was included in the drug’s approval package.  (The pharmacology review is referred to as the “Lee” reference in Teva’s PGR.) 

The DDI study that Corcept undertook, which is referred to as the Nguyen-Minze study, was published in 2017.  The study showed that co-administration of ketoconazole and mifepristone increased the concentration of mifepristone by approximately one third (28% Cmax and 38% AUC).  The study further showed that co-administering ketacozonole with upwards of 600 mg/day was purportedly well-tolerated and effective. 

The drug hit the market, however, before the Nguyen-Minze study was completed.  Accordingly, the FDA initially required that the Korlym® label include a warning about co-administration with CYP3A inhibitors.  Without the benefit of the study, which had yet to be completed, the FDA purportedly (according to Teva) adopted a conservative approach that recommended not exceeding 300 mg/day of Korlym® for patients also taking a CYP3A inhibitor, such as ketoconazole.  According to Teva, the FDA also contemplated that the label would be amended once the DDI study was completed.

During prosecution of the ’214 patent, Corcept made two principle argues to convince the Patent Office to grant the patent.  First, Corcept argued that the results of the DDI study were unexpected—namely, it was not expected that administering 600 mg/day of mifepristone with ketoconazole would be discovered to be both effective and well-tolerated.  Corcept cited the “Greenblatt effect,” which was based upon an article authored by David Greenblatt.  Greenblatt’s article suggested that coadministration of a CYP3A inhibitor with a CYP3A substrate could increase the concentration of the substrate by 7 to 20 times.  Yet, because the Nguyen-Minze study showed that the concentration of mifepristone (a CYP3A substrate) when coadministered with ketoconazole only increased by about 3 times, Corcept argued this was unexpected.  Corcept characterized this purported “failure to observe the Greenblatt effect” as evidence of non-obviousness.  As further evidence, Corcept also argued that the impact of ketoconazole on co-administered drugs is essentially unpredictable.  Therefore, Corcept argued that its discovery that a dose-reduction to 600 mg of mifepristone when taken with ketoconazole was well-tolerated warranted patent protection.

In its PGR, Teva attempts to dismantle this argument.  Teva argues that the results of the Nguyen-Minze study were not necessarily unexpected.  Rather, those results were simply not known before the study was conducted.  Indeed, that is precisely why the FDA instructed Corcept to conduct a DDI study involving CYP3A inhibitors and mifepristone—for the purpose of more accurately determining the safe and effective dosage at which mifepristone could be co-administered with ketoconazole. 

Teva’s PGR includes a declaration from Dr. Greenblatt himself, the same Dr. Greenblatt whom Corcept cited during prosecution of the ‘214 patent.  Dr. Greenblatt now testifies that the purported “Greenblatt effect” is not a hard-and-fast rule.  Rather, drug-drug-interactions allegedly always have an element of unpredictability, which is why empirical testing is typically required to pin down the precise results of a particular DDI.  But because a DDI is inherently unpredictable does not, according to Teva, mean that running a test to make it known necessarily warrants a patent.   

From the perspective of patent law, this is a fairly powerful argument.  Corcept was awarded a patent by identifying a purported principle of drug-drug interactions based upon an article authored by a certain Dr. Greenblatt.  Corcept used that alleged effect, which it coined as the “Greenblatt effect,” to argue that its discovery in the ‘214 patent application warranted grant of a patent.  Now, in its attempt to invalidate Corcept’s patent, Teva has marshaled the testimony of that very same Dr. Greenblatt to undercut the very arguments made by Corcept during prosecution of the patent in the first instance.   

This is no doubt strong evidence, but it is also not necessarily a deathblow.  Exactly how a typical scientist would have read Dr. Greenblatt’s earlier article is still relevant, despite what Dr. Greenblatt now testifies.  (Also, from a practical view, Corcept’s response to Teva’s petition shows that it has the unfortunate exercise of affirming Greenblatt’s expertise for certain opinions, such as the Greenblatt effect, while attempting to impeach his expertise for other opinions.) 

In countering Teva’s PGR, Corcept come close to arguing that because drug-drug interactions are always unpredictable, then any outcome that Corcept discovered would have warranted patent-protection.  Teva cites legal authority holding that is not the law.  Rather, the results of the Nguyen-Minze study were simply a matter of routine optimization.  In other words, according to Teva, it was simply a matter of running a test according to well-known clinical protocols to discover the optimal level of mifepristone that can be safely administered with ketoconazole.

Patent law has long recognized that discoveries resulting from routine optimization do not, generally, warrant a patent.  Corcept claims that its pathway to the Nguyen-Minze study was anything but routine, but does not necessarily highlight any particular breakthroughs in the process that would be tantamount to defending the “inventiveness” of its patent.  Either way, this could be a fact-issue that the PTAB chooses to resolve after institution, once more evidence is gathered.  (Corcept also strangely argues that it first had to confirm that ketoconazole could be administered with even low levels of mifepristone, which begs the question why the drug was approved with that recommendation in the first place.)

Corcept’s second argument during prosecution of the ‘214 patent was that the Korlym® label itself taught away from administering 600 mg of mifepristone with ketoconazole.  This was because the label recommended that when the drug is co-administered with CYP3A inhibitors, the dose of the drug should be reduced to 300 mg a day.  Corcept included declarations from numerous doctors testifying that, in light of the label’s recommendation, they would be wary of prescribing more than 300 mg a day when co-administered with ketoconazole.  

Teva’s PGR also attacks this argument.  Teva points out that Korlym®’s label recommends limiting mifepristone to 300 mg/day when co-administered with ketoconazole, but it is not a black-box warning.  The label went further and indicated that, based on a particular patient, titrating above 300 mg may be warranted.  Moreover, the recommendation itself was not grounded in a clinical study.  On the contrary, the approval package for Korlym® showed that there was no clinical data on the drug-drug-interactions between mifepristone and ketoconazole.  In fact, that is precisely why the FDA instructed Corcept to run a test to gather that data.  Thus, according to Teva, rather than teaching away from administering a higher dosage of mifepristone with ketoconazole, the Korlym® label in combination with the FDA approval package actually encouraged a person of skill in the art to figure out the highest dose of mifepristone that could still be safely and effectively administered with ketoconazole.   

In response to Teva’s PGR, Corcept emphasizes the gravity of the risks posed by co-administration of mifepristone with ketoconazole.  The prior art showed that increasing the plasma concentration of mifepristone could cause side effects such as adrenal insufficiency and hypokalemia.  The prior art also showed that CYP3A inhibitors themselves carried potentially life-threatening toxicity risks.  Given these risks, according to Corcept, scientific disclosures in the prior art previously recommended co-administering mifepristone with ketoconazole only with extreme caution, only when necessary, and not more than 300 mg/day of mifepristone. 

Corcept also counters Teva’s argument that the inherent unpredictability of the interaction between mifepristone and CYP3A inhibitors warrants invalidation of the ‘214 patent.  In particular, Corcept argues that to successfully invalidate the ‘214 patent, Teva must show that the prior art revealed a “reasonable expectation of success” that co-administration of mifepristone and ketoconazole at 600 mg/day would be safe and effective.  But, according to Corcept, the prior art doesn’t show that.  Rather, the prior art shows the drug-drug interactions between ketoconazole and 600 mg of mifepristone was unpredictable.  Indeed, Dr. Greenblatt testifies to that in connection with Teva’s PGR.  And if it was unpredictable, according to Corcept, then there could not be any reasonable expectation of success.  And without any reasonable expectation of success evidenced by the prior art, Corcept argues that Teva’s PGR must be denied institution. 

This is an important argument, and in many ways, it forms the crux of Teva’s PGR.  This dispute illustrates an interesting intersection of two relatively cornerstone principles in the law of patent obviousness.  On the one hand, Corcept argues that if the science is unpredictable, then a discovery in that field cannot be obvious.  On the other hand, Teva argues that the science may have been unpredictable, but sorting it out was a routine matter, which does make it obvious.  Both parties arguments have firm grounding patent law, and thus, there is some uncertainty how the PTAB will eventually come down. 

But Corcept may be overstating its case a bit.  It was not unpredictable that strong CYP3A inhibitors increased the concentration of co-administered CYP3A substrates.  The only question was by how much.  Patent law does not necessarily hold a patent should be granted to anyone running a test to make that which was previously unknown now certain.  Indeed, that is Teva’s point—the precise dosages when co-administered mifepristone and ketoconazole were unsafe may itself have been unpredictable, but figuring it out was simply a matter of “routine optimization.”  Corcept argues that Teva’s routine-optimization argument is a ruse because all of the prior art indicated that “extreme caution” should be taken when co-administering both drugs and that the dosage of mifepristone should not exceed 300 mg. 

But that is not necessarily accurate.  Indeed, the FDA’s pharmacology review for Korlym® actually instructed Corcept to conduct a study to figure out if higher levels of mifepristone could be co-administered with ketoconazole.  In other words, Corcept argues that the prior art admonitions about the toxicity of co-administering ketoconozale and mifepristone would have stopped any scientist from exploring a mifepristone dose higher than 300 mg.  But the FDA itself wasn’t stopped from doing so.  Rather, the FDA itself suggested this should be explored.  The FDA’s pharmacology review for Korlym® is in some ways a stick in the wheel of Corcept’s principle rebuttal to Teva’s PGR.

Corcept’s final response to Teva’s PGR is that the petition should be denied institution because Teva is making the very same arguments to invalidate the ‘214 patent in the pending district court case, and therefore, the PGR itself is inefficient.  Corcept argues the district court case is scheduled to go to trial in August 2020, whereas the PGR is not scheduled to issue a final decision until November 2020. 

This argument should not be discounted.  There is precedent for the PTAB exercising its discretion to deny institution of an IPR or PGR when there is parallel district court action that is likely to go to trial before the PTAB renders its decision.  Corcept’s analysis does, however, overlook certain facts.  There is no schedule in place yet guaranteeing trial in August 2020, and even if there was, it is unlikely the court will render its decision until several months after that.  Under Corcept’s logic, a PGR would never be efficient when there is a pending parallel district court case, but that has hardly been the PTAB’s practice.