Can Exelixis keep out generics until 2030 even if it loses on the ‘776 patent?
Zachary Silbersher
I previously blogged about the Exelixis patent dispute. The Cabometyx patent trial occurred last week (May 16). A recent note on Seeking Alpha indicates that comments from the court during the trial suggests that Exelixis may lose on the ‘776 patent, but win on the ‘473 patent. (I was one of two patent experts on the Truist call referenced in the note.) If that happens, will Exelixis still be able to secure freedom from generics through 2030 based upon the ‘439, ‘440 and ‘015 patents asserted in its second lawsuit?
The ‘776 patent
The ‘776 is a polymorph patent of cabozantinib, and. The patent requires showing one of three measurements to show infringement: (1) at least four peaks from solid state 13-C NMR spectrum; (2) powder x-ray diffraction pattern with specific values; (3) powder x-ray diffraction with pattern in FIG. 8 of the patent. Based on statements made within the pre-trial order, Exelixis appears to have focused at trial on the first measurement – namely, solid state 13-C NMR spectrum. That may because it is the easiest to prove. For instance, FIG. 8 from the ‘776 patent has roughly 18 peaks.
In advance of the trial, there was some discussion that MSN purportedly developed its own proprietary polymorph, Form S, that will be used in MSN’s ANDA product. Based upon letters sent to the court during discovery (see e.g., Dkt. 61), the parties appeared to generally agree that MSN’s Form S is likely embodied within MSN’s own foreign patent, namely, WO2018/104954 (the ‘954 patent).
The claims of MSN’s ‘954 patent identify the powder x-ray powder diffraction peaks for Form S as follows: 6.6, 7.6, 8.2, 10.0, 10.5, 12.5, 13.3, 15.1 , 16.4, 18.2, 19.0, 20.2, 20.6, 21.6, 22.6, 25.1 , 26.1 , 27.0, 28.1 , 29.8 and 32.8 ± 0.2 degrees two theta. By contrast, the Exelixis ‘776 patent identifies the following required powder x-ray powder diffraction peaks: 2θ values at 20.9±0.2 °2θ (no overlap) and 21.9±0.2 °2θ (no overlap), and two or more 2θ values selected from: 6.4±0.2 °2θ (overlap), 9.1±0.2 °2θ (no overlap), 12.0 (no overlap). Comparing these two shows that there is likely only overlap once.
What does this mean? Proving patent infringement does not involve comparing one patent to another. And MSN’s defense to the allegation that its ANDA product will infringe the ‘776 patent was not that it happens to have its own patent. Yet, if we assume that MSN’s ANDA product includes Form S, and if we assume that Form S is really embodied in ‘954 patent, then the comparison above provides some transparency into how difficult it likely was for Exelixis to prove infringement of the ‘776 patent at trial. If there was only one overlapping peak in the x-ray powder diffraction between the ‘776 patent and the peaks characterizing MSN’s polymorph, then when it came to the ‘776 patent, Exelixis may already been behind the eight-ball going into the trial.
The ’439, ‘440 and ‘015 patents
If Exelixis wins on the ‘473 patent, but loses on the ‘776 patent, then that presumably means that Caboymetyx® will have freedom from MSN’s generic at least until expiration of the ‘473 patent—namely 2026. (That conclusion is contingent on the outcome of any appeal—and there will definitely be an appeal. Most likely, there will be two appeals—MSN will appeal the ‘473 patent, and Exelixis will appeal the ‘776 patent.)
Yet, Exelixis has also commenced a separate suit that asserts three additional patents—namely, the ‘439, ‘440 and ‘015 patents. Those patents expire in 2030. Accordingly, if Exelixis loses on the ‘776 patent, will that decision read-through to a loss on any chance of exclusivity until 2030?
The ’439, ‘440 and ‘115 patents each share the same specification as the ‘776 patent. Accordingly, that might suggest that if Exelixis cannot prove infringement of the ‘776 patent, that decision may read through to the ‘439, ‘440 and ‘015 patents. Yet, a closer look at these patents suggests otherwise.
The ‘439 and ‘440 patents are not polymorph patents in the same way that the ‘776 patent is. Instead, they are technically directed to a crystalline malate salt of cabozantinib. The patents are generally broader than the ‘776 patent. They do not require x-ray powder diffraction peaks or peaks from solid state 13-C NMR spectrum to show infringement.
Separately, the ‘015 patent is directed to a method of treating cancer with the crystalline malate salt, and therefore, infringement of the ‘015 patent is likely to rise and fall with findings for the ‘439 and ‘440 patents.
Altogether this means it may be much easier for Exelixis to show infringement of ‘439 and ‘440 patents as compared to the ‘776 patent. In short, even if Exelixis loses on the ’776 patent, read-through to the ‘439 and ‘440 patents should not be presumed. There is a very real universe where Exelixis can prove infringement of the ‘439 and ‘440 patents, even if it could not do so for the ‘776 patent.
On the other hand, precisely because the ‘439 and ‘440 patents are broader than the ‘776 patent, that may also mean it will be easier for MSN to invalidate these patents. The ‘776 patent is a very narrow patent—it requires multiple precise x-ray powder diffraction peaks or NMR spectrum peaks to show infringement. This is likely one of the reasons that MSN dropped any challenge to the validity of the ‘776 patent at trial, and relied exclusively on proving non-infringement.
The examination histories for the ‘439 and ‘440 patents show that Exelixis acquired these patents by presenting purported evidence of unexpected results. Specifically, Exelixis claimed it was unexpected and surprising that the claimed crystalline malate salt in the ‘439 and ‘440 patents has superior dissolution properties compared to amorphouscabozantinib malate salt. Exelixis also claimed it was unexpected that the claimed crystalline malate salt also had enhanced solubility compared to the free base of cabozantinib or other salts of cabozantinib.
During examination of the patents, Exelixis submitted experimental data and a declaration from a scientist to back up these claims. The Patent Examiner at the Patent Office accepted Exelixis’ representations and experimental data, but does not thoroughly vet Exelixis’ claims (which is not unusual for these types of patents.) Accordingly, MSN will likely try to poke holes in Exelixis’ experiments. MSN will also attempt to undermine Exelixis’ contention that the dissolution properties of the claimed crystalline malate salt were really surprising or unexpected. Rather, MSN will try to prove that persons of skill would have expected crystalline malate salt of cabozantinib to have superior dissolution properties. Indeed, the parties are likely currently conducting discovery on these very issues the second pending lawsuit.
Accordingly, whether or not Exelixis can keep out generics until 2030—namely, when the ‘439, ‘440 and ‘115 patents expire—likely depends on whether Exelixis’ representations to the Patent Office—namely, that claimed crystalline malate salt of cabozantinib has superior and unexpected dissolution properties—actually holds up. Indeed, if the case reaches trial in May 2023, this issue will likely be the centerpiece of that trial.