From our Foreign Correspondents --
While many patiently await developments in the pending U.S. interference proceedings relating to the CRISPR patents in the U.S., matters are progressing in Europe. The Opposition Division (OD) of the European Patent Office (EPO) has just issued (on June 29, 2021) its written decision in the case of EP3241902, owned by the University of California/Berkeley, the University of Vienna, and Emmanuelle Charpentier (collectively termed "CVC"). While the headline decision of the OD to revoke the patent was delivered orally at the hearing on April 13, 2021, the OD has now provided its detailed reasoning behind the decision.
Formally, EP3241902 was revoked for lack of inventive step. However, pivotal in this decision was the finding by the OD that the patent was not entitled to priority from at least the first priority claim of May 25, 2012 (US 61/652,086; "P1"). In view of this, the inventors' own landmark publication from June 28, 2012 ("Jinek 2012") became prior art.
Readers may recall that the Broad Institute's European patent EP2771468 was previously revoked for lack of novelty as a result of lack of priority entitlement. However, the basis for that decision on priority was very different from the present case. Broad's priority claim fell due to a formal defect in the priority claim (as a result of removal of an inventor at the PCT stage), meaning that the applicant for the patent was not the same entity (or successor in title) as the applicant for the priority application.
In contrast, CVC's patent was denied priority for two more substantive reasons, relating to the content of P1 vis-à-vis the claims of the granted patent, namely:
1) P1 lacked a sufficient (enabling) disclosure to support the claims of the patent; and
2) P1 did not disclose the "same invention" as claimed in EP3241902.
To give a little background, EP3241902 had various claims directed to "chimeric Cas9 proteins" and their uses. In the "chimeric Cas9 proteins," a Cas9 enzyme with "reduced nuclease activity" is fused to a heterologous effector domain (e.g., transcriptional activator/repressor, methylase, etc.). The Cas9 component is intended to act as a DNA-targeting protein, delivering the effector domain to a desired target site, somewhat akin to the use of TALEN and zinc finger nucleases well-known in the art. Because EP3241902's earliest priority date of May 25, 2012 antedated Jinek 2012, a successful challenge of the priority claim was required in order to make Jinek 2012 available as prior art and, ultimately, to provide a basis for revocation.
In the first preliminary opinion issued on February 13, 2020 along with a summons to oral proceedings, the OD initially considered that the patent was entitled to its earliest priority date, and therefore Jinek 2012 was not available as prior art. In arriving at this preliminary view, the OD was persuaded by CVC's arguments that the sufficiency was evidenced by the fact that several research groups were able to carry out aspects of the claimed invention following on from the Jinek 2012 publication.
In the run up to oral proceedings, which were cancelled due to COVID, the opponents filed further submissions focused heavily on the lack of an enabling disclosure in P1, and, moreover, that the disclosure in P1 was extremely limited compared to Jinek 2012. Furthermore, the success of the research groups reported in the post-published articles was the result of significant research endeavor, building on the more extensive disclosure in Jinek 2012 than was found in the P1 specification.
In view of these submissions, the OD issued a revised preliminary opinion on September 29, 2020, in which the OD reconsidered its position on the enablement in view of P1. The OD provisionally concluded that the claimed invention was not plausible in view of P1 which, inter alia, did not disclose how to reduce nuclease activity of Cas9. Because the disclosure in P1 did not render the claimed subject matter plausible, under EPO jurisprudence CVC could not rely on post-filing evidence to remedy the lack of enablement. The OD deferred venturing an opinion on novelty or inventive step until entitlement to priority had been discussed fully at the oral proceedings rescheduled for April 12-14, 2021.
The oral proceedings were scheduled to take place by videoconference over three days; this is atypically long for opposition hearings, reflecting the complexity of the case and number of parties (albeit fairly typical for the European CRISPR hearings). However, by the end of the second day, the patent had been revoked in its entirety. The minutes accompanying the decision reveal that the majority of the discussion centered around entitlement to priority. The Main Request and 10 Auxiliary Requests (representing a series of "fallback" amended claim sets) filed by CVC were all found to lack entitlement to priority and were therefore obvious in view of a combination of the TALEN and zinc finger nuclease systems well-known from the art and the disclosure of Jinek 2012.
Examining the decision further, there are two distinct strands to the findings of the OD on entitlement to priority:
The first relates to the concept of Cas9 with "reduced nuclease activity". While P1 mentioned in broad terms the concept of Cas9 with "reduced nuclease activity," there was no material disclosure in the P1 specification of how the skilled person might actually produce such a modified Cas9 without an undue burden. OD found that the disclosure in P1 was vague and speculative regarding the regions of Cas9 conferring nuclease activity against a target DNA (later discovered to be the RuvC and HNH domains), and gave no meaningful guidance to the skilled person on how nuclease activity might be modified. Moreover, the OD further found that the sole example in P1 was ill-suited to identify Cas9 variants with reduced nuclease activity. There was considerable debate about the level of common general knowledge of the skilled person, and the extent to which this might supplement the disclosure of P1. The OD did not arrive at a settled view on the common general knowledge, but it concluded that at best the skilled person would be able to produce a completely nuclease-inactive Cas9 (so-called "dead Cas9") without an undue burden. The use of a dead Cas9 was also consistent with the TALEN and zinc finger nuclease systems known from the art, which do not have any intrinsic nuclease activity. Thus, inasmuch as the claims related to reduced but not eliminated nuclease activity, they were not enabled by P1.
Secondly, the OD considered whether Auxiliary Request 2, with claims limited to Cas9 with "substantially no nuclease activity", was entitled to priority from P1. Priority was again denied but for a quite different reason. The claimed subject matter was held not to relate to the "same invention" as disclosed in P1, as required under Art. 87 EPC. The "same invention" requirement is assessed using the strict "direct and unambiguous" disclosure test favored by the EPO (see G 2/98). Readers may have encountered essentially this same strict test in the more familiar context of assessing basis for amendments during EP prosecution. In brief, the OD found that while the claims related to a modified but otherwise whole Cas9 protein linked to a heterologous domain, the disclosure in P1 was limited to a portion of Cas9 (i.e., not the whole protein) linked to a heterologous domain. P1 disclosed in general terms various hypothetical embodiments of chimeric Cas9s comprising an "RNA-binding portion" derived from Cas9 linked to a heterologous domain, and in the OD's view this "RNA-binding portion" was always disclosed in P1 as a sub-portion of Cas9 (i.e., excluding the "activity portion" of Cas9). This issue was compounded by the absence of any actual example of a chimeric Cas9 protein in P1. While the "same invention" issue arose with respect to Auxiliary Request 2, it applied equally to all of the requests filed by CVC.
CVC were given the opportunity to propose further amendments in an attempt to overcome the priority issues, but they failed to do so. CVC have filed a notice of appeal against the decision, and the issuance of the written decision sets a four month deadline for them to file their complete grounds of appeal.
Thus, another CRISPR patent has been revoked in EPO opposition proceedings because of priority issues. While the Broad case drew attention to the strict formal requirements of priority at the EPO, the present case serves as a reminder of the particularly strict requirements of the EPO that the priority document discloses the "same invention" as the patent. Regarding enablement, the EPO has a reputation for being comparatively patentee friendly, placing a correspondingly onerous burden on opponents to prove "serious doubts substantiated by verifiable facts," with any benefit of the doubt typically going to the patentee. This can be a difficult challenge for opponents, but where a priority document is thin on experimental work and largely speculative, as seems to be the case here, priority can fall on this ground.
EP3241902 is one of a number of patents in the name of CVC currently in the opposition and appeal process at the EPO. EP2800811 (the parent of EP3241902), directed to Cas9 and single-guide RNA for DNA cleavage, was maintained in amended form by the OD and that decision is currently under appeal. EP3401400 is directed to use of Cas9 in eukaryotes, and oral proceedings before the OD are scheduled to begin on November 29, 2021. It is likely that the CRISPR patents will continue to face such challenges in view of the nature of the invention and its powerful and flexible applications in a number of commercially important areas of genetic modification.
