The Court of Appeal’s decision largely focused on the validity of the patent when arriving at its conclusion. In particular, the main turning point for the reversal of the first instance decision was whether or not it would be obvious apply the in vitro method of a prior art document, “Gȍransson” (also referred to as “D6”) to an in situ method involving cell or tissue samples.
The claimed method
At a simplified level, claim 1 of the patent relates to a method of detecting a number of analytes in a cell or tissue sample by incubating the sample with a number of detection reagents, each detection reagents including (i) a probe reagent being capable of binding to one of the analytes, and being conjugated to (ii) a subsequence that is designed to be hybridised with a decoder probe. The decoder probes are designed to hybridise with a subsequence of the detection reagent and also include detectable label with a signal signature.
The method specifies detecting the subsequences in a temporally-sequential manner by hybridising a set of decoder probes with the subsequence, detecting the signal signature, removing the signal signature and repeating the hybridisation and removal sequence with different sets of decoder probes.
In this way, the method enables the detection and visualisation of multiple targets in a single cell or tissue sample. The specific wording of claim 1 is as follows:
A method for detecting a plurality of analytes in a cell or tissue sample, comprising:
(a) mounting the cell or tissue sample on a solid support;
(b) contacting the cell or tissue sample with a composition comprising a plurality of detection reagents, the plurality of detection reagents comprising a plurality of subpopulations of detection reagents;
(c) incubating the cell or tissue sample together with the plurality of detection reagents for a sufficient amount of time to allow binding of the plurality of detection reagents to the analytes;
wherein each subpopulation of the plurality of detection reagents targets a different analyte, wherein each of the plurality of detection reagents comprises: a probe reagent targeting an analyte of the plurality of analytes and one or a plurality of pre-determined subsequences, wherein the probe reagent and the one or the plurality of pre-determined subsequences are conjugated together;
(d) detecting in a temporally-sequential manner the one or the plurality of pre-determined subsequences, wherein the detecting comprises:
(i) hybridizing a set of decoder probes with a subsequence of the detection reagents, wherein the set of decoder probes comprises a plurality of subpopulations of decoder probes and wherein each subpopulation of the decoder probes comprises a detectable label, each detectable label producing a signal signature;
(ii) detecting the signal signature produced by the hybridization of the set of decoder probes;
(iii) removing the signal signature; and
(iv) repeating (i) and (iii) using a different set of decoder probes to detect other subsequences of the detection reagents, thereby producing a temporal order of the signal signatures unique for each subpopulation of the plurality of detection reagents;
(e) using the temporal order of the signal signatures corresponding to the one or the plurality of the pre-determined subsequences of the detection reagent to identify a subpopulation of the detection reagents, thereby detecting the plurality of analytes in the cell or tissue sample.
Inventive step
The Court of Appeal considered Gȍransson “would have been of keen interest to a person skilled in the art who … was seeking to develop high-throughput optical multiplexing methods for detecting target molecules in a sample.” While the Court of First Instance discussed the concept of closest prior art in relation to Article 56 EPC, the Court of Appeal did not refer to Article 56 EPC and has not explicitly used the EPO’s term of closest prior art. It is not clear whether this is an intentional deviation or not.
Gȍransson describes forming genomic DNA circles from specific genomic DNA sequences isolated from blood samples and selector probes (i) and the DNA circles are amplified by RCA (rolling-circle amplification) (ii) or otherwise enriched (iii) to generate amplified single molecules (ASMs). The ASMs are then immobilised and a random array is generated on a microscopy glass slide. The ASMs immobilised in the array are decoded by sequential hybridisation of sandwich probes, tag probes (red or blue) and a general probe after they have been incubated for one hour on shake at 55°C.
The Court of Appeal considered that the only difference between claim 1 of the patent and Gȍransson is the use of ASMs generated from genomic DNA isolated from blood samples rather than cell or tissue samples. The Court of Appeal concluded that “after successful application of an in vitro multiplex method for the detection of ASMs, the next step was to consider transferring the method to an in situ environment.”
The Court of Appeal justified this conclusion from (i) journal article references in Gȍransson that refer to in situ genotyping and (ii) a prior art reference (Stougaard et al., B30) describing a method for the detection of non-polyadenylated RNA molecules using “a new probe format” (“Turtle Probes”), which was initially carried out in vitro in “a controlled environment” and, after successful implementation was also tested in situ with positive results.”
In contrast, the first instance decision considered that “no such inducement has been presented.” In particular, the first instance decision refers to Nanostring’s expert declaration, which allegedly admits “obvious differences” between “DNA fixed on a slide” (as in Gȍransson) and “a fixed tissue sample on a slide” (as in the patent claim). The first instance decision interprets the opinion of Nanostring’s expert as meaning the person skilled in the art would have no insurmountable objections to applying the teaching from Göransson to a cell or tissue sample (and would thus see a “very high expectation of success”). The Court of First Instance decided this opinion is based on a retrospective view (ex post facto analysis) with knowledge of the invention.
There is no reference to Nanostring’s expert opinion (or 10X Genomics’ expert opinion) in the Court of Appeal decision. Indeed, it is not clear to what extent, if at all, the Court of Appeal, considered the expert witness opinions.
As such, the Court of Appeal concluded that since “it is more likely than not that the patent at issue will prove to be invalid in proceedings on the merits due to a lack of inventive step, there is no sufficient basis for the issuance of a preliminary injunction.”
The reference to “more likely than not” in the conclusion stems from the Court of Appeal’s earlier statement that “a sufficient degree of certainty is lacking if the court considers it on the balance of probabilities to be more likely than not that the patent is not valid” (referring to the “sufficient degree of certainty” that the patent in question is valid to grant the provisional measure in R. 211.2 RoP, see section 5 a) of the decision, pp. 27 – 28). The written decision acknowledges that setting the standard of proof too high or too low would unfairly damage one or the other party in agreement with the first instance decision.
Other points
The Court of Appeal’s approach to claim construction followed Article 69 EPC and the Protocol on its interpretation. In essence, the claims must not be limited to their strict literal interpretation, but the description and drawings should be used as explanatory aids for the interpretation. However, the claims do not serve only as a guideline. Overall, the Court of Appeal’s construction of the claim appears to follow the EPO’s “broadest reasonable interpretation”, even if this approach was not explicitly mentioned.
With respect to novelty, the Court of Appeal agreed with the court of first instance that Gȍransson was not novelty destroying for claim 1, but for slightly different reasons. The Court of Appeal did not consider that the removal of the sandwich probes (the detection reagents of patent claim 1) in Gȍransson between each hybridsation with the tag probes (decoder probe of patent claim 1) to be a difference between claim 1 of the patent in suit because of the Court of Appeal’s claim construction.