Tuesday, November 17, 2015

Can a Compound Claim Be Supported by an Inventive Process?

Amgen Canada Inc v Apotex Inc (NOC) 2015 FC 1261 Hughes J
            1,341,537 / filgrastim / NEUPOGEN

In Amgen / filgrastim Hughes J held Claim 43 of the Amgen’s ‘537 patent to be invalid for obviousness, though it survived novelty and utility attacks. This post focuses on Hughes J’s obviousness analysis, which seems to me difficult to reconcile with Sanofi 2008 SCC 61.

Human granulocyte colony-stimulating factor (G-CSF) stimulates the growth of white blood cells. Naturally occurring G-CSF was first isolated and purified in small quantities by Welte and others at Sloan-Kettering Institute [16]. However, Welte had not determined any of the amino acid sequence [18]. If the DNA sequence of the naturally G-CSF could be determined with sufficient precision, G-CSF could be produced in commercial quantifies by recombinant DNA technology. Amgen was ultimately successful in doing so, leading to the invention of filgrastim, sold by Amgen as NEUPOGEN, which is used to boost the production of white blood cells and therefore help prevent infection in people whose immune system is compromised as a result of chemotherapy.

Claim 43, the only claim at issue, claims a 175 amino acid polypeptide which is functionally equivalent to naturally occurring G-CSF, in that it is effective to stimulate white blood cell production. While the claimed amino acid sequence is very close to that of naturally occurring G-CSF, it differs in that it begins with Met (Methionine), which the naturally occurring sequences does not [77]. It also potentially differs in the details of the subsequent amino acid, as the process used by Amgen to sequence the naturally occurring factor is not guaranteed to replicate it precisely, and the exact naturally occurring sequence is still not known. While the claimed sequence is certainly very close to that of the naturally occurring sequence – the first 40 amino acids are the same and the rest are so close as to not make any functional difference [26] – it is “quite possibly but not certainly the sequence of the natural product” [80]. Consequently, Hughes J held that the claimed sequence was not anticipated by the naturally occurring product that had been isolated by Welte et al. (I must say that it is not clear to me what evidential burden was applied by Hughes J. As this was an NOC proceeding and Apotex had clearly put the issue in play, Amgen bore the burden of proving the allegation is not justified [41] by showing the claimed sequence was probably not the same as the naturally-occurring sequence. The fact that it is not “certainly” the same is beside the point. Perhaps showing that it is “quite possibly” the same is equivalent to showing it is probably not the same? In any event, it appears that Hughes J would have held it to be novel on the basis of the leading Met alone.)

Having concluded that the claimed sequence was novel, the next question was whether it was obvious. There was clearly no invention in conceiving of the invention as Welte had identified the desirability of isolation and cloning of the gene in order to allow large-scale production of G-CSF [95], [105]. The only question was whether Amgen’s sequencing of G-CSF was itself inventive.

I have two concerns regarding Hughes J’s analysis. Consider the following two paragraphs:

[93] Techniques for taking the various steps undertaken by Amgen were known. There were a variety of choices to be made at each step and each step had to be carefully undertaken. A wrong choice or improperly conducted step could lead to failure. However, Amgen did not utilize any hitherto unknown step or technique.

[101] There was a high degree of skill required and risk involved in what Amgen undertook. The steps were routine in the sense that they were carried out by skilled persons operating with the science as it was known at the time. This amounts to . . .“skilled work” . . . and not to the “creative work” necessary to deserve patent protection.

It seems to me that Hughes J’s conclusion that the invention was obvious even though there was “a high degree of skill required and risk involved” is difficult to reconcile with the standard statements that an invention is not obvious unless it would be “more or less self-evident” or “very plain” that what was being tried ought to work: Sanofi 2008 SCC 61 [65], [92]. As Barnes J said in the Sanofi FC decision which was ultimately affirmed by the SCC “Having to try different methods, though they be well-known, in order to discover which one will yield the desired result cannot mean that the desired result. . . was obvious” 2005 FC 390 [80]. Moreover, while the steps may have been routine, the proper question goes not to the individual steps, but to the course of inquiry as a whole. The SCC noted that “prolonged or arduous trial and error would not be considered routine [37] (and similarly [69]), even if the individual steps may be. The reason is that “A patient searcher is as much entitled to the benefits of a monopoly as someone who hits upon an invention by some lucky chance or an inspiration”: Halocarbon [1979] 2 SCR 929, 944 (and see also 2005 FC 390 [84]). As Lord Diplock has explained in Dann's Patent, [1971] RPC 425, 451:

The task of finding such a strain of micro-organism calls for the exercise of technical proficiency and is laborious and very costly, for the odds against success are large. It is not easy to see what inventive step, as distinct from the mere exercise of proficiency and practice, is involved in this kind of research, but the result of success in it is a new product useful to humanity which does not exist in nature. If such research is to be encouraged in a competitive society, the monetary rewards of success must be assured to those who undertake the expense; and the means of doing so in this and in most other countries with comparable social systems is by according to the successful discoverer of the new product the controlled and limited monopoly granted for inventions under the national patent laws.

I have a second distinct concern with Hughes J’s analysis, which is evident in the following statement:

[106] In this case, Welte found the protein and said to the readers of his paper to go out and make it in quantity. Amgen did that. Perhaps Amgen did so using an inventive process, and perhaps it is entitled to a patent claiming such a process or processes. I note that there are several claims in the ‘537 patent directed to processes. The end product, which is simply the protein made by whatever process, was not itself inventive. Welte clearly and unambiguously pointed to that protein, leaving to others to devise new processes, or use old processes, to get that product in quantity.

Here, Hughes J seems to be saying that even if the process used by Amgen was inventive and the claimed invention itself was new, Amgen is not entitled to a patent on the compound itself as a matter of law, but only on the inventive process. This is in contrast to the law set out in Lundbeck [2009] UKHL 12 aff’g [2008] EWCA Civ 311. Lord Hoffmann’s explanation in the EWCA [26], [27] is succinct:

The [trial] judge held that claim 1 and claim 3 (which is dependent on claim 1) were insufficient. His reasoning was that claim 1, being a claim to the (+) enantiomer as a product, was a claim to a monopoly of that product however made: see section 60(1)(a) of the 1977 Act. But Lundbeck's inventive idea was not to discover that the enantiomer existed and had a medicinal effect. Everyone knew that the two enantiomers existed and that one or other or both had a medicinal effect. What Lundbeck discovered was one way of making it. But that did not entitle them to a monopoly of every way of making it.

I can understand and sympathise with the judge's instinctive reaction to the inherent breadth of a product claim. Indeed, as I shall in due course show, he is not the first to have registered such a protest. But in my opinion his reasoning is not justified either by the statute or the authorities. In an ordinary product claim, the product is the invention. It is sufficiently enabled if the specification and common general knowledge enables the skilled person to make it. One method is enough.

This was the central issue on appeal, and the UKHL affirmed. I am not sure that precisely this point has been addressed in Canadian law, but the facts in Sanofi were very similar. The claims at issue were to the dextro-rotatory isomer of clopidogrel 2005 FC 390 [25]. The racemic compound was known in the prior art [28], the method of resolving the racemate into its enantiomers was inventive [84], and the claim to the compound per se was held to be valid. See also Hughes J’s own decision in Janssen-Ortho / levofloxacin 2006 FC 1234 [115] aff’d 2007 FCA 217, with very similar facts to Sanofi. Both Sanofi and Levoflaxin might be distinguished on the basis that in those cases it was not known whether the claimed compound would be beneficial before it was actually synthesized (that is, it was not known whether either enantiomer would have properties superior to the racemate), whereas in this case it was known that purified C-GSF was desirable. So it might be said that in both Sanofi and Levoflaxin the invention lay in deciding to make the attempt; and indeed, in Sanofi Barnes J did so hold [81]. But Barnes J also held that the inventive step in the separation itself was enough to render the claimed compounds non-obvious, even though the claims were not restricted to the process [80]. Thus, there is at the very least a significant tension between Hughes J’s and the result in Sanofi.

Given the controversy identified by Hoffmann J, there is no doubt a real debate to be had as to whether the Lundbeck rule is sound as a matter of policy. I am inclined to think that the UK courts were right, largely for a practical reason illustrated by this case. If we accept that prolonged and arduous trial and error would not be considered routine, even if the individual steps are not themselves inventive, then it might be possible that synthesis of a new compound which is known to be desirable would be non-obvious even though there is no specific patentable process at all. In such a case, a process or product-by-process claim will not be available, and if a compound claim is not allowed either, then an inventor will have no way to claim the fruits of a non-obvious invention. That cannot be right: prolonged and arduous trials by a patient searcher must be rewarded to provide an adequate incentive for valuable discoveries that would not be undertaken but for the lure of a patent.


  1. You note that the exact naturally occurring sequence is still not known, which fits with the treatment this fact gets in the decision. This should be startling, and one would hope that this state of affairs might lead to some serious reservations about whether the claimed sequence was obvious. If we still don’t know what the sequence is of the active naturally occurring protein, how is it that a particular claimed sequence of an active protein can be self evident. The conclusion on obviousness in this case flows from an analytical framework which holds that something is obvious if the path to get to it is obvious. You make very good points about how the path to elucidating the native sequence may not have been all that clear in this case. Perhaps more to the point though, the established fact of the matter here is that we still don’t know where that path leads – we apparently don’t know the actual natural active protein sequence some 30 years after the patent application was filed. In my view, there could hardly be a better illustration of the perils of hindsight in an obviousness analysis. The path to your destination will always be clear once you get there, but to quote Yogi Berra “You've got to be very careful if you don't know where you are going, because you might not get there.”

  2. It seems that Canadian law (via Commissioner's decisions 1359, 1367) has indeed dealt with the issue of the patentability of an obviously desirable product whereby merely a new process has been developed to produce that product.