The first substantive Australian decision in the global dispute over Amgen's PCSK9 antibody patents has now been handed down by the Australian Patent Office in Sanofi v Amgen Inc.  APO 67.1 Sanofi challenged the validity of five of Amgen's patent applications directed to PCSK9 antibodies and was unsuccessful on all grounds.
Since this is the first Australian decision in this dispute, this article will firstly summarise the Australian patent family and the relevant law under which the decision was made before discussing the decision itself.
The opposed applications
The opposed applications stem from international patent application no. PCT/US2008/074097. The initial Australian national phase application has granted (Patent No. 2008288791) and its term was extended under Australia's pharmaceutical patent term extension provisions based on the regulatory approval of REPATHA (evolocumab).
Amgen filed several divisional applications, five of which were accepted and subsequently opposed by Sanofi in 2016:
Application No. 2013203751
Application No. 2013203748
Application No. 2013203689
Application No. 2013203685
Application No. 2013203677
The opposed applications are titled "Antigen Binding Proteins to Proprotein Convertase Subtilisin Kexin Type 9 (PCSK9)" and cover Amgen's monoclonal antibody, evolocumab, which targets PCSK9 and lowers low-density lipoprotein (LDL) cholesterol levels. Evolocumab was first approved by the Australian Therapeutic Goods Administration in 2015 for the treatment of hypercholesterolaemia. If granted, the opposed applications may also be eligible for an extension of term based on the regulatory approval of evolocumab.
The applicable law
The opposed applications are all subject to the Patents Act 1990 as it existed prior to the introduction of the Intellectual Property Laws Amendment (Raising the Bar) Act 2012, which came into effect in 2013. The so called "Raising the Bar" amendments were introduced by the Australian Parliament with the express intention of aligning Australia's written description requirements with those of its major trading partners, particularly Europe and the US. Under the current Act, the requirements of support and sufficiency apply, meaning that, as in Europe, the scope of the claims must be commensurate with the technical contribution to the art, and the specification must enable a skilled person to perform the invention across the full scope of the claims without undue burden or further invention.
Under the "old" Act, however, the requirements of "full description" and "fair basis" apply. The standards set by full description and fair basis are much lower than those set by support and sufficiency, and challenging a patent on these grounds has been notoriously difficult (and more often than not, unsuccessful). To meet the full description requirement, the specification needs only to enable the skilled addressee to produce something within each claim without new inventions or additions or prolonged study of matters presenting initial difficulty. To meet the fair basis requirement, the specification must provide a "real and reasonably clear disclosure" of the claimed invention.
The old Act also applies a lower inventive step threshold compared to the current Act. However, although inventive step was initially relied upon by Sanofi as a ground of opposition, it was not pressed at the hearing.2
Oppositions conducted under the new Act allow the Commissioner of Patents to refuse a patent application if satisfied on the balance of probabilities that a ground has been made out.3 Under the old Act, however, the onus of proof lies with the opponent – in this case, Sanofi – who must establish that it is clear or practically certain that a valid patent cannot be granted.
The Delegate in Sanofi v Amgen broadly grouped the disputed claims into three classes, namely:
i) epitope claims, which define an isolated monoclonal antibody by its ability to bind an epitope of PCSK9, the epitope comprising nominated residues;
ii) residue claims, which define an isolated monoclonal antibody by its ability to bind one or more specific residues of PCSK9;
iii) competition claims, which define an isolated monoclonal antibody by its ability to compete for binding with a structurally-defined antibody.4
The claims also include functional language referring to the ability of the antibody to block or reduce binding of PCSK9 to the LDL receptor (LDLR). The Delegate then set out a detailed construction of certain terms in the claims that were critical to Sanofi's opposition.
Based on expert evidence and the disclosure of the specification, the Delegate construed the term "epitope" to include antigen residues that directly contribute to non-covalent interactions with the antibody, as well as residues on the surface of the antigen that are covered by the antibody.5 Therefore, the phrase "binds an epitope of hPCSK9 comprising one or more of amino acid residues 207, 208, 162, 164, 167, or 123 of SEQ ID NO: 1" in one of claims was construed to mean that the claimed antibody interacts with a region on the surface of human PCSK9 that includes one or more of the nominated PCSK9 residues. The nominated PCSK9 residue(s) do not necessarily form non-covalent interactions with the antibody and the epitope may include other unidentified residues.6
Blocks binding of PCSK9 to LDLR
Based on context provided by the specification, the reference to "blocking" in the claims was construed to mean an interference of the interaction between PCSK9 and LDLR.7 Blocking of the PCSK9 and LDLR interaction does not require 100% elimination of the interaction, but merely a reduction in the interaction between PCSK9 and LDLR.
The term "compete" was specifically defined in the specification, and the Delegate had no trouble construing it in the claims, observing that the phrase "competes for binding to PCSK9" refers to the ability of one antibody to contest the other when binding to PCSK9. The specification was found to clearly set out a range of competitive binding assays that can be used to determine whether two antigen binding proteins compete with each other as required by the claim.8
The Delegate accepted Sanofi's submission that competition between two antibodies can occur via direct and indirect means. Two antibodies may share the same or an overlapping epitope or may bind to different epitopes and interfere with each other's binding due to steric hindrance.9
Sanofi opposed the claims of the applications for lack of clarity. The issues raised in their submissions essentially related to the use of inexact language in the claims, for instance terms such as binds, blocks, reduces, neutralizing and competes. The Delegate rejected these submissions, finding that each term could be given meaning and that the claims provide a workable standard.10
Sanofi asserted that the claims of each application were not fairly based on the matter described in the specification.
The question of fair basis has been expressed by Australia's High Court as whether there is a real and reasonably clear disclosure in the body of the specification of what is claimed, so that the alleged invention is broadly, that is in a general sense, described in the body of the specification.11
Sanofi argued that only two antibodies disclosed in the applications were actually made, tested and shown to block the binding of PCSK9 to LDLR, and thereby lower plasma LDL levels. Sanofi asserted that it is those two antibodies for which the specification provides a real and reasonably clear disclosure.12
The Delegate rejected those arguments and pointed to statements in the specification which described the invention in broader terms. Although generic and not tied to specific examples, those paragraphs indicated to the Delegate that the invention extends beyond the specific antibodies isolated and characterised in the applications.13 The invention as described by the specification was considered to include antigen binding proteins that bind to the same or an overlapping region of PCSK9 as bound by the EGFa domain of LDLR or the two exemplified antibodies.14
With regard to the epitope and residue claims, Sanofi asserted that there is no proof that the antibodies of the invention bind to one or more of the identified residues. However, the Delegate emphasised that fair basis is a consideration of the disclosure provided in the specification, namely what the body of the specification read as a whole describes as the invention. It was not necessary for there to be scientific proof of non-covalent binding between the claimed antibodies and the identified amino acid residues.15
The Delegate accepted that the specification does not demonstrate that the amino acid residues identified as part of the interaction interface are directly involved in non-covalent interactions that effect binding between PCSK9 and LDLR or the two characterised antibodies.16 Instead, the specification described X-ray crystallography experiments identifying those residues on the antigen that are located closest to the antibody when the two molecules are bound. The Delegate considered it a reasonable extrapolation to infer that amino acid residues within the identified region are involved in the non-covalent interactions that effect binding between PCSK9 and antibody.17 Accordingly, the Delegate was satisfied that the specification provides a real and reasonably clear disclosure of the antibodies encompassed by the epitope and residue claims.
As for the competition claims, the Delegate again pointed to statements in the specification which described the invention in broad and general terms, and to paragraphs disclosing means for identifying competitively binding antibodies. Having construed the invention in these broad terms, the Delegate was satisfied that the specification also provides a real and reasonably clear disclosure of the antibodies encompassed by the competition claims. Consequently, this ground of opposition failed.
Full description (sufficiency)
The test for sufficiency of the description has been articulated by Australia's High Court as whether the disclosure of the specification will enable the addressee to
produce something within each claim without new inventions or additions or prolonged study of matters presenting initial difficulty.18
Sanofi submitted that the opposed applications do not disclose an antibody that falls within the scope of any of the claims, and that a skilled person could not reproduce the antibodies disclosed in the applications. With regard to the epitope claims, Sanofi argued that the applications do not disclose a single antibody that binds an epitope comprising the residues recited in the claims.
The Delegate re-framed the test for sufficiency by asking whether, based on the disclosure provided, the addressee will be able to produce an antibody that binds:
a) an epitope that comprises stated amino acid residues; or
b) to the specific amino acid residues specified;
or whether to do so would require new inventions or additions or prolonged study of matters presenting initial difficulty.
With regard to the epitope claims, the Delegate was satisfied that the specification described the binding site or interaction interface between PCSK9 and LDLR, and that the specification showed how two antibodies interact with this region to block binding between PCSK9 and LDLR. The Delegate again noted that the epitope, as construed earlier, will include specific amino acids that directly contact the antibody and also amino acids that are covered by the antibody. In that context, the Delegate considered that the residues of PCSK9 that the specification demonstrates with crystallographic experiments to be within the region covered by the antibody, can be considered the epitope, and therefore the epitope claims are fully described.19
With regard to the residue claims, the experts for both parties agreed that the term "binds to" means that the claimed antibody forms a non-covalent interaction with at least one of the nominated residues of PCSK9. But the parties' experts presented opposing views as to whether the residues specified in the claims are in fact directly involved in binding. Sanofi did not present any evidence that the two exemplified antibodies would not bind at least one of the residues set out in the claims.20 Amgen's expert, however, performed an analysis using crystal data in the specification and concluded that specific residues identified in the claims very likely form non-covalent interactions with the exemplified antibodies or with LDLR.21
The Delegate accepted Amgen's submission that the specification discloses the interaction interface or epitope, and that the residues identified form non-covalent interactions between PCSK9 and the antibodies.22 The question then became whether the addressee could take this information, along with the other information provided in the specification, to generate antibodies that fall within the scope of the claims.
Sanofi submitted that, to make antibodies that will bind to the relevant residues (or epitope comprising the relevant residues), the skilled person would have to undertake one of two research projects. The first was said to require making a biosimilar of the antibodies disclosed in the application using transgenic techniques. The second approach would be to seek to obtain antibodies either by hyperimmunization of transgenic mice or by other means such as phage display and then carry out experiments to characterise the antibodies.
Amgen, on the other hand, submitted that the state of antibody arts was advanced and mature at the priority date and that armed with the teachings of the application it would be routine for the addressee to make antibodies of the claimed invention. The Delegate favoured the proposition presented by Amgen, and in particular, that antibodies within the scope of the claims could be produced using well understood mammalian expression vector methodologies such as cloning the CDRs of the exemplified antibodies into the framework region of a known antibody, and that this approach would not require new inventions or additions or prolonged study of matters presenting difficulty.23 The Delegate also noted that the second "research project" described by Sanofi is analogous to the situation described in Warner-Lambert Company LLC v Apotex Pty Limited (No 2)  FCAFC 26, where the Full Court of Australia observed that:
it is beside the point that, in undertaking the task, the person skilled in the art might need to apply considerable skill, effort and resources, or that the work might be complex, time-consuming and expensive. Work of this kind would not be of the character captured by the second limb of the test in Kimberly-Clark (i.e., work requiring prolonged study of matters presenting initial difficulty). Once the invention has been described, it does not lack sufficiency (in terms of description) because the person skilled in the art needs to take further steps which are known to be required and which are, for that person, routine in character. Put another way, the need for such steps does not mean that the description has not enabled the person skilled in the art to work the invention. In terms of s 40(2)(a), the invention has been "fully described".24
The Delegate was not satisfied that the work required to produce one antibody embodying each claim using the information provided in the specification requires anything more than what is routine in the art, even if such work may be complex, time consuming and expensive.25 Similarly, in relation to the competition claims, the Delegate found that the skilled addressee could use well-established techniques to produce a library of antibodies that is then screen using standard techniques to assess for competition with the reference antibodies.26 Consequently, this ground of opposition also failed.
The Patents Act 1990 also requires the complete specification to describe "the best method known to the applicant of performing the invention."27
Sanofi alleged that the specification failed to disclose the residues of PCSK9 with which the antibody will form non-covalent interactions or that form the epitope of the claimed antibodies. They submitted that by withholding this information, Amgen concealed the best method by which to achieve the result which constitutes the invention.
The Delegate rejected this submission, finding that the specification discloses exemplary antibodies that represent the invention and provides information that would allow the skilled addressee to produce antibodies with the same CDRs, and therefore binding properties, as these antibodies. The Delegate also noted that there was no evidence that Amgen knew of a better method than what is disclosed in the specification.28
Sanofi also submitted that the invention as claimed failed to achieve what was "promised" by the specification and therefore lacked utility.29 Sanofi argued that none of the claims are limited to isolated monoclonal antibodies that: a) lower, maintain or prevent an increase in plasma cholesterol of the subject to which they are administered or are useful as a diagnostic tool (The Promise); or b) have a biological effect of achieving The Promise. They asserted that only two exemplified antibodies have the picomolar affinity for PCSK9 and the resultant ability to block binding of LDLR to PCSK9 to the extent that the antibody is capable of lowering plasma LDL levels.
The Delegate did not agree with Sanofi's characterisation of the "promise" as being limited only to those antigen binding proteins that will be capable of lowering plasma LDL levels. Rather, the specification was found to more broadly disclose antigen binding proteins that bind to particular regions of PCSK9 to prevent binding to LDLR. The claims, by virtue of the functional characteristics defined in each of the epitope, residue or competition claims, were considered to necessarily encompass those monoclonal antibodies that achieve the more broadly stated promise of the invention.30 As such, Sanofi's opposition on this ground was unsuccessful.
The global dispute over Amgen's PCSK9 antibody patents has highlighted the different approach taken by patent offices around the world in assessing the validity of functionally defined antibodies. The decision of the Australian Patent Office described in this article was made under the Patents Act 1990 prior to the introduction of amendments that raised the written description requirements, and so it is not directly applicable to patent applications that are subject to the current Patents Act 1990. But the decision still provides useful guidance on issues such as best method, claim construction, the structural determinants of binding specificity and the nature of work that is routine in the art, including work that is complex, time consuming and expensive. The guidance provided on these and other issues will be relevant to applications that are subject to the current Patents Act 1990.
1 'Sanofi v Amgen'.
2 Sanofi v Amgen at .
3 Section 60(3A).
4 Sanofi v Amgen at .
5 Sanofi v Amgen at .
6 Sanofi v Amgen at .
7 Sanofi v Amgen at .
8 Sanofi v Amgen at .
9 Sanofi v Amgen at .
10 Sanofi v Amgen at .
11 Lockwood Security Products Pty Ltd v Doric Products Pty Ltd  58 at .
12 Sanofi v Amgen at .
13 Sanofi v Amgen at  and .
14 Sanofi v Amgen at .
15 Sanofi v Amgen at .
16 Sanofi v Amgen at .
17 Sanofi v Amgen at .
18 Kimberly-Clark Australia Pty Ltd v Arico Trading International Pty Limited  HCA 8 at .
19 Sanofi v Amgen at .
20 Sanofi v Amgen at .
21 Sanofi v Amgen at -.
22 Sanofi v Amgen at .
23 Sanofi v Amgen at .
24 Warner-Lambert Company LLC v Apotex Pty Limited (No 2)  FCAFC 26 at .
25 Sanofi v Amgen at .
26 Sanofi v Amgen at .
27 Section 40(2)(a).
28 Sanofi v Amgen at .
29 Patents Act 1990, s 18(1).
30 Sanofi v Amgen at .
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.