26 March 2017

Enforcing Patents on Processes for Producing Biologics – Easier Said Than Done?

Seeking answersA biological medicine, or biologic, is a medicine that contains one or more active substances made by a biological process or derived from a biological source.  Compared to conventional pharmaceuticals, which are made by chemical synthesis using different organic and/or inorganic compounds, biologics are generally much larger molecules that are derived from the living cells of micro-organisms or animals by utilizing the metabolic processes of the organisms themselves.  In many cases, cells are genetically engineered, using recombinant DNA technology, to co-opt their expression capabilities and turn them into tiny ‘factories’ for producing the desired biologic molecules.  The first such substance to be approved for therapeutic use was synthetic ‘human’ insulin, developed by Genentech and first marketed by Ely Lilly in 1982.  Since then, biologics have become increasingly important, and hundreds of biologic medicines are now in use, including therapeutic proteins, DNA vaccines, monoclonal antibodies, and fusion proteins.

The chemical structures of traditional ‘small molecule’ pharmaceuticals are usually well-defined.  Biologics, on the other hand, are often difficult, and sometimes impossible, to characterise structurally due to their size and complexity.  In many cases, some components of a finished biologic may be unknown.  Protecting biologics with patents can therefore be challenging, given the potential difficulty in claiming the active biologic molecule in terms that are sufficiently precise to meet legal requirements for the invention to be clearly defined, while also being broad enough to encompass variations that may have the same therapeutic properties, and satisfying the tests for novelty, inventive step and subject-matter eligibility.

The complexity of biologics also means that they are more difficult to produce than small molecule pharmaceuticals.  Minor differences in the production process, raw materials, temperature, pH, or cell line can result in a significant alteration in a medicine’s quality, efficacy, or safety.  While this creates challenges in manufacturing and quality control, it also creates opportunities for the protection of intellectual property, by securing patents on methods of producing biologic medicines either in place of, or in addition to, trying to claim the active molecules themselves in some form.

Process claims are all very well if the patentee can established that a suspected infringer is using the patented method of production.  But what happens when the suspect company keeps its manufacturing processes secret, as indeed is most commonly the case?  Well, then the patentee may need to convince a judge to issue an order for preliminary discovery, compelling the accused infringer to disclose relevant information about its processes.

This is exactly what Pfizer recently sought to do in preparation for possible patent litigation against Samsung Bioepis AU (‘SBA’) in relation to Pfizer’s product ENBREL (active ingredient etanercept), which is a biological medicine used in the treatment of autoimmune diseases such as rheumatoid arthritis, juvenile rheumatoid arthritis and psoriatic arthritis.  Unfortunately for Pfizer, Justice Stephen Burley has denied its application for discovery, ruling its evidence insufficient to establish a ‘reasonable belief’, as opposed to a ‘mere suspicion’, that SBA is infringing its patents relating to a process for producing etanercept: Pfizer Ireland Pharmaceuticals v Samsung Bioepis AU Pty Ltd [2017] FCA 285.

Background and Biosimilars

Pfizer’s ENBREL product has been commercially available in Australia since 2003.  Throughout that time, Pfizer Australia Pty Ltd has been the sponsor on the Australian Register of Therapeutic Goods (ARTG)  of etanercept for the products marketed under the name ENBREL.  Pfizer has a number of Australian patents covering processes for producing etanercept.

In July 2016 two pharmaceutical products containing etanercept as their active ingredient were registered on the ARTG under the name BRENZYS, with the sponsorship of SBA. 

BRENZYS is a biosimilar of ENBREL, which is essentially the biologic equivalent of a generic pharmaceutical.  However, whereas generic small molecule compounds can generally be shown to be identical in structure to the original product, biosimilars are (as the name suggests, and for the reasons outlined above) typically only similar to the original in a range of relevant respects.

A biosimilar can be registered on the ARTG based on evidence of its similarity to an existing registered biologic.  While this requires comprehensive comparability studies, it nonetheless involves less effort and expense than is required to establish safety and efficacy to register a new biologic.  The Australian Therapeutic Goods Administration (TGA) has prepared guidelines which require that a biosimilar and its reference medicine have similar physicochemical, biological, immunological, efficacy and safety characteristics.  The TGA has adopted European quality guidelines which acknowledge that:

The manufacturer developing a biosimilar product would normally not have access to all information that could allow an exhaustive comparison with the reference medicinal product, particularly with regards to the manufacturing process.  Nevertheless, the analytical data submitted should be such that firm conclusions on the physicochemical and biological similarity between the reference medicinal product and the biosimilar can be made.

In applying for registration on the ARTG of its BRENZYS biosimilar, SBA prepared and submitted various documentation, including documents relating to certain processes used to manufacture the BRENZYS product.  It is these documents that Pfizer was seeking to inspect via preliminary discovery.  They are, unsurprisingly, confidential.

‘Reasonable Belief’

For a court, compelling a party to divulge its secrets, even under suitable undertakings as to confidentiality, for the purposes of potential litigation that may not eventuate, is not an insignificant step.  It is not enough that the applicant for discovery merely suspects that the information will confirm its right to sue.  The Federal Court Rules (specifically, Rule 7.23) require that the applicant ‘reasonably believes’ that it has a right to relief.

While ‘reasonable belief’ is not necessarily a high threshold, it is one that must be assessed objectively, and ‘there must be some tangible support that takes the existence of the alleged right beyond mere “belief” or “assertion” by the applicant’: Reeve v Aqualast Pty Ltd [2012] FCA 679 at [65]. 


Pfizer led evidence in the form of affidavits from Dr Neysi Ibarra, who is the Director and Group Leader of Process Development, Manufacturing Science and Technology of Pfizer Ireland Pharmaceuticals.  This evidence was to the effect that Pfizer currently produces etanercept using a process that falls within the scope of its patents, that this process does not require the use of foetal bovine serum, and that this is the only such serum-free process known by Dr Ibarra to be in use for making etanercept.  This evidence was said to be significant because SBA’s product information for BRENZYS states that its biosimilar etanercept is manufactured using a serum-free process.

In short, Pfizer was asking the court to infer that it had very good reason to believe that SBA’s process may be infringing its patents because that process is serum-free, and the only viable serum-free process known (and used) by Pfizer is one that is covered by the patents.  Pfizer also argued that BRENZYS is a biosimilar of ENBREL, and it is well-known that the characteristics of biologics are very sensitive to the process by which they are made, further supporting an inference that BRENZYS is made by the same process as ENBRYL.  Furthermore, certain measurable characteristics of the two products (‘glycosylation profiles’) are very similar.  Finally, Pfizer noted that SBA’s Korean parent company had both motive and opportunity to take advantage of the information in Pfizer’s published patent applications to copy its process.

SBA led its own evidence in the form of affidavits from three Australian experts in bioengineering and biologic medicines, as well as an affidavit from the head of Cell Engineering at SBA’s Korean parent company.  This evidence challenged Pfizer’s argument that the similarities in characteristics between ENBREL and BRENZYS are a sufficient basis to infer that they are manufactured by the same process.  Additionally, SBA argued that Dr Ibarra’s evidence was insufficient to establish that Pfizer’s own process falls within the scope of its patents.


In order for the court to find that Pfizer held a ‘reasonable belief’ that one or more of its patents was being infringed, it would need to be satisfied of the likelihood that a number of related propositions were simultaneously true:
  1. that the Pfizer patents were consulted in the course of development of BRENZYS;
  2. that Dr Ibarra’s lack of awareness of any serum-free process being used for making etanercept, other than Pfizer’s own process, is evidence that SBA’s process is likely to be the same;
  3. that the similarities in characteristics of ENBREL and BRENZYS is evidence that they are made by the same process; and
  4. that Pfizer’s process of manufacturing ENBREL in fact falls within the scope of its patents.
The court found it plausible that the Pfizer patents were identified in searches during the development of BRENZYS, but this falls short of establishing that any manufacturing process described in the patents was copied.  The court further found that Dr Ibarra’s absence of knowledge of other serum-free process being in-use for making etanercept was not persuasive evidence that any such processes are known or used, particularly given that methods used by different manufacturers are usually confidential.

On the matter of similarities, Pfizer’s own evidence worked against it as much as in its favour.  By its own admission, it had changed processes in 2007, from a serum-based method to a serum-free process.  There was evidence that various characteristics of the ENBREL product were measurably different before and after this change.  And yet, despite these differences, the ‘after’ product was relevantly biosimilar to the ‘before’ product such that the TGA was satisfied that it could continue to be marketed under SBA’s existing ARTG registration.  The court thus found that Dr Ibarra’s evidence did not ‘rise above speculation ... that the similarities observed might mean that the BRENZYS Process is similar to the Pfizer Process’ and that these ‘coincidences are cogently explained by a far more available inference; that the end products are biosimilar’ (at [97]).

Finally, the court was not satisfied that Dr Ibarra’s evidence was sufficient to establish that Pfizer’s own process falls within the scope of its patents.  The difficulty here seems to have been that Pfizer was reluctant to disclose details of its manufacturing process despite the fact that it wanted the court to compel SBA to do exactly the same thing.  As a result, Dr Ibarra’s evidence amounted to an opinion that the Pfizer process is covered by the patent claims, but without stating important facts – namely the actual features of the process – upon which that opinion was based.  This is a no-no for expert evidence under the Federal Court rules of evidence!  While the court was prepared to waive strict application of the rules in the circumstances of an application for preliminary discovery, it found that the utility of the evidence ‘in assessing the strength of the inference that Pfizer asks to have drawn in its favour is considerably weakened by the absence of detail from which inferences may safely be drawn’ (at [128]).

The court thus denied Pfizer’s application for preliminary discovery, stating (at [155]):

Ultimately, I am not satisfied that, taken as a whole, the evidence of Dr Ibarra provides a reasonable basis for Pfizer’s “belief” in patent infringement by SBA, as opposed to mere suspicion, notwithstanding that the evidence may in an application of this type leave something to surmise or conjecture...

Conclusion – ‘Reasonable Belief’ and Biosimilars

Determining whether or not claims directed to a process are being infringed can be difficult if the process in question is not publicly observable.  Manufacturing processes are a classic case-in-point, however similar challenges can arise, for example, in relation to computer-implemented or electronic processes that take place on secured servers, or in software and/or hardware devices with technological or physical protection measures employed to prevent reverse-engineering.

In all such cases, it is necessary initially to try to determine the nature of the process being used on the basis of externally-observable characteristics or ‘signatures’.  In the case of electronic or software processes this might be a matter of comparing outputs of the suspected infringement with outputs of the patented process for a variety of different inputs.  In appropriate circumstances it may be possible for the initial suspicion to be confirmed to a state of near-certainty.

As this case demonstrates, however, there may be particular problems with establishing a ‘reasonable belief’ that infringement is occurring in the case of patented processes for making biologics.  The very characteristics that make biosimilarity difficult to quantify – i.e. that similar physicochemical, biological, immunological, efficacy and safety characteristics may be observed despite variations in the structure of the complex biologic molecules – also made it difficult in this case, at least, for Pfizer to convince the court that biosimilarity is a basis for anything more than a ‘mere suspicion’ of infringement.

Considering the frequency with which patents on biologics rely on process claims to protect products whose structures may be known only partially, barriers to obtaining preliminary discovery in order to confirm (or refute) suspicions of infringement may be particularly detrimental to enforcement in this field of technology.  Given the importance of the issue to Pfizer, I will not be at all surprised if this decision is appealed.


Post a Comment

Copyright © 2014
Creative Commons License
The Patentology Blog by Dr Mark A Summerfield is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Australia License.