After intense discussion between regulators and the biotechnology industry, consensus was reached on the need for data to establish the clinical equivalence of therapeutic proteins
The patients for many of the first recombinant DNA derived therapeutic proteins such as erythropoietin (EPO), insulin, growth hormones, and cytokines have expired which opens the possibility for marketing non-innovator versions of these products. For small molecule drugs the requirement to obtain a marketing authorisation for generic versions is modest. The manufacturer of a generic only need show the product to contain an identical active substance and to have comparable pharmacokinetics to the originator product. Formal clinical efficacy and safety studies are not necessary.
However this generic model cannot be applied to therapeutic proteins. Their active substance is often a mixture of isoforms and/or modified proteins and rarely a single molecular entity. Their molecular size and complexity is 100- to 1000- fold greater than small molecule drugs and current analytical tools do not allow for their full characterisation.
After intense discussion-between regulators, the biotechnology industry and the companies planning to introduce non-innovator versions of protein drugs- a consensus was reached on the need for data to establish the clinical equivalence of therapeutic proteins.
To differentiate them from the generic regulatory pathway, the term ‘biosimilar’ was introduced in the European Union (EU). In 2004 the EU was the first to adopt legislation to establish a comprehensive regulatory pathway for bringing biosimilars to market. The European Medicines Agency (EMEA) and its scientific Committee for Medicinal Products for Human Use (CHMP) developed guidance documents to provide more detail on the requirements. In the US the Food and Drug Administration (FDA) is expected to establish similar pathways within about 12 months.
The regulatory guidance published by the EMEA/CHMP concerning biosimilars has three levels. There is an overarching guideline, introducing the concept of similar biological products and outlining the basic principles. Three general guidelines were issued. The Guideline on similar biological medicinal products containing biotechnology-derived proteins as active substance: quality issues, lays down the quality requirements fora biological medicinal product claiming to be similar to another one already marketed. The Guideline on similar biological medicinal products containing biotechnology- derived proteins as active substance: non-clinical and clinical issues, describes the animal and clinical studies required for a biological medicinal product claiming to be a biosimilar. For most biosimilars comparative clinical trials are considered to be necessary to demonstrate clinical comparability. The third general guideline Immunogenicity assessment of biotechnology- derived therapeutic proteins, concerns biologics in general but has major implications for the marketing authorisation of biosimilars.
Four guidelines, specific to product class, were issued for the development of biosimilars containing recombinant erythropoietin (EPO), somatotropin, human insulin, and human granulocyte colony-stimulating factor (G-CSF). Two additional product-class specific guidelines concerning low molecular weight heparins and interferon alpha are expected to be released in 2009.
As of September 2009, the EU has allowed the market introduction of six different biosimilars, four of which are marketed by more than one company using different brand names. Two marketing authorisation requests for biosimilar somatropins, Omnitrope and Valtropin, were approved in the EU. Omnitrope is a biosimilar version of Genotropin. Valtropin is produced in Saccharomyces cerevisiae, a yeast, and is a biosimilar version of Humatrope synthesised in Escherichia coli.
Five biosimilar epoetins produced by two companies have been approved in the EU. Abseamed, Binocrit, and Epoetin alfa Hexal produced by Rentschler Biotechnologie GmbH are biosimilar versions of the reference product Eprex. Two additional biosimilar epoetins, Retacrit and Silap, are manufactured by Norbitec GmbH, again with Eprex as the reference product.
Two biosimilar filgrastims (human G-CSF) with Neupogen (Amgen) as reference product have been approved in the EU. The first is marketed by three different companies under four different brand names: Biograstim, (CT Arzneimittel), Filgrastim ratiopharm and Ratiograstim (Ratiopharm) and TevaGrastim (Teva Generics). The latest biosimilar to be approved by the European Commission is filgrastim produced by Hexal Biotech Forschungs GmbH and marketed by two different companies under two different brand names, Filgrastim Hexal and Zarzio.
UnAlpheon, a biosimilar version of Roferon-A (interferon alfa-2a), was rejected by the EMEA. The reasons the EMEA gave included quality and clinical differences between Alpheon and the reference product, inadequate data on the stability of the active substance, inadequate validation of the process for the finished process, and insufficient validation of immunogenicity testing.
Another biosimilar application for three different human insulin formulations with Humulin as reference product was withdrawn by the company Marvel. Based on the review of the data, the CHMP had some major concerns and was of the provisional opinion that Insulin Human Rapid Marvel, Insulin Human Long Marvel and Insulin Human 30/70 Mix Marvel could not have been approved for the treatment of diabetes mellitus. The main concerns of the CHMP were that the comparability of the Marvel insulins and the Humulin insulins had not been shown.
Although there is regulatory requirement for clinical data for the authorisation of biosimilars, the studies in patients are too limited to reveal the full long-term safety profile of these products. All manufacturers of biosimilars have to submit risk assessment and safety plans which include safety monitoring, including immunogenicity in patients under chronic treatment. Post approval commitments (PACs) require the setup of Phase IV and observational trials after marketing authorisation to collect data on daily clinical practice. To enable adverse effects to be linked to specific products automatic substitution with biosimilars is banned in some countries. In some other countries either the health authorities or the professional organisation advise only to change the treatment with biologics with the consent of the treating physician or pharmacist, as recommended by the EMEA.
Interpreting safety data
The lack of sufficient background data and standardisation of diagnostic assays hampers the interpretation of the safety data of biosimilars. A good example of the problem is the recent halt of a clinical trial of Binocrit in chronic renal failure using the subcutaneous route of administration, since two patients out of 337 developed neutralising anti-EPO antibodies which, in one case, was associated with pure red cell aplasia (PRCA). The patient who developed PRCA, confirmed by bone marrow analysis, was an 81 year old male in Germany. He developed antibodies about six weeks after start of treatment.
A few months later a low level of neutralising antibodies was detected in a Russian male patient, who was 79 years of age. The sample was taken four months after the start of treatment, after the routine sample taken one month earlier was found to be positive for binding antibodies. There was no sign of reduced response to epoetin and no drop in haemoglobin in this patient. The patient died a few days later. Because of this observation of neutralising antibodies and the PRCA case in Germany the trial was interrupted, as requested by the Data Safety Monitoring Board and the principal investigator of the study, to exclude a specific immunogenicity problem with epoetin product when
Although this patient could be considered to have been at risk to develop PRCA, its likelihood is difficult to ascertain. The natural history of PRCA associated with epoetin treatment is unknown, because all reported cases were identified by their presentation of clinical symptoms or decreased response to treatment. Nothing is known about how the antibody response develops before the symptoms occur and what the predictive value is of a single observation of neutralising antibodies is for the development of PRCA.
The sera of the Russian patient were also tested by different laboratories using different assay formats and reporting different levels of neutralising anti-EPO antibodies. Because of the lack of standardisation any prediction based on historical data is impossible.
According to the European Public Assessment Reports published by the EMEA, Binocrit contains less aggregation compared to Eprex and we have confirmed this finding in our laboratory. Aggregates are the most important risk factor for immunogenicity. That makes a structural problem with the product itself unlikely. This suggests some mishap with the packaging or storage, in my opinion, to be the most likely cause of the two cases of immunogenicity.
Need for standardisation
The use of biosimilars is increasing and the lessons learned from Binocrit demonstrate the need for standardisation of assays and more prospective safety studies.