Håkan Mellstedt, MD, PhD
Department of Oncology-Pathology
Karolinska Institute, Stockholm, Sweden
The cost of drugs is an essential part of total healthcare expenditures, especially in oncology. During the last two decades there has been an exceptional increase of new oncology drugs on the market: of both chemical drugs and biotherapeutics. There is no doubt that this development has led to improved cancer healthcare, with a better quality
of life and increased overall prognosis for cancer patients.
When a drug patent expires, follow-on drugs can be introduced to the market, and provided safety and efficacy are not a major concern, this can greatly reduce healthcare costs. Indeed, this has been a great success for chemical generics. ‘Biosimilars’ is the collective name for protein drugs, approved after patent expiration of the originator biotherapeutics. Protein drugs are important in the therapeutic arsenal for cancer, both for supportive care (for example, erythropoietin, granulocyte-colony stimulating factor (G-CSF)) and for targeted therapy (for example, monoclonal antibodies).
Protein drugs are complex molecules compared to chemical drugs; production is therefore complicated and the production platform may greatly influence the final product. Even if the primary and secondary structures of biosimilars are similar or comparable to the originator drug, tertiary and quaternary structures (such as carbohydrate moieties and other post-translational modifications) are likely to differ, which might have clinical safety and efficacy consequences.
The first biosimilar to be introduced into the field of oncology was erythropoietin, which has so far been approved for chemotherapy-induced anaemia. The efficacy of biosimilars for this indication is considered to be equivalent to the original protein drugs, but not necessarily at the same protein concentrations. Erythropoietins have also been used in oncology outside approved indications, such as in cancer-related anaemia, fatigue, and attempts in the treatment
of myelodysplastic syndromes.
For clinical approval, erythropoietin biosimilars were tested in chronic renal failure and also in an acceptable number of patients with chemotherapy-induced anaemia. Short-term use for chemotherapy-induced anaemia is probably not a safety issue. Erythropoietins have been approved for long-term use in chronic renal failure, but have not yet been tested for long-term use in other indications such as myelodysplastic syndrome; therefore there may be a safety concern for long-term usage in hemato-oncology.
In 2009, a clinical trial that included subcutaneous administration of intravenous formulation in chronic renal failure patients was discontinued due to cases of pure red cell aplasia (PRCA). Long-term subcutaneous administration of a protein drug is an effective way to induce anti-protein antibodies, and this aspect might be of concern for patients using erythropoietins on a long-term basis, where sufficient long-term documentation on approval is unavailable. The number of patients treated must be sufficiently large to detect the few cases where antibodies develop and result in severe clinical consequences, such as PRCA.
Antibody induction with clinical consequences is probably only witnessed in a fraction of patients, and after long-term use of erythropoietin. In myelodysplastic syndromes, antibodies against erythropoietin may aggravate the dysplasia of the erythroid lineage, and it is therefore important to have sufficient documentation of a biosimilar drug when extrapolating to other indications. Results from clinical studies of reference drugs cannot always be translated for use in a biosimilar drug that has not been tested for that indication.
Granulocyte-colony stimulating factor
The main indication for G-CSF is chemotherapy-induced anaemia (85%) and there are currently three unglycosylated G-CSF biosimilars approved for clinical use. These three drugs have been clinically tested in chemotherapy-induced neutropenia (CIN) and the clinical documentation for the three G-CSF biosimilars varies for this application. For one of the G-CSF biosimilars, the clinical documentation in CIN seems adequate; for the second, the number of patients is low, but fulfils the criteria established by EMA G-CSF guidelines;1 for the latest approved G-CSF biosimilar, the conclusion reached by EMA is somewhat surprising. For this biosimilar, there was slight statistically significant inferior efficacy with regard to effects on the duration of severe neutropenia (DSN) compared to the reference drug.
The conclusion reached by EMA was that the study met the criteria for demonstrating clinical equivalence in DSN and the small statistically significant increase in DSN was not deemed clinically significant. Statistical calculations for evaluation of efficacy of drugs are normally important to reach an adequate conclusion and a statistically significant difference is considered to indicate that the groups being compared are not equal. Furthermore, there were additional differences between the two groups, as the numbers of patients with severe neutropenia in cycle 1 favoured the reference drug. Overall, the clinical efficacy of the biosimilar drug is considered slightly inferior compared to the reference drug.
Moreover, this particular biosimilar had an additional toxicity profile regarding musculoskeletal and joint side effects. Furthermore, around 3% of the sample developed G-CSF antibodies, compared with none in the reference group.3 The clinical significance of anti-G-CSF antibodies is currently unknown, but it cannot be excluded that they may induce chronic neutropenia with clinical consequences, such as PRCA. Data on this aforementioned G-CSF illustrate how a biosimilar product, tested for comparability in pre-clinical systems, can exhibit some efficacy and safety concerns when further tested in the clinic. It should also be considered that the patient population receiving the G-CSF biosimilar was treated for a restricted time period, under immune-suppressed conditions.
Although G-CSF biosimilars are acceptable in treating CIN, there are some documented concerns both with regard to efficacy and safety in this application. These uncertainties might be accentuated when indications are extrapolated. A major concern for extrapolation is stem cell mobilisation in healthy donors. If musculoskeletal side effects and antibody induction is noted in patients on chemotherapy, these side effects are not acceptable in healthy donors. Moreover, no results are currently available on the quality of the mobilised stem cells, but if the quality is not as expected, it may put patients at risk.
Therefore, careful, long-term observations are required when G-CSF bioisimilars are used in healthy donors. G-CSF can also be used in the treatment of chronic neutropenia and myelodysplastic syndromes. Unlike the reference drugs, G-CSF biosimilars have not been subject to long-term clinical evaluations for these indications. Patients using them should be monitored in a pharmacovigilance programme or preferably included in clinical trials.
In conclusion, there may be differences in efficacy and safety between G-CSF biosimilars, which clinicians need to be aware of. This can lead to confusion, as many physicians have been informed that biosimilars should be comparable and exchangeable with the reference drug. A thorough pharmacovigilance programme focussing on G-CSF biosimilars, including the originator drugs, is necessary, with regular reports provided to the healthcare professional societies. Caution should be taken when using G-CSF biosimilars for extrapolated indications.
The patents of the first therapeutic monoclonal antibodies (Herceptin and Rituximab) will soon expire (in 2014 and 2015 respectively) and the first biosimilars are then expected to enter the market.
In November 2009, the EMA approved draft guidelines for monoclonal antibody biosimilars, which was sent for external consultation. Responses are to be submitted before 31 May, 2012 and it is expected that the guidelines will be approved the following year. The guidelines for erythropoietin and G-CSF are straightforward, and clinical trials have been suggested. Extrapolation has not been a major concern for erythropoietin but has to some extent for G-CSF.
Characterising and comparing monoclonal antibody biosimilars to the reference drug will be difficult, and production of biosimilar antibodies is significantly more complicated than small bioisimilars. Major changes in the manufacturing process contribute to differences in glycosylation, aggregation and protein folding, which may alter immunogenicity, antibody targeting and therapeutic activity.
The epitope of the target antigen should not differ from the comparator drug, as the function of antibodies is epitope dependent. The amino-acid sequence should also be identical to the comparator antibody. There are questions to consider: can a biosimilar antibody recognising a slightly different epitope of the target antigen be considered the same antibody? Do the DNA sequences of the amino acids of the CRD regions of VH and VL (the antigen binding region) have to be identical in the reference and the biosimilar antibody?
As previously mentioned, glycosylation pattern and other post-translational modifications for erythropoietin and G-CSF might have a clinical impact on efficacy and safety. The amino-acid sequence and glycosylation pattern of the CH2 region of the Fc part influence Fc-receptor binding and antibody-dependent cell-mediated cytotoxicity (ADCC) activity. For example, small differences in fucosylation have been shown to significantly affect in vitro ADCC, which might have consequences in vivo. The glycosylation pattern of the Fc not only determines ADCC but also CDC (complement dependent cytotoxicity) and the pharmacokinetic properties of
the antibodies. It may not be possible to copy identically using post-translational modifications.
The clinical consequences of possible differences between monoclonal antibody biosimilars and the reference antibody are difficult to establish. Biosimilar antibodies are already in clinical testing: how should the clinical study be designed? The probable aim of the biosimilar manufacturer is to demonstrate non-inferiority rather the superiority. Should response be evaluated for the antibody alone, or in combination with chemotherapy? Should tumour response, progression-free survival, and/or overall survival be measured? In the case of anti-CD20 antigens, which indication should be used: chronic lymphocytic leukaemia, follicular lymphoma, or diffuse large B cell lymphoma? All these tumours have different biology and different responses to anti-CD20 MAb. Extrapolation will be difficult due to differences in clinical behaviour, as well as different responses to MAb therapy.
All these aspects should be taken into consideration when evaluating a new monoclonal antibody biosimilar. The treating physician has to be aware of advantages and disadvantages of different antibodies to make the best decision, and they should be aware that a biosimilar antibody is not completely exchangeable for the reference antibody. However, this development may lead to novel and improved antibodies.
There has been tremendous progress in the development of new anticancer drugs, including biotherapeutics and biosimilars. This is a positive development; however, we should be aware that such protein drugs are difficult to reproduce and may have different activities even if, superficially, they may look the same. These considerations put major responsibilities on authorities approving drugs, companies producing and marketing drugs and physicians treating their patients. It is necessary to collect data from all treated patients to gather information on the safety and efficacy of new protein drugs: both reference drugs and biosimilars. This is a particular consideration for biosimilars, as at the time of approval the clinical information is restricted.