Science and Regulatory
 Authorisation of Multi-source Biopharmaceuticals
A Multisource Approach Towards Biotechnology-Derived Products
— The EGA Industry Position —
EGA Annual Meeting on Regulatory Affairs
London, January 25, 2000
Dr. Erich Kohler
Head of Development Biotechnology
BioGeneriX AG
Ladies and Gentlemen,
It is a great pleasure for me to open the last session of this conference, dealing with the regulatory situation of biotechnology-derived products. I want to thank the organisers for giving me this opportunity, especially Nadene McClay and Greg Perry of the European Generic Medicines Association, and Mrs. Singer and her colleagues of Management Forum for her patience while waiting for my overheads and supporting documents.
Well, why talk about biotech products in the context of generic applications? There are no biotech generics existing today, for a number of reasons:
- Biologics have historically been regulated apart from chemical drugs, and so have biotechnology-derived products
- Biologics have been complex substances or mixtures, difficult to characterise or standardise or even reproduce
- Source materials have been variable and potentially contaminated, effecting in individual impurity profiles
- Analytical tools have neither provided proof of structure, nor allowed for conclusion on efficacy.
Thus, in contrast to chemical drugs, biologics are defined by their manufacturing process. Data from process validation and batch consistency are at least as important as product characterisation and data from release testing. Consequently, even minor changes in the manufacturing process must be looked at as potential triggers for a change in safety or efficacy, which could, strictly speaking, only be ruled out by conducting additional clinical studies.
During the next couple of minutes we will come back to these considerations, and see whether they also apply to biotechnology-derived products. But first of all it is necessary that we agree on what products are to be considered biotech-derived, at least for this session. I suggest that we stick to those products for which the Centralised European Licensing Procedure was established a few years ago.
As you know, this procedure is based on a single application to be filed by a sponsor with the EMEA, the European Medicinal Evaluation Agency. This will lead to an opinion issued by the CPMP, the Committee for Proprietary Medicinal Products. Finally, a single marketing authorisation can be obtained for all countries of the European Union.
The Centralised Procedure is mandatory for:
- Products developed by recombinant DNA-technology
- Products developed by controlled expression of genes coding for biologically active proteins in prokaryotes and eukaryotes, including transformed mammalian cells
- Products developed by hybridoma or monoclonal antibody methods
In addition, the Centralised Procedure may be used on request of the sponsor for products containing a new active substance, products whose administration or indication or manufacturing process constitutes a significant innovation, and for new products derived from human blood or human plasma.
Let us have a closer look at the classes of biotechnology-derived products, and how eventual changes in their manufacturing processes and the aspect of comparability are currently regulated.
For monoclonal antibodies things are quite clear. The existing CPMP Guideline on Production and Quality Control of Monoclonal Antibodies describes the conditions to prove product equivalence after changes made in the manufacturing process. Examples of such changes are:
- Change in culture conditions
- Change in purification procedure
- Transition of in vivo to in vitro production
- Modification of the monoclonal antibody molecule
The guideline stipulates that essential identity of the products before and after the change be demonstrated by means of physico-chemical and biological characterisation of the antibodies, by characterisation of the cell lines, validation of the manufacturing process, and by pharmacological studies. Clinical studies performed with the former monoclonal antibody can then be accepted for the new version.
A prerequisite, however, is that manufacture is based on the same Master Cell Bank. Otherwise, the guideline states, clinical trials have to be carried out with the second form of the antibody. As this scenario would apply to another manufacturer trying to develop an equivalent product, generic or multisource approaches are currently not being looked at for monoclonal antibodies, at least not to my knowledge.
The situation is different with recombinant DNA-derived products. The existing CPMP Guidelines on Production and Quality Control of Medicinal Products Derived by Recombinant DNA-Technology, on Production and Quality Control of Cytokine Products Derived by Biotechnological Processes or on Analysis of the Expression Construct in Cells Used for Production of rDNA-Derived Protein Products do not specifically address the implications of process changes, and therefore leave more room for case-by-case assessment.
This flexibility has over the years been applied when defining the requirements for the proof of structural equivalence and batch consistency after changes made for a given product. In the US, the Food and Drug Administration also recognised that changes in the manufacturing process, equipment or facilities would not necessarily result in changes in the biotech product itself, or in the need for additional clinical studies.
We will have the US situation analysed right after my talk, so I will not go any further into this direction. However, it is important to note for the moment that in 1996 a Guidance was issued by FDA concerning demonstration of comparability of biological and biotechnology-derived products.
This guidance covers biological and biotech products regulated by CBER (Center for Biologics Evaluation and Research), as well as the protein drugs regulated by CDER (Center for Drug Evaluation and Research), such as human growth hormone or human insulin. Comparability testing programs are outlined with the usual progression of complexity, which may include physico-chemical analysis, in vitro or in vivo bioassays, pharmacokinetics or toxicity studies in animals, and eventually human pharmacological studies.
However, all this refers to the assessment of comparability of a given product following changes made in its manufacturing process. It does not address the comparability of an existing recombinant DNA-derived product and a counterpart newly developed by another manufacturer.
On the other hand, in summer 1998 EMEA issued a Concept Paper on the Development of a CPMP Guideline on Comparability of Biotechnology-Derived Products. According to this document, the goal was set to establish a guideline which would - besides the implications of changes with a given product - cover also the more general question on how to compare two biotech products from different manufacturers. To be more focused: how to compare two recombinant proteins from different manufacturers.
It is suggested that such guideline shall establish the requirements for:
- Physico-chemical and biological tests to demonstrate structural equivalence
- Identification and assessment of potential impact of changes made in the production process on the quality of a product
- Discussion to what extent quality tests are sufficient to demonstrate the comparability of two recombinant proteins, so that only limited toxicological or clinical studies are necessary
So, the focus of a potential multisource approach is not on most recent biotech innovations like antisense-drugs or gene therapy, but on recombinant DNA-derived proteins. They represent a somewhat homogenous class of products when thinking of their development: isolation of the genetic information, establishment of a cell bank system, fermentation and purification, formulation as finished product.
Several important recombinant proteins are already or will be coming off patent during the next couple of years. This will further stimulate the question of whether equivalence can be demonstrated between a newly developed and an already marketed biotech product. Some examples are given on this overhead:
- Human insulin
- Human growth hormone
- Interferon alfa
- Interferon gamma
- Tissue plasminogen activator
- Erythropoietin
- Interleukin 2
- Granulocyte colony stimulating factor
I think it is not really impossible to completely describe the structure and molecular characteristics of some of them, or impossible to not only assess their activity but to conclude on their efficacy in humans. Taking human growth hormone as an example, we actually face a generic situation already today. Depending on the country, half of a dozen products can be available from different manufacturers.
In Germany, 5 products are currently marketed in various dosage strengths. 4 of them originate from E.coli, two of which from standard E. coli K12, and the other two from special E. coli strains. One product is manufactured using a transformed mouse cell line. The somatropin contained in all products, however, has not only the same 191 amino acid sequence, identical with human pituitary growth hormone, but also the same potency of 3 IU per mg and the same clinical effects.
What can we conclude from this? From differently constructed vectors and differently expressing host cells the same protein may be obtained, as demonstrated by highly sensitive characterisation methods. Applying different manufacturing processes and different purification protocols, apparently equivalent products arise with the same safety and efficacy profiles when administed to patients.
And now, what is to be done when another manufacturer shows data on his somatropin: expressed in E. coli K12, identical in amino acid sequence with the pituitary growth hormone and the products already established in the marketplace, and equipotent to them as demonstrated by means of bioassay and pharmacokinetic studies. Would it really contribute to patient safety if all knowledge about the already available products were put aside, and instead of assessing comparability a full stand-alone clinical development were required?
I do not want to be misunderstood. I do not argue in favour of a minimalistic development program. I agree that every sponsor must establish a comprehensive data base for his product. This means to understand the expression system, to validate the manufacturing process, to characterise the protein, and to know what it will effect therapeutically. This may well include also pre-clinical analyses or data from humans, but not necessarily a full program in each and every case. I want to argue in favour of a thorough, scientifically sound case-by-case assessment.
The example with human growth hormone, you may say, is too straightforward. Somatropin is not glycosylated, and the protein can be well characterised. But we can also look at a more complex biotech product such as recombinant human erythropoietin. You have listed here the main characteristics of EPO:
- Glycoprotein with 165 amino acids
- Molecular weight approx. 31 kD
- 2 disulfide bonds between Cys 7-161 and Cys 29-33
- 4 complex carbohydrate side chains at Asn 24, 38, 83, and Ser 126
- Specific activity > 120 000 IU/mg
Two different EPO products are currently approved in Europe which were developed independently from each other more than 10 years ago. Both are manufactured using Chinese Hamster Ovary-cells, but the expression systems used are not the same. One EPO is derived from genomic DNA, the other from a cDNA construction. Fermentation and purification procedures are quite different, and take place at different locations, of course.
The two drug substances are even named differently, Epoetin alfa and Epoetin beta. But when you analyse them, differences in their protein characteristics will not become apparent. They show us the same amino acid sequence, the same peptide maps and isoelectric focusing pattern, the same graphs in spectroscopy, and so forth. You will not find indications of different conformation or secondary structure.
Due to the heavy glycosylation both substances are intrinsically micro-heterogenous. With today’s analytical tools it is possible to detect variations in, for example, distribution of di-, tri- and tetra-antennary structures, or repeats of carbohydrate components within an antenna. But this may be observed even on the level of batch-to-batch variability for each individual EPO. Anyhow, for these two products their micro-heterogeneity seems to have no impact on clinical safety or efficacy, because for a decade now they have been reported to be interchangeable in their clinical use.
Last year, a monograph was issued in the European Pharmacopoeia, describing analytical methods and pharmaceutical requirements for EPO drug substance manufactured by recombinant DNA-technology. The equivalence, or identity, of Epoetin alfa and beta has triggered that the Biological Reference Preparation necessary for establishing the monograph actually became a 50:50 blend of both EPO’s.
In my view, such a monograph on a biotech product makes only sense when the analytical tools are recognised to be elegant enough to detect molecular changes that might impact safety and efficacy, as with other product classes. If I believe that my characterisation program will leave room for unobserved molecular features whose change will result in an altered clinical profile, then there is no basis at all for comparability, neither for a multisource approach nor for process changes with a given product.
Again, I do not argue that everything that looks like EPO could be administered to patients without a thorough understanding of what would be the effect. But applying current state of the art it is possible for one manufacturer to optimise his expression vector or to establish a new Master Cell Bank in the course of development, and then to prove chemical, pharmaceutical and biological equivalence, so that pre-clinical and clinical studies performed with the former protein can be accepted for the new protein. Why should not the same principles apply when a second manufacturer uses the same genetic information to construct an analogous host cell system able to produce the same active moiety?
The different impurity profiles, you may say, could constitute a safety risk, or the different profiles of product-related substances. Well, on the one hand it is recognised that one manufacturer may scale-up his manufacturing process, change to a serum-free fermentation medium or optimise the purification steps, and may rule out potential clinical impact with a limited equivalence study. Why should not, on the other hand, a second manufacturer setting up an analogous manufacturing process at another site have the same opportunity, without being obliged a priori to complete a full pre-clinical and clinical development program?
I am aware that such a multisource approach can easily reach its limits. It is obvious that equal potency should be established, preferably using a bioassay which is known to reflect the efficacy of the product in humans. This cannot always be achieved, especially when the exact mechanism of action is not yet known. But I want to point out that it seems illogical to allow for a comparability assessment in case of two proteins obtained before and after a significant process change, and to rule out up-front such an assessment for two equally well-characterised multisource proteins.
In both the cases of process changes for a given product and a multisource approach, the impact on the impurity profile, on product-related substances, on potential immunogenicity and, finally, on safety and efficacy should be assessed on a case-by-case basis. The examples listed in the EMEA Concept Paper, of which the potential effect on product quality shall be identified, cannot be assigned to either of the two cases, but in my eyes are relevant for both:
- New production site
- Pilot versus full-scale production
- Different cell substrate or culture media
- Change in the purification scheme
- Change in equipment and facilities
- Different impurity profile
- Neo-antigenicity of the new product
Consequently, the anticipated CPMP Guideline on Comparability of Biotechnology-Derived Products should not in the first instance differentiate between process changes for a given product and a multisource approach. It should rather define the requirements for a submission under the Centralised Procedure dependent on the significance of the change in the manufacturing process or the complexity of the active moiety.
This means that for a heavily glycosylated protein appearing in a range of isoforms the amount of studies and documentation beyond physico-chemical and biological characterisation required from a second manufacturer to demonstrate comparability will be greater than for a small peptide. This is similar to today’s different requirements with a given product for a new Master Cell Bank or manufacturing site versus a change in buffers used during production or for shelf-life extension.
I want to conclude that a truly generic approach, analogous to conventional chemical drug products, would not be appropriate for products manufactured by means of recombinant DNA technology. However, significant improvements have been made in recent years regarding consistency of manufacture, control of impurities and analytical procedures.
Hence, a multisource approach for recombinant proteins should no longer be a priori ruled out. Comparability should be assessed looking at the scientific data and arguments provided by the sponsor, and the requirements for regulatory approval be defined on a case-by-case basis. This would help bring important, and eventually improved, biotech products on the market more efficiently and expeditiously.
Thank you very much for your patience and attention.
<< Back
|
