The Best Targets for Biogenerics

Developing biogeneric products is a hot topic, and the subject of many recent technical publications. Unfortunately, the lack of a comprehensive and definitive list of target products for biogeneric manufacturers has fostered ambiguity in many of the public discussions. Sorting through all the protein products that have been approved or licensed is an unwieldy and confusing task.

Charles E. DiLiberti

This analysis provides a comprehensive, 75-item list of all likely protein and peptide targets based on reasonable selection criteria. This approach eliminates many products whose sales volume could never justify a biogeneric copy. It also eliminates products whose unique properties exclude them from the mainstream discussion of biogenerics. We hope that reviewing the characteristics of these plausible biogeneric targets will hasten their development and marketing.

BIOGENERIC TARGET PRODUCTS: THE REGULATORY PATHWAYS

The first step in assembling a list of potential biogeneric products is to examine the FDA definitions and regulatory pathways for drugs and biologicals. Many names have been proposed to describe such products, including biogenerics, biosimilars, generic biologics, and follow-on protein products. Definitions vary, but most convey similar concepts. For our purposes, FDA's definition of follow-on protein product will suffice: "a protein product which is intended to be a similar version or duplicate of an already approved or licensed protein product."¹

Innovative protein products can be brought to the US marketplace via any of four possible regulatory pathways. The pathway that the target product employs impacts the pathway that the biogeneric product uses. The four pathways are reviewed as a first step in establishing reasonable criteria to limit the final list of biogeneric targets.

New Drug Application and Biologics Licensing Application Pathways

The New Drug Application, or NDA pathway, was established in 1938 by the Food Drug and Cosmetic Act (FD&C Act).² This pathway applies to products that are classified as drugs according to the FD&C Act definition:

"The term 'drug' means (A) articles recognized in the official United States Pharmacopoeia, official Homeopathic Pharmacopoeia of the United States, or official National Formulary, or any supplement to any of them; and (B) articles intended for use in the diagnosis, cure, mitigation, treatment, or prevention of disease in man or other animals; and (C) articles (other than food) intended to affect the structure or any function of the body of man or other animals; and (D) articles intended for use as a component of any article specified in clause (A), (B), or (C)...."³

This FD&C Act definition of drug, particularly at sections B and C above, is obviously quite broad and applicable to virtually any conceivable pharmaceutical product. Biogeneric mimics of target protein products approved as NDAs have a number of clear, well-established regulatory paths to follow, including potentially, the 505(b)(1) or full NDA pathway, the 505(b)(2) or limited NDA pathway, or the 505(j) or ANDA (Abbreviated New Drug Application) pathway.^4,5,6 Of these three potential pathways, the most often suggested for biogenerics is the 505(b)(2) pathway, which involves reliance, in part, upon data for which the applicant has no right of reference. What is meant by the regulatory term "no right of reference" is simply that the applicant does not own the data or have written authorization from the data's owner to use the data. The 505(b)(2) pathway thus allows an applicant to rely on data from another party that may not be connected in any way with the applicant or that may not want to cooperate with the applicant (e.g., the target product manufacturer). The former situation can arise when an applicant relies on a published scientific article for which the author of the article has no connection with the applicant. The latter situation can arise when an applicant relies on the fact that the target product was approved as safe and effective. This last point is the underlying principle by which generic drugs are approved — showing that the generic drug is equivalent to an approved brand drug and so possesses all of the safety and efficacy properties proved in the original studies that were submitted by the brand manufacturer.

The Public Health Services Act (PHSA),⁷ (later amended by the Food and Drug Administration Modernization Act of 1997 [FDAMA]) provides an alternative regulatory pathway for pharmaceuticals classified as biological products.⁸ Such products are licensed, but not approved. The PHSA defines "biological product" as follows:

"In this section, the term 'biological product' means a virus, therapeutic serum, toxin, antitoxin, vaccine, blood, blood component or derivative, allergenic product, or analogous product, or arsphenamine or derivative of arsphenamine (or any other trivalent organic arsenic compound), applicable to the prevention, treatment, or cure of a disease or condition of human beings."⁹

This definition is clearly much narrower than the definition of drug in the FD&C Act. Therefore, biological products are also drugs, but not all drugs are biological products. In other words, the universe of all biological products is a subset of the universe of all drugs.

Looking naively at these definitions and not knowing anything about the manner in which the products are actually registered in the US, most people would instinctively classify recombinant protein products that are not vaccines, toxins, or antitoxins, as drugs, rather than biological products. However, the FDA actually classified some recombinant protein products as drugs and others as biological products. It did this by using administrative rulemaking to create a broad interpretation of the definition of biological product in the PHS Act.

"600.3(h). Biological product means any virus, therapeutic serum, toxin, antitoxin, or analogous product applicable to the prevention, treatment or cure of diseases or injuries of man:

1. A virus is interpreted to be a product containing the minute living cause of an infectious disease and includes but is not limited to filterable viruses, bacteria, rickettsia, fungi, and protozoa.

2. A therapeutic serum is a product obtained from blood by removing the clot or clot components and the blood cells.

3. A toxin is a product containing a soluble substance poisonous to laboratory animals or to man in doses of 1 milliliter or less (or equivalent in weight) of the product, and having the property, following the injection of non-fatal doses into an animal, of causing to be produced therein another soluble substance which specifically neutralizes the poisonous substance and which is demonstrable in the serum of the animal thus immunized.

4. An antitoxin is a product containing the soluble substance in serum or other body fluid of an immunized animal, which specifically neutralizes the toxin against which the animal is immune.

5. A product is analogous:

(i) To a virus if prepared from or with a virus or agent actually or potentially infectious, without regard to the degree of virulence or toxicogenicity of the specific strain used.

(ii) To a therapeutic serum, if composed of whole blood or plasma or containing some organic constituent or product other than a hormone or an amino acid, derived from whole blood, plasma, or serum.

(iii) To a toxin or antitoxin, if intended, irrespective of its source of origin, to be applicable to the prevention, treatment, or cure of disease or injuries of man through a specific immune process."¹⁰

The broad interpretation of the term "virus" provides clear evidence of the latitude FDA took in interpreting the statute through the rulemaking process. Applying a similarly broad interpretation of the PHSA definition of "biological product," FDA classified many protein products that are now targets for biogeneric mimics as "biological products."

Nevertheless, many other biogeneric targets were classified as drugs and approved as NDAs under the FD&C Act. The implications for manufacturers seeking to register a biogeneric mimic of a given target protein product is significant; the regulatory pathway is clear if the target product was approved as an NDA, but it is not nearly as clear if the target product was licensed as a biological product. Although some ideas about how a manufacturer potentially could seek registration for a biogeneric mimic of a licensed biological product have been proposed, these pathways have not been tested.

The well-characterized biotechnology products that were licensed as BLAs and are typically included in the public discussion of biogenerics were originally under the jurisdiction of FDA's Center for Biological Evaluation and Research (CBER). On June 30, 2003, these products were transferred to the jurisdiction of the Center for Drug Evaluation and Research (CDER). This move reflects the drug-like qualities of biotech products. The more "classical" biological products, such as gene therapy products and blood products, i.e., those that fit the literal definition in the PHSA, have remained under CBER's jurisdiction. (The only exceptions are two botulinum toxin preparations.)

Medical Device Pathway

The third pathway to market for innovative drugs, which is often overlooked because of the limited number of products to which it applies, involves filing the product as a medical device, under the FD&C Act. Examples of such products include gelatin hemostatic wound-care products (e.g., Surgifoam, Gelfoam). These products are excluded from the discussion of biogeneric targets because the protein in these products acts in a nonspecific manner.

DESI Pathway

The fourth pathway available is to market a product containing an old, pre-1938 drug without an approval or license. Some selected active ingredients in use before 1938 are still under Drug Efficacy Study Implementation (DESI) review. This program was initiated to review the efficacy data of old drug products that had been marketed before FDA registration was required and thereby ensure that such older products did indeed work as claimed. Drug products containing active ingredients under DESI review may still be marketed legally as long as the manufacturers follow GMP regulations.

This pathway is also often overlooked because of the limited number of products involved. We believe this pathway was used to market such topical wound debridement products as Xenaderm (containing the protease trypsin) and Accuzyme (containing the protease papain), although this is difficult to confirm given the virtual absence of publicly available regulatory information on these products. Any DESI target products that met the financial criteria were excluded from the final list because the protein component was only a small fraction of the total amount of active ingredients present, and because such products have not been part of the mainstream discussion of biogenerics.

METHODS

Financial Inclusion Criteria

For a branded protein pharmaceutical product to have any value to a potential biogeneric manufacturer, it must have an adequate potential for economic return to justify the significant investment required to develop the biogeneric mimic. We arbitrarily selected $20 million as the minimum annual product sales for any branded product to justify developing even the simplest of biogeneric mimics. This figure errs, if anything, on the side of including some products that may not be economically feasible targets, but it was decided that providing a comprehensive list was of greater value than providing an incomplete, but highly selective list.

The starting data set for this analysis, therefore, consisted of a list of all pharmaceutical products (including protein and small-molecule products) for which the 12 full months' worth of US sales data from September 2004 to August 2005 was at least $20 million.¹¹ From this preliminary list, all protein and peptide products were identified and retained, while all small molecule products were removed from consideration.

Exclusion Criteria

Once the list of all protein and peptide products with annual sales of at least $20 million was prepared, products that fell into the categories listed in Table 1 were removed. The specific reasons for removing each category from consideration are provided in the table, but the more general rationale is that these types of products have not been part of the mainstream discussions of biogenerics. These excluded products represent either classical biological products that clearly fit within the literal PHSA definition of "biological products" and have remained under the jurisdiction of CBER, or highly specialized situations not amenable to generalized regulatory guidance. Excluding such products from this analysis is not meant to suggest that they cannot be genericized, but simply highlights the fact that they have not been considered in the recent discussions of biogeneric products.

Table 1. Classes of products excluded from consideration

The products remaining on the list at this point were consistent with the focus of recent public discussions, which have been about developing biogenerics for well-characterized (or at least well-characterizable) recombinant DNA and natural source target products. (For more on the discussions see the sidebar.)

Recent Public Discussion

Data Consolidation

The final adjustment to the list of biogeneric targets was to group redundant entries together as if they were single products. The first criterion used for this grouping was to consolidate entries for products that were deemed therapeutically equivalent (A-rated) in FDA's Orange Book.¹² The second criterion was to consolidate all products from a single manufacturer containing the same active ingredient; the rationale was that the same active ingredient from the same manufacturer, though used in different finished products, ought to be the same.

The result of these consolidations, therefore, was a list in which each entry contains a unique active ingredient. This approach was taken in part to avoid undue weighting of some products, particularly insulins, for which several distinct formulations are available from a given manufacturer using a single active ingredient.

Product Attributes Each product on the final list of biogeneric target products was examined for a variety of characteristics, including filing mechanism (NDA or BLA), number of amino acid residues, major chemical modification (glycosylation, PEGylation), source (synthetic, rDNA, or native DNA), expression system (where appropriate), and whether the product bears a current A-rating (indicating therapeutic equivalence) in FDA's Orange Book.

Most of this information was derived from package inserts for the products, although some was obtained from other sources, such as publicly disclosed product approval or licensing information available through FDA's Freedom of Information Act (FOI) Office, and scientific publications.

Table 2a. Biogeneric target products with annual US sales > $20 million

The number of amino acid residues reported was either the actual count (if available), or for a few products, as indicated with boldface type, was estimated from the reported molecular weight. For products known to associate as multimers (e.g., insulin hexamers), the number of amino acids listed is the number of amino acids in unique sequences. Therefore, for homomultimers, the number of amino acids reported is simply the number of amino acids in the monomer. For heteromultimers (e.g., heterodimeric pituitary hormones, such as follitropin), the number of amino acids reported is the sum of the number in each distinct subunit. In the case of follitropin, for example, the alpha subunit contains 92 amino acid residues, and the beta subunit contains 111 amino acid residues, for a total of 203 amino acid residues.

Table 2b. Biogeneric target products with annual US sales > $20 million

RESULTS

The protein and peptide products meeting our criteria for potential biogeneric targets are presented in Table 2 in ascending order of the number of amino acids in unique sequences in the product. This order very crudely approximates ascending order of product complexity.

Table 2c. Biogeneric target products with annual US sales > $20 million

A total of 75 target products were found that met all of the criteria specified, of which 34 had been filed as NDAs and 41 as BLAs. There is substantial overlap in the molecular sizes of the NDA and BLA products: the NDA products ranged in size from 7 to about 691 amino acids, and the BLA products from 109 to about 2,322 amino acids. About half of the products were not chemically modified, three were PEGylated, and the rest were glycosylated. These products have been sorted by chemical modification and filing pathway in Table 3.

Table 3. Breakdown of products by pathway and chemical modification

These results show that target protein products, without major chemical modifications, have been filed by both the NDA and the BLA pathways. Likewise, glycosylated target products have been filed by both pathways. Although no PEGylated target products in this final list were filed by the NDA pathway: this pathway does not preclude PEGylated products. Indeed, there are at least two PEGylated protein products not reported here (due to insufficient sales volume) that have been approved via the NDA pathway (Pegademase Bovine [Adagen], and Pegvisomant [Somavert]).

Table 4 shows the results from the final list tabulated by source and by filing pathway. Approximately 15 percent of the products were produced synthetically. These products ranged in size from 7 to 36 amino acids, and all were approved via the NDA route. The recombinant DNA products accounted for most (about 83 percent) of the total number of products on the final list, and ranged in size from 29 to about 2,330 amino acids. Many recombinant DNA products were filed via each of the two major pathways. Only two products were derived from native DNA (i.e., non-recombinant systems). These were 58 and about 295 amino acids in length, and were both filed as NDAs.

Table 4. Breakdown of products by pathway and source

Table 5 shows the products classified by expression system and by filing pathway. Although mammalian cell culture production methods were the most common (about 45 percent of the total), microbial systems (yeast, and especially bacterial) were nearly as prevalent (39 percent of the total). Only one natural extract (aprotinin, derived from bovine lung) was found among the final data set. With the exception of the target products produced synthetically, which were filed only through the NDA pathway, there was substantial overlap between the expression systems used for NDA and BLA products.

Table 5. Breakdown of products by pathway and expression system

DISCUSSION

From this analysis, it is clear that potential biogeneric target products span a wide range of characteristics ranging from very small homogeneous peptides to large heterogeneous glycoproteins. Presented with this information, there is little point in trying to create an artificial boundary between products that are small enough to call peptides and larger compounds that might be called proteins. As protein size increases, complexity tends to follow, though not in any direct way, and certainly not in any discontinuous way that would warrant a peptide–protein boundary of sorts. Therefore, the regulatory requirements should be graded similarly to reflect product complexity, without any sharp lines of demarcation. Likewise, while some may argue that the smallest entries in Table 2a should not be called proteins, they are, nevertheless, included for completeness, as well as to illustrate the broad continuum of target products available.

Although more of the target products were approved as BLAs, a substantial number (45 percent) were approved as NDAs. Thus, even if the regulatory pathway for the registration of biogeneric mimics of BLA target products is not entirely clear at this time, there are still numerous protein products approved as NDAs. Biogeneric firms can target those knowing that a clear regulatory pathway exists.

There is substantial overlap between the characteristics of target products approved as NDAs and the characteristics of those approved as BLAs. The notion that the products that were licensed as BLAs somehow fit the statutory definition of "biological product" any better than many of the target products approved as NDAs is puzzling. All of the products in this analysis clearly fit the broader FD&C Act definition of "drug." This overlap between target products registered through the two pathways also argues in favor of a graded scientific approach to the development of biogeneric mimics of both types of target products rather than two distinct, dissimilar scientific approaches.

There are two groups of products that could conceivably be targeted by biogeneric companies, but which have not been part of the mainstream discussion of biogenerics. These two groups are, first, what could be termed "classical" biological products (i.e., those less well characterized products that clearly fit the literal PHSA definition and remain under CBER's jurisdiction, such as vaccines, toxins, antitoxins, blood products, etc.), and second, some highly specialized protein products not registered via the NDA or BLA pathways (such as wound care products). Because these two groups of products have not undergone substantial public discussion, guidances may take some time to develop. Therefore, the most urgent need is to issue guidances that address the development of biogeneric mimics of well-characterized protein and peptide products such as those listed in Table 2. Although only a few of these products have as yet received Orange Book A ratings, (and are thus, deemed to be fully interchangeable with other products), the technology exists to establish therapeutic equivalence and/or comparability between biogenerics and the corresponding target protein products listed in Table 2.

ACKNOWLEDGEMENTS

The author wishes to express his gratitude to Mr. Nicholas Tantillo and Ms. Christine Mundkur for their invaluable assistance and regulatory guidance.

Charles E. DiLiberti is vice president of scientific affairs at Barr Laboratories,Inc., 400 Chestnut Ridge Road,Woodcliff Lake, NJ,07677, cdiliberti@barrlabs.com

REFERENCES

1. Webber K. Scientific considerations related to developing follow-on protein products. FDA Public Workshop; 2004 Sept. 14-15; Rockville, MD

2. The Food Drug and Cosmetics Act. 1938 June 25.

3. The Food, Drug, and Cosmetics Act, 21 USC 321(g)(1) Definition of drug.

4. The Food, Drug and Cosmetic Act Section 505(b)(1).

5. The Food, Drug and Cosmetic Act Section 505(b)(2).

6. The Food, Drug and Cosmetic Act Section 505(j).

7. Public Health Service Act. Title 42 United States Code. Chapter 6.

8. Food and Drug Administration Modernization Act of 1997.

9. US Code. Title 42, Chapter 6A, Subchapter II, Part F. Subpart 1. Regulation of biological products. 42 USC 262(i).

10. Code of Federal Regulations. Title 21—Food and Drugs. Chapter I Subchapter F — Biologics. Part 600 – Biological Products: General. 21 CFR 600.3(h).

11. IMS Health National Sales Perspective Audit, IMS Health, Fairfield, CT.

12. FDA, CDER, Office of Pharmaceutical Science, Office of Generic Drugs. Approved drug products with therapeutic equivalence evaluations. Known as the Orange Book.

13. FDA Public Docket 2004N-0355; Available at: www.fda.gov/ohrms/dockets/dockets/04n0355/04n0355.htm.

14. EMEA Final Guidelines; Available at: www.emea.eu.int/htms/human/biosimilarfin.htm.