Development and Regulatory Challenges for Peptide Therapeutics

Opportunities and Challenges of Intracellularly Targeted Peptide Drug Discovery (Tomi Sawyer)

Historically, peptide drug discovery (Figure 1) began with a major focus on receptor targets (eg, G protein–coupled receptors [GPCRs]) during the 1970s and this first wave has continued onward to present times. In the late 1980s and into the 1990s, there were noteworthy achievements in the identification of peptides that exhibited cellular permeability and/or the ability to modulate intracellular target activity. This second wave was exemplified by the macrocyclic natural product peptide cyclosporin, several classes of cell-penetrating peptides (CPPs) such as transactivating transcriptional activator (TAT)-based synthetic CPPs, and a worldwide campaign on HIV drug discovery that include many novel synthetic peptidomimetic inhibitors of HIV protease. Over the past 3 decades, powerful and innovative technologies enabling extraordinary diversity (10⁶-10¹²) of peptides through phage display, mRNA-display, DNA-encoded and other synthetic libraries are significantly driving a third wave of macrocyclic peptide therapeutics. Importantly, it is well-recognized nowadays that peptides are a compelling modality for both “classic” targets as well as “undruggable” targets versus small molecules that may have inability to bind to large, flat binding surfaces or antibodies that have limited cell permeability to access intracellular targets. ^{19 -27}

OPEN IN VIEWER

Figure 1.

Some historical milestones in peptide drug discovery, including 3 major waves (see text for details).

It is recognized that there has been increased biotech/pharma R&D efforts focused on macrocyclic peptides to challenge varying targets ranging from GPCRs to other classes of receptors, the cell membrane (eg, antibiotics, CPPs) and intracellular proteases, protein–protein interactions, and protein–nucleic acid interactions. In terms of recently FDA-approved peptide drugs (Figure 2), such target space is well exemplified across many therapeutic indications, including endocrine, gastrointestinal, and infectious diseases as well as cancer. Interestingly, there is a significant effort focused on macrocyclic peptides and peptidomimetics ^{19 -27} that are being advanced, and unquestionably these will further increase relative to the third wave of peptide drug discovery to build upon what has already been accomplished. Yet, a remaining challenge for peptide therapeutics is drug delivery. Albeit, sustained-acting peptides for receptor targets (eg, GLP-1 analogs) have been advanced via methods such as fatty acid conjugation or PEGylation to provide subcutaneous injectable therapeutics, there has been very limited success relative to oral delivery. Nevertheless, it is noteworthy to highlight the success of peptidomimetics with respect to oral delivery as well as recent achievements ²⁸ with the GLP-1 agonist peptide Rybelsus (FDA-approved 2019) which spurring renewed enthusiasm to further investigate the potential of oral delivery technologies for peptides.

OPEN IN VIEWER

Figure 2.

Some Food and Drug Administration (FDA)–approved peptide/peptidomimetic drugs (2005-2015) and future strategies (see text for details).

Perhaps the most challenging, but yet compelling, opportunities for peptide drug discovery is that related to intracellular target space. Many therapeutic targets in the cell have been considered essentially undruggable by conventional small molecules and although quite likely druggable by antibodies there has been very limited success to confer cellular permeability to antibodies or therapeutic proteins. Therefore, peptides, and more precisely, macrocyclic peptides have become highly favored within the molecular armamentarium of biotech/pharma to become a new therapeutic modality to tackle intracellular target space in innovative ways. ^{19 -27} Some noteworthy examples include cyclosporine-A (CsA), ²⁹ cyclic CPPs, ³⁰ and ATSP-7041 ³¹ (Figure 3).

OPEN IN VIEWER

Figure 3.

Some macrocyclic peptides exemplifying cell permeability by varying mechanisms (see text for details).

Unquestionably, the natural product peptide CsA was a monumental breakthrough for peptide drug discovery relative to its cell permeability properties and ability to modulate intracellular target function. Albeit its large-size, the specific structural features of CsA, such as N-methylation of several backbone amides as well as a high degree of lipophilicity, have been determined to contribute to its passive permeability. In fact, CsA is regarded as a benchmark for both structure-based design and diversity-driven strategies to advance novel macrocyclic peptides of this general class. ^23,25,32 Cyclic CPPs are a relatively new class of macrocyclic peptides which have translated past learnings from linear CPPs (eg, Tat and poly-Arg) in terms of cationic partitioning into cell membranes and endocytosis to achieve cellular uptake and release from endosomes into the cytosol. ^22,30,33 Such cyclic CPPs may be used as carriers or be designed to incorporate a peptide pharmacophore for an intracellular therapeutic target. Stapled α−helical peptides have also arisen as a promising new class of macrocyclic peptides ^{24,26,34 -36} and have recently advanced into the clinic (ie, ALRN-6924). ³⁷ ATSP-7041 ^{31,38 -40} (a progenitor of ALRN-6924) is a benchmark stapled α−helical peptide to explore cellular permeability that seems to fit into a third mechanism which has been dubbed lipophilic partitioning that is different from the passive transport properties of CsA and the cationic partitioning (and endosomal pathway) of cyclic CPPs. These 3 general mechanisms of cellular uptake are recognized to effectively enable peptide entry into the cell to subsequently engage and modulate the functional properties of a therapeutic target (Figure 4). Nevertheless, there are many other factors which may significantly impact the efficacy of such peptides, including intracellular metabolism as well as the extracellular concentration of the peptide and how metabolism, excretion, and serum albumin binding may collectively contribute to exposure levels and PK properties.

OPEN IN VIEWER

Figure 4.

Model for cell uptake and intracellular fate of 3 classes of macrocyclic peptides (see text for details).

To further understand both peptide cell permeability and drug delivery, there is great need for robust screening tools (computational and experimental) which can facilitate predictive design rules and iterative testing thereof. As expected, there exists a framework of existing screening tools for passive permeability (eg, transcellular permeability using cellular monolayers and proteolytic stability in varying biological matrices) as well as emerging computational and biophysical methods, ^{41 -51} including those that address exposed polar surface properties, nuclear resonance/mass spectrometry (NMR/MS) analysis of intramolecular H-bonding, radius of gyration, and aqueous/lipid phase partitioning (Table 2). Cell-based screening methods ^{52 -56} include MS and/or fluorescence microscopy for unlabeled or fluorescently labeled peptides, respectively, and such may also be used to investigate cell uptake and correlate with cell-based functional properties. More recently, the development of a peptide cell permeability screening using a HaloTag-based assay dubbed CAPA (chloroalkane penetration assay) has provide quantitative and high-throughput analysis to understand the structure-permeability relationship of a diverse set of peptides. Overall, such screening tools will significantly enhance the understanding of peptide cell permeability and the development of intracellularly targeted peptide therapeutics.

OPEN IN VIEWER

Table 2.

Experimental and Computational Screening Tools to Understand Peptide Permeability (See Text for Details).

Permeability screening tool	Some key features and properties evaluated
Exposed polar surface area (EPSA)	Experimental EPSA values are determined using supercritical fluid chromatography Low EPSA values have been shown to correlate with high passive permeability and predicted oral bioavailability
ΔG(insertion)	ΔG(insertion) is a calculated value that refers to the solvation free difference for transferring peptide in a low-dielectric conformation (LDC) from water to a low-dielectric environment (membrane-like) and is predictive of passive permeability
Parallel artificial membrane permeability assay (PAMPA)	PAMPA uses mixtures of phospholipids infused into lipophilic microfilters with net negative charge (surrogate model system to correlate with experimental oral bioavailability)
Lipid: water phase partitioning	Both octanol: water phase partitioning and recent modification to incorporate fatty acids and pH gradient as shown for guanidinium-rich peptides to be predictive of energy-independent translocation
Cell monolayer transcytosis	Caco-2 cells, or other cell types, are used to measure passive permeability from donor to acceptor compartments to correlate with in vivo oral bioavailability
Cell-based chloroalkane penetration assay (CAPA)	CAPA measures cytosolic delivery of chloroalkyl-modified peptides in cells that stably expresses a haloenzyme-GFP fusion protein
Radius of gyration (R-gyr)	R-gyr is calculated as the root-mean-square distance between a peptide’s atoms and its center of gravity R-gyr is an alternative property for MW for beyond-rule-of-5 (bRo5)molecules
NMR analysis of intramolecular versus solvent H-bonding	Solution NMR studies using hydrogen/deuterium (H/D) exchange to determine peptide backbone amide temperature coefficients and intramolecular hydrogen bonding
Label-free mass spectrometric and fluorescently tagged cell uptake analysis	Direct measurement methods for cell uptake of peptides using MS methods and/or imaging studies (eg, fluorescence correlation microscopy) to quantitate intracellular exposure

Genotoxicity Assessment of Peptide/Protein Therapeutics—Perspectives From an HESI Working Group (Zhanna Sobol)

Health and Environmental Sciences Institute’s Genetic Toxicology Technical Committee developed a publication describing the “Genotoxicity assessment of peptide/protein-related biotherapeutics: points to consider before testing.” ⁵⁷ The goal of this publication was to identify specific challenges in genetic toxicology testing of peptide/protein-related biotherapeutics and provide recommendation for best practice approaches. Earlier initiatives surveyed genotoxicity testing of approved large molecule drugs and showed that many protein/peptide drugs were tested in the standard genotoxicity test battery and almost all were negative. ^58,59 The working group sought to understand when standard genotoxicity testing added value to the risk assessment of the molecule and when the testing was unnecessary.

Relevant regulatory guidelines such as ICH S6 and ICH S2 contain gaps related to genotoxicity assessment of protein/peptide therapeutics. International Council for Harmonisation S2 guideline describes the standard genotoxicity test battery (in vitro bacterial reverse mutation assay, in vitro test for chromosome damage, and in vivo tests for chromosome damage) but does not address limitations for testing of proteins/peptides because this guideline was intended for small molecule therapeutics. Although ICH S6 states that genotoxicity testing is not relevant for biotherapeutics/biologics, it stipulates that testing may be needed if there is a cause for concern. The ambiguity around what constitutes a cause for concern often leads to conduct of the standard genotoxicity test battery. Synthetic peptides in particular fall in a gray area between small molecule and biologic and are often tested in the standard genetic toxicology battery even when there is no cause of concern. When synthetic peptides contain only natural amino acids, they are not expected to behave differently toward DNA than recombinant proteins. Likewise, if synthetic peptides contain modified amino acids that have been previously tested, the inclusion of these modified amino acids in a new sequence is not expected to alter the reactivity of the resulting peptide toward DNA. The potential limitations of testing protein/peptide therapeutics according to ICH S2 guideline include lack of uptake in in vitro systems, histidine release confounding the bacterial mutation assay, and various confounding effects on the in vivo assays. Nongenotoxic mechanisms that can lead to induction of micronuclei in vivo include stimulation of erythropoiesis, nucleotide pool imbalance that is secondary to decrease in folate levels and hyper/hypothermia. ⁶⁰ The potential confounding effects on in vivo genotoxicity tests have been documented for protein/peptide therapeutics such as recombinant human erythropoietin leading to acceleration of erythroblastic maturation or GLP1/GIP dual inhibitor causing a decrease in folate levels as exaggerated pharmacology of the therapeutic. A positive response in the micronucleus assay as a result of such confounding factors would not be indicative of reactivity with DNA.

The cause for concern with protein/peptide therapeutics is often the inclusion of chemical components, such as linkers, that could potentially be released in vivo and lead to genotoxicity. Exposure to the released small molecule component is expected to be low, like impurities in small molecule therapeutics. Since ICH M7 contains a framework for assessment and control of genotoxic impurities, risk assessment concepts described in this guideline are useful when addressing the genotoxic risk of released small molecule components in protein/peptide therapeutics. Although ICH M7 is not applicable to active components or in vivo metabolites, the concept of threshold of toxicological concern (TTC) for mutagenic carcinogens and guidance for in-silico mutagenicity assessment can be useful. For example, if the maximum in vivo exposure to the released small molecule component is below the TTC, then it should not be considered a cause for concern even if its mutagenic potential is unknown. However, if exposure is expected to be above the TTC, then an in-silico assessment of the chemical structure can identify structural alerts for mutagenicity. The presence of structural alerts can warrant testing of the small molecule component; however, selecting the correct chemical species for testing can be challenging as it is important to understand the penultimate form of the small molecule component that will be released upon treatment with the whole therapeutic. The genotoxicity evaluation of the whole therapeutic and products of cleavage could potentially be evaluated as part of the general toxicity studies at doses considered to be compatible with pharmacological activity and therapeutic indication.

Numerous factors weigh into determining the value of genotoxicity testing for a particular protein/peptide therapeutic. The most important consideration is whether the results of the genotoxicity tests provide added information for risk assessment or be uninterpretable.

Regulatory Background

Understanding the regulatory history of peptide products can be useful to understanding the complications and questions that have arisen regarding nonclinical study requirements for peptide therapeutics. The FDA Center for Drug Evaluation and Research (CDER) regulates synthetic, recombinant, and endogenous peptide products as drugs under the Food, Drug, and Cosmetic Act (FD&C). ⁶¹ The 2009 Biologics Price Competition and Innovation Act (BPCI Act) created an abbreviated licensure pathway in section 351(k) of the Public Health Service (PHS) Act for biological products shown to be biosimilar to, or interchangeable with, an FDA-licensed biological reference product. ⁶² This also amended the statutory definition of a “biological product” in the PHS Act to include a protein, except any chemically synthesized polypeptide (csPP). At the time of the November 2019 symposium, the statutory definitions for peptides, csPPs, proteins, and biological products discussed were based on the descriptions provided in the December 2018 Draft Guidance for Industry New and Revised Draft Q&As on Biosimilar Development and the BPCI Act Revision 2. ⁶³ Consistent with the scientific literature, FDA interprets the term “protein” in the statutory definition of a biological product to not include “peptides.” In the simplest context, a peptide is considered to be any polymer composed of less than or equal to 40 amino acids, regardless of how it is made. Thus, “peptides” and “chemically synthesized polypeptides” are not “biological products” and will be regulated as drugs under the FD&C Act in the United States, unless they otherwise meet the statutory definition of a “biological product” (eg, peptide vaccine). Under the 2018 draft Q&A Revision 2, the proposed statutory definition of a csPP was any α amino acid polymer that is both ¹ made entirely by chemical synthesis and ² greater than 40 amino acids, but less than 100 amino acids. Subsequently, a protein was defined as any α amino acid polymer with a specific defined sequence that is greater than 40 amino acids in size “(except any chemically synthesized polypeptide).” In early 2020, section 605 of the Further Consolidated Appropriations Act ⁶⁴ further amended the definition of a “biological product” under section 351(k) of the PHS Act by removing the parenthetical “(except any chemically synthesized polypeptide)” from the statutory definition of a protein that was added by the BPCI Act. The final ruling on the definition of a protein (Excerpt 1) focuses strictly on the number of amino acids in the polymer in order to “provide regulatory certainty and provide a bright-line interpretation of the term ‘protein,’” and went into effect on March 23, 2020. In other words, amino acid polymers less than or equal to 40 amino acids will be regulated as peptide drugs and any amino acid polymers greater than 40 amino acids will be regulated as protein biologics in the United States as of March 23, 2020. ⁶⁵

Excerpt 1: Final Statutory Definition of “Protein,” as stated in 21 CFR Part 600 paragraph (h) ⁶  (Food and Drug Administration, HHS, 2020)

⁶A protein is any α amino acid polymer with a specific, defined sequence that is greater than 40 amino acids in size. When 2 or more amino acid chains in an amino acid polymer are associated with each other in a manner that occurs in nature, the size of the amino acid polymer for purposes of this paragraph (h)⁶ will be based on the total number of amino acids in those chains, and will not be limited to the number of amino acids in a contiguous sequence.

International Council for Harmonisation M3(R2) was written for the development of pharmaceuticals and is intended to be a “guide for drug development.” Since peptides are regulated as drugs in the United States, ICH M3(R2) has often been applied in the past to guide peptide therapeutic nonclinical programs even though peptides are not mentioned specifically. Within the scope of ICH M3(R2), consultation of ICH guidances for specific product areas is recommended if they exist, such as biotechnology-derived products which are covered under ICH S6. The scope of ICH M3(R2) under Section 1.3 indicates that only the timing of nonclinical studies described in ICH M3(R2) applies to biotechnology-derived products, and that the most appropriate type of nonclinical studies for these products are described in the 1997 ICH S6 Guideline for Preclinical Safety Evaluation for Biotechnology-Derived Pharmaceuticals (ICH S6). Revisions to ICH S6 have since been made to include updates on species selection, study design, immunogenicity, reproductive and developmental toxicity, and carcinogenicity, which are described in the current widely used ICH S6(R1), published in June of 2011. ⁶⁶

The definition of biotechnology-derived products ICH S6(R1) has not always been interpreted the same way by scientists working with peptide therapeutics. According to the scope of ICH S6(R1) under Section 1.3, biotechnology-derived peptide products are those purified from plants, animals, or cell cultures. Paragraph 2 of Section 1.3 goes on to state that the principles outlined in ICH S6(R1) may also apply to chemically synthesized peptides. Thus, ICH S6(R1) has also been applied to peptide therapeutics in the past. However, potential safety concerns for synthesis-related processes are not discussed in ICH S6(R1). Thus, it is important to note that ICH S6(R1) will not always be considered by regulators to apply to all chemically synthesized peptide therapeutics.

It is notable that neither ICH M3(R2) nor ICH S6(R1) specifically apply to synthetic or conjugated peptides, with many peptide-based products essentially falling somewhere between both guidances. The nonclinical studies recommended by each guidance have several notable differences. Thus, disagreements as to the most appropriate ICH guidance to apply to synthesized and conjugated peptide products have arisen, particularly between industry and regulatory scientists. When considering the culmination of imperfectly fitted guidances, revisions, terminology refinement, and interpretation differences compounded by transitions between Review Centers and changes in regulatory processes, it becomes understandable that lack of clarity for nonclinical expectations for peptide therapeutics has led to confusion and frustration for both investigators and regulators. Historically, differences in interpretation and application of guidances ICH M3(R2) and ICH S6(R1) have resulted in additional information and study requests from Review Divisions, which have been associated with development delays and clinical holds for investigators.

Clarifying the Peptide Therapeutic Drug Space

Both synthetic and recombinant peptide products are regulated as drugs in the United States by CDER. Thus, strictly speaking, all peptide products may be subject to ICH M3(R2) principles. Peptides purified from tissues or cells, including recombinant peptides produced in cell culture, clearly meet the definition for biotechnology-derived drug products and are subject to recommendations outlined by ICH S6(R1). In contrast, synthetic peptides are not biotechnology-derived products and may be subject to ICH M3(R2) recommendations. Nevertheless, there are some scenarios where application of ICH S6(R1) principles is acceptable for synthetic peptides. For example, safety assessment recommendations outlined by ICH S6(R1) may be sufficient for synthetic peptides that contain only natural proteinogenic amino acids, particularly if there are no significant safety concerns related to the synthesis process, such as synthesis-related impurities. Proteins are generally regulated as biotechnology-derived biologics by the FDA Center for Biologics Evaluation and Research.

Proteins, biotechnology-derived peptides and synthetic peptide products can be conjugated; however, conjugation generally does not affect determination as to whether a product will be regulated as a drug or biologic. Conjugated moieties can include nonpeptide components, such as fatty acids, lipids, nucleic acids, or chemical entities, including small molecules, toxins, or metal nanoparticles, such as bionanomaterials. Conjugation to nonpeptide components is made possible using linkers, which can be small molecules that have inorganic and/or organic chemistries that are static, hydrolysable, or cleavable. Antibody-drug conjugates are an example of protein biologics conjugated to nonpeptide components. Safety concerns related to the conjugation process, the linker, or other non-amino acid components may be subject to additional studies. Since potential safety concerns for conjugated products or conjugation-related processes are not discussed in ICH S6(R1), application of ICH M3(R2) to conjugated products has been implemented in the past. For example, if a nonnatural or small molecule new molecular entity (NME) is part of a conjugated product, applications of ICH M3(R2) principles may be deemed appropriate.

International Council for Harmonisation M3(R2) Versus ICH S6(R1) Nonclinical Study Recommendations

The primary differences between ICH M3(R2) and ICH S6(R1) in nonclinical study recommendations pertain to the duration and type of general toxicology, metabolism, and genetic toxicity studies. Although both guidances recommend general toxicology studies in 2 species (1 rodent and 1 nonrodent), the maximum duration is 6 months for both species under ICH S6(R1), but 9 months for nonrodents under ICH M3(R2). Furthermore, studies longer than 1 month may only be necessary in one species under ICH S6(R1), depending on the 1-month toxicology findings and the species pharmacodynamic activity. Historically, both approaches have been applied to peptide products with acceptability of abbreviated nonclinical programs determined on a case-by-case basis. However, the factors contributing to acceptance of abbreviated studies in place of longer term nonrodent toxicity studies (eg, 6 months vs 9 months) for some peptide products remain unclear. Given the large costs of animal studies, greater understanding of regulatory expectations regarding the duration of nonrodent toxicity studies could be very beneficial to many drug developers in the field of peptide therapeutics. Furthermore, since drug development costs are often transferred to prescription costs for patients, which can affect patient accessibility to needed medications, further discussions regarding the scientific rationale and clarity concerning regulatory expectations are warranted.

Metabolism of amino acid polymers involves catabolism into individual amino acid components. For peptides composed of only naturally occurring amino acids, there are no safety concerns after catabolism. Thus, ICH S6(R1) does not recommend metabolism studies for biotechnology-derived products. However, for peptide products containing nonnatural amino acids (NNAAs), of which there are nearly 600 to date, or other nonnatural components, such as linkers or NMEs, the potential for safety concerns of catabolic products may not always be clear and ICH M3(R2) principles may be applied. In most cases, catabolism would likely result in nonnatural components bound to an adjacent amino acid. Nevertheless, there may be instances where it may be appropriate to assess free forms of nonnatural components or reactive ends, such as when hydrolysable or cleavable linkers are used in conjugated products. A risk-based rationale may also warrant evaluation of catabolic characterization and/or PK of nonpeptide components on a case-by-case basis. Examples include a positive in vivo genetic toxicity result, unusual or unexpected adverse toxicity compared to other members of a drug class, and scenarios when linkers or nonpeptide moieties are part of a prodrug leaving group. Although hazard analysis of some nonnatural components commonly used in peptide products may have been previously evaluated by regulators, without a right of reference to the GLP study results, new studies may be expected.

A significant difference between ICH M3(R2) and ICH S6(R1) nonclinical study recommendations is genetic toxicity safety assessments. According to ICH S6(R1), genetic toxicity studies are not warranted for biotechnology-derived products. However, ICH M3(R2) recommends that a full battery of genetic toxicity studies is completed according to ICH S2(R1) guidelines prior to phase 2 clinical trials. This battery includes an in vitro mutation assay in bacteria and an in vivo rodent assay with chromosomal damage assessment in either 2 tissues (option 2) or in 1 tissue along with an in vitro cytogenetic test for chromosomal damage in mammalian cells (option 1).

Discussion of Genotoxic Potential for Peptide Therapeutics

In an effort to explore the genotoxicity potential of peptide products from a conservative perspective, the following question was asked: is there a scientific basis for reassurance of nongenotoxicity without testing? Subsequently, this leads one to ask: are there peptides that test positive in genotoxicity assays? To address these questions, a broad literature search was conducted and evidence of 9 peptides and amino acid polymers with positive results in genotoxicity assays were identified (Table 3), which were categorized into 2 groups.

OPEN IN VIEWER

Table 3.

Literature Results of Peptides and Proteins With Positive Genotoxicity Assay Results.

		Name	CAS #	# Amino acids (aa)	Nonnatural aa	Cell permeable	Secondary structure	In vitro cytogenetic test in mammalian cells	In vivo rodent assay
Group I	1	Nodularin	118399-22-7	5	Yes	Yes	Cyclic	Positive	Positive
	2	Microcystin-LR	101043-37-2	7	Yes	Yes	Cyclic	Positive	Positive
	3	Bleomycin	9041-93-4	9	Yesa	Yes	Linear	Positive	Positive
	4	α-amanitin	23109-05-9	8	Yesa	Yes	Cyclic	Positive	Positive
	5	Melittin	37231-28-0	26	No	Yes	Helical dimer	Positive	Positive
Group II	6	Parathyroid hormone (PTH1-34)	83139-29 120945-80-4	34	No	Receptor-mediated internalization	Helical peptide	Positive in osteoblasts (3 weeks)	Positive (10 weeks)
	7	GLP-1/GIP dual agonist		40	Yes	Receptor-mediated internalization likely	ND	Negative	Positive (13 weeks)
	8	β amyloid αβ(1-42)	107761-42-2	42	No	Internalized	Helical/beta-sheet	Likely negative (no positive result found)	Positive
	9	Epoetin α	11096-26-7	165	No	Receptor-mediated internalization	4 α helices	Negative	Non-GLP: Positiveb (acute) GLP: Negative

Abbreviations: GLP-1, glucagon-like peptide-1; ND, not determined.

^a Nonnatural amino acid in humans, but natural in bacteria and some archaea.

^b Not great dose-dependence. May be within historical controls (% MNPCE/PCE below 0.5%). Ten-fold below induction of positive control (mitomycin C) induction.

Group I includes the 5 peptides nodularin, ⁶⁹ microcystin, ⁶⁹ bleomycin, ⁷⁰ α-amanitin, ⁷¹ and melittin, ⁷² which have various tertiary structures and all tested positive for both clastogenicity in mammalian cell in vitro cytogenetic tests and genotoxicity in rodent in vivo assays. In fact, it may be argued that the genotoxic potential was predicted by the in vitro clastogenicity assay, which was then confirmed by the in vivo rodent assay. It is notable that all 5 group I peptides are cell permeable, ^73,74 but only 4 of the 5 contain NNAAs. Interestingly, not only are each of the group I peptides clastogenic genotoxins, but they are also known natural toxins.

Group II includes 2 peptides, parathyroid hormone (PTH1-34) ⁷⁵ and a GLP-1/GIP dual agonist, ⁷⁶ and 2 proteins, β amyloid αβ ^{1 -41,77,78} and epoetin α, ⁷⁹ that were not associated with positivity in standard GLP clastogenicity in vitro assays but were associated with positive results in the in vivo rodent assay. However, for each group II amino acid polymer, the positive in vivo micronucleus assay results were secondary to downstream pharmacodynamic activity and were likely false positives. Thus, all 4 are decidedly not genotoxins. Nevertheless, it is important to acknowledge that they may still be nongenotoxic carcinogens, which is assessed in chronic carcinogenicity assays addressed in both ICH M3(R2) and S6(R1) guidances.

Comparison of Perspectives

When considering the potential for whether or not new peptide products could be identified as group I clastogenic genotoxins, philosophies coming from liberal and conservative perspectives were explored. From the liberal perspective, it may be argued that there is no concern for inappropriate clinical exposure because toxins are likely to be identified early in development and either eliminated early on or used according to the toxic characteristics, such as with antibody-drug combination products for cancer treatment. However, from the conservative perspective, it may be argued that there remains a need for some sort of reassurance of nongenotoxicity for new peptide products in absence of conducting a battery of genotoxicity studies.

To determine whether structural characteristics could be associated with predicting genotoxicity risk, types of structural modifications present in the group I clastogenic peptides were explored. Although cell penetration alone would not be expected to drive genotoxicity, it is noticeable that all of the group I peptides are cell penetrable. Another noticeable structural characteristic is that 4 of 5 contain NNAAs. At least 593 NNAAs have been reported in the literature ⁸⁰ and the lack of genotoxic potential of each one has not been verified. Other nonnatural modifications, such as nonnatural linkers or other nonnatural components, may also raise concerns of genotoxic risk for peptide products. It is also notable that naturally occurring chemical modifications, such as ferrous chelation ⁸¹ and advanced glycation end products, have been reported to be genotoxic in the literature. ⁸² Advanced glycation end products are produced by reaction of glucose with primary amino groups, which are likely to be increased under hyperglycemic physiological conditions. ⁸³

In order to provide reassurance of the lack of genotoxicity without additional testing for each NNAA, nonnatural component, and potential catabolic products containing the NNAA or nonnatural components, a public database of experimental GLP data could be created. Furthermore, such a database could potentially be used to construct an in silico assessment tool using a 3-D spectroscopic data-activity relationship (SDAR) modeling approach ⁸⁴ in order to predict the genotoxicity potential of NNAAs and peptide components that are otherwise too large for QSAR assessment. Such an in silico assessment could then be used alongside QSAR analyses of synthesis-related and conjugation-related impurities to determine a more complete prediction of genotoxic risk for peptide products. It is intriguing to imagine that perhaps the combination of in silico approaches, such as impurity QSAR analyses and a hypothetical NNAA SDAR analysis, could represent the first step in genotoxic assessment, such that positive alerts could be followed up with additional nonclinical studies, but the absence of alerts may not. However, taking on such an endeavor would likely be costly and the potential for success in reducing potential genotoxicity concerns and number of additional genotoxicity studies requested by regulators remains uncertain.

Summary of Current Regulatory Expectations for Peptide Products

When considering which regulatory guidance to apply to peptide products, it is important to remember that ICH S6 does not specifically apply to synthetic peptide products in that they are not considered to be biotechnology-derived products and regulators may apply ICH M3R2 principles. On the other hand, not all studies outlined by ICH M3R2 may be relevant for all synthetic peptide products. It is important to keep in mind that CDER Review Divisions may make nonclinical study requirement determinations on a case-by-case basis, particularly for inclusion of genotoxicity studies and acceptance of 6-month nonrodent general toxicology studies in place of a 9-month study. Furthermore, in silico evaluation of synthesis-related and conjugation-related impurities can be expected, including follow-up of positive alerts according to ICH M7 guidelines and identification and/or qualification of impurities according to ICH Q3A and/or Q3B thresholds. Additionally, in vitro and/or in vivo genetic toxicology studies may be expected. For conjugated products, additional studies that may be expected by CDER Review Divisions include assessment of reactive end potential, assessment of linker-related or nonnatural catabolites, and PK assessments of nonpeptide components. The need for submission of catabolic characterization may be based on risk-based rationale, which could include a positive in vivo genetic toxicity result, unusual or unexpected adverse toxicity compared to other members of a drug class, and/or the presence of linker of nonpeptide moieties that are part of a prodrug leaving group. Given that it is not always clear which regulatory guidance or which parts of each guidance should be applied in nonclinical development programs for all peptide products, and given that nonclinical requirements may be determined on a case-by-case basis, it is wise to contact the applicable CDER Review Division with questions regarding design of the nonclinical program design. Submission of rational supporting the proposed nonclinical studies may also be useful in determining the most appropriate path forward with the least degree of potential for future delays in regard to the nonclinical package.