Industry Survey and BLA Reviews

To understand the current ADME practices of TPs, an industry-wide survey was conducted with Innovation and Quality member companies between March and May 2021. A total of 26 major questions along with 13 subquestions were prepared by the IQ working group, broadly covering absorption, distribution, metabolism, excretion, and regulatory strategy of TPs. Only one collated response was accepted from each representative company. Briefly, the questionnaire started with requesting the information on the route of administration (intravenous, subcutaneous, and others) of TPs. The questions around absorption were focused on studies conducted to understand the mechanism of absorption and in vivo models used to investigate pharmacokinetics of TPs. Distribution questions gathered information on the studies conducted to understand tissue distribution of TPs in preclinical species/humans and the utility of the data in pharmacokinetics/pharmacodynamics (PK/PD) modeling. The metabolism/biotransformation questions asked if participating companies perform biotransformation studies and about the utility of biotransformation data in characterizing the disposition of TPs. Although excretion is usually not characterized for TPs, the questions were included to know more about the general practices across the industry. The last section assessed if the participating companies had any specific regulatory strategies around ADME characterization, regulatory interaction experience, and the requirement for additional guidance. The questions included were either multiple-choice answers, dichotomous (e.g., yes/no), or with open responses. The data collected from the survey were blinded and processed by the IQ secretariate prior to distribution to the working group. The survey responses were analyzed by the working group to derive inferences. An overall picture of ADME assessments of TPs is shown in Fig. 2. Considering the diversity of modalities and complexity of ADME processes, the survey didn’t attempt to focus on a specific modality within TPs, but rather considered TPs as a broader category. We also reviewed the regulatory submission packages of U.S. Food and Drug Administration– approved TPs between 2011-2020, focusing on ADME data (Table 1), to understand if such investigations have been used in submission packages (Supplemental Tables 2 and 3). Since ADME of ADCs was discussed and reviewed by another IQ working group earlier (Kraynov et al., 2016; Li et al., 2021), we excluded ADCs in the survey.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Percent response from the participating companies on conducting studies to understand mechanism of absorption, distribution, biotransformation, and elimination routes for therapeutic proteins.

Absorption

Absorption is characterized for new drugs with the intent to ensure the drug achieves the desired systemic circulation or tissue concentration, depending on the routes of administration. Typically, absorption is described as oral absorption from the gastrointestinal system or topical absorption through skin for NCEs. The absorption assessment for TPs can include several routes of administration, including subcutaneous, intramuscular, inhalation, or local tissue administration, such as intravitreal administration to the eye. Based on the survey responses, the most frequently used routes of administration of TPs are intravenous (68%), followed by subcutaneous (23%), and others (e.g., intravitreal and intratumoral, 9%) (Fig. 3A).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Results from the survey on most frequently used route of administration (A) and nonclinical species used for PK assessment (B). Although wild-type and transgenic mice are used for PK evaluation, nonhuman primate is the most frequently used preclinical species for PK assessment.

Early in preclinical development, the pharmacokinetics of TPs are evaluated in animals. TPs are often evaluated after a single dose intravenous administration and for many cases this is usually conducted in nonhuman primates (NHPs) (20 out of 22 responses, 91%), followed by wild-type mice (16 out of 22 responses, 73%), and FcRn transgenic mice (13 out of 22 responses, 59%) (Fig. 3B). As NHP physiology is similar to human (Glassman et al., 2015), single species allometry from NHP data has been used to successfully predict human pharmacokinetics (PK) of monoclonal antibodies and other TPs in the absence of target-mediated drug disposition (TMDD) (Avery et al., 2016). Transgenic mice, which express human FcRn, have also now been demonstrated to successfully predict linear (i.e., non-TMDD) human PK of monoclonal antibodies (Avery et al., 2016). In addition, other preclinical species, such as rats, murine tumor models, and severe combined immunodeficient beige mice, may be used to investigate absorption, guide molecule design, and/or to understand the mechanism of action.

With subcutaneous administration becoming more common for TPs, especially as it may enhance convenience for patients, a number of groups have tried to assess which preclinical species can be used to predict subcutaneous bioavailability. Richter et al. investigated sites of administration in minipigs to evaluate the optimal subcutaneous site that translates to humans (Richter et al., 2020). A monoclonal antibody was dosed at four subcutaneous regions (interscapular area, flank, behind ear, and inguinal area) and showed there was a difference in bioavailability from the four sites. The rate of absorption is key in selecting the subcutaneous dose site, and the observed rate varied across the different minipig dose sites. The bioavailability from the injection behind the ear and interscapular area best correlated with the human PK data. In addition, the thickness of the subcutaneous tissue in minipigs should be taken into consideration when conducting PK studies (Richter et al., 2020).

In the human setting, site of injection can, in some instances, affect the bioavailability and achieved concentrations in serum or plasma. Examples such as dulaglutide, a peptide-Fc (crystallizable fragment) fusion protein, exemplified no obvious differences in exposure after administration to the abdomen, upper arm, or thigh (Geiser et al., 2016). However, for some monoclonal antibodies, statistically significant differences have been observed when comparing injection to the thigh or abdomen (Bittner et al., 2018). Known examples include secukinumab (Bruin et al., 2020), ixekizumab (Talz label), bococizumab (Wang et al., 2017), and mepolizumab (Ortega et al., 2014). For golimumab, statistically nonsignificant differences were observed (Xu et al., 2010). There are cases of significant differences in terms of exposure based on the site of injection, exemplified by human growth hormone (hGH) (Beshyah et al., 1991). Investigators assessed the absorption after a single subcutaneous injection of hGH to the abdomen versus thigh in healthy young adults. The hGH was administered on two distinct occasions, first into the thigh and then into the abdomen on a subsequent visit. Absorption of hGH was significantly higher after subcutaneous injection to the abdomen compared with the thigh. It was unclear why this was the case, but it showed that absorption assessment should be done, especially with TPs that have narrow therapeutic windows.

Most companies surveyed do not investigate mechanisms of absorption (91%) for TPs. Among the two companies who did, one conducted bioavailability studies and evaluated in vivo structure/function relationships, whereas the other may use techniques such as photoacoustic imaging to study the absorption of peptides and monoclonal antibodies. Additionally, this company also investigated the correlation between bioavailability and biophysical properties, which include hydrophobicity, aggregation potential, nonspecific binding potential, charge patches, and binding to FcRn. These types of investigations impacting absorption are likely being conducted during the molecular design and selection stage, as well as lead optimization of a TP in discovery, and not necessarily as part of the development package for a regulatory submission.

Several factors can impact the absorption and PK profiles of TPs in animals and humans. These include bioavailability, anti-drug antibody (ADA) formation, and target meditated drug disposition.

1. Bioavailability: Determining the bioavailability from the subcutaneous dose site demonstrates that a portion of the TP is absorbed into the blood circulation via the lymphatics or directly into the blood stream and redistributed to the site of action. The bioavailability of monoclonal antibodies can range from 50% to 100% (Deng et al., 2012). Differences observed can be molecule, physiology, and route dependent (Collins et al., 2020). The absorption of antibodies is known to be FcRn dependent, and this may also impact subcutaneous bioavailability in humans (Glassman and Balthasar, 2019). There are challenges in developing preclinical models across species to accurately predict bioavailability in humans. For some companies, the decision to use the subcutaneous route early in program development means that true bioavailability (i.e., compared with intravenous administration) is not necessary. However, formulation or device modifications during later clinical development will require relative bioavailability assessments.

2. ADAs: An essential element of the ADME data package is to evaluate the potential for ADA formation to the TP (Chirmule et al., 2012). The administration of TPs to animals and humans may trigger an immunogenic response, which may lead to an impact on the exposure, bioavailability, and efficacy of the TP. The presence of ADAs may also interfere with the bioanalytical method used to measure the PK of the TP and thus requires careful consideration during method development (Thway et al., 2013; Ryman and Meibohm, 2017). Overall, although the incidence of ADA formation in animals does not always translate to humans, it is important to understand immunogenicity in both animals and humans as part of the TP regulatory submission package to correctly interpret toxicology study results and ensure human safety and efficacy, respectively.

3. TMDD: This is a phenomenon whereby the amount and turnover of the target can influence the absorption or pharmacokinetic profile of the TP significantly. This occurs often with TPs because the kinetics of binding, and especially the disassociation rates involved between target and TP, are often much slower than physiologic processes governing target behavior. When conducting PK studies with TPs, several dose levels of TP may be needed to fully evaluate the exposure-response relationship in the presence of TMDD, where it may also be important to assess if TMDD can be saturated with high doses. The extent of TMDD may vary between nonclinical species and humans (Glassman et al., 2015) and should be carefully considered. Mechanistic mathematical models can be developed to account for this phenomenon to support initial human dose projections.

In summary, for absorption, although the intravenous route is still the most frequent route of administration, per the IQ survey (Fig. 3A), other routes of administration are being explored and used for TPs. For more recently approved TPs and TPs under development, subcutaneous administration is becoming the desired route across many different therapeutic indications (Turner and Balu-Iyer, 2018; Collins et al., 2020). The absorption and full pharmacokinetic profile of the TP are the foundation for understanding and achieving adequate exposure for therapeutic benefits. Absorption assessment remains a key component to the overall ADME data package for delivery of TPs via nonintravenous routes.

Distribution

Tissue biodistribution studies contribute to the knowledge about the disposition of TPs, which is critical to understand efficacy (site of action/target tissue) and safety (off-target tissue) from lead selection to clinical development. Physicochemical properties, such as molecular size, isoelectric point, glycosylation, and target binding affinity, play a role in distribution of TPs and hence these factors should be carefully considered for selection of TP for development (Tabrizi et al., 2010).

As shown in Table 1, only 20 out of the 91 BLAs (22%) submitted from 2011 to 2020 contained preclinical distribution data as measured in mice, rats, guinea pigs, or monkeys. Interestingly, the majority of the companies in the survey (77%, 17 out of 22, Fig. 2) indicated that they analyze biodistribution to target tissues, for instance by measuring target tissue drug concentrations in preclinical species (Fig. 4). The survey also revealed that biodistribution of TPs to target tissue in human subjects is analyzed by 50% of the companies (Fig. 4). The discrepancy between distribution data presented in BLAs (22%) and the survey (77%) might suggest that many preclinical distribution studies are exploratory in nature and are not being submitted in the BLAs. However, the survey results also partly indicate a trend that companies are increasing their efforts to understand the distribution processes of TPs.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

Percent response on conducting preclinical and clinical biodistribution studies for TPs.

Many analytical techniques are available to study biodistribution of TPs. Survey results indicated that the tools used to evaluate preclinical and human biodistribution to target tissues are similar (Table 2). Examples of such techniques and methods include ligand binding assays (LBAs), liquid chromatography-tandem mass spectrometry (LC-MS/MS), and imaging technologies utilizing fluorescence or radiolabeling approaches. For instance, quantitative whole-body autoradiography (QWBA) methods with ³H-labeled proteins can give valuable quantitative insights into the whole-body distribution in rodents or NHPs. Other radiotracers, such as ¹¹¹In, ¹²⁵I, and ⁸⁹Zr, are frequently employed, particularly for in-life imaging (Conner et al., 2020). Fc-containing TPs are most commonly labeled with ¹²⁵I, which might be mainly due to early development of labeling chemistry involving the attachment of iodine atoms to tyrosine residues in proteins (Cohen et al., 1956). Recent advancement of labeling chemistry and imaging technologies have provided more opportunities to characterize biodistribution of therapeutic drugs. Although traditional QWBA methods require an animal per time point, positron emission tomography (PET) or single-photon emission computed tomography techniques enable image collection from live subjects at multiple time points, finding applicability in not only nonclinical species but also in humans (Gomes et al., 2011). In addition, PET combined with computed tomography gives robust contrast and spatial resolution.

Generally, the need for sampling of target tissue(s) in a clinical setting is only judged on a case-by-case basis depending on accessibility, volumes of biopsy, etc. Understanding the TP concentrations in target tissues in humans is critical to predict optimal dose and dosing regimen design. Examples of sampling techniques include open flow microperfusion or simply withdrawal of blister fluid in case of TPs developed for psoriasis or withdrawal of synovial fluid in patients with rheumatoid arthritis (Choy and Panayi, 2001; Dragatin et al., 2016). Furthermore, cerebrospinal fluid (CSF) sampling can be of interest for evaluation of central nervous system distribution. As an example, CSF-to-serum ratio of less than 0.5% for five independent antibodies in rats and NHPs were measured and agree well with CSF-to-total serum immunoglobulin G (IgG) ratios of 0.1%–0.2% in healthy humans as reported by several groups (Naessens et al., 2018; Wang et al., 2018; Brys et al., 2019).

In the case of conducting biodistribution studies with TPs, the contribution of TMDD should be clearly distinguished from nonspecific distribution mechanisms and quantified accordingly. Some recommendations can be made in this respect:

1. Nonspecific biodistribution can be studied by in vitro, for example, assessing interactions with human embryonic kidney (HEK) cells (Datta-Mannan et al., 2015), as well as in animals that do not express the target of the TP or the TP is not crossreactive with the target expressed in that species.

2. TMDD can be very well evaluated with single and multiple dose PK in nonclinical species (if TP has similar binding properties). In nonclinical species, dose–exposure-response relationships to determine regions of nonlinearity can be correlated with biodistribution results (i.e., receptor occupancy by flow cytometry).

3. A good understanding of target expression in tissues together with robust multicompartmental modeling might be pivotal for a robust understanding of PK/PD relationships and translation to human predictions. The survey results indicated most companies (82%) use distribution results in translational PK/PD modeling for efficacy and safety predictions. This nicely demonstrates the increasing role of modeling and simulation strategies that use TP biodistribution data. More details on the utility of biodistribution data in PK/PD modeling is presented in the companion manuscript of this working group (Ball et al., 2022).

There are numerous examples of preclinical and clinical investigation of biodistribution of TPs. Some enable human dose setting, whereas others eventually lead to United States product insert and/or summary of product characteristics labeling. Three examples are given below.

1. Biodistribution of various recombinant antibody platforms was assessed with respect to tumor accumulation and kinetics in xenograft mice models (Schneider et al., 2009). Four different antibody-based therapeutic platforms, monoclonal antibody (mAb), diabody, single-chain fragment variable (scFv), and novel mini-antibody were labeled with PET tracers and evaluated in prostate cancer cell xenograft mouse studies. The authors conclude that larger antibody frameworks (IgG and mini-antibody) gave higher tumor uptake levels than smaller frameworks [diabody and single-chain fragment variable (scFv)]. These larger frameworks may be more suitable for radioimmunotherapy applications, whereas the smaller ones may be suitable for radiodiagnostic applications.

2. Active transplacental transport of IgG from mother to infant is mediated by FcRn in the placenta, mainly during the second and third trimesters of pregnancy (Pentsuk and van der Laan, 2009). As a result, various alternative formats have been assessed to avoid significant exposure of the drug to the developing fetus. One example is the PEGylated, Fc-free antitumor necrosis factor fragment antigen binding (Fab) called certolizumab pegol, approved for the treatment of various autoimmune diseases. Due to the lack of an IgG Fc region, certolizumab pegol does not bind FcRn and is consequently not expected to undergo transfer across the placenta. This was confirmed in a clinical study in which 16 pregnant women were assessed at term for transfer to their babies. None to minimal exposure of certolizumab pegol in babies was observed, suggesting lack of placental transfer during pregnancy (Mariette et al., 2018). A more recent example explored temporal and quantitative placental transfer of humanized IgG2Δa format in cynomolgus monkeys (Catlin et al., 2020). Such efforts with more novel formats would also improve the assessment of teratogenic potential of TPs and enhance our ability to extrapolate relevance of these data to humans.

3. Based on recent review of 19 selected mAb therapeutics, distribution to breast milk was typically at low concentrations and tended to peak within 48 hours, although maximum levels could occur up to 14 days from infusion (LaHue et al., 2020). The relative infant dose, a metric comparing the infant with maternal drug dose (<10% is generally considered safe), was evaluated for three mAbs and was <5.3%. Importantly, a total of 368 infants were followed for ≥6 months after exposure to breast milk of mothers treated with mAbs; none experienced reported developmental delay or serious infections. In conclusion, the data confirmed low mAb drug transfer to breast milk, but further studies are needed (keeping in mind that data were largely derived from case reports), especially with regard to longer-term effects on infant immunity and childhood development.

In summary, a robust knowledge of biodistribution properties of TPs is beneficial for enabling reliable predictions of in vivo PK and target tissue distribution. Characterization of the tissue exposure to TPs can be challenging, but efforts in this respect at the early stages of development can be helpful to reveal mechanisms governing TP pharmacology and/or toxicity. Continued research in the field of TP biodistribution is strongly encouraged. Specific areas of focus include deeper understanding of physicochemical properties and their relationship to distribution of TPs, target expression profiles in preclinical species and humans, and continued improvements in biosampling in the clinical setting.

Metabolism

There is increasing interest in gaining insights into the metabolism of TPs. Depending on the intended purpose, specific studies on biotransformation, a general terminology used in this paper to describe metabolism, catabolism, and other modifications and degradations, may be considered at various stages of drug discovery and development.

Biotransformation is generally the process of conversion of xenobiotics, i.e., nonendogenous compounds, to products that are suitable for elimination from the body. Understanding of the biotransformation of NCEs is essential, as it could lead to the formation of pharmacologically or toxicologically active metabolites or lead to drug-drug interaction potential, which often triggers efficacy and/or safety concerns. Therefore, biotransformation of small molecules is continuously monitored from preclinical studies through clinical development.

Biotransformation of TPs includes hydrolysis of amide bonds in the protein backbone, metabolism of amino acid side chains, and other post-translational modifications (PTMs) such as glycation. Biotransformation may alter the PK, ADME, and efficacy, but rarely increases the toxicity of the TPs. Hydrolysis of amide bonds by proteases in vivo is a major clearance pathway of TPs. Biotransformation products (BPs) from proteolysis of TPs may lose their pharmacological activity since the target binding site may be cleaved or geometrically changed. In cases of hydrolysis occurring on one or more amino acids at the C- or N-terminus of a TP, BPs may still retain partial or full pharmacokinetic properties and/or pharmacological activities. Proteolytic reactions are expected to occur in lysosomes and endosomes after endocytosis, in the proximal tubule postglomerular filtration, and in the circulatory system or extracellular fluid by proteases. For example, therapeutic mAbs are mainly catabolized in lysosomes inside cells by resident proteases. Some TPs, particularly the recombinant endogenous peptide hormones such as glucagon-like peptide 1 and gastric inhibitory polypeptide, are very sensitive to the proteases and peptidases in the circulatory system (Mentlein et al., 1993; Werle and Bernkop-Schnurch, 2006). Their plasma half-life can be extremely short in vivo due to both proteolysis in the blood and glomerular filtration. To mitigate fast clearance by glomerular filtration, recombinant endogenous proteins, peptide hormones, or their analogs can be fused or conjugated to large proteins such as mAb (Rangwala et al., 2019), Fc domains (Glaesner et al., 2010; Hall et al., 2010; Hecht et al., 2012; Lee and Lee, 2017), human albumin (Baggio et al., 2004; Matthews et al., 2008), and polyethylene glycol (PEG) (Jevsevar et al., 2010) to increase the molecular sizes. However, the protein or peptide moiety in a fusion or conjugate TP can still be susceptible to cleavages induced by proteases and peptidases. To minimize potential proteolysis of fusion or conjugate TPs, different strategies can be used (Walles et al., 2022). Integrated analytical strategies, often involving LBA and hybrid immunoaffinity liquid chromatography-mass spectrometry (LC-MS) platforms, have been adopted gradually over time to gain better understanding of biotransformation of TPs (Kaur et al., 2020).

Biotransformation of amino acid side chains, such as oxidation, deamidation, and isomerization, as well as other PTMs like glycation, may impact critical properties of the molecule. For example, oxidation of key methionine residues on human IgG weakens the FcRn or Fc-γ receptor binding affinity, which can negatively affect the PK and/or effector functions (Gao et al., 2015). Similarly, modification of the amino acid side chain in the complementarity-determining region of a mAb may ablate the antigen binding affinity and result in decreased pharmacokinetics and activities (Bults et al., 2016; Dashivets et al., 2016; Menke-van der Houven van Oordt et al., 2016). Glycation of serum albumin is a known modification that occurs in circulation, especially in disease states such as diabetes. Recent studies have shown that the interaction of glycated albumin, particularly at position K525, leads to a reduction in affinity for FcRn (Leblanc et al., 2019). Molecules involving albumin fusions may require a deeper characterization to understand the PK impact in certain clinical populations. Some evidence suggested possible differences in biotransformation across certain patients (Liu et al., 2018).

Traditionally, biotransformation studies for TPs are not required for regulatory filing. In general, unlike NCEs, safety risk due to biotransformation of the TP is believed to be minimal. The risk for inhibition or induction of metabolizing enzymes by the TPs is considered relatively low and thus not monitored in general, unless the mechanism of action is known to impact cytochrome P450 (CYP) expression, e.g., Interleukin (IL)-6 (Schmitt et al., 2011). Among the regulatory submission packages analyzed, for the majority of the approved biotherapeutics, no biotransformation data were included.

In line with the observation that biotransformation is in general not conducted in regulatory filings of TPs, specific guidance from the health authorities on this subject has been limited. In this regard, the ICH Topic S6 guideline and the European Medicines Agency guideline on the clinical investigation of the pharmacokinetics of TPs are often cited. Under the ICH Topic S6 guideline issued in 1998, there was no expectation in conducting conventional preclinical metabolism studies. Nevertheless, stability testing in the matrix used for bioanalysis (e.g., plasma or serum) is recommended to understand the behavior of the biopharmaceutical and the possible influence of binding proteins to the pharmacodynamic effect. The more recent European Medicines Agency clinical guidance for PKs of TPs suggests a case-by-case consideration for in vitro metabolism studies (European Medicines Agency, 2007), recognizing that it could be relevant for certain protein drugs, particularly conjugated and altered protein sequence drugs, where the cleavage products could still contribute to the overall pharmacological activity.

Although biotransformation of the TP is typically not required for regulatory submission, interestingly, the survey results indicate that there is an almost even split among the responding companies—54% perform metabolism/biotransformation studies, whereas 46% do not (Fig. 2). It is worth noting that these numbers may largely depend on the drug modalities in the pipeline of each company. Typically, biotransformation is not performed for monoclonal antibodies, whereas specific studies may be conducted for fusion proteins, protein/peptide conjugates, and proteins with in vivo modification liability or unexpected PK behaviors. It is believed that the biotransformation assessments are most likely conducted in the discovery and early development phases to help optimize and select the TP. A detailed overview of biotransformation studies to address the liabilities of TPs can be found in the companion manuscript (Walles et al., 2022).

The majority of the survey companies (82%) have not established in vitro-in vivo correlation with regards to metabolism of TPs. However, certain reaction-driven PTMs in buffer (deamidation or isomerization) may be used to predict in vivo liability and serum incubation to predict in vivo stability. Most companies (77%) do not perform cross-species metabolite profiling, which is typically done for small molecules. Analytical platforms used for biotransformation evaluation so far are primarily mass spectrometry and LBA. The majority of the companies (86%) do not use radiolabeled TPs to elucidate biotransformation pathways. Quantification of active metabolites/catabolites is not usually pursued. In general, there has not been a well-established strategy in studying biotransformation yet, with studies being conducted on a fit-for-purpose basis. Depending on the individual company and the drug modalities in its pipeline, the biotransformation data could be used for gaining insights into the areas of clearance mechanism, PK/PD disconnects, safety, regulatory, and others, as depicted in Fig. 5.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

Percent responses on utility of biotransformation data of TPs. Among the companies participated, respondents use biotransformation data for explaining PK/PD disconnect, clearance mechanism, and safety aspects.

Excretion

For smaller MW proteins (∼<60 kDa), kidney is the primary elimination organ with high levels of renal filtration and subsequent degradation after proximal tubule reabsorption (Maack et al., 1979). For instance, fibroblast growth factor 21 is a small protein with a MW of ∼19 kDa used in the treatment of type 2 diabetes. It has a very short half-life in rodents and NHPs, which is partially attributed to the renal excretion (Hager et al., 2013). IL-2 is a cytokine with 15.6 kDa in MW, which plays an important role in the immune system. The terminal half-life of recombinant IL-2 was 85 minutes in humans with the renal filtration as the major clearance pathway (Konrad et al., 1990). Interferon-α is a group of cytokines with MWs ranging from 17.5 to 23 kDa, which are also cleared mainly by renal filtration (Wills, 1990).

Several strategies to improve the pharmacokinetic properties of lower molecular weight TPs, specifically to maximize the half-life in the body, have been described in the Metabolism section. PEG polymers are perceived to be nontoxic, nonimmunogenic, nonantigenic, highly soluble in water, and have been approved by agencies for use in humans when chemically conjugated to low molecular weight TPs, e.g. certolizumab pegol (Cimzia) or pegvisomant (Somavert). However, elimination of PEGylated proteins is not yet well understood. Data for certolizumab pegol, a Fab conjugated with a 40kDa PEG moiety, has shown that its PEG moiety is almost entirely excreted unchanged in the urine once cleaved from the Fab (Nesbitt et al., 2007).

The survey sought responses about studies conducted to understand excretion of TPs. The majority of companies do not determine elimination routes of TPs either in preclinical species (82%) (Fig. 2) or in humans (91%). These studies are generally performed for fusion and PEGylated proteins (e.g., cytokines, growth factors), nanobodies, peptides, and ADCs. Reviewed data from BLA submissions suggest that elimination studies, when conducted, use rodents or NHPs. For higher MW TPs like antibodies, the renal excretion plays a minimum role in their clearance as their sizes are too large to be filtered by glomerulus. However, higher MW proteins can be excreted by the kidney under certain disease conditions, such as nephrotic syndrome and glomerular proteinuria (Waldmann et al., 1972). Similar to renal excretion, biliary excretion plays a very minor role in the elimination of large MW proteins, such as IgGs (Thomas and Balthasar, 2019). Instead, both active (receptor-mediated endocytosis) and passive (fluid-phase endocytosis) mechanisms of protein transport from the vascular endothelium to the underlying tissue and subsequent intracellular degradation are typically major elimination pathways for large proteins (Wang et al., 2008).

Regulatory Strategy

With regard to regulatory strategies for TPs, most companies (86%) do not have an internal regulatory strategy specifically for ADME of TPs (Fig. 6A). Regulatory interactions from most companies (73%) did not lead to any additional ADME-focused studies for TPs (Fig. 6B). Interactions that led to ADME studies for TPs were primarily related to drug-drug interactions (three companies). Based on the survey results, the expectation on the need for further regulatory guidance on ADME of TPs, on a scale of 1 (none needed) to 10 (yes, absolutely need more), the median value of the response was 4.5 and the mean was 4.2 (Fig. 6C).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 6.

Results from the survey on regulatory strategy within the company (A), regulatory interaction experience (B), and need for additional guidance (C).