The Effect of Gastrointestinal Graft-Versus-Host Disease and Diarrhea on the Pharmacokinetic Profile of Sotrovimab in Hematopoietic Stem Cell Transplant Recipients

Abstract

Background

Monoclonal antibodies (mAbs) are utilized broadly to treat cancer and infectious diseases, and mAb exposure (serum concentration over time) is one predictor of overall treatment efficacy. Herein, we present findings from a clinical trial evaluating the pharmacokinetics of the long-acting mAb sotrovimab targeting severe acute respiratory syndrome coronavirus 2 in hematopoietic cell transplant (HCT) recipients.

Methods

All participants received an intravenous infusion of sotrovimab within 1 week prior to initiating the pretransplant preparative regimen. The serum concentration of sotrovimab was measured longitudinally for up to 24 weeks posttransplant.

Results

Compared to non-HCT participants, we found that mAb clearance was 10% and 26% higher in autologous and allogeneic HCT recipients, respectively. Overall sotrovimab exposure was approximately 15% lower in HCT recipients compared to non-HCT recipients. Exposure was significantly reduced in HCT recipients who developed diarrhea and lower gastrointestinal graft-versus-host disease (GVHD) posttransplant.

Conclusions

These data show that sotrovimab exposure may be reduced in HCT recipients, possibly related to increased gastrointestinal clearance in patients with GVHD. This phenomenon has implications for dose selection and duration of efficacy with sotrovimab and potentially other mAbs in this vulnerable patient population. Thus, mAb dose regimens developed in non-HCT populations may have to be optimized when applied to HCT populations.

The elimination half-life of immunoglobulin G (IgG) is commonly observed to be approximately 15–20 days [1, 2]. The impact of this property is illustrated by the need for monthly dosing of palivizumab, a commercially available monoclonal antibody (mAb) used to prevent respiratory syncytial virus (RSV) in high-risk infants, or of intravenous immunoglobulin in patients with hypogammaglobulinemia. The half-life of IgG is largely determined by 2 clearance mechanisms, with approximately 85% degraded intracellularly in lysosomes and 10%–15% excreted in stool [3, 4]. IgG can be rescued from lysosomal degradation and recycled back into circulation by binding to the neonatal Fc receptor (FcRn) at an acidic pH in endosomes and dissociating from FcRn at a physiologic pH in blood or tissue [5].

However, there are emerging data that mAb pharmacokinetics (PK) can vary significantly in specific target patient populations [6–10]. Hematopoietic cell transplant (HCT) recipients, because of their increased risk of complications from infectious diseases, are one such population where PK may vary and where mAbs are used for prophylaxis and/or treatment. In this population, vaccination is generally delayed, and vaccine responses are diminished [11, 12]. Immunoprophylaxis with mAbs provides an alternative or supplement to vaccination to protect against infectious diseases. Yet, the PK of mAbs is understudied in this population. In 2001, a small retrospective analysis of 2 separate clinical studies found that allogeneic HCT recipients were exposed to less than half the levels of the anti-RSV mAb palivizumab after intravenous (IV) administration, compared to autologous HCT recipients [13]. This early observation led us to hypothesize that mAbs are cleared more rapidly after HCT, particularly after HCT from allogeneic donors where recipients are at increased risk of gastrointestinal (GI) protein loss due to graft-versus-host disease (GVHD).

We conducted a phase 1 clinical trial to measure mAb exposure in HCT recipients (the COVIDMAB study, NCT05135650). In this study, patients >18 years of age undergoing any HCT received sotrovimab intravenously as preexposure prophylaxis within 1 week prior to the start of the pretransplant preparative regimen. Sotrovimab is a fully human neutralizing anti–severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) mAb that contains a 2-amino-acid Fc-modification (“LS”) that is designed to increase half-life by boosting binding affinity to FcRn [14–16]. At the time of study initiation, sotrovimab (IV 500 mg) had been granted an Emergency Use Authorization (EUA) by the United States Food and Drug Administration (FDA) for the treatment of mild-to-moderate coronavirus disease 2019 (COVID-19) in adult and pediatric patients (≥12 years of age and ≥40 kg) who tested positive for SARS-CoV-2 and were at a high risk of progression to severe COVID-19. The EUA was deauthorized in April 2022 given reduced in vitro neutralization of sotrovimab against the BA.2 variant with unclear impact on efficacy. Serum was collected longitudinally for 6 months. Participants also filled out online symptom surveys and electronic medical records were reviewed. From this study, we analyzed the effects of covariates, including diarrhea and GVHD severity, on sotrovimab clearance and exposure posttransplant.

MATERIALS AND METHODS

Study Design

This study complies with all relevant ethical regulations and was reviewed and approved by the Fred Hutchinson Cancer Center Institutional Review Board. The COVIDMAB study (NCT05135650) was designed as a prospective, open-label, single-arm study that enrolled patients prior to HCT at the Fred Hutchinson Cancer Center. Participants received one 500-mg dose IV of sotrovimab within 1–7 days prior to the start of conditioning. The inclusion criteria were (1) age at least 18 years; (2) undergoing HCT (any donor or stem cell source, including autologous or cord blood transplant); (3) and ability to provide informed consent. The exclusion criteria were (1) positive polymerase chain reaction (PCR) result for SARS-CoV-2 within 4 weeks of scheduled conditioning; (2) signs or symptoms of uncontrolled active infection; (3) pregnancy or breastfeeding; (4) previous known allergies to any component of sotrovimab; (5) previous anaphylaxis or severe hypersensitivity reaction, including angioedema, to a mAb; (6) participant of other clinical studies that preclude the use of other investigational compounds; and (7) participants who, in the judgment of the investigator, would be unlikely or unable to comply with the requirements of the protocol or unlikely to survive to the end of study. Participants completed a weekly symptom questionnaire prior to and weekly after mAb administration for a total of 24 weeks. Immediately prior to mAb administration and at the end of infusion, weekly up to week 4, monthly up to week 12, and at week 24 after mAb administration, blood was drawn by venipuncture (10–20 mL per blood draw). Nasal swabs were collected for SARS-CoV-2 PCR as needed based on respiratory symptoms (answering yes to 1 or more of upper or lower respiratory symptoms) within 48 hours of symptom onset and up to week 36. If positive for SARS-CoV-2, weekly nasal swabs were obtained longitudinally for SARS-CoV-2 PCR until negative or until week 36.

GVHD and Diarrhea Severity Scale

GVHD assessments were performed and documented in the electronic medical record weekly for allogeneic HCT recipients as part of their standard transplant care. An overall grade for GVHD was assigned based on consensus criteria [17]. Lower GI GVHD was graded by the clinical transplant team based on center-specific criteria as 0 for no diarrhea; 1 for 500–1000 mL/day or 3–4 episodes/day attributable to GVHD; 2 for 1001–1500 mL/day or 5–7 episodes/day attributable to GVHD; 3 for >1501–2000 mL/day or 8–10 episodes/day attributable to GVHD; and 4 for melena or frank blood in stool or severe abdominal pain attributable to GVHD. Diarrhea severity was self-graded by the participant on weekly symptom surveys as 0 for none, 1 for mild, 2 for moderate, and 3 for severe.

Bioanalytical Pharmacokinetic Assay

Blood was allowed to clot for 30 minutes at room temperature before centrifuging at 1500g for 15 minutes. Serum was aliquoted and frozen at −80°C for storage until samples could be analyzed in batches. A validated electrochemiluminescence (ECL) Immunoassay was performed at Resolian (Malvern, Pennsylvania) to quantify sotrovimab in human serum samples. AbD34205 mu IgG2a (an anti-LS mAb) was coated onto MSD High Bind plates (ECL capable). Sotrovimab in standards, controls, and samples were captured onto the coated plate. After thorough washing of the wells to remove unbound mAb, Ruthenylated AbD42688 rFab was added to the wells. Following incubation with detection reagent, plates were washed, followed by the addition of freshly prepared 1% formaldehyde for fixation and MSD read buffer. The assay plates were read using an MSD ECL plate reader.

Anti-drug Antibody Assay

A validated ECL-based method was used for the detection of anti-sotrovimab antibodies in serum (Resolian). This assay was a bridging assay performed on an MSD platform. Test samples, positive controls, and negative controls were acidified, neutralized, and incubated with Bt-sotrovimab and sTag-sotrovimab conjugates. The samples were then plated in duplicate wells of a streptavidin-coated MSD plate. Any anti-drug antibody that bound to both reagents formed a complex that was captured by the streptavidin-coated MSD plate and generated an electrochemiluminescent signal. The signal intensity was proportional to the amount of anti-drug antibody present in the positive control and test samples.

SARS-CoV-2 Real-Time Reverse-Transcription PCR Assay

Nasal swabs were self-collected using a mid-nasal foam swab. Each swab was mixed with 2 mL of 1× phosphate-buffered saline (PBS) and vortexed for 30 seconds. After a 10-minute rest to reduce aerosols, the liquid was then aliquoted and frozen at −80°C for storage until samples could be analyzed by PCR in batches. The FDA-authorized Panther Fusion SARS-CoV-2 real-time reverse-transcription PCR assay was used to detect SARS-CoV-2 in nasal swabs [18]. The limit of detection of this assay is 0.01 50% tissue culture infectious dose/mL.

Pharmacokinetics Analysis

Nonlinear mixed-effects modeling (NONMEM, version 7.3.0) was used for PK model development. Perl and Perl-speaks-NONMEM (version 4.6.0) were used for model evaluation and automatic covariate model building. First-order conditional estimation with interaction (FOCE + I) [19], with proportional and additive proportional models for residual variability, with standard development and evaluation process were applied [20]. R (version 3.2.5) was used for data visualization and analyses. The population PK model of sotrovimab in non-HCT patients was published previously [21]. Sotrovimab PK was best described by a 2-compartment model with linear elimination, and this model was used as the base model for this analysis. Covariate relationships were explored using a stepwise covariate modeling approach [22]. Various covariate relationships were tested including linear and nonlinear (“hockey-stick,” exponential, and power) relationships. Forward inclusion and backward exclusion steps were conducted with a cutoff of change in objective function value (OFV) of at least 3.84 (P ≤ .05, for 1 degree of freedom) and 6.83 (P ≤ .01, for 1 degree of freedom), respectively. Additionally, scientific plausibility and clinical relevance of the covariates were also taken into account. The covariates evaluated in the analysis were demography (age, race, ethnicity, body weight, height, body mass index, sex), diarrhea status as a categorial variable (yes or no), GVHD status as a categorical covariate (yes or no), and lower GI GVHD severity and diarrhea severity (defined in the previous section “GVHD and diarrhea severity scale”) as continuous covariates. Both diarrhea and lower GI GVHD status and severity were time-varying covariates whereas the rest of the covariates were based on baseline values. Statistically significant covariate effects were included in the final model if the relationship was also considered clinically plausible.

Statistical Analyses

Optimal PK models were determined with statistical significance in NONMEM using the FOCE + I. The difference in the minimum OFV between hierarchical models was assumed to be χ² distributed with degrees of freedom equal to the difference in the number of parameters between models. The comparison of full versus reduced models using the log-likelihood criterion was considered as 1 major factor together with goodness-of-fit plots and visual predictive checks in the model selection process.

RESULTS

Study Population and Clinical Characteristics

Participants were enrolled in the COVIDMAB study between January to April 2022 and followed up to October 2022. Enrollment was halted in May 2022 due to the emergence of SARS-CoV-2 variants to which sotrovimab demonstrated reduced in vitro neutralization potency. Of the overall 20 participants enrolled, 15 (75%) received an allogeneic HCT and 5 (25%) received an autologous HCT (Table 1). All transplants were performed for treatment of an underlying malignancy. The most common diagnosis among allogeneic HCT recipients was acute leukemia (5/15 [33%]) and myelodysplastic syndrome (5/15 [33%]), whereas the most common diagnosis among autologous HCT recipients was multiple myeloma (3/5 [60%]). Conditioning chemotherapy included nonmyeloablative and reduced-intensity and myeloablative regimens (Table 1).

After transplant, 55% of participants developed grades 2–4 acute GVHD, with half of participants having lower GI involvement (Supplementary Table 1 and Figure 1). Furthermore, 60% of participants self-reported diarrhea posttransplant (Supplementary Table 1 and Figure 1). During the study, 3 participants developed respiratory symptoms and tested positive for SARS-CoV-2 by PCR of nasal swab specimens at day 95, 101, and 128 posttransplant (Supplementary Table 1). All 3 infections occurred in mid-to-late June, coinciding with a rise in incidence of the SARS-CoV-2 BA.5 variant in the community. Symptoms were mild and limited to the upper respiratory tract and virus was no longer detectable within 7 days for 2 participants and 25 days for the third participant. None of the participants required hospitalization for COVID-19. One participant voluntarily withdrew early from the study after 2 weeks, citing a difficult transplant course and not having enough time to complete study tasks. One participant died at week 26 posttransplant due to complications from their transplant, including bone marrow failure, GVHD, and GI hemorrhage. None of the participants developed anti-drug antibodies (Supplementary Table 1).

Population Pharmacokinetics Model

Serum concentrations of sotrovimab were measured longitudinally up to week 24 posttransplant using a PK assay designed to specifically detect sotrovimab based on its “LS” Fc modification (Figure 1). A 2-compartment population PK model with central and peripheral compartments best fit the data (Supplementary Figure 1). Body weight at baseline was a significant covariate on sotrovimab central volume of distribution and clearance (Table 2). Given the constraints of the sample size, the number of observations for each severity level of diarrhea or lower GI GVHD was sparse and constituted <10% of total observations (Supplementary Table 1). Consequently, for the primary analysis, all severity grades for each condition were collapsed into 1 group to represent patients with any severity level of diarrhea or lower GI GVHD (Supplementary Table 1). Diarrhea, overall GVHD, and lower GI GVHD were identified as significant covariates on antibody clearance. Overall GVHD and lower GI GVHD were significantly correlated (P < .01). To address multicollinearity, the OFVs for models incorporating overall GVHD or lower GI GVHD were compared, and there was no significant difference between the 2 models [23]. Given that lower GI GVHD had greater clinical relevance and scientific plausibility as a covariate for antibody clearance compared to overall GVHD (which includes skin GVHD), lower GI GVHD was selected for use in the final model.

The relationship between the sotrovimab PK parameters with diarrhea and lower GI GVHD was next examined in both models. Increased sotrovimab clearance (Figure 2A) and decreased half-life (Figure 2B) were significantly associated with the presence of diarrhea and lower GI GVHD. Although there were relatively fewer observations for severe diarrhea and severe lower GI GVHD (Supplementary Table 1), there was a trend toward increased sotrovimab clearance and decreased sotrovimab half-life with worse diarrhea and greater lower GI GVHD severity (Supplementary Figure 2).

Sotrovimab Clearance After HCT

To compare sotrovimab clearance between HCT recipients and nontransplant participants, PK data were compared to estimates from non-HCT volunteers studied previously who received sotrovimab in separate clinical trials [21]. Clearance in non-HCT patients was 0.097 L/day based on previous population PK analysis. In contrast, the clearance in autologous and allogeneic HCT recipients was 0.104 and 0.122 L/day, respectively (Table 3). The observed half-life of sotrovimab was shorter in allogeneic versus autologous HCT recipients and the non-HCT population (49.4 vs 63.9 and 60.7 days, respectively) (Table 2).

Effect of Covariates on Sotrovimab Exposure After HCT

To further examine the clinical relevance of the effect of covariates on sotrovimab exposure, a series of simulations was conducted to reflect different clinical scenarios. Sotrovimab exposure was measured by the area under the curve of simulated sotrovimab concentrations from time 0 to week 24 (AUC_0-24weeks). In comparison to non-HCT recipients, autologous HCT recipients had 3% lower sotrovimab exposure, whereas allogeneic HCT recipients had 15% lower sotrovimab exposure (Figure 3). To compare how adjusting 1 covariate at a time affects exposure to sotrovimab, a simulated scenario in which allogeneic HCT recipients weighed 86 kg and did not develop diarrhea or GVHD posttransplant was selected as the reference (Figure 3). Compared to this reference scenario, autologous HCT recipients had 14% higher sotrovimab exposure. A decrease in body weight by 10 kg led to a 7% increase in sotrovimab exposure. The development of diarrhea or lower GI GVHD posttransplant led to a 45% and a 50% decrease in sotrovimab exposure, respectively.

Additional scenarios were simulated to assess the impact of time of onset and duration of diarrhea and lower GI GVHD on sotrovimab exposure (Figure 4A). To simulate chronic diarrhea and lower GI GVHD, scenarios in which diarrhea and lower GI GVHD persisted for 12, 16, or 20 weeks’ duration in allogeneic HCT recipients were generated and found to lead to median sotrovimab AUC_0-24weeks of 3064, 2592, and 1906 μg × day/mL, respectively (Figure 4B). The scenario without diarrhea or lower GI GVHD after allogeneic HCT had a median sotrovimab AUC_0-24weeks of 3656 μg × day/mL (Figure 3). To simulate acute diarrhea and lower GI GVHD, scenarios with diarrhea and lower GI GVHD of just 1 week duration and starting 4, 8, or 12 weeks posttransplant led to median sotrovimab AUC_0-24weeks of 3317, 3096, and 2741 μg × day/mL, respectively (Figure 4B). A stronger impact was observed for duration of diarrhea and lower GI GVHD than the time of onset of diarrhea and lower GI GVHD.

DISCUSSION

In the COVIDMAB clinical trial, we observed decreased sotrovimab exposure in allogeneic compared to autologous HCT recipients. Higher body weight and the development of diarrhea or lower GI GVHD were all associated with significantly decreased mAb exposure among HCT recipients. In a cross-study comparison, we also found that HCT patients had up to an 18% increase in clearance compared to nontransplant patients. Together, these data indicate that sotrovimab PK can vary significantly, depending on multiple factors associated with HCT. The PK of mAbs has generally been extrapolated from studies in nonimmunocompromised individuals. As a result, the PK of mAbs has been largely understudied in this vulnerable target population with a variety of diseases that may affect mAb clearance. In infection prevention and treatment trials, altered mAb PKs should be considered for the rational selection of mAb dose and dose frequency and the evaluation of efficacy, especially in immunocompromised patient populations who may not demonstrate a robust response to vaccination and therefore in whom mAb prophylaxis may be of particular benefit.

Although the exact mechanism of accelerated mAb clearance observed in our study is unclear, the association between diarrhea in the context of GI GVHD is suggestive of GI mucosal injury with inflammation and increased intestinal permeability driving mAb elimination and leading to decreased exposure. GI mucosal barrier loss and subsequent wasting of protein, including antibodies, is one potential explanation for the increased sotrovimab clearance observed in allogeneic HCT recipients with lower GI GVHD. Regimen-related toxicity, especially from conditioning regimens associated with severe mucositis, could also contribute to increased intestinal permeability and protein wasting. The bioanalytical PK assay utilized in this study was designed and validated to quantify sotrovimab levels in serum, and its performance on stool samples is unknown. Future studies are needed to determine if mAb clearance can be analyzed from stool specimens. Increased clearance of mAb has been observed in mouse models of GI inflammation leading to protein-losing enteropathy [24]. Interestingly, a recent clinical trial of the human immunodeficiency virus mAb VRC01 also demonstrated faster mAb clearance and decreased mAb exposure in participants with higher levels of serum intestinal fatty acid binding protein, a biomarker of intestinal epithelial permeability [25]. Another potential mechanism of increased sotrovimab clearance includes alterations in FcRn expression. Inflammatory conditions can enhance the clearance of mAbs by altering FcRn expression on vascular endothelium [6–8]. For example, C-reactive protein level, as a relatively nonspecific indicator of systemic inflammation, has previously been identified as a weak predictor of mAb clearance in nontransplant patients with inflammatory conditions [9, 10].

In this study of immunocompromised HCT patients, none of the participants developed anti-drug antibodies posttransplant, indicating that endogenous antibodies to sotrovimab were not responsible for increased clearance. Another theoretical mechanism for mAb clearance is through degradation of immune complexes formed between SARS-CoV-2 antigen and sotrovimab during infection. Only 3 participants acquired SARS-CoV-2 during the study, and 1 of the 3 participants with COVID-19 was an autologous HCT recipient. Since we observed the opposite phenomenon in this group (decreased clearance compared to allogeneic HCT recipients) and since these infections occurred relatively late after transplant, it is unlikely that SARS-CoV-2 infection had a significant impact on the PK results of this study.

In the present study, a simulation approach was employed to demonstrate how modulating the timing of post-HCT complications could impact PK exposure to sotrovimab in different potential clinical scenarios. Chronic diarrhea and lower GI GVHD had the greatest effect on reducing sotrovimab exposure over time. Breakthrough infection occurred in 3 participants between days 95 and 128 posttransplant. Based on circulating patterns of SARS-CoV-2 in the community, these were likely caused by the BA.5 variant, which has escape mutations that decrease the in vitro neutralizing activity of sotrovimab, although the correlation of reduced in vitro potency with clinical efficacy of sotrovimab is unclear [26, 27]. Our findings suggest that both the PK profile and a changing SARS-CoV-2 variant landscape can contribute ultimately to clinical efficacy for a given mAb dose and dosing regimen.

Limitations of this study include the small sample size, which led to relatively low precision in estimating the effects of covariates. Due to the sample size, the models and the data presented herein focused on GI GVHD as a cause of diarrhea, and a subgroup analysis could not be performed to examine whether different pretransplant preparative regimens were also associated with increased mAb elimination due to intestinal injury. Our covariate analysis did include type of transplant at the level of autologous versus allogeneic transplant, but, due to low sample size, we did not perform a covariate analysis on different types of allogeneic transplantation (eg, umbilical cord blood, matched related, matched unrelated, and mismatched unrelated transplants). The study was halted prematurely as a result of the emergence of variants with reduced in vitro neutralization to sotrovimab. Further validation in a larger cohort and generalizability to other mAbs is warranted, including mAbs that do not carry half-life–extending Fc modifications. Even though the FDA has withdrawn the EUA for sotrovimab for the early treatment of COVID-19, the results of this study are clinically relevant beyond sotrovimab. The decreased exposure to mAb in HCT recipients with diarrhea with and without GVHD has broader implications for other antibody-based therapies. Clinical trials of other mAbs with extended half-lives are being planned or are currently underway for RSV (NCT04484935) and SARS-CoV-2 (NCT05648110). Their efficacy in vulnerable populations, particularly HCT recipients, will rely on a more refined understanding of how transplant-specific complications may affect the PK profile of mAbs.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online (http://jid.oxfordjournals.org/). Supplementary materials consist of data provided by the author that are published to benefit the reader. The posted materials are not copyedited. The contents of all supplementary data are the sole responsibility of the authors. Questions or messages regarding errors should be addressed to the author.

Notes

Author contributions. J. B., J. H., A. W., and M. B. conceived and designed the clinical study, analyzed the data, and wrote the manuscript. Q. W. and A. N. analyzed data and helped write the manuscript. L. K., M. L., R. B., J. N., and E. F. coordinated and conducted the clinical study. M. M. helped design the clinical study and write the manuscript. J. W. coordinated regulatory and ethical affairs for the study. T. S.-A., S. F., A. L. G., and J. C. coordinated the laboratory analysis and data management. All authors reviewed and edited the manuscript.

Acknowledgments. We thank Chris Davis and Ryan Bascom for assistance with extracting data on GVHD from the electronic medical record; Phil Pang and Wendy Yeh for providing sotrovimab; Deidre Hill, Haiyan Ping, and Angela Gautreau for administrative support; Lucus Wassira and Karin Albert for programming support; Morgan Gapara for laboratory management support; Graham Craig for data management support; and the members of the Boeckh Laboratory and Boonyaratanakornkit Laboratory for timely sample processing and helpful discussions.

Data availability. Model codes are available upon request.

Financial support. This study was supported by a New Investigator Award from the American Society for Transplantation and Cellular Therapy (to J. B.) and the Amy Strelzer Manasevit Award from the National Marrow Donor Program/Be The Match (to J. B.). A portion of this research is a Supported Collaborative Study funded by GSK and Vir Biotechnology (to J. B., A. W., and M. B.).

References

Author notes