On the Kinetics of Human Skin Mast Cell Depletion and Repopulation Following <scp>KIT</scp> Inhibition

Mast cells (MCs) are long-lived tissue resident immune cells located at barrier sites close to the interface with the environment. MCs have evolved to respond to numerous intrinsic and extrinsic stimuli and drive powerful inflammatory and allergic responses by releasing a wide array of mediators. MCs exist in small numbers as circulating immature progenitors (MCp) and only reach full maturity in tissues where they can reside for years [1]. Despite their sessile nature during homeostasis, MCs can dynamically infiltrate and efflux from tissues in response to inflammatory cues [2]. MCs are highly reliant on stem cell factor (SCF)-dependent signaling by the KIT receptor for many of their functions, including their survival, and genetic or pharmacological KIT suppression leads to their depletion from tissues [3, 4]. However, MCps are not reliant on KIT signaling for survival, providing an opportunity to study the kinetics of MC depletion and repopulation in tissues following KIT inhibition [5, 6]. We collected longitudinal skin biopsies from 20 patients with the two most common forms of chronic inducible urticaria (CIndU)—a MC-driven disease—10 from subjects with symptomatic dermographism (SD) and 10 with cold urticaria (ColdU). Samples were collected prior to and following a single dose of the anti-KIT monoclonal antibody barzolvolimab (NCT05405660) [3]. In this study, barzolvolimab led to strong decreases in serum tryptase—a MC-specific marker of MC load- (Figure 1A) and profound clinical responses as measured by the patient-reported urticaria control test (UCT) and by provocation testing (Figure 1B, Figure S1) [3]. Immunohistochemical (IHC) staining of skin MCs with tryptase or CD117 (KIT)-specific antibodies showed dermal MC density consistent with normal skin at week 0 (Figure 1C, Figure S2). As early as 1 week following drug administration, MC numbers rapidly decreased, reaching approximately 90% reduction after 8 weeks and with similar kinetics as reductions in serum tryptase and clinical symptoms (Figure S3). Interestingly, CD117+ MCs decreased more rapidly at earlier timepoints (60% at W1 and 82% at W4) than tryptase+ MCs (21% at W1 and 59% at W4) and appeared to begin recovery sooner. Kinetics of CD117+ MC reduction mirrored serum tryptase and clinical symptoms more closely than tryptase+ MCs. While the nature of this discrepancy is unclear, it may reflect the presence of phagocytes that clear apoptotic MCs and retain tryptase staining transiently. MCs only recovered partially 36 weeks after dosing (31% for tryptase and 38% for CD117 from baseline), suggesting that MCs can repopulate skin, albeit very slowly. However, by Week 36, barzolvolimab had long cleared from patients (Figure S4), serum tryptase had recovered to near baseline levels, and clinical symptoms had fully returned (Figures 1B and 2B–D) indicating that mature and reactive MCs capable of driving full disease symptoms were present in significant numbers. We further investigated this phenomenon using RNA sequencing on the same serial biopsies from a subset of patients (6 SD and 4 ColdU). MC-specific RNA signatures showed a similar pattern of depletion with 38%, 81%, and 91% reductions at 1, 4, and 8 weeks after treatment, respectively. In contrast to IHC data, MC transcripts completely recovered by Week 36 (Figure 2A) and the kinetics closely mirrored that of clinical symptoms and serum tryptase (Figure 2B–D). The apparent discordance in MC recovery between IHC and RNA might lie in the methodology employed for enumeration; while MCs in the upper dermis were counted in the stained sections, RNA from the full biopsy was captured by RNA sequencing, which would include both subcutaneous and newly infiltrating MCs. Our data demonstrate that KIT-dependent MC reduction is fully reversible and that mature and reactive MCs repopulate skin from MC progenitors within several weeks following their depletion. Diego Alvarado, Margo Heath-Chiozzi, Tibor Keler, Elizabeth Crowley, Diane Young, Philip Golden, Dorothea Terhorst-Molawi, and Martin Metz designed the study and experiments. Diego Alvarado, Deena M. Maurer, Thomas Hawthorne, Lawrence J. Thomas, Dorothea Terhorst-Molawi, Melba Munoz, and Martin Metz performed the experiments and conducted data analysis. Diego Alvarado drafted the manuscript. We would like to thank Dr. Marcus Maurer, without whom this work would not have been possible. We are thankful to patients and clinical staff for participating in the clinical trial described herein. We would like to thank Celldex employees for the critical review of the manuscript. Diego Alvarado, Deena M. Maurer, Philip Golden, Thomas Hawthorne, Lawrence J. Thomas, Margo Heath-Chiozzi, Tibor Keler, Elizabeth Crowley and Diane Young are full-time employees of Celldex Therapeutics and may own stock. Martin Metz has received honoraria as a speaker and/or advisor from: AbbVie, Advanz, ALK-Abello, Allegria, Almirall, Amgen, Argenx, AstraZeneca, Astria, Attovia, Bambusa, Berlin-Chemie, Blueprint, Celldex, Celltrion, DeepApple, Escient, FatiAbGen, Galderma, Granular, GSK, Incyte, Jasper, Japan Tob Inc., Lilly, Lycia, Novartis, Pfizer, Pharvaris, Regeneron, Sanofi, Santa Ana Bio, Septerna, Teva, Third Harmonic Bio, Vifor. Melba Munoz has received honoraria as a speaker and/or advisor from: Biocryst, Blueprint, Celltrion, Jasper. Dorothea Terhorst-Molawi has received research funds/was advisor for Celldex, Moxie, Novartis, Sanofi. Research data is not shared. Figure S1: Kinetics of clinical activity in patients with CIndU following a single dose of barzolvolimab. Clinical response was measured by provocation testing to known triggers, using a Temptest in patients with ColdU (n = 10) and a Frictest in patients with SD (n = 10). Full recovery of symptoms is observed by Week 36 (n = 5 in ColdU and n = 4 in SD). Patients with a negative Temptest were assigned a value of 3°C, as the instrument's lower limit is 4°C. Data shown are means with SEMs. Figure S2: Representative longitudinal series of tryptase staining from biopsies taken from a patient with SD. Figure S3: CD117+ MCs are depleted more rapidly than tryptase+ MCs. Baseline-normalized CD117+ and tryptase+ MCs (green and purple, respectively) are shown relative to serum tryptase (black) and clinical response measured by provocation testing (blue for ColdU, left figure and orange for SD, right figure). Kinetics of CD117+ depletion are more closely correlated with serum tryptase and reduction in provocation thresholds than tryptase+ MCs. Data shown are means with SEMs. Figure S4: Barzolvolimab pharmacokinetics in patients with CIndU (N = 20) following a single 3 mg/kg intravenous dose. Drug levels reach approximately 1 μg/mL by Week 12, as MC markers and clinical symptoms start to recover. Barzolvolimab is fully cleared after Week 18. Data are shown as the geometric mean and geometric standard deviation. Dashed line represents the lower limit of assay quantitation. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.