Safety and reactogenicity of a controlled human infection model of sand fly-transmitted cutaneous leishmaniasis

Cutaneous leishmaniasis (CL) is a vector-borne disease transmitted by phlebotomine sand flies, with substantial global burden and no licensed human vaccine. Clinical outcomes vary by Leishmania species; L. major lesions often self-resolve, whereas other species can cause chronic or metastatic disease. Vector control alone is insufficient, and vaccination is a key public health goal. Despite decades of preclinical work, few candidates have reached clinical trials, underscoring the need for new approaches to evaluate vaccine efficacy and understand human protective immunity. Controlled human infection models (CHIMs) have accelerated vaccine development for other pathogens and can provide rapid, cost-effective efficacy data. The study aimed to establish a safe and effective CHIM for sand fly-transmitted L. major CL that incorporates natural vector transmission, to estimate take rate (proportion developing CL lesions) and assess safety, and to characterize early local immune responses.

Prior controlled infection efforts in leishmaniasis included historical leishmanization and early needle-based challenges, but these did not account for the immunomodulatory role of sand fly transmission. Evidence suggests vaccines protective against needle challenge may fail against natural transmission. CHIMs are established for malaria, schistosomiasis, and hookworm, demonstrating utility in de-risking and accelerating vaccine pipelines. WHO has prioritized Leishmania vaccines, with recognized potential health impact and feasible demand and affordability. However, only a few candidates (for example, ChAd63-KH, live attenuated L. major centrin-deleted) are in clinical development. The literature highlights the importance of vector factors, early host–parasite interactions, and myeloid cell involvement in CL pathogenesis, as well as gaps in understanding early human lesion immunology that this study addresses.

Study design: Open-label, observational first-in-human sand fly transmission CHIM (LEISH_Challenge; NCT04512742) conducted at the University of York Translational Research Facility. Fourteen healthy, Leishmania-naive adults (18–50 years; eight female, six male; predominantly White) were enrolled between November 2021 and August 2022 and exposed between 24 January and 12 August 2022. Inclusion criteria included no history of Leishmania exposure and willingness to avoid travel to L. major-endemic regions; negative HIV, HBV, HCV serology; absence of significant atopy or active skin disease. Ethics approvals were obtained and written informed consent was provided. Vector and parasite: Laboratory-reared Phlebotomus duboscqi (Senegal origin) maintained under standard insectary conditions. Flies were infected via membrane feeding with rabbit blood containing 10^6 L. major MHOM/IL/2019/MRC-02 (cGMP-manufactured) promastigotes per mL. Infection rates >90% were confirmed by dissection. On challenge day, five female infected flies were placed in a bespoke biting chamber on the volar forearm for 30 minutes. Procedures and cohorts: Participants were observed during biting; biting failure was defined by absence of biting sensation, observed activity, immediate dermoscopic/photographic bite signs, and blood in fly abdomens. Cohort 1 (n=6) used a 6-mm aperture; Cohort 2 (n=8) used smaller apertures (3–5 mm) to minimize lesion size/scarring and had earlier biopsies. Lesions were monitored by clinical exam and dermoscopy. Therapeutic lesion biopsies were planned when lesions reached ~2–3 mm diameter; excision biopsy at later times in Cohort 1 (median day 34) vs earlier 6–8 mm punch in Cohort 2 (median day 18). Parasitology and histology: Biopsies were split for histology (50%), qPCR (25%), and immunology (25%). qPCR targeted kinetoplast minicircle DNA (SYBR Green; standard curve), reporting parasites per mg of tissue. FFPE sections underwent H&E and immunohistochemistry (IHC) for immune markers (CD3, CD4, CD8, CD14, CD68, CD20, CD66b) and Leishmania Oligopeptidase B (OPB); nuclei stained with YOYO-1. Imaging used Zeiss Axioscan; quantification via StrataQuest. Spatial transcriptomics: Visium (10x Genomics) FFPE workflow on biopsies from three participants (LC001, LC003, LC008), generating 20,241 spots across 12 sections. Data processed with Space Ranger; analysis in Seurat with SCTransform, integration, and clustering. Cell type deconvolution used cell2location with a skin scRNA-seq reference (Reynolds et al.). Differential expression analysis (Wilcoxon with Bonferroni correction) and pathway analysis (STRING, g:Profiler). Spatial co-localization inferred by Pearson correlations of predicted abundances. Safety monitoring and PROs: Adverse events graded; labs (FBC, LFTs, U&E, CRP) at baseline and follow-up. Participants completed electronic visual analogue scale (VAS) diaries (itch, pain, erythema, swelling, malaise, myalgia, fever, nausea) and quality-of-life instruments (DLQI; GAD-7). Statistics: Observational exploratory study; pragmatic sample size to estimate attack/take rate with lower 95% CI ~60%. Nonparametric tests where appropriate (Mann–Whitney; Spearman’s correlation); Student’s t-test where indicated; no blinding; automated pipelines for image analyses.

• Take rate and lesion development: In Cohort 1 (6-mm aperture), 5/6 had confirmed bites and developed clinically compatible CL lesions; per-protocol take rate 83% (95% CI: 0.44, 0.97), rising to 100% (95% CI: 0.57, 1) among those with confirmed bites. Parasite loads by qPCR were positive in 5/5 biopsies with wide variance (median 1,218 parasites/mg; range 255–27,547/mg). In Cohort 2 (3–5 mm apertures, earlier biopsy), 2/8 were bite failures. Among the remaining six with confirmed bites, one lesion resolved spontaneously by day 42 without biopsy; four showed parasitological confirmation (qPCR and/or IHC), and one was clinically compatible but qPCR/IHC negative with non-focal histology. Take rate estimated at 50% (95% CI: 0.22, 0.78) overall, 67% (95% CI: 0.3, 0.90) with confirmed bites. Overall across both cohorts: 64% (9/14; 95% CI: 0.39, 0.84) for all participants and 82% (9/11; 95% CI: 0.52, 0.95) among those with confirmed bites.
• Recurrence and treatment: Of 10 biopsied participants, 7 (70%) required no further treatment. Three (LC001, LC004, LC016) developed lesions at the biopsy site 4–8 months post-biopsy; second biopsies confirmed parasitism (30,800; 184; 1,658 parasites/mg at 255, 282, and 126 days post-bite, respectively). All were successfully treated with cryotherapy; one participant (LC004) had a second recurrence at 14 months, resolving with additional cryotherapy. Potential contributors included an unexcised bite site, local trauma, and intralesional steroid.
• Scarring: Scarring occurred in all biopsied participants; generally milder with smaller apertures and earlier punch biopsy (Cohort 2) versus excision (Cohort 1). Quantified scar area at final follow-up was significantly smaller in Cohort 2 (P = 0.0005). Two wound infections occurred and may have contributed to scarring; both resolved with treatment.
• Safety and tolerability: No grade 3 or serious adverse events. One grade 1 event (scar exudate) and two grade 2 events (wound infections). No clinically significant lab abnormalities; minimal inflammatory responses remained within normal ranges. VAS symptom scores were low and comparable to uninfected sand fly bites. Mean DLQI change 1.92 ± 2.54 (below MCID); GAD-7 change −0.17 ± 1.94.
• Immunohistology: Lesions showed dense lympho-histiocytic dermal infiltrates, acanthosis, patchy hyperkeratosis, occasional unorganized granulomas; compact organized granulomas were absent. CD4+ and CD8+ T cells were present at variable ratios; CD8:CD4 ratio positively correlated with lesion duration (Spearman r = 0.7881; 95% CI: 0.37–0.94; P = 0.0034). CD14+ monocytes and CD14+CD68+ monocyte-derived macrophages often outnumbered CD68+ dermal macrophages; parasites were largely confined to CD14+, CD68+, and CD14+CD68+ cells. CD66b+ neutrophils were infrequent and mostly near ulceration; rarely parasitized. CD20+ B cells were sparse, more abundant in recurrent lesions.
• Spatial transcriptomics: Identified a leucocyte-rich lesion core (cluster 2) enriched for myeloid cells (DC2, migratory DCs, monocytes/macrophages) and T cell subsets (Tc, Th, Tregs), with elevated interferon-inducible genes (e.g., CXCL9), antigen processing/presentation (HLA-DRA, HLA-DPA1, CD74), metalloproteinases (MMP2, MMP9, TIMP1), and cytokines/chemokines (CCL5, CCL18, CCL19, CCL21), versus an ulcer cluster associated with epidermal remodeling. Within the core, discrete spatial niches were observed, including CXCL9-high and CCL19-high regions and fibroblast-associated zones (e.g., MMP2, GREM1), indicating functional compartmentalization. Differential expression identified 134 genes (FDR 5%, log2FC >1.5) distinguishing core (80 up) from ulcer (54 up).

The study demonstrates that a CHIM using natural sand fly transmission of L. major is feasible, effective, and safe, achieving take rates comparable to other CHIMs while avoiding an excessive force of infection. The approach provides a platform to generate early, cost-effective human efficacy data for CL vaccine candidates, potentially enabling trials with around 50 participants if a 70% vaccine efficacy target product profile is assumed and lesion/no lesion used as a dichotomous endpoint. Protocol parameters (e.g., number of infected flies, exposure duration, biopsy timing) can be tuned to adjust take rate and lesion size to suit vaccine, drug, or mechanistic objectives. Earlier punch biopsy and smaller exposure apertures appear to reduce scarring, though lesion size and earlier sampling can decrease apparent take rate. The immunological analyses reveal early lesion architectures and cell type distributions, highlighting predominant parasitism of monocyte/macrophage populations, increased CD8 T cell prevalence with time/recurrence, and discrete cytokine/chemokine-defined spatial niches. These findings separate antiparasitic inflammatory responses in the core from wound-healing processes in the ulcer, offering insights into pathogenesis and potential correlates of protection or pathology that can inform vaccine design and host-directed therapies.

This work establishes the first sand fly-transmitted L. major CHIM in humans as safe, acceptable, and sufficiently effective for interventional studies. It enables vaccine candidate selection based on human efficacy, evaluation of pre- and post-exposure therapies, and detailed study of early immune mechanisms in CL. Spatially resolved analyses uncover functionally distinct immune niches and compartmentalization within lesions. Future research should optimize protocol parameters to balance efficacy readouts with cosmesis, include more diverse populations and settings, explore different biopsy timepoints (including innate phases and late pathology), and leverage continuous endpoints (e.g., parasite load dynamics, scar outcomes) where appropriate. Given evidence of cross-species protection, this CHIM may aid development of vaccines for multiple leishmaniasis manifestations, including visceral disease.

• Small, single-center exploratory study not powered for between-group comparisons; pragmatic design limits inferential strength.
• Early therapeutic biopsy confounds natural lesion progression and limits assessment of self-cure; absence of a control arm complicates attribution of scarring to CL versus biopsy.
• Apparent take rate is sensitive to lesion size, biopsy timing, and strict parasitological confirmation criteria; cohort differences in procedures affected rates.
• All participants were White; generalizability across races, ethnicities, and environmental backgrounds requires further study.
• Biopsy timing focused on early lesions; innate and late-stage immunopathology warrant targeted sampling windows.
• Recurrences occurred in a minority and were treatable; intralesional steroid may precipitate recurrence and should be avoided; long-term relapse risk appears low but cannot be fully excluded.
• Implementation in endemic settings may face logistical and vector-related constraints; model parameters may need adaptation.