In this issue of Blood, Boiko and colleagues,1 using temporal single-cell RNA sequencing of mouse blood, demonstrate dysregulated signatures of interleukin-17 (IL-17) and colony-stimulating factor 1 (CSF-1) expression in monocyte subsets during chronic graft-versus-host disease (cGVHD). Analogous signals were found in patients with cGVHD.
cGVHD still remains the leading cause of failures after allogeneic hematopoietic cell transplantation (HCT), impacting up to 30% to 70% of patients. Clinical symptoms are varied and may affect almost every organ system. The most dreaded symptoms are due to fibrotic reactions involving the skin (cutaneous sclerosis) and lungs (bronchiolitis obliterans). The pathophysiology of cGVHD has been proposed to occur in 3 phases: early inflammation caused by tissue injuries in the epithelial organs (phase 1), initiated dysregulation of immunity in the lymphohematopoietic systems (phase 2), and abnormal fibrosis (phase 3).2 Multiple immune pathways are disrupted during these phases, including biased T-cell differentiation where enhanced IL-17–secreting CD4/CD8 T cells (Th/Tc17), T follicular helper cells, and deficient regulatory T cells generate pathological responses. Likewise, pathological downstream germinal center B-cell differentiation is associated with auto- and alloantibody production, resulting in lung fibrosis. Macrophage sequestration dependent on CSF-1 receptor and alternative differentiation in target tissue appear to be involved in phase 3, resulting in fibrotic skin and lung cGVHD. Macrophage-mediated fibrosis is also IL-17-dependent, the mechanism of which is currently unresolved.
Recently, the Food and Drug Administration approved the novel agents ibrutinib, ruxolitinib, belumosudil, and axatilimab for steroid-refractory or dependent cGVHD.3 Ibrutinib demonstrated 21% complete response (CR) and 45% partial response (PR) in a trial of patients with steroid-refractory/dependent cGVHD (https://clinicaltrials.gov/ NCT02195869).4 Ruxolitinib had a 6.7% CR and 43% PR in the phase 3 randomized trial (REACH3, #NCT03112603).5 Belumosudil had a best overall response rate (ORR) of 74% (62%-84%) in the pivotal phase 2 randomized study (ROCKstar study, #NCT03640481).6 Axatilimab, the newest agent, is an anti-CSF-1R monoclonal antibody and exhibited 74% ORR (#NCT04710576).7 All of these agents showed similar ORR in steroid refractory/dependent cGVHD despite different mechanisms of action. However, the proportion of CR was still low and not satisfactory.
So, how we can select the best treatment for each patient with cGVHD? We are in a maze of treatment choices for patients with cGVHD, with only a trial-and-error approach. Various biological approaches based on cellular or protein or multiomics levels are being explored. Unfortunately, a flow cytometry–based approach failed to generate helpful biomarkers to date.8 The proteomic biomarkers such as CXCL9, DKK3, and MPP3 are not yet specific to driver pathways and are therefore helpful for the selection of therapeutic strategies.9
Boiko and colleagues here report unbiased single-cell sequencing-based approaches in preclinical and clinical settings to rationally select targeted patient-specific cGVHD therapy. Specific immune signatures in mouse blood were identified after HCT that showed dysregulated monocyte differentiation characterized as Ly6Clo clusters and their precursor Ly6Chi. These cells exhibit IL-17–related activation and chemotactic molecules that may directly affect cell homing to target tissues. IL-17 depletion using genetically IL-17 knockout mice upregulated transcription factors related to the resting monocyte/macrophage phenotype. From these results, IL-17 was shown to promote monocyte differentiation into tissue macrophages. Second, high Csf1 gene expressions in terminally differentiated neutrophils that were mobilized by G-CSF were determined as a potentially unappreciated source of CSF-1 in cGVHD. IL-17 contributes to neutrophil maturation; therefore, IL-17–dependent neutrophils may promote peripheral monocyte differentiation into tissue-infiltrating macrophages that cause tissue fibrosis. This unexpected finding raises the question whether mature neutrophils dependent on G-CSF may contribute to cGVHD development. Anti-CSF-1R blockades decreased not only Ly6Clo monocytes that expressed high CSF-1R but also some Ly6Chi monocytes in murine blood. Gene expression alterations were also observed upstream of these monocyte clusters. These novel findings were not expected. Third, the IL-17 signatures as human nonclassical CD14−CD16+ monocytes (derived from mouse Ly6Clo monocytes) were detectable in 7 of 30 patients at day +100 post-HCT. These signatures were present in half of patients with cGVHD but not in healthy volunteers or in most patients without cGVHD. The CSF-1R dysregulated signatures (derived from mouse Ly6Chi and Ly6Clo monocytes) as CD14+CD16− classical and CD14−CD16+ nonclassical monocytes were also detectable in 4 (Ly6Chi signature) and 14 (Ly6Clo signature) of 30 patients at day +100 post-HCT. Furthermore, 7 of 10 patients were positive for at least 1 signature at cGVHD onset. Surprisingly, these proportions were similar to those that responded to belumosudil or axatilimab treatment.6,7
This study could identify pathway-specific benefits for cGVHD patients with IL-17 or CSF1-R target therapies. At the same time, some other cellular and gene expression changes cannot be included because this study focused on IL-17 and CSF-1. Also, these signatures were observed in only a subset of patients, implicating the complexity of cGVHD pathophysiology. Despite remaining questions, Boiko and colleagues’ finding may help delineate potential responders and nonresponders to relevant therapeutics.
Conflict-of-interest disclosure: Y.M. declares no competing financial interests.
