bioRxiv2025
BackgroundImmune checkpoint blockade (ICB) achieves durable responses in approximately half of patients with advanced melanoma, but the mechanistic basis for resistance in the remaining patients remains incompletely defined. AXL tyrosine kinase has emerged as a candidate resistance mediator, yet clinical development of AXL inhibitors has yielded heterogeneous results in unselected patient populations, suggesting that cell-type and tumor microenvironment context may critically govern therapeutic response. MethodsWe characterized AXL expression across cell types in ICB-resistant melanoma using publicly available single-cell RNA sequencing datasets and The Cancer Genome Atlas (TCGA). In vivo efficacy was assessed in the YUMM1.7 PD-1-resistant syngeneic melanoma model using three pharmacologically distinct AXL inhibitors (warfarin, bemcentinib, and cabozantinib) as monotherapy and in combination with anti-PD-1 therapy. Tumor-associated macrophage (TAM) context-dependency was established using anti-CSF1R and anti-F4/80 depletion strategies. Functional PD-L1:PD-1 checkpoint interactions were quantified in tumor sections by immune Forster Resonance Energy Transfer (iFRET). TAM secretome reprogramming was characterized by 40-plex Luminex immunoassay in polarized RAW264.7 macrophages. ResultsAXL expression in ICB-resistant melanoma was predominantly localized to TAMs rather than tumor cells by single-cell analysis, with AXL+ TAMs distributed across both M1-like and M2-like phenotypic compartments. AXL inhibition significantly reduced tumor burden and synergized with anti-PD-1 therapy in vivo; however, efficacy was abolished by depletion of the monocyte-derived myeloid compartment (anti-CSF1R) and enhanced by depletion of tissue-resident TAMs (anti-F4/80), establishing TAM-context dependency. iFRET revealed a paradoxical gain of PD-L1:PD-1 interaction efficiency in anti-PD-1-treated resistant tumors, a functional resistance signature detectable by iFRET but not by PD-L1 expression that was reversed by AXL combination therapy. In vitro secretome profiling demonstrated that combination therapy reprograms macrophage secretome in a polarization-context-dependent manner, amplifying pro-inflammatory cytokines including IL-6 and GM-CSF in M2-like macrophages while selectively dampening T cell chemokine production. ConclusionsThese findings establish AXL as a TAM-resident immune target in ICB-resistant melanoma whose therapeutic relevance is governed by tumor-immune micronenvironment (TiME) macrophage composition rather than tumor cell AXL expression. TAM polarization contexture represents a candidate stratification axis for AXL inhibitor-based combination strategies, and iFRET-measured checkpoint interaction offers a functional complement to PD-L1 expression for monitoring resistance and response. Key MessagesO_ST_ABSWhat is already known on this topicC_ST_ABSO_LIAXL tyrosine kinase has been studied predominantly as a tumor-intrinsic mesenchymal marker and therapeutic target in melanoma, with prior clinical development of AXL inhibitors predicated on tumor cell AXL expression as the primary biomarker; the contribution of TAM-expressed AXL to ICB resistance and the dependence of therapeutic response on TiME composition have not been defined. C_LI What this study addsO_LIAXL expression in ICB-resistant melanoma is predominantly localized to tumor-associated macrophages distributed across both M1-like and M2-like phenotypic compartments, reframing AXL as a TAM immune target whose therapeutic relevance is governed by TiME macrophage composition rather than tumor cell AXL expression. C_LIO_LICombination AXL inhibition and anti-PD-1 reprograms the macrophage secretome in a polarization-context-dependent manner and reverses a paradoxical gain of functional PD-L1:PD-1 checkpoint interaction in resistant tumors -- a resistance signature detectable by spatial iFRET but not by PD-L1 expression. C_LI How this study might affect research, practice or policyO_LIThe context-dependent effects of AXL inhibition (immunostimulatory in M2-heavy TiMEs, potentially counterproductive in M1-heavy TiMEs) indicate that TAM polarization profiling should be incorporated into clinical trial design for AXL inhibitor-based combinations, and that unselected enrollment may obscure meaningful efficacy signals in the subset most likely to benefit. C_LIO_LIiFRET-based quantification of functional checkpoint interaction represents a candidate dynamic biomarker for ICB resistance monitoring that is orthogonal to PD-L1 immunohistochemistry and may identify resistant patients who retain immunosuppressive checkpoint interactions that can be disrupted by AXL targeting. C_LI Graphical Abstract O_FIG O_LINKSMALLFIG WIDTH=170 HEIGHT=200 SRC="FIGDIR/small/667666v2_ufig1.gif" ALT="Figure 1"> View larger version (90K): org.highwire.dtl.DTLVardef@521de1org.highwire.dtl.DTLVardef@127a933org.highwire.dtl.DTLVardef@d56132org.highwire.dtl.DTLVardef@e03ef9_HPS_FORMAT_FIGEXP M_FIG C_FIG