Reprogramming CXCL10-mediated NK cell trafficking to enhance immunotherapy in osteosarcoma

Main text

Osteosarcoma remains the most common primary malignant bone tumor in children and adolescents. Despite aggressive multimodal therapy, outcomes for relapsed or refractory disease remain dismal, with 5-year overall survival below 20%.¹ Immunotherapeutic strategies have shown limited success in osteosarcoma, in part due to low tumor mutational burden, limited neoantigen availability, and sparse T cell infiltration.² In this context, natural killer (NK) cells represent an attractive alternative, as they mediate antigen-independent cytotoxicity. However, a major barrier to NK cell therapy in solid tumors has been inefficient trafficking and retention within the tumor microenvironment.³

In a recent study, Eguchi et al.⁴ identify tumor-derived CXCL10 as a critical determinant of NK cell recruitment and therapeutic efficacy in osteosarcoma. The authors demonstrate that ex vivo-expanded NK cells expressing CXCR3 exhibit enhanced migration toward osteosarcoma cells engineered to secrete CXCL9, CXCL10, or CXCL11, with CXCL10 showing the most potent effect. In xenograft models, CXCL10-expressing tumors displayed significantly greater NK cell infiltration and improved response to adoptive NK transfer, particularly when combined with NKTR-255, a long-acting IL-15 receptor agonist. The combination enhanced tumor control and prolonged survival selectively in CXCL10-positive tumors. Single-cell profiling further revealed induction of interferon-responsive and apoptotic pathways, alongside enrichment of TGF-β signaling within tumor-associated macrophages. Rather than simply improving NK cytotoxicity, this work reframes NK cell therapy in osteosarcoma by highlighting chemokine-guided trafficking as a gatekeeping step for therapeutic success. These findings suggest that tumor chemokine gradients, particularly CXCL10-driven CXCR3 signaling, determine whether adoptively transferred NK cells can effectively localize to and persist within the tumor microenvironment.

Importantly, several translational considerations merit discussion. First, the in vivo studies were conducted in NSG mice, which lack functional adaptive immunity and key innate immune compartments. While appropriate for isolating NK-mediated effects, these models do not fully capture the complexity of tumor-immune interactions in patients. The durability of response, potential immune crosstalk, and long-term remodeling of the tumor microenvironment remain to be defined in more immunologically intact systems. Second, tumor cell retroviral transduction to enforce CXCL10 expression represents a proof-of-concept model rather than a clinically deployable strategy. Translation will require alternative approaches to induce or augment intra-tumoral CXCL10 expression safely and effectively.

Notably, CXCL10 is not uniformly absent in osteosarcoma. Prior immunohistochemical analyses of osteosarcoma specimens have reported detectable CXCL10 expression in a substantial proportion of tumors, approaching 80%–90% in some cohorts, although intensity and distribution vary considerably.⁵

This heterogeneity suggests that endogenous CXCL10 levels may already stratify patients into chemokine-high and chemokine-low subsets. The critical question is not whether CXCL10 is present but whether it reaches a threshold sufficient to drive meaningful NK infiltration. Prospective studies correlating baseline chemokine expression with NK density and clinical outcomes will be essential.

At present, there is no clinically validated companion diagnostic assay for CXCL10 expression. While immunohistochemistry and RNA-based platforms are available, assay standardization, scoring criteria, and prospective validation would be required before CXCL10 could be implemented as a stratification biomarker in clinical trials.^6,7 Development of a reproducible, clinically deployable CXCL10 assay will therefore be an important step toward biomarker-enriched trial design. Beyond osteosarcoma, elevated CXCL10 expression has been observed in other malignancies, including melanoma, ovarian cancer, and subsets of lung and breast cancers, where it correlates with immune infiltration.^8,9,10,11 These observations raise the possibility that chemokine-guided NK trafficking may represent a broader immunotherapeutic paradigm rather than a disease-specific phenomenon.

For tumors with insufficient endogenous CXCL10 expression, strategies to induce chemokine production warrant exploration. Oncolytic viruses engineered to express CXCL10 could locally convert immunologically “cold” tumors into NK-attracting niches, although their typically localized delivery may limit efficacy in metastatic disease. Additional approaches may include interferon pathway activation, innate immune agonists, radiation-induced chemokine release, or epigenetic modulation to enhance CXCL10 transcription within tumor cells. These strategies could be integrated with cytokine support such as NKTR-255 to maximize NK persistence and effector function. The identification of TGF-β signaling enrichment following combination therapy further highlights the dynamic interplay between response and resistance. TGF-β is a well-recognized suppressor of NK cell activity and promoter of immunoregulatory macrophage phenotypes. Its upregulation in treated tumors suggests the emergence of compensatory immunosuppressive mechanisms and provides a rational basis for future trials incorporating TGF-β blockade or myeloid reprogramming strategies alongside NK-based therapy.

Collectively, the study by Eguchi et al. shifts the conceptual framework of NK immunotherapy in osteosarcoma from solely enhancing cytotoxicity to strategically engineering tumor homing. CXCL10-driven chemokine signaling emerges as a critical determinant of NK cell infiltration and therapeutic efficacy, particularly when paired with IL-15 pathway support. Successful clinical translation will depend on validated biomarker assays, refined patient selection strategies, and feasible methods to induce or augment tumor chemokine expression. By integrating chemokine modulation, cytokine support, and microenvironment-targeted approaches, this work lays the foundation for more precise and effective NK cell-based immunotherapy not only in osteosarcoma but also in other solid tumors. A schematic summarizing the mechanism of the combination treatment is presented in Figure 1.

Figure 1.

A model summarizing CXCL10 expression levels, NK cell recruitment, and their impact on immunotherapy response and tumor cell behavior in osteosarcoma

Declaration of interests

The authors declare no competing interests.