Adaptive immune remodeling and tumor-intrinsic STING signaling shape response to trastuzumab deruxtecan
SITC 2026
Julia A Steele, Stella Feliciello, Susan R Opalenik, Jacey L Marshall, Chad R Potts, Justin A Cartailler, Paula I Gonzalez-Ericsson, Violeta Sanchez, Neil C Chada, Lucas D Mani, Yingying Yu, Stamatini Fragkogianni, Lucia Speroni, Metamia Ciampricotti, Ben H Park, Eben L Rosenthal, P Brent Ferrell, John T Wilson, Justin M Balko
Background
The immunologic effects of trastuzumab deruxtecan (T-DXd) are incompletely defined, partly due to the lack of immunocompetent models that support trastuzumab binding. This limitation has constrained mechanistic understanding and development of rational combination strategies.
Methods
We developed HER2, a minimally modified murine HER2 receptor that permits trastuzumab binding without inducing xenoantigen-driven immune responses in vivo. Using syngeneic, EMT6 HER2 mammary tumor models, we evaluated T-DXd activity, immune composition by mass cytometry, and combinations with immune checkpoint inhibitors (ICIs). Tumor-intrinsic STING function was assessed using CRISPR-mediated knockout models and pharmacologic STING agonism. Findings were complemented with RNA sequencing and clinical outcomes from patients (n = 626) with metastatic breast cancer treated with topoisomerase I inhibitor-based antibody-drug conjugates. Overall survival (OS) was defined as time from treatment initiation to death from any cause. Time to next treatment survival (TTNT) was defined as time from treatment initiation to the start of the next treatment regimen or death from any cause.
Results
T-DXd induced robust antitumor activity in EMT6 HER2 tumors, achieving 50–60% complete responses and prolonged survival. Of complete responders, 73% demonstrated protection from tumor rechallenge, which was dependent on CD8 T cells and MHC-I competency, but independent of HER2 expression.
T-DXd significantly remodeled the tumor microenvironment, increasing the proportion of plasmacytoid dendritic cells, CD62L NK cells, and CD8 and CD4 T cells, while reducing neutrophil, monocyte, and CD206 macrophage populations. CD8 effector T cells showed increased activation (ICOS, granzyme B) and inhibitory receptor upregulation (LAG-3 and TIM-3), identifying additional targetable pathways. ICI combinations improved complete response rates to 70% (αLAG-3 combination), 80% (αTIM-3 combination), and 90% (αPD-1 combination) versus 50% with T-DXd monotherapy.
T-DXd also activated tumor-intrinsic STING signaling, consistent with treatment-induced DNA damage, in vitro and in vivo. STING knockout abrogated complete responses and markedly reduced tumor control, whereas STING agonism enhanced tumor suppression and survival. Moreover, in a real-world cohort of metastatic breast cancer patients treated with topoisomerase I inhibitor-based antibody-drug conjugates, high pretreatment STING1 expression (n = 315) was associated with improved TTNT (median 10.0 vs. 8.2 months; HR 0.76, p = 0.029) and OS (median 17.8 vs. 13.4 months; HR 0.72, p = 0.003) compared to low STING1 expression (n = 311).
Conclusions
T-DXd drives immune remodeling, promotes CD8 T cell–dependent antitumor immunity, and increases targetable immune checkpoint molecule expression. Furthermore, tumor-intrinsic STING signaling contributes to treatment efficacy and correlates with clinical outcomes, providing a mechanistic and translational framework for combinations with ICIs and STING-directed therapies.
Related publications