Anti-tumor activity of trastuzumab deruxtecan in pediatric solid tumors with variable HER2 expression

Abstract

Trastuzumab deruxtecan (T-DXd) is a HER2-targeting antibody-drug conjugate (ADC) with efficacy across adult cancers exhibiting variable HER2 expression. Prior studies demonstrating HER2 expression in osteosarcoma (OS) motivated a clinical trial of T-DXd in pediatric and adolescent/young adults with OS but was terminated early for inactivity. We evaluated the activity of T-DXd using OS patient-derived xenograft (PDX) models and found a 22% objective response rate despite no detectable HER2 expression across PDXs tested. To further assess non-HER2 mediated activity, we evaluated the activity of T-DXd across 31 pediatric cancer cell lines and found OS to be amongst the most resistant to T-DXd, as well as unconjugated deruxtecan, providing a potential explanation for the negative results observed in the clinical trial of T-DXd in OS. T-DXd evaluation in PDX models representing pediatric histologies with greater intrinsic sensitivity to deruxtecan, including pediatric renal tumors and desmoplastic small round cell tumor (DSRCT), revealed both HER2-enhanced activity, as well as substantial non-HER2 mediated activity as evidenced by equipotent activity using an isotype-matched control ADC. Together, these results underscore translational opportunities for ADC therapeutics in tumor histologies with high sensitivity to the payload, and where enhanced tumor delivery may be mediated by antibody-targeted mechanisms as well as macromolecular characteristics of ADCs (e.g., enhanced permeability and retention effect) and tumor microenvironmental factors (e.g., proteolytic payload release). Our findings challenge the role of HER2 as a biomarker predictive of T-DXd response in pediatric cancers and support further biomarker-agnostic clinical development of T-DXd in DSRCT and pediatric renal tumors. .

Figure 1.. Anti-tumor activity of T-DXd in osteosarcoma (OST) PDX models.

(A) Waterfall plot of fold changes in tumor volume relative to baseline in 10 discrete OST PDX models treated with vehicle control or T-DXd at 10 mg/kg IV weekly for 2 doses. All animals were evaluated at Day 21 (endpoint ± 3 days). T-DXd resulted in significant improvement in median tumor volume compared to vehicle (p=0.0009, Mann-Whitney) with an overall disease control rate of 61% and an objective response in 22% of the OST PDX (n = 4/18 with PR). (B) Kaplan Meier analysis demonstrating significant disease control in T-DXd-treated models compared to vehicle (p=0.006, log-rank). (C) Representative micrographs of 4 T-DXd-resistant OST PDX (MSKOST-85018, MSKOST-11890, MSKOST-41876, and MSKOST-53898) and 4 T-DXd-sensitive OST PDX (MSKOST-25998, MSKOST-45826, MSKOST-64473 and MSKOST-11721) probed with HER2 antibody clone CB11 by IHC. All OST PDX models lacked detectable membranous staining (IHC H-score = 0) (scale bar = 20 μm). (D) Immunoblot evaluation of HER2 expression in 4 representative OST PDX models (MSKOST-11721, MSKOST-11890, MSKOST-85018 and MSKOST-86347) using a HER2-amplified cell line (Calu-3) as a positive control. Western blots were probed with two HER2 antibodies (clones 29D8 and CB11) and ACTB was used as a loading control. (E) Mass spectrometry-based proteomic analysis of OST PDX tumor samples treated with T-DXd or vehicle control. No HER2 was observed in the probed OST PDX tumor samples.

Figure 2.. Ex vivo activity of T-DXd and in vivo HER2 expression across disease histologies.

(A) Correlation plot from ex vivo cell line screen of 28 distinct pediatric solid tumor-derived cell lines evaluating the anti-proliferative effect of T-DXd vs isotype control ADC (two-tailed analysis, Pearson correlation r= 0.9875, R²= 0.9751, **** p < 0.0001). The data point for Calu-3 cells is shown in red. (B) Graphical plot of the composite IHC H-scores determined from the HER2 IHC expression levels across custom-made pediatric PDX TMAs using anti-HER2/ERBB2 clone CB11 (Cell Signaling) (mean ± SEM). (C) Representative micrographs from individual PDX models within the pediatric PDX TMAs probed by IHC with HER2 antibody clone CB11 (malignant rhabdoid tumor PDX MSKMRT-29749, hepatoblastoma PDX MSKHEP-11073, and Wilms tumor PDX MSKREN-96914; and desmoplastic small round cell tumor (DSRCT) PDXs: MSKSAR-28381, MSKSAR-22043 and MSKSAR-99599) (scale bar = 10 μm).

Figure 3.. Anti-tumor activity of T-DXd across HER2-expressing and HER2-negative PDX models.

(A) Representative micrographs of HER2-expressing PDX models (Wilms tumor PDX MSKREN-94456, DSRCT PDX MSKSAR-28381, and malignant rhabdoid tumor PDX MSKMRT-29749), HER2-negative PDX models (Wilms tumor PDX MSKREN-31827), and a HER2-negative DSRCT cell line-derived xenograft (CDX) (SK-DSRCT-2) probed with HER2 antibody clone CB11, with corresponding IHC H-Scores (scale bar = 20 μm). (B-D) Relative tumor volume graphs of HER2-expressing PDX models treated with T-DXd (magenta) or isotype control ADC (cyan) at 1 mg/kg or 5 mg/kg IV weekly for four doses vs vehicle control (black), using MSKREN-94456, MSKSAR-28381, and MSKMRT-29749. (E-F) Relative tumor volume graphs of HER2-negative PDX models treated with T-DXd (magenta) or isotype control ADC (cyan) at 1 mg/kg or 5 mg/kg IV weekly for two doses vs vehicle control (black), using MSKREN-31827 and SK-DSRCT-2. Data plots are presented as fold change in tumor volume relative to baseline. (PD: progressive disease; SD: stable disease; PR: partial response; CR: complete response).