Targeted immunotherapy against distinct cancer-associated fibroblasts overcomes treatment resistance in refractory HER2+ breast tumors

Abstract

About 50% of human epidermal growth factor receptor 2 (HER2)+ breast cancer patients do not benefit from HER2-targeted therapy and almost 20% of them relapse after treatment. Here, we conduct a detailed analysis of two independent cohorts of HER2+ breast cancer patients treated with trastuzumab to elucidate the mechanisms of resistance to anti-HER2 monoclonal antibodies. In addition, we develop a fully humanized immunocompetent model of HER2+ breast cancer recapitulating ex vivo the biological processes that associate with patients' response to treatment. Thanks to these two approaches, we uncover a population of TGF-beta-activated cancer-associated fibroblasts (CAF) specific from tumors resistant to therapy. The presence of this cellular subset related to previously described myofibroblastic (CAF-S1) and podoplanin+ CAF subtypes in breast cancer associates with low IL2 activity. Correspondingly, we find that stroma-targeted stimulation of IL2 pathway in unresponsive tumors restores trastuzumab anti-cancer efficiency. Overall, our study underscores the therapeutic potential of exploiting the tumor microenvironment to identify and overcome mechanisms of resistance to anti-cancer treatment.

Fig. 1. Stromal traits define resistance to trastuzumab.

a Percentage of CD45+ cells in HER2+ BC tumors comparing relapse-free (blue) to relapsing patients (red) after trastuzumab-based therapy. Central mark indicates the median, box extends from the 25 to 75th percentiles, whiskers represent the maximum and minimum data point. Number of patients in each group is indicated. b Violin plots of immune/stroma enrichment score ratio for patients with pCR (blue) vs. RD (red) (GSE50948) and for relapsing (red) vs. relapse-free patients (blue) (GSE55348). Number of patients in each group is indicated. c–e Gene-set enrichment analyses (GSEA) in tumor samples collected before treatment of (c) sCAF, (d) CAF-S1, and (e) pCAF gene expression signatures comparing pCR vs. RD (upper row) and relapsing vs. relapse-free (lower row) HER2+ BC after trastuzumab treatment. Rel relapsing patients, RF relapse-free patients, pCR pathological complete response, RD residual disease, ES enrichment score, NES normalized enrichment score, FDR false discovery rate. Two-sided, unpaired t test p values (P) are indicated for (a, b). GSEA nominal p value (P) and FDR-adjusted p value are indicated for (c–e). Source data are provided as a Source Data file.

Fig. 2. HER2+ 3DiBC recapitulates immune exclusion and resistance to trastuzumab.

a Confocal bioimaging of HER2+ 3DiBC tumor (left panel). Magnification of inside/outside 3D tumor co-culture interface illustrates immune exclusion from the tumor (right upper panel). Dashed lines delineate tumor inner and outer compartments. Representative of n = 3 HER2+ 3DiBCs. Scale bars: 50 μm. 3D tumor Z-axis is indicated (right lower panel). Red: breast cancer cells (BCCs; HCC1954); Green: breast fibroblasts (BFs); Cyan: immune cells (ICs). b Bioluminescent tracking of HCC1954 in HER2+ 3DiBCs with increasing amounts of BFs. n = 3 biologically independent experiments. Values are mean ± s.d. c Bioluminescent tracking of HCC1954 in HER2+ 3DiBCs with or without ICs. n = 14 (HCC1954/BF1; light gray) and n = 12 (HCC1954/BF2; dark gray) HER2+ 3DiBCs examined per condition, from 7 biologically independent experiments. Values are mean ± s.e.m. d Bioluminescent tracking of HCC1954 in HER2+ 3DiBCs with ICs (red), TTZ (gray) or ICs + TTZ (blue) in absence (left panel) or in presence of BFs (right panel; HCC1954/BFs in a 1/5 ratio). IgG1 was used as control for TTZ. n = 9 HER2+ 3DiBCs examined per condition, from 3 biologically independent experiments. Values are mean ± s.e.m. e ICs abundance in TTZ-treated HER2+ 3DiBCs in presence or absence of BFs. n = 7 biologically independent experiments. Values are mean ± s.d. f, g Representative bioimaging of treated HER2+ 3DiBC from (e). Scale bars: 50 μm. f Micrograph of ICs infiltrating the tumor in absence of BFs representative of n = 3 independent samples. g Micrograph of immune exclusion in presence of BFs representative of n = 3 independent samples. Dashed lines delineate tumor inner and outer compartments in (f, g). Red breast cancer cells (BCCs; HCC1954), Green breast fibroblasts (BFs), Cyan immune cells (ICs). TTZ trastuzumab, RLU relative luminescence units. Two-sided, unpaired t test p values (P) are indicated for (b–e); ns indicates non-significant. Source data are provided as a Source Data file.

Fig. 3. Response to IL2 enhances anti-cancer immunity.

a, b GSEA of NK-IL2RS gene set comparing (a) pCR vs. RD and (b) relapsing vs. relapse-free HER2+ BC after TTZ treatment. c Kaplan–Meier curve shows relapse-free survival (RFS) of HER2+ BC patients treated with TTZ presenting low (red) or high (blue) expression levels of NK-IL2RS. Hazard ratio (HR) and log-rank test p value (P) are indicated. Number of patients in each group is indicated. d Multivariate Cox regression model analysis and p values of NK-IL2RS, adjusted by stage, grade, age and lymph nodes positivity. e CD56+ cell number in HER2+ BC tumor cores comparing relapse-free (blue) to relapsing patients (red) after TTZ-based therapy. Central mark indicates the median, box extends from the 25 to 75th percentiles, whiskers represent the maximum and minimum data point. Number of patients in each group is indicated. f Cell-specific relative fluorescence intensity (RFI) of ICs (blue), HCC1954 (red) and BFs (green) in HER2+ 3DiBCs after 72 h treatment with TTZ or TTZ + rhIL2. n = 3 biologically independent experiments. Values are mean ± s.d. g 72 h follow-up of cell-specific RFI from (f). TTZ-treated HER2+ 3DiBCs show increased abundance of ICs after 12/24 h (blue line), decreased presence of HCC1954 over time (red line) in presence of rhIL2. n = 3 biologically independent experiments. Values are mean ± s.d. h Micrographs of HER2+ 3DiBC IMARIS analysis from (f). Left panel: immune excluded tumor in absence of rhIL2. Right panel: Immune cells abundance is associated with a strong reduction of epithelial cancer cells in presence of rhIL2. Red breast cancer cells (BCCs; HCC1954), Green breast fibroblasts (BFs), Gray immune cells (ICs). Scale bar: 100 μm. Rel relapsing patients, RF relapse-free patients, pCR pathological complete response, RD residual disease, ES enrichment score, NES normalized enrichment score, FDR false discovery rate, TTZ trastuzumab. GSEA nominal p value (P) and FDR-adjusted p value are indicated for (a, b). Two-sided log-rank test (c, d) and two-sided, unpaired t test (e–g) p values (P) are indicated; ***P < 0.001; **P < 0.01; *P < 0.05; ns non-significant. Source data and exact p values for (g) are provided as a Source Data file.

Fig. 4. FAP-IL2v restores anti-cancer immunity.

a FAP expression in HER2+ BC tumors comparing relapse-free (blue) to relapsing patients (red) after TTZ-based therapy. Central mark indicates the median, box extends from the 25 to 75th percentiles, whiskers represent the maximum and minimum data point. Number of patients in each group is indicated. b Micrographs of FAP expression in HER2+ BC human tumor samples from (a). T tumor, S stroma. Scale bar: 100 μm. Representative of n = 22 tumors. c Correlation between FAP expression levels and CAF-S1, pCAF or NK-IL2RS in HER2+ BC patients (n = 48) from GSE50948 dataset. Correlation coefficients (R) and Spearman p values are indicated. d Scheme depicting HER2+ 3DiBCs expansion and treatment. Left panel: representative co-culture in a well from 96-well plate. e Micrograph of FAP expression in HER2+ 3DiBCs (HCC1954/BF2). Scale bars: 100 μm. Representative of n = 3 independent HER2+ 3DiBCs. f Bioluminescent tracking of HCC1954 in HER2+ 3DiBCs treated as indicated in absence of ICs. IgG1 was used as control for TTZ. n = 9 (HCC1954/BF1; blue) and n = 6 (HCC1954/BF2; red) HER2+ 3DiBCs examined per condition, from 3 biologically independent experiments. Values are mean ± s.e.m. g Bioluminescent tracking of HCC1954 in HER2+ 3DiBCs treated as indicated in presence of ICs. IgG1 was used as control for TTZ. n = 6 (HCC1954/BF1; blue) and n = 9 (HCC1954/BF2; red) HER2+ 3DiBCs examined per condition, from 3 biologically independent experiments. Values are mean ± s.e.m. h Representative micrographs of HER2+ 3DiBCs from (g). Left panel: TTZ/DP47-IL2v treatment. Right panel: TTZ/FAP-IL2v treatment. Arrows indicate ICs. Scale bar: 20 μm. Red breast cancer cells (BCCs; HCC1954), Green breast fibroblasts (BFs), Cyan immune cells (ICs). Rel relapsing patients, RF relapse-free patients, TTZ trastuzumab, RLU relative luminescence units. Two-sided, unpaired t test p values (P) are indicated for (a, f, g); ns indicates non-significant. Source data are provided as a Source Data file.

Fig. 5. FAP-IL2v enhances ADCC through NK activation.

a, b Relative gene expression levels of (a) CD45 and NCR1, (b) PRF1, GZMB and IFNG measured by qRT-PCR in HER2+ 3DiBCs (HCC1954/BF2) treated as indicated. IgG1 was used as control for TTZ. n = 3 biologically independent experiments. Values are mean ± s.d. c Flow cytometry analysis using gating panels from Supplementary Fig. 4e and showing percentages of NK cells, CD16+ NK cells and NKT cells in infiltrating ICs from HER2+ 3DiBCs (HCC1954/BF2) treated with trastuzumab (T) or trastuzumab + FAP-IL2v (T/F) compared to control (IgG1; C). n = 6 biologically independent experiments. d Flow cytometry profiles illustrating data from (c). e GSEA of CD56+ NK and CD56+CD3+ NKT cells gene signatures comparing pCR vs. RD (left panels) and Rel vs. RF (right panels) HER2+ BC after TTZ-based therapy. f Growth kinetics of macroscopic tumors derived from subcutaneous injection of HCC1954 cells into nude mice treated with DP47-IL2v (dashed blue; n = 11), FAP-IL2v (dashed red; n = 12), TTZ (gray; n = 8), TTZ/DP47-IL2v (blue; n = 12), TTZ/FAP-IL2v (red; n = 11) compared to control (IgG1; dashed gray; n = 7). Values are mean ± s.e.m. g Bioluminescent tracking of HCC1954 in HER2+ 3DiBCs treated with TTZ, FAP-IL2v, DP47-IL2v, and/or blocking antibody against CD16 in presence of NK cells. n = 4 (HCC1954/BF1; blue) and n = 9 (HCC1954/BF2; red) HER2+ 3DiBCs examined per condition, from 3 biologically independent experiments. Values are mean ± s.e.m. pCR pathological complete response, RD residual disease, Rel relapsing patients, RF relapse-free patients, ES enrichment score, NES normalized enrichment score, FDR false discovery rate, TTZ trastuzumab, RLU relative luminescence units. Two-sided, unpaired (a, b, f, g) and paired (c) t test p values (P) are indicated. GSEA nominal p value (P) and FDR-adjusted p value are indicated for (e). Source data are provided as a Source Data file.

Fig. 6. FAP is an actionable biomarker of TGF-beta activity in breast cancer microenvironment.

a Correlation between FAP, TGFB and TBRSs expression levels in HER2+ BC patients (n = 51) from GSE55348 dataset (from up to down, left to right: macrophage, endothelial, T cell, fibroblast, NK cell). Correlation coefficients (R) and Spearman p values are indicated. b Levels of FAP protein (upper panel) in BFs untreated or treated with recombinant human (rh) TGF-Beta1. Bottom panel shows β-Actin protein levels as normalization control. Representative of n = 3 biologically independent experiments. c, d GSEA of F-TBRS gene-set comparing (c) pCR vs. RD and (d) recurring vs. relapse-free HER2+ BC after TTZ treatment. e Bioluminescent tracking of TGF-beta pathway activation in BFs co-cultured with increasing amounts of HCC1954, in presence of TGFb inh. (red) compared to untreated control (Ctrl, blue). n = 3 biologically independent experiments. Values are mean ± s.d. f Level of FAP protein (upper panel) in BFs, BFs co-cultured with BCCs (HCC1954) untreated or treated with TGFb inh. Bottom panel shows β-Actin protein levels as normalization control. Representative of n = 3 biologically independent experiments. g Mean anti-FAP immunoFluorescence Intensity (FI) detected on BFs co-cultured with HCC1954 in presence (blue) or absence (red) of TGFb inh. compared to controls without anti-FAP antibody (gray). n = 25 regions examined per condition, from 5 biologically independent experiments. Values are mean ± s.e.m. h Representative micrographs from (g). Left panel: anti-FAP antibody identifies BFs. Right panel: loss of anti-FAP binding to BFs upon TGFb inh. treatment. Scale bar: 20 μm. Gray HCC1954 (BCCs), Green BFs, Red anti-FAP antibody immunofluorescence, Yellow FAP/BFs colocalization. i Bioluminescent tracking of HCC1954 in HER2+ 3DiBCs with BF1 (blue) or BF2 (red) in presence of ICs and treated as indicated. n = 21 (TTZ/DP47-IL2v), n = 24 (TTZ/FAP-IL2v), n = 21 (TTZ/DP47-IL2v/TGFb inh), n = 24 (TTZ/FAP-IL2v/TGFb inh.) HER2+ 3DiBCs examined per condition, from 6 biologically independent experiments. Values are mean ± s.e.m. Rel relapsing patients, RF relapse-free patients, pCR pathological complete response, RD residual disease, ES enrichment score, NES normalized enrichment score, FDR false discovery rate, TGFb inh. TGF-beta pathway inhibitor, RLU relative luminescence units, TTZ Trastuzumab. GSEA nominal p value (P) and FDR-adjusted p value are indicated for (c, d). Two-sided, unpaired t test p values (P) are indicated for (e, g, i); ****P < 0.0001; ***P < 0.001. Source data and exact p values for (e) are provided as a Source Data file.