Immune microenvironment dynamics of HER2 overexpressing breast cancer under dual anti-HER2 blockade

Introduction The clinical prognosis of the HER2-overexpressing (HER2-OE) subtype of breast cancer (BC) is influenced by the immune infiltrate of the tumor. Specifically, monocytic cells, which are promoters of pro-tumoral immunosuppression, and NK cells, whose basal cytotoxic function may be enhanced with therapeutic antibodies. One of the standards of care for HER2+ BC patients includes the combination of the anti-HER2 antibodies trastuzumab and pertuzumab. This dual combination was a breakthrough against trastuzumab resistance; however, this regimen does not yield complete clinical benefit for a large fraction of patients. Further therapy refinement is still hampered by the lack of knowledge on the immune mechanism of action of this antibody-based dual HER2 blockade. Methods To explore how the dual antibody challenge influences the phenotype and function of immune cells infiltrating the HER2-OE BC microenvironment, we developed in vitro 3D heterotypic cell models of this subtype. The models comprised aggregates of HER2+ BC cell lines and human peripheral blood mononuclear cells. Cells were co-encapsulated in a chemically inert alginate hydrogel and maintained in agitation-based culture system for up to 7 days. Results The 3D models of the HER2-OE immune microenvironment retained original BC molecular features; the preservation of the NK cell compartment was achieved upon optimization of culture time and cytokine supplementation. Challenging the models with the standard-of-care combination of trastuzumab and pertuzumab resulted in enhanced immune cytotoxicity compared with trastuzumab alone. Features of the response to therapy within the immune tumor microenvironment were recapitulated, including induction of an immune effector state with NK cell activation, enhanced cell apoptosis and decline of immunosuppressive PD-L1+ immune cells. Conclusions This work presents a unique human 3D model for the study of immune effects of anti-HER2 biologicals, which can be used to test novel therapy regimens and improve anti-tumor immune function.

cultures with PBMCs, by flow cytometry, revealed a significant decline in the immune cell population from day 4 to day 7 of culture in both co-cultures.
B and C: Cell populations were identified and quantified according to the gating strategy supplied in D).Bars represent mean ± S.D. from N = 6 independent experiments, performed with different immune donors.Pairwise statistical comparisons between indicated groups performed with unpaired t-test; *, p<0.05; **, p<0.01.

Supplementary Figure 2. Flow cytometry gating strategy.
A) Gating strategy used in flow cytometry analysis of cancer and immune cells.Cancer and immune cells were gated on the basis of CD45 expression (I), and then dead (DAPI + ) cells were removed from each population before further analysis (II).For immune (CD45 + ) cells, lymphocytic and monocytic cells were separated based on size and complexity (forward scatter and side scatter) (III).Lymphocytes were either directly analyzed for the expression of NK cell markers (IV) (Panel 1) or split into CD3 + and CD3 -populations (V) (Panel 2).Using Panel 2, CD3 -lymphocytes were analyzed for the quantification of different NK cell subsets (VI).In both panels, NK cells were identified within the lymphocyte gate by positive expression of CD56 (CD56 + ) (either dim or bright) (IV and VI).Myeloid cells were further analyzed for the expression of CD14 (VII).Quantification of PD-L1 + cells was performed after removing dead cells from CD45 + or CD45 -populations, and according to gate drawn using a Fluorescence-Minus-One control (red peak on PD-L1 plot) (VIII).Panel 1 was used for Figure 3 and Panel 2 was used for Figures 5 and 6 and Supp.Fig. 4, 5 and 6.B) NK cell gating (stage VI of the gating strategy) for four representative biological replicates of dual anti-HER2 antibody challenge on encapsulated co-cultures of SKBR3 and PBMCs (quantified in Fig. 5), depicting NK cell populations before co-culture (day 0, upper row) and after 4 days of antibody challenge (day 4, bottom row).The 4-day dual anti-HER2 antibody challenge induced a modulation of the NK cell compartment.Co-cultures were supplemented with IL-15.C) Assessment of the percentage of biological replicates in which SKBR3 cell death (measured by LDH leakage) was significantly higher in direct co-culture of NK cells and SKBR3 cancer cell aggregates, with or without antibody challenge, at different E:T ratios demonstrated that NK cell basal cytotoxic ability (lighter shade) could be boosted in presence of Tmab (darker shade).Percentage of biological replicates in which tumor cell death was significantly (p<0.05)higher: in the presence of NK than in negative control with only cancer cells (no antibody) or in the presence of NK + Tmab than in the presence of NK cells only (Tmab).The percentage of immune cells from distinct donors that responded in at least one of the ratios tested (At least one ratio) is also plotted.NK cells from 6 different donors were assayed independently.

D)
Assessment of the percentage of biological replicates in which SKBR3 (yellow) or HCC1954 (blue) cell death (measured by LDH leakage after 4 days of challenge in 2D co-culture), at the indicated E:T ratios, was significantly (p<0.05)higher in the presence of NK cells than negative control with only cancer cells, revealed that NK cells exhibit basal cytotoxicity against both HER2 + breast cancer cell lines in 2D.The percentage of biological replicates displaying increased cytotoxicity in at least one of the ratios tested (At least one ratio) is also plotted.Plot indicates percentage of biological replicates in which tumor cell death was significantly (p<0.05)higher in in the presence of NK cells than negative control with only cancer cells, evaluated by unpaired t-test for each of the E:T ratios, for each of the cell lines.NK cells from 2 different donors, assayed independently.

E)
Assessment of the percentage of biological replicates in which SKBR3 (yellow) or HCC1954 (blue) cell death (measured by LDH leakage after 4 days of challenge in 2D), at the indicated E:T ratios, was significantly (p<0.05)higher in the presence of NK cells + Tmab than in the presence of NK cells only demonstrated that HCC1954 cells are less susceptible than SKBR3 to trastuzumab-induced, NKmediated ADCC in 2D.The percentage of biological replicates displaying increased cytotoxicity in at least one of the ratios tested (At least one ratio) is also plotted.Plot indicates percentage of biological replicates in which tumor cell death was significantly (p<0.05)higher in the single antibody challenge than without antibody, evaluated by unpaired t-test for each of the E:T ratios, for each of the cell lines.NK cells from 2 different donors, assayed independently.different immune donors.Pairwise statistical comparisons, relative to co-cultures with IL-15 and without antibody challenge, evaluated by a paired t-test: all comparisons are non-significant (p>0.05).expression in the myeloid population upon 4 days of single and dual anti-HER2 antibody challenge.Co-cultures supplemented with IL-15.D) Percentage of PD-L1 + live lymphoid cells (PD-L1 + in the total SSC low /FSC low lymphoid cell population), detected by flow cytometry.The 4-day single and dual anti-HER2 antibody challenge induced an upregulation of PD-L1 in the lymphoid compartment.Co-cultures supplemented with IL-15.

E)
Percentage of live myeloid cells (SSC hi /FSC hi in the CD45 + DAPI -total viable immune cell population, left Y axis, black) and the percentage of CD14 + monocytic cells (CD14 + in the total SSC hi /FSC hi CD45 + DAPI -live myeloid cells, right Y axis, blue), detected by flow cytometry.The 4day single and dual anti-HER2 antibody challenge induced a decline in the proportion of myeloid cells within the immune compartment, particularly the CD14 + monocytic cells.Co-cultures supplemented with IL-15.
factors detected in all experimental replicates and conditions tested are highlighted in green shade.ANG: angiogenin; OSM: oncostatin M; TPO: thrombopoietin.B) Semi-quantitative chemiluminescent detection of cytokines and chemokines in the supernatant of this model, represented as log2-transformed fold change of the "dual antibody challenge" over "no antibody", reveals heterogeneous soluble factor secretion phenotypes across distinct PBMC donors.N = 4 independent experiments, performed with different immune donors.
NK cells A) Direct cytotoxicity assay performed with SKBR3 aggregates and NK cells at different effector (NK) : target (cancer cell) ratios (E:T; 2:1, blue; 5:1, green; 10:1, purple), by assessment of LDH leakage after 4 days of challenge.Tested conditions included negative (cancer cells only) control; positive (cancer cells with lysis buffer) control, to determine maximal cancer cell death (max death); no challenge and single (trastuzumab, Tmab) antibody challenge, assayed at the different E:T.Data from one experiment representative of those depicted in (C); bars represent mean ± S.D. of five technical replicates.Pairwise statistical comparisons between indicated groups (*) or between the indicated condition and negative control (#) performed with unpaired t-test: ##, p<0.01, ***, p<0.001.B) Assessment of LDH leakage derived from spontaneous or maximum NK cell death in the context of the direct cytotoxicity assay described in (A) indicates that the contribution of NK cell death to the final assay readout is residual.Bars represent mean ± S.D. of five technical replicates.