Activation of a cGAS-STING-mediated immune response predicts response to neoadjuvant chemotherapy in early breast cancer
In this clinical study, we have validated a 44 gene expression based DDIR signature as a predictor of response to neoadjuvant chemotherapy in breast cancer, which can be integrated into clinical workflows, and potentially identifies a subgroup of cancers responsive to anthracycline treatment.
Breast cancer is increasingly treated with neoadjuvant chemotherapy, prior to surgery, with the aim of reducing tumour size, enabling less extensive surgery to be carried out. In addition, response to neoadjuvant systemic therapy is highly prognostic for the individual patient and allows the tailoring of further systemic therapy after surgery. Although there are a broad range of molecular subtypes of breast cancer, currently chemotherapy regimens are largely selected regardless of tumour subtypes, and there are no good biomarkers to predict response to specific treatment regimens.
From a clinician’s point of view this is frustrating – it would be valuable to identify patients likely to respond to chemotherapy given before surgery and achieve a complete pathological response (pCR), with no viable invasive tumour within the breast. Previously, our colleagues had developed a 44 gene expression-based assay that identifies tumours with a DNA repair deficiency, that are most likely to respond to chemotherapy, which primarily works via damaging DNA. This signature is characterised by up-regulation of a group of genes associated with immune activation and has been termed the DNA damage immune response (DDIR assay)1. The DDIR assay, which works across all molecular subtypes of breast cancer, has been retrospectively validated to predict response to DNA damaging chemotherapy in both the neoadjuvant and adjuvant settings, and furthermore has been associated with favourable disease-free and overall survival in patients with triple negative breast cancer treated with adjuvant anthracycline/cyclophosphamide regimens2.
Previous work from our lab had shown that this reflected activation of the cGAS-STING pathway, with IRF3 activation resulting in upregulation of chemokines such as CXCL10 and CCL5, driving immune infiltration in these tumours, as illustrated in Figure 13.
To determine the feasibility of using this signature in clinical practice, we carried out a prospective clinical study in patients receiving standard of care neoadjuvant chemotherapy. The study aimed to determine whether we could integrate the assay into our clinical workflows, but also to determine the ability of the assay to predict response to neoadjuvant chemotherapy, and potentially to explore the biology underlying treatment response or non-response by collecting serial samples during and after treatment. Studies like this can be challenging to recruit to, as there is no direct benefit to patients taking part, but there is a need to undergo the discomfort and inconvenience of additional biopsies during treatment. As always, we were taken aback with the altruism shown by our patients, and 50 patients from two hospitals in Northern Ireland agreed to take part in this study.
We showed that it was clearly possible to use the assay with FFPE core biopsy samples taken at diagnosis. Furthermore, the assay did indeed clearly predict response to neoadjuvant chemotherapy, with an odds ratio for pCR of 6.60 (95% CI 1.334-32.2, p=0.0217) for DDIR positive patients compared to DDIR negative tumours. Interestingly, in our study the assay was once again shown to outperform the ability of tumour infiltrating lymphocytes (TILs) to predict treatment response in this cohort, consistent with previous reports in triple negative breast cancer2.
After completing the study and assessing the primary feasibility endpoints, we used RNA sequencing of the serial tumour samples to explore the underlying biology of both DDIR positive and negative tumours. To do this, we were able to utilise the novel claraT total mRNA report developed by our collaborators at Almac Diagnostic Services, which provides a comprehensive overview of tumour biology using both gene expression signatures and single gene analytes, categorised by the Hallmarks of Cancer. Using transcriptomic data, we clearly showed that DDIR positive tumours had significantly higher scores for signatures representing both innate and adaptive immune cell populations, with concomitant upregulation of immune checkpoint genes including CD274 (PD-L1) and IDO1. This “inflamed yet immunosuppressed” microenvironment apparent in DDIR positive tumours suggests that the signature may in fact identify a cohort of patients likely to benefit from treatment with immune checkpoint inhibitors. This would be clinically very valuable as in breast cancer at present, there is no good predictive biomarker to identify patients likely to respond to immune modulation.
Furthermore, using biopsies obtained after three cycles of DNA damaging neoadjuvant chemotherapy, we were able to show that tumours which were DDIR negative at baseline demonstrated an increase in infiltration of adaptive immune cells, together with a marked increase in immune biologies including interferon activation and also increased immune checkpoint gene expression. Whilst this trend didn’t reach statistical significance (only a small number of paired samples were available for assessment), this data is in keeping with previously reported findings in the literature suggesting that treatment with DNA damaging chemotherapy (with agents such as anthracyclines) may be excellent immune priming agents, able to convert immune cold tumours to immune hot, making them the optimal companion to immune checkpoint blockade therapies4,5.
Finally, there is increasing interest in the use of anthracycline-sparing chemotherapy regimens in treating breast cancer, particularly in HER2+ve disease6. Our data however suggests that there are some breast cancers which have an innate DNA repair defect, which can be identified using the DDIR assay, and are particularly sensitive to DNA damaging agents like anthracyclines. Thus, the DDIR assay may potentially act as a biomarker to identify a cohort of patients where the use of anthracycline-sparing regimens should not be considered.
In conclusion, we have demonstrated that it is possible to integrate the 44-gene DDIR assay into standard clinical workflows, with assay results available in a timely fashion to facilitate neoadjuvant chemotherapy decision-making. This biomarker potentially has wide clinical utility in identifying patients where anthracycline-sparing chemotherapy should not be omitted, and in identifying patients where immune checkpoint blockade may be useful. We look forward to carrying out further validation work using this assay in the clinical setting in the near future.
- Mulligan JM, Hill LA, et al. Identification and validation of an anthracycline/cyclophosphamide-based chemotherapy response assay in breast cancer. Journal of the National Cancer Institute 2014;106(1):djt335.
- Sharma P, Barlow WE, et al. Validation of the DNA Damage Immune Response Signature in Patients With Triple-Negative Breast Cancer From the SWOG 9313c Trial. Journal of Clinical Oncology : 2019;37(36):3484-3492
- Parkes EE, Walker SM, et al. Activation of STING-Dependent Innate Immune Signaling By S-Phase-Specific DNA Damage in Breast Cancer. Journal of the National Cancer Institute 2017;109(1):djw199
- Wilkinson RDA, McCabe N, et al. Topoisomerase II inhibitors induce cGAS-STING dependent inflammation resulting in cytokine induction and immune checkpoint activation. bioRxiv 2019:764662.
- Voorwerk L, Slagter M, et al. Immune induction strategies in metastatic triple-negative breast cancer to enhance the sensitivity to PD-1 blockade: the TONIC trial. Nat Med 2019;25(6):920-28.
- Tarantino P, Tolaney SM, et al. Anthracyclines for Human Epidermal Growth Factor Receptor 2–Positive Breast Cancer: Are We Ready to Let Them Go? Journal of Clinical Oncology 2021;39(32):3541-45.