Effect of pembrolizumab on biomarkers related to intratumoral immunity, proliferation and apoptosis in early breast cancer
Up to thirty percent of patients with early breast cancer (BC) will eventually develop metastases. This results in an ongoing search for treatments to improve the long-term survival of these patients. Immune checkpoint inhibitors (ICI) are one promising type of therapeutic targets and have shown good results in malignancies as melanoma and non-small cell lung cancer. In BC, and especially in triple negative BC (TNBC), a clinical benefit for the addition of ICI to standard therapy has been shown in both the metastatic and early stage of disease. However, only a small subset of patients experience durable and long-term disease control by ICI, with many tumors unlikely to respond. In addition, measurable anti-tumor activity may appear slower in immune therapies than in cytotoxic therapies and lesions may enlarge before shrinking.
Therefore, we designed a window-of-opportunity trial (BioKey) to evaluate the short-term effects of the ICI pembrolizumab in patients with early BC with biomarker determination by comparing multiple markers pre-treatment and at surgery.
Furthermore, a better understanding of the tumor immune microenvironment (TIME) of TNBC and the other molecular subtypes of BC tumors is paramount to identify and implement new biomarkers in ICI BC treatment. Changes related to intratumoral immunity and proliferation may provide evidence for biological activity of pembrolizumab.
The vast majority of studies investigating ICI in patients with BC have focused on TNBC. In our retrospective study, we were able to show that the TIME is very similar between TNBC and HR-negative HER2-positive BC tumors by performing an in-depth characterization of the TIME. A comparative analysis of TNBC and HER2-positive BC revealed no significant differences regarding sTIL and the majority of the investigated immune cell subpopulations (CD3, CD4, CD8, CD73, PD-1, PD-L1). We nevertheless identified significantly higher levels of CD68+ and FOXP3+ cells in TNBC as compared to HER2-positive tumors. Additionally, we performed a similar comparison using publicly available transcriptomics data. Altogether, the results show a comparable TIME in both groups, with possible implications for the use of ICI in patients with HR-negative HER2-positive breast tumors.
BioKey was a single center, prospective, open-label, non-randomized study in patients with a non-metastatic newly diagnosed primary invasive carcinoma of the breast. All patients were treated with one intravenous administration of 200 mg of pembrolizumab at 10 ± 4 days before surgery. The study consisted of 2 cohorts with a total of 54 patients; cohort A included 39 patients who were scheduled for upfront surgery and cohort B included 15 patients with estimated residual tumor size of at least 10 mm on imaging after neo-adjuvant chemotherapy. Patients were monitored carefully for the development of adverse events, which were actively reviewed at 10 ± 4 days and 30 ± 7 days after administration of the study medication.
Single-cell transcriptome, T-cell receptor and proteome profiling was used to compare paired pre- versus on-treatment biopsies. Pre-treatment, one-third of tumors contained PD1-expressing T-cells, which clonally expanded upon anti-PD1. This expansion mainly involved CD8+ T-cells with increased expression of cytotoxic-activity, immune-cell homing and exhaustion markers, as well as CD4+ T-cells with improved T-helper-1 and follicular-helper activity. We designated these as patients with clonotype expansion or ‘expanders’ (E), whereas patients lacking clonotype expansion were considered ‘non-expanders’ (NE). Pre-treatment, immunoregulatory dendritic cells and several macrophage phenotypes expressing PD-L1, and cancer cells exhibiting MHC class I/II expression correlated positively with T-cell expansion. We depicted an immune landscape consisting of various immunophenotypes and associated gene sets being either positively or negatively correlated with T-cell expansion following anti-PD1. Our data suggest novel biomarkers or potential immunotherapeutic modalities to improve clinical benefit from immune checkpoint blockade.
Finally, we investigated the TIME of the expander phenotype with central pathology. We observed that i) higher levels of sTIL, PD-1+ sTIL, PD-L1+ and FOXP3+ cells in the pre-treatment biopsies, ii) higher levels of sTIL, PD-1+ sTIL, PD-L1+, CD3+, CD4+, CD8+, CD68+, FOXP3+ cells and Ki67 in the post-treatment surgical specimen, as well as, iii) higher delta values (surgical specimen – biopsy) of sTIL, CD3+, CD4+, CD8+, CD68+ and Ki67 are associated with the expander phenotype. We were able to detect associations between various cell populations of the tumor immune microenvironment evaluated before and after a single dose of pembrolizumab and the T-cell expansion phenotype previously defined at the single-cell level.
In a chapter we discuss the immune related adverse events (irAEs) after a single dose of pembrolizumab. Twenty-two patients (41%) experienced at least one adverse event in the follow-up period. In general, the irAEs were mild (grade I or II), and in most cases, no intervention was needed. The observed AEs were common AEs in the context of ICI and were mainly related to gastrointestinal disorders, skin and subcutaneous disorders and hyperthyroidism.
Two patients developed late irAEs, which they reported 115 and 101 days after administration of the study medication. The first patient was diagnosed with sarcoidosis and uveitis and the second patient with severe hepatitis. Both patients initially developed hyperthyroidism which later evolved into hypothyroidism. Our observations reinforce the importance of clinical vigilance for late irAEs after ICI discontinuation and imply that merely a single dose of ICI can have a long-lasting immunomodulating effect.
In conclusion, we observed that one in three patients demonstrated pronounced clonotype expansion of PD-1-expressing proliferative T-cells along CD8+ or CD4+ trajectories in the post-treatment samples after one dose of pembrolizumab. We hypothesize that this T cell clonotype expansion is a mechanism of response to pembrolizumab. This pronounced clonotype expansion was not limited to patients with TNBC, but was also observed in patients with HR-positive HER2 positive and HR positive BC.
Clonotype expansion is clearly measurable via single-cell profiling. The immunohistochemistry data showed associations between the clonotype expansion status, predefined based on single cell profiling, and levels of sTIL and expression of various other immune markers. Nonetheless, we were not able to predict the clonotype expansion status of patients by immunohistochemistry alone. A new neo-adjuvant clinical trial has been designed and will start soon to allow the validation of our findings and to evaluate if T-cIell clonotype expansion is indeed associated with clinical benefit to ICI.