Targeting B7-H3 in patients with metastatic triple-negative breast cancer and brain metastases: an opportunity to “revive” immunotherapy as a treatment strategy?

Abstract

Approximately one-third of patients with metastatic triple-negative breast cancer (TNBC) develop brain metastases during their lifetime. Notably, up to 90% of TNBC brain metastases express B7-H3, a molecule that contributes to an immunosuppressive microenvironment—positively correlating with regulatory T cells and negatively with CD8+T cells. We propose that B7-H3-targeted antibody-drug conjugates (ADCs) may help convert these immunologically “cold” tumors into “hot” ones, thereby enhancing the responsiveness of patients with metastatic TNBC and brain metastases to immunotherapy. The ability of ADCs to penetrate a blood-brain barrier that is disrupted by brain metastases and/or associated local therapies generates considerable excitement for this potential approach.

We further describe a search of the clinicaltrials.gov website, which yielded eight phase I/II trials examining the efficacy of B7-H3 ADCs, 12.5% (n=1/8) of which exclude patients with brain metastases and 62.5% (n=5/8) of which include patients with asymptomatic brain metastases. Given the natural progression of metastatic TNBC and the high incidence of brain metastases, future studies of B7-H3-targeted therapies should consider the broad inclusion of patients with brain metastases. Additionally, the potential of B7-H3-targeted ADCs to prevent brain metastases warrants investigation in both preclinical models and clinical trials.

Main text

Brain metastases occur in approximately one-third of patients with metastatic triple-negative breast cancer (mTNBC) and confer a poor prognosis, highlighting the need for novel therapeutic strategies in this population. Early immunotherapy clinical trials for patients with heavily pretreated and biomarker unselected metastatic breast cancer failed to demonstrate efficacy. However, when used in the first-line setting for patients with metastatic TNBC and programmed death-ligand 1 (PD-L1) positive disease, pembrolizumab in combination with chemotherapy significantly prolonged both progression-free (PFS) and overall survival. Although this is a significant advancement in standard-of-care therapy for patients with mTNBC, only ~40% of patients are candidates for immunotherapy based on expression of the predictive biomarker PD-L1.¹ Further, the activity of immunotherapy for patients with mTNBC and brain metastases remains under-explored. Less than 5% (n=33/750) of patients in the KEYNOTE-355² and 7% (n=86/1271) in the IMpassion130³ clinical trials had stable treated brain metastases, respectively; and no patients with active brain metastases were included in either study.^{2 3}

While immunotherapy trials for brain metastases from breast cancer are limited, the addition of atezolizumab to pertuzumab and high-dose trastuzumab in a heavily pretreated patient population with HER2+brain metastases has been explored. However, this trial closed prematurely due to futility after enrolling 19 patients (NCT03417544).⁴ In the TNBC setting, the addition of immunotherapy to standard brain radiotherapy (eg, NCT03449238, NCT03807765) has been evaluated. Ahmed et al evaluated the safety and efficacy of nivolumab combined with stereotactic radiosurgery (NCT03807765) in 12 patients across all subtypes of breast cancer (half of whom had TNBC) with ≤10 brain metastases.⁵ The rate of local and distant intracranial control at 12 months was 94% and 33%, respectively. The median time to distant intracranial progression was 7.4 months, which is intriguing given that all patients received at least one prior line of systemic therapy and one-third of patients received ≥2 prior lines. It is notable that no cases of radiation necrosis were identified with a median follow-up of 56 months.⁵ While this innovative trial design mitigates risks associated with potential brain metastases progression, intracranial response cannot be evaluated given that brain metastases are no longer measurable after radiation therapy. In addition, a phase 2 trial evaluated the use of pembrolizumab in patients with a variety of solid tumors who had brain metastases either previously untreated (n=9) or previously treated but recurrent and progressive (n=48) brain metastases.⁶ Among 11 patients with TNBC, 4 were reported to have intracranial benefit (all of whom had stable disease as a best intracranial response).⁶ In this trial, patients were quite heavily pretreated (median three lines of systemic therapy); unfortunately, by the time brain metastases occur among patients with mTNBC, most patients would already have been exposed to pembrolizumab for PD-L1-positive disease.⁷ Thus, a new immunotherapy strategy for patients with mTNBC and brain metastases would need to overcome immunotherapy resistance in PD-L1 positive populations, and ideally offer hope of activity in PD-L1 negative populations.

Programmed cell death protein 1 (PD-1) is a cell-surface protein primarily expressed on activated T cells, B cells, and some myeloid cells. Its main ligand PD-L1 is expressed on tumor and various immune cells within the tumor microenvironment. The interaction between PD-L1 and PD-1 induces the inhibitory pathways to T cells, leading to a reduction in proliferation, cytokine secretion, and cytotoxic function. Thus, monoclonal antibodies blocking this pathway restore T-cell activation and enhance antitumor immune response. Resistance to immune checkpoint inhibitors (ICIs) stems from both intrinsic and acquired mechanisms, which was recently reviewed by Li et al.⁸

Several predictive biomarkers of response to immunotherapy have already been identified, including the presence of stromal tumor-infiltrating lymphocytes (sTILs), high tumor mutational burden (TMB) and the expression of the PD-L1 protein.⁹ The presence of sTILs in TNBC is associated with improved prognosis and more favorable response to ICIs.⁹ In addition, a higher TMB (the total number of somatic mutations per coding area of a tumor genome) has been associated with an increased likelihood of response to ICIs across various metastatic solid tumors. A study of patients with resected brain metastases from a variety of solid tumors, including breast, melanoma, and lung, demonstrated that a low TMB (<5.2 mutations/Mb) was associated with significantly shorter time to local progression, as well shorter time to local re-occurrence when treated with ICIs postoperatively, compared with patients with a high TMB.¹⁰ While data specific to mTNBC and brain metastases are still limited, pembrolizumab is broadly approved for patients with metastatic solid tumors with high TMB.

An emerging therapeutic target for patients with mTNBC is B7-H3, a member of the PD-1 family of the immune checkpoint family. B7-H3 is overexpressed in many solid tumors, including TNBC, lung cancer, melanoma and prostate cancer, while its expression in normal tissue remains relatively low, thereby increasing its therapeutic potential.¹¹ In contrast to PD-L1, expression of B7-H3 in solid tumors has been associated with resistance to immunotherapy; however, the biology of B7-H3 is not so clearly or consistently immunosuppressive as PD-L1 and its downstream signaling cascade has not been definitively established.^{12 13} In addition, the oncogenic effects of B7-H3 may be more predominantly immune independent (eg, epithelial-mesenchymal transition and shift to a glycolytic metabolism), which in turn have been associated with increased propensity for metastasis, treatment resistance, and poor prognosis.^{14 15}

Nevertheless, preclinical studies targeting B7-H3 in TNBC models have successfully demonstrated tumor regression, and the addition of an ICI enhanced these effects.¹¹ In addition, clinical efficacy has been demonstrated in a phase I/II study of enoblituzumab (anti-B7-H3 antibody) and pembrolizumab among 133 patients with non-small cell lung cancer, head and neck squamous cell carcinoma, urothelial cancer, and melanoma.¹⁶ Such studies raise the tantalizing possibility that B7-H3 could help convert immunologically “cold” tumors into “hot” tumors, thereby improving sensitivity to immunotherapy. Brain metastases are considered “cold” due to their low TIL infiltration and high levels of tumor-associated macrophages, which inactivate PD-L1 on T cells and generate an immune suppressive environment, among other reasons.¹⁷

The observation that B7-H3 is highly expressed, not only in primary breast cancers but also brain metastases, is of significant clinical interest. In a study by Joshi et al, B7-H3 positivity, as determined via immunohistochemistry, was observed in 82% of primary breast cancer and 85% of brain metastases.¹⁸ Approximately 90% of the 22 triple-negative brain metastases expressed B7-H3.¹⁸ Since B7-H3 may contribute to an immune-suppressive microenvironment, characterized by its positive correlation with regulatory T cells and negative association with CD8+T cells in breast cancer (particularly in TNBC),¹⁸ there is a strong rationale to explore combinations of B7-H3-targeted antibody drug conjugates (ADCs) with immunotherapy for the treatment of brain metastases.

The ability of ADCs to penetrate a blood-brain barrier that is disrupted by brain metastases and/or associated local therapies generates considerable excitement for this potential approach. For example, in the DESTINY-Breast12 trial, the HER2-targeted ADC trastuzumab deruxtecan demonstrated an impressive intracranial response rate of 62.3% among patients with HER2+active brain metastases.¹⁹ In addition, B7-H3-directed chimeric antigen receptor-T cells have demonstrated promising activity in patients with primary central nervous system (CNS) tumors.20,23

Multiple phase I/II clinical trials using B7-H3 ADCs are currently ongoing, but published results are not yet available. In a search of the clinicaltrials.gov website, we identified that most breast cancer trials focusing on B7-H3 either exclude patients with brain metastases or have restrictive enrollment stipulations. A search of “breast cancer” and “B7-H3” yielded eight studies, 12.5% (n=1/8) of which exclude patients with brain metastases and 62.5% (n=5/8) include patients with brain metastases that are asymptomatic. Further, all but one of these five trials mandates stable CNS disease. Such enrollment restrictions are problematic as most brain metastases from breast cancer are symptomatic due to the absence of routine screening programs for early detection in the real-world setting. Given the high expression of B7-H3 in brain metastases and the well-known intracranial efficacy of ADCs, the scientific community has an opportunity to improve patient outcomes by conducting well-designed trials that broadly include patients with both stable and active brain metastases.

In addition to therapeutic potential in mTNBC, B7-H3 targeted therapies may warrant evaluation in patients with early-stage disease in the future. As recently demonstrated for patients with early-stage melanoma, neoadjuvant therapy with dual anti-lymphocyte-activation gene 3(LAG-3) therapy (relatlimab) and anti-PD-1 (nivolumab) agents increased the rates of major pathological response (MPR) and improved PFS compared with nivolumab monotherapy.¹⁹ Importantly, higher expression levels of B7-H3 were observed in patients who did not achieve MPR, and this correlated with a shorter PFS,²⁴ further highlighting B7-H3 as a potentially actionable target to overcome resistance to immunotherapy.

In conclusion, several studies, both translational and clinical, have demonstrated the potential of using ADCs and ICIs in the treatment of patients with mTNBC brain metastases. Given the natural history of TNBC and its pattern of progression, future studies should evaluate B7-H3 targeted therapies in patients with brain metastases and explore their role in future prevention strategies.