Opening the Door for Immune Oncology Studies in Invasive Lobular Breast Cancer

Invasive lobular breast cancer (ILC) is the second most common histological breast cancer subtype, after invasive ductal breast cancer (IDC). ILC accounts for approximately 10% to 15% of all breast cancers, with approximately 24 000–36 000 cases annually in the US only. The hallmark of ILC is loss of E-cadherin (CDH1), resulting in discohesive cancer cells that infiltrate the breast stroma in a single-file pattern (1). Unique clinical features that differentiate ILC from IDC include more frequent multifocal disease, growth patterns often resulting in higher stage at presentation, positive resection margins necessitating completion mastectomy (2–5), and metastatic sites (6,7). Despite better prognostic factors (eg, more estrogen receptor [ER]+, and lower levels of the proliferation marker Ki67), there is increasing evidence that patients with ILC have worse long-term outcome compared with those with IDC (8–10). Finally, recent studies by The Cancer Genome Atlas (TCGA) (11) and Rationale Therapy for Breast Cancer (12) consortia have demonstrated differences between the genetic make-up of ILC and IDC and—of relevance here—have consistently identified an “immune-related” ILC transcriptomic subtype.

Breast cancer has long been considered a relatively weak immunogenic tumor. However, there is growing evidence that a subset of breast cancer is associated with substantial tumor-infiltrating lymphocytes (TILs), and that their enumeration has prognostic and predictive value. A recent review summarized the analysis of TILs in thousands of breast tumors (13), with increased TILs being associated with improved DFS for patients with Triple Negative Breast Cancer and human epidermal growth factor receptor 2 (HER2)–positive disease. In addition, TILs are associated with higher response rates to neoadjuvant therapy (14), and initial indications are that immunotherapy using pembrolizumab can be effective in a subset of breast cancer patients (15). Until recently, there has been limited evidence describing a role for TILs in ER+ disease. A study by Bense et al. (16) applied CIBERSORT, a computational estimation of relative proportions of 22 subsets of immune cells (17), to publicly available gene expression profiles of 7270 breast tumors and demonstrated correlations similar to those observed in other solid tumors—for example, CD8+ T-cell exhaustion with poor outcome and follicular helper T-cell (Tfh) and increased M1 macrophage signature scores with improved outcome. Similar findings were reported by Ali et al., who used the same deconvolution algorithm analyzing almost 11 000 breast tumors (18). Interestingly, they also described associations between regulatory T cells and poor prognosis in ER+ (and ER-) tumors. These studies not only provide evidence for a role of the host immune system in ER+ breast cancer, but they also represent the largest studies to date describing immune infiltration in relationship to receptor status and molecular subtypes. However, none of these studies compare the immune microenvironment between IDC and ILC.

The first comprehensive analysis of immune cell composition in ILC and its association with clinico-pathological variables is published in this issue of the Journal by Desmedt et al. (19). The authors analyzed samples from a retrospective series of ILC (n = 614) (20) and from BIG 02-98 (n = 149 ILC, n = 807 IDC) (21). For the retrospective series, TILs were assessed by three pathologists, following protocols described by Salgado et al. (22). For BIG 02-98, previously reported data were re-analyzed to compare TILs in IDC vs ILC (21). The retrospective ILC series showed relatively low levels (5%) of TILs in ILC; however, there was a wide range of TIL levels, and 15% of the tumors had greater than 10% TILs. High TIL levels were associated with poor prognostic markers. An association between the TIL number and Ki67 was also seen in BIG 02-98, in both IDC and ILC; however, an association with age and grade was restricted to ILC and IDC, respectively. An additional analysis showed an association of high TIL counts with the mixed, nonclassical subtype, and low TIL counts associated with the alveolar variant. These interesting associations reinforce the need to understand the molecular basis for different histological ILC variants. It will be especially interesting to include pleomorphic ILC (pILC) into future studies, as this is the most aggressive ILC variant, with unique genetic and clinical attributes (23–25).

The authors had previously performed molecular characterization of their retrospective ILC cohort (20), which allowed for correlative analysis of TILs and transcriptomic/genomic data. In addition to associations with some copy number aberrations, ERBB3 mutations were associated with lower TILs, whereas mutations in TP53, ARID1A, and BRCA2 were associated with increased TILs. These results provide the foundation for additional studies linking the molecular make-up of ILC with TILs, especially with respect to the identification of a potential subset of ILC with high tumor mutation burden. This is a priority, given the recent approval of a test for high microsatellite instability (MSI) or mismatch repair (MMR) deficiency as a biomarker for response to pembrolizumab. There is limited information on MSI or MMR deficiencies in ILC, but an older study suggested that some ILCs have relatively high levels of MSI (26), and thus the analysis of MSI/MMR status, and tumor mutation burden in relationship to TIL infiltration, deserves further analysis. Such studies should also address the expression of critical regulators of immune checkpoints, as Thompson et al. recently showed differences in PD-L1 expression between IDC and ILC (27).

Using an independent data set (“Nottingham series”; n = 159 ILC, n = 468 IDC) (28–30), the authors also quantified immune cell subtypes in the intratumoral, adjacent, or distant stromal compartments, comparing ILCs with IDCs. This analysis showed that ILCs are characterized by relatively lower numbers of intratumoral CD8+ cells and adjacent stromal CD3+ cells, as well as lower numbers of CD68+ cells in all compartments. In addition, FOXP3+ cell numbers were lower in the adjacent and distant stroma, but not intratumorally. Using CIBERSORT data from TCGA and METABRIC (18), the authors identified fewer follicular helper and gamma delta T cells, and lower M0, M1, and, to a lesser extent, M2 macrophages in ILC, but more memory B cells, monocytes, and CD8+ T cells. The CIBERSORT analysis, however, doesn’t provide data on the spatial distribution of TILs, and there is increasing evidence that this might be critical for their prognostic power. This is illustrated by a recent study analyzing 1178 ER+ breast tumors from the ATAC trial (31)—there was no prognostic value in the abundance of TILs; however, immune scores based on the spatial heterogeneity of TILs were prognostic, especially for long-term outcome. Desmedt et al. (19) found limited prognostic value of TILs, and the data from the ATAC study provide further strong rationale for additional outcome studies incorporating spatial information.

In summary, there is increasing evidence that the number, location, and phenotypes of intratumoral immune cell infiltrates might correlate with prognosis in particular breast cancer subtypes, indicating the complexity of immune involvement in breast cancer. Furthermore, the disparities in TILs observed by the authors in ILC and IDC, and the apparent differences in the clinical implications of TILs for these histologic subtypes, suggests that there are critical variables that have yet to be elucidated. Future descriptive, mechanistic, and clinical studies, using a variety of complementary approaches, will need to consider histological subtypes, in addition to molecular subtypes, to draw increasingly nuanced conclusions regarding the impact of the immune contexture on breast cancer prognosis and therapy response. The current study by Desmedt et al (19). paves the road for such studies, ultimately offering the field a starting point for future studies on the impact of the immune system in ILC.

Notes

Affiliations of authors: Department of Pharmacology and Chemical Biology (SO) and Section of Breast Surgery, Division of Surgical Oncology, Department of Surgery (PFM), Womens Cancer Research Center, UPMC Hillman Cancer Center, Magee Womens Research Institute, University of Pittsburgh, Pittsburgh, PA; Department of Pathology (PCL) and Department of Immunology (TCB, DAAV), University of Pittsburgh School of Medicine, Pittsburgh, PA; NSABP/NRG Oncology, Pittsburgh, PA (PCL); Tumor Microenvironment Center, UPMC Hillman Cancer Center, Pittsburgh, PA (TCB, DAAV).

The authors have no conflicts of interest to disclose.

Author notes

See the Notes section for the full list of authors’ affiliations.