Abstract

   Breast cancer is a heterogeneous disease with distinct molecular
   subtypes that disproportionately affects Black women. Immune cells are
   a key component of the tumor microenvironment, influencing tumor growth
   and treatment outcomes. Here, we explored immune landscape differences
   between TNBC and non-TNBC subtypes, assessing any race-specific
   patterns. TNBC showed higher infiltration of B-cells, Treg cells, Th1
   cells, and CD8+ cells, and fewer mast cells than non-TNBC. Race-wise
   comparisons revealed that White TNBC had more Th1 cells than Black
   TNBC, while Black non-TNBC exhibited higher NK and Treg cells but lower
   DCs. KEGG pathway analysis identified immunosuppression in TNBC, with
   Black patients exhibiting the same regardless of molecular subtype.
   Higher TAM and lower T-cell infiltration were linked to metastatic
   disease. In White patients, lower immune cells (particularly T-cells,
   DCs, and NK cells) correlated with more metastasis, but not in Black
   patients. These race- and subtype-specific immune differences may guide
   tailored immunotherapies.

   Subject terms: Breast cancer, Cancer microenvironment, Tumour
   immunology

Introduction

   Breast cancer (BC) is the most frequently diagnosed malignancy in women
   worldwide^[42]1. Approximately 316,950 women will be diagnosed with BC
   this year in the United States alone, and nearly 42,170 will succumb to
   the disease^[43]2. At the molecular level, BC is broadly classified
   into four molecular subtypes, luminal A [estrogen and progesterone
   receptor positive (HR^+), human epidermal growth factor receptor 2
   negative (HER2^−)], luminal B (HR^+ HER2^−/+ with high Ki67), HER2
   enriched (HR^− HER2^+), and triple-negative breast cancer (TNBC or HR^-
   HER2^−)^[44]3,[45]4. Among these, luminal subtypes have the best
   prognosis, followed by the HER2-enriched subtype due to the
   availability of targeted therapies. TNBC has the worst prognosis
   because of its highly aggressive nature and limited availability of
   effective therapeutic options^[46]5,[47]6.

   Significant health disparities are observed in patients diagnosed with
   BC; especially Black women appear to bear the greatest burden. Although
   the BC death rate has declined by approximately 44% over the past
   decade, it has remained largely unchanged in Black women for the past
   several decades. Black women experience 38% higher mortality compared
   to their White counterparts despite having a 5% lower incidence
   rate^[48]6,[49]7. Moreover, they are diagnosed with BC at a younger age
   and have nearly twice the frequency of diagnosis with the TNBC subtype
   compared to White women^[50]8,[51]9. Black women diagnosed with the
   luminal subtype are also more likely to relapse after endocrine
   therapy, further adding to their elevated mortality rates^[52]10.
   Clearly, there is a critical need to deepen our understanding of the
   molecular pathobiology of BC and develop new mechanism-based therapies
   to improve its clinical management and address existing disparities.

   Immune system-targeted therapies have gained high prominence in recent
   years and have shown great promise for the clinical management of
   various cancers^[53]11–[54]13. Immune cells are a crucial component of
   the tumor microenvironment (TME) that play a significant role in tumor
   development and therapeutic outcomes^[55]14,[56]15. Cancer and other
   stromal cells residing in TME modulate the immune cell infiltration and
   their functional phenotypes in a way that creates an immunosuppressive
   microenvironment, helping the tumor cells evade immune killing.
   Understanding these aspects can help develop insights into BC
   pathobiology and holds promise for immune therapeutic management. In
   the present study, we analyzed the immune landscape of breast tumors
   and compared the differences across race or BC molecular subtypes. We
   also determined the immune signaling pathways that were differentially
   regulated in these tumors and explored the relationship between immune
   infiltration and tumor metastasis.

Results

TNBC and non-TNBC subtypes exhibit differential tumor immune microenvironment

   To investigate the differences in the immune landscape between TNBC and
   non-TNBC molecular subtypes, we performed immune cell profiling on 81
   BC tumor samples (55 non-TNBC and 26 TNBC) using the nanoString
   PanCancer immune panel on the nCounter® pro analysis system (Fig.
   [57]1A). The immune cell abundance score was determined for each sample
   by the Rosalind software based on average expression values of their
   corresponding marker genes. Unsupervised clustering based on the
   abundance score of different immune cell types grouped all the samples
   into four major clusters (BC Clusters 1–4) in the heat map, with
   Cluster 1 and Cluster 4 exhibiting low and high tumor-infiltrating
   lymphocytes (TILs), respectively, and Clusters 2 and 3 having moderate
   TILs (Fig. [58]1B). Immune cells were also clustered into three major
   clusters (IC-Clusters 1–3). IC-Cluster 1 consisted of cells of myeloid
   lineage, such as tumor-associated macrophages (TAMs), neutrophils, and
   dendritic cells (DC), except for mast cells, which were placed alone in
   IC-Cluster 3. All the immune cells of lymphoid lineage, like T-cells,
   B-cells, regulatory T-cells (Tregs), CD8+ T-cells, Th1 cells, and
   natural killer (NK) cells, were grouped in IC-Cluster 2 (Fig. [59]1B).
   Notably, the distribution of TNBC and non-TNBC samples within the
   BC-Clusters revealed that the moderate and high TILs clusters contained
   a higher percentage of TNBC, while a greater percentage of non-TNBC
   were placed in the low TILs cluster (Fig. [60]1C).

Fig. 1. Distinct tumor immune microenvironments in TNBC and non-TNBC BC
subtypes.

   [61]Fig. 1
   [62]Open in a new tab

   A Graphical presentation of the working strategy used for immune cell
   abundance analysis in the tumor samples. B Heat map showing the
   unsupervised hierarchical clustering of immune cell types and TNBC and
   non-TNBC cases based on immune abundance scores and patterns. C
   Percentage distribution of TNBC and non-TNBC samples in low, moderate,
   and high immune infiltration clusters obtained from the heatmap.
   Scatter plot showing the Pearson correlation coefficient (r) of mast
   cells with DC (D), TAM with neutrophils (E), B-cells with T-cells (F),
   and B-cells with Treg cells (G) infiltration, calculated from the
   immune abundance scores across total samples. (H and I) Scatter plot
   showing the Pearson correlation coefficient (r) for infiltration of
   B-cells with Treg cells (H), NK cells with Treg cells, CD8+ T-cells,
   and Th1 cells, calculated from the immune abundance scores in TNBC and
   non-TNBC samples. Comparison of the immune abundance score among TNBC
   and non-TNBC samples for overall immune infiltration represented by
   CD45 positive cells, B-cell, mast cells, T-cells, TAM, neutrophils, DC,
   and NK cells (J), Treg cells, Th1 cells, and CD8+ T-cells (K). A
   non-parametric Mann–Whitney test was used to determine the p-value. A
   p-value < 0.05 was considered statistically significant.

   Since in our unsupervised clustering analysis, mast cells did not
   cluster with other cells of the myeloid lineage, we examined their
   correlative abundance with other immune cell types using the Pearson
   correlation analysis. A significant positive correlation (r = 0.32,
   p = 0.003) was observed with DC infiltration only (Fig. [63]1D) in both
   TNBC (r = 0.40, p = 0.039) and non-TNBC (r = 0.45, p < 0.001) (Fig.
   [64]S1A). Among other cell types of myeloid origin, we observed the
   strongest significant correlation between TAM and neutrophils
   (r = 0.61, p < 0.001) (Fig. [65]1E) in both TNBC (r = 0.66, p < 0.001)
   and non-TNBC (r = 0.60, p < 0.001) (Fig. [66]S1B). A significant
   positive correlation was also observed between TAM and DC (r = 0.58,
   p < 0.001) but not between DC and neutrophils (r = 0.18, p = 0.099)
   (Fig. [67]S1C, D). Among lymphoid-origin immune cells, the highest
   significant correlation was observed between B-cells and T-cells
   (r = 0.85, p < 0.001) (Fig. [68]1F) in both TNBC (r = 0.88, p < 0.001)
   and non-TNBC (r = 0.80, p < 0.001) (Fig. [69]S1E). Among the T cell
   subtypes, Tregs showed the highest correlation with B-cells (r = 0.66,
   p < 0.001) compared to Th1 (r = 0.48, p < 0.001) and CD8 + T-cells
   (r = 0.57, p < 0.001) (Figs. [70]1G and [71]S1F), which was notably
   stronger in TNBC (r = 0.84, p < 0.001) than in non-TNBC (r = 0.48,
   p < 0.001) (Fig. [72]1H). Besides, NK cells correlated strongly with
   Tregs in TNBC, while they showed better correlations with CD8 + T-cells
   and Th1 cells in non-TNBC (Fig. [73]1I). In overall abundance score
   comparisons between TNBC and non-TNBC, a significantly higher
   (p = 0.047) presence of lymphocytes represented by CD45 positivity was
   observed in TNBC than in non-TNBC. TNBC also had significantly higher
   B-cell infiltration (p = 0.025) and a strong trend for higher T-cells
   (p = 0.051) and TAM (p = 0.066) infiltration as well; however, the mast
   cell infiltration was significantly lower in TNBC than in non-TNBC
   (p < 0.001) (Fig. [74]1J). Among T-cell subpopulations, Tregs
   (p = 0.002), Th1 cells (p = 0.023), and CD8+ T-cells (p = 0.004)
   infiltration was significantly higher in TNBC compared to non-TNBC
   (Fig. [75]1K). When we compared immune cell infiltration in non-TNBC
   subtypes between hormone receptor-positive (HR+) and hormone
   receptor-negative (HR−), HR− samples had a significantly higher Treg
   cell infiltration (p = 0.010) than HR+ samples (Fig. [76]S1G).

Racial variations in the abundance of the TILs in TNBC and non-TNBC samples

   Next, we examined whether the pattern of immune cell infiltration in
   TNBC and non-TNBC subtypes varied between White and Black patients. For
   this, we performed the unsupervised clustering of TNBC cases from White
   and Black patients and examined the sample distribution in clusters
   with varying levels of TILs. Analysis created two distinct clusters of
   TNBC with relatively high TILs (TNBC-Cluster 1) and low TILs
   (TNBC-Cluster 2). The immune cells were clustered into four clusters
   (IC Clusters 1–4), with IC-Cluster 1 having Mast cells, IC-Cluster 2
   having other myeloid cells, IC-Cluster 3 having DC, and IC-Cluster 4
   having all other immune cells of lymphoid lineage (Fig. [77]2A).
   However, TNBC from Black and White patients showed equal distribution
   in both high and low TILs clusters (Fig. [78]2B). Moreover, we also did
   not observe significant differences in the immune cell abundance scores
   between the Black and White patients (Fig. [79]2C). Interestingly,
   among the T-cell subpopulations, we observed a significantly higher
   abundance of Th1 cells (p = 0.008) in White TNBC patients along with a
   trend for higher CD8+ T-cell abundance (p = 0.054) compared to the
   Black TNBC. No significant difference in the abundance of Treg cells
   (p = 0.462) was observed (Fig. [80]2D). Since the infiltration of Th1
   cells and CD8+ T-cells was higher in White than Black patients, we
   checked the correlation of these cells with other infiltrating cells.
   We found that Th1 cell infiltration was significantly correlated with
   B-cells (r = 0.78, p < 0.001), NK cells (r = 0.76, p = 0.001) and DC
   infiltration (r = 0.54, p = 0.046) in Black patients, while this
   correlation was not present in the White patients (B-cells r = 0.11,
   p = 0.713; NK cells r = 0.14, p = 0.656; DC r = 0.18, p = 0.557) (Fig.
   [81]2E). The CD8 + T-cells also showed significant positive correlation
   with B-cells in Black patients (r = 0.70, p = 0.004) and with NK cells
   in both White (r = 0.71, p = 0.008) and Black patients (r = 0.85,
   p < 0.001) (Fig. [82]S2A).

Fig. 2. Black TNBC samples show reduced infiltration of Th1 and CD8+ T-cells.

   [83]Fig. 2
   [84]Open in a new tab

   A Heat map showing the unsupervised clustering of Black and White TNBC
   samples and immune cell types based on immune abundance scores in the
   tumor. B Percentage distribution of Black and White TNBC samples in low
   and high immune infiltration clusters from the heatmap. Comparison of
   the immune abundance score among Black and White TNBC samples for total
   immune cells (CD45+), TAM, B-cells, T-cells, NK cells, neutrophils,
   mast cells, and DC (C), and in T-cells subtype cells i.e., Th1, CD8+
   T-cells and Treg cells (D), non-parametric Mann–Whitney test was used
   to determine the p-value. E Scatter plot showing the Pearson
   correlation coefficient (r) for the association between infiltration of
   Th1 cells with B-cells, NK cells, and DC in Black and White TNBC
   samples; the correlation was calculated from the abundance score of
   immune cells in the tumor tissues. A p-value < 0.05 was considered
   statistically significant.

   In the non-TNBC samples from Black and White patients, the unsupervised
   hierarchical clustering created two distinct clusters, non-TNBC-Cluster
   1 and 2, with relatively high and low TILs, respectively. Immune cells
   clustered into two major clusters, with all lymphoid origin cells being
   clustered in IC-Cluster 1, and TAM and DC of myeloid lineage clustering
   together in IC-Cluster 2. Neutrophils, NK cells, and mast cells were
   placed independently in IC-Cluster 3, IC-Cluster 4, and IC-Cluster 5,
   respectively (Fig. [85]3A). A slightly higher distribution (53.57%) of
   White non-TNBC was observed in the high TILs cluster, while the reverse
   was true for Black non-TNBC (48.15%) (Fig. [86]3B). When abundance
   scores for immune cell types were compared between Black and White
   non-TNBC, we found NK cells to be significantly higher in Black
   patients (p = 0.005), whereas DC cells were significantly lower
   (p = 0.021) in these patients, compared to the White patients (Fig.
   [87]3C). Among the T-cell subtypes, the Tregs were significantly higher
   (p = 0.011) in Black non-TNBC compared to White patients (Fig. [88]3D).
   When we compared immune cell infiltration in HR+ and HR- subtypes in
   Black versus White non-TNBC patients, a significantly higher
   infiltration of NK cells (p = 0.003) and Tregs (p = 0.016) in Black HR+
   subtype was observed. Further, a trend for lower DC (p = 0.064) in HR+
   and lower infiltration of mast cells (p = 0.057) and CD8+ T-cells
   (p = 0.057) in HR- subtype in Black non-TNBC were also observed as
   compared to their respective white non-TNBC (Fig. [89]3E).

Fig. 3. Black non-TNBC tumors have higher Treg cell infiltration.

   [90]Fig. 3
   [91]Open in a new tab

   A Heat map showing the unsupervised clustering of Black and White
   non-TNBC samples and immune cell types based on immune abundance scores
   in the tumor. B Percentage distribution of Black and White non-TNBC
   samples in low and high immune infiltration clusters from the heatmap.
   Comparison of the immune abundance score among Black and White non-TNBC
   samples for total immune cells (CD45+), TAM, NK cells, DC, B-cells,
   T-cells, mast cells, and neutrophils (C), and in T-cells subtype cells,
   i.e., Th1, CD8 + T-cells, and Treg cells (D). E Fold change in immune
   abundance score and corresponding p-values in Black samples compared to
   White counterpart in HR^+ and HR^− BC subtypes. The p-value was
   calculated using the non-parametric Mann-Whitney test. A p-value < 0.05
   was considered statistically significant.

TNBC tumors exhibit greater immunosuppression compared to non-TNBC tumors
despite higher TILs

   From a total of 770 immune-related genes in the panel, we observed
   differential expression of 152 genes between TNBC and non-TNBC samples,
   of which 122 were upregulated and 30 downregulated (Fig. [92]4A, [93]B,
   and Supplementary Table [94]S1). The KEGG pathway enrichment analysis
   using Enrichr web-based tool showed the primary immunodeficiency and
   PD-L1 expression and PD-1 checkpoint pathways to be among the top
   upregulated pathways in TNBC, whereas complement and coagulation
   cascade, c-type signaling, and Th17 cell differentiation pathways were
   among the top downregulated pathways (Fig. [95]4C). Immunosuppressive
   genes such as FOXP3, PDCD1, ICOS, LAG3, CTLA4, IDO1 were significantly
   over-expressed in TNBC, while genes involved in immune activation such
   as GATA3 and IKBKB were significantly repressed (Fig. [96]4D). These
   findings indicate that although TNBC have relatively higher TILs than
   non-TNBC, they exhibit higher immunosuppression and are thus good
   candidate for immune checkpoint therapies.

Fig. 4. TNBC tumors exhibit greater immunosuppression than non-TNBC tumors.

   [97]Fig. 4
   [98]Open in a new tab

   A Volcano plot representing the up-regulated (red dots) and
   down-regulated (blue dots) genes in TNBC compared to the non-TNBC
   samples. B Heat map representing the expression pattern of the
   differentially expressed genes among the TNBC and non-TNBC samples. C
   KEGG pathway analysis of the differentially expressed genes showing
   up-regulated and down-regulated pathways in TNBC compared to non-TNBC
   samples. D Representation of selected up-regulated genes associated
   with immunosuppression (FOXP3, PDCD1, ICOS, LAG3, CTLA4, IDO1) and
   down-regulated genes associated with immune activation (GATA3, IKBKB)
   in TNBC compared to non-TNBC samples. The p-value was calculated using
   the non-parametric Mann-Whitney test. A p-value < 0.05 was considered
   statistically significant.

TNBC and non-TNBC in Black patients exhibit greater immunosuppression than in
white patients

   In separate race-specific comparisons within TNBC, we found a total of
   95 differentially expressed genes, of which 80 were upregulated and 15
   downregulated in Black patients compared to the White patients (Fig.
   [99]5A). The heatmap shows the gene expression pattern across all
   samples (Fig. [100]5B and Supplementary Table [101]S2). The KEGG
   pathway enrichment analysis showed cytokine-cytokine receptor
   interaction, and primary immunodeficiency pathways to be among the top
   upregulated pathways in Black TNBC compared to White TNBC, whereas TNF,
   IL-17, and NOD-like receptor signaling pathways were among the top
   downregulated pathways (Fig. [102]5C). Race-wise comparison of non-TNBC
   identified a total of 123 differentially-expressed genes, of which 82
   were upregulated and 41 downregulated (Fig. [103]5D, [104]E, and
   Supplementary Table [105]S3). The pathway prediction analysis suggested
   upregulation of NK cell-mediated cytotoxicity and cancer and primary
   immunodeficiency pathways in Black patients whereas complement and
   coagulation cascades, NFκB signaling, and antigen processing and
   presentation pathways were among the top downregulated pathways in
   these patients (Fig. [106]5F). Genes regulating Th1 cells inflammatory
   response, TBX21 (p = 0.009) and Th1 cells differentiation, CD44
   (p = 0.009), were significantly upregulated in White TNBC compared to
   Black TNBC, while genes involved in the suppression of Th1
   differentiation, CRP (p = 0.020), was significantly upregulated in
   Black TNBC (Fig. [107]5G). Besides, we observed a downregulation of the
   complement pathway-related genes, C1R (p = 0.024), C6 (p < 0.001), C3
   (p = 0.026), and CFD (0.005) in Black non-TNBC patients, compared to
   the respective subtype from white patients (Fig. [108]5H).

Fig. 5. Black patients exhibit greater immunosuppression compared to White
patients.

   [109]Fig. 5
   [110]Open in a new tab

   A Volcano plot representing the up-regulated (red dots) and
   down-regulated (blue dots) genes in Black TNBC samples compared to the
   White TNBC. B Heat map representing the expression pattern of the
   differentially expressed genes among the Black and White TNBC samples.
   C KEGG pathway analysis of the differentially expressed genes showing
   up-regulated and down-regulated pathways in Black patients compared to
   White patients in TNBC. D Volcano plot representing the up-regulated
   (red dots) and down-regulated (blue dots) genes in Black non-TNBC
   samples compared to the White non-TNBC. E Heat map representing the
   expression pattern of the differentially expressed genes among the
   Black non-TNBC and White non-TNBC samples. F KEGG pathway analysis of
   the differentially expressed genes showing up-regulated and
   down-regulated pathways in Black patients compared to White patients in
   non-TNBC samples. G Representative genes supporting Th1 function (TBK21
   and CD44) were down-regulated, and Th1 suppression (CRP) was
   up-regulated in Black TNBC compared to White TNBC patients. H
   Complement pathway-associated genes were down-regulated in Black
   non-TNBC compared to White non-TNBC patients. The p-value was
   calculated using the non-parametric Mann–Whitney test. A p-value < 0.05
   was considered statistically significant.

Race- and subtype-specific association of specific immune cell type abundance
with breast cancer metastasis

   Having observed race- and subtype-specific differences in the tumor
   immune landscape, we sought to investigate the association of the
   abundance of different immune cell types with BC metastasis. Our
   analysis revealed that metastasis-positive cases had significantly
   higher TAM infiltration (p = 0.002) in the TME in both non-TNBC
   (p = 0.046) and TNBC subtypes (p = 0.047) (Fig. [111]6A). In addition,
   we found that T-cell abundance was significantly lower in metastatic
   cases (p = 0.038), however, this association was significant in the
   non-TNBC group (p = 0.048) but not in TNBC (p = 0.216) (Fig. [112]6B).
   Further analysis of T-cell subsets indicated a strong trend for lower
   CD8+ T-cells in metastatic BC (p = 0.055), especially metastatic
   non-TNBC (p = 0.053) (Fig. [113]6C). In race-wise correlation, we
   observed that metastasis was significantly associated with the lower
   overall abundance of immune cells in White patients as indicated by
   CD45 score (p = 0.030). Further, in these patients, T-cells
   (p = 0.023), DC (p = 0.048), and NK-cells (p = 0.012) had a lower
   presence in cases with confirmed metastasis than those with
   non-metastatic disease. No significant association of immune cell
   abundance with metastatic disease was reported in Black patients (Fig.
   [114]6D).

Fig. 6. Association of immune cell abundance with BC metastasis in subtype
and race.

   [115]Fig. 6
   [116]Open in a new tab

   A Relative abundance of TAM in total BC, non-TNBC and TNBC samples,
   showing the metastatic tumors had significantly higher abundance of
   TAMs. B Relative abundance of T-cells in total BC, non-TNBC, and TNBC
   samples, showing the lower abundance of T-cells was significantly
   associated with metastasis in non-TNBC samples but not in TNBC samples.
   C Relative abundance of CD8+ T-cells in total BC, non-TNBC and TNBC
   samples, showing a trend for lower CD8+ T-cells abundance in non-TNBC
   tumors was associated with metastasis. D Relative abundance of total
   immune cells (CD45), T-cells, DC, and NK-cells in White and Black
   samples, showing their association with metastasis. The p-value was
   calculated using the non-parametric Mann–Whitney test. A p-value < 0.05
   was considered statistically significant.

Discussion

   The type and abundance of immune cells present in the TME play a
   crucial role in influencing both the cancer progression and the
   treatment outcomes. Moreover, it could also be a good predictor for
   response to immune checkpoint therapies and help in treatment
   planning^[117]16–[118]18. Immune cells normally recognize and eliminate
   abnormal cells, including cancer cells. However, cancer cells often
   develop mechanisms to evade immune detection by creating an
   immunosuppressive environment that reduces the efficacy of the immune
   response. In our study, we observed a differential immune landscape
   between TNBC and non-TNBC subtypes along with race-specific differences
   in White and Black patients. Cells of myeloid and lymphoid lineages
   clustered together, with some cells showing a unique clustering
   pattern. The infiltration of immune cells in the TME is influenced by
   tumor stromal composition, antigen presentation, vascularization, etc.,
   but it is mainly regulated by chemokine secretion by cells present in
   the tumor^[119]19–[120]21. Certain chemokines serve as chemoattractant
   for myeloid lineage cells, such as CCL2, CXCL2, and β-defensin, which
   recruit macrophages, neutrophils, and DCs,
   respectively^[121]22,[122]23, whereas CCL19, CCL21, and CXCL13 are
   associated with lymphoid cell (T-cells and B-cells)
   infiltration^[123]24–[124]26. Another chemokine, CCL5, is associated
   with the recruitment of T-cells and NK-cells in the TME^[125]27.

   The presence and recruitment of immune cell types in the TME is also
   greatly influenced by the crosstalk among them^[126]22,[127]28,[128]29.
   We observed the highest correlation of TAM with neutrophils among the
   myeloid lineage cells and B-cells with T-cells among cells of the
   lymphoid lineage. Further, B-cells and NK cells were strongly
   correlated with Treg cells in TNBC, while the NK cells showed a better
   correlation with Th1 and CD8+ T-cells in non-TNBC, suggesting that
   tumor cell-intrinsic properties may also influence the crosstalk among
   immune cells and thus TME composition. Studies have shown that B-cells
   present in tumors can convert T helper cells into T regulatory
   cells^[129]30,[130]31. Conversely, the Tregs can also suppress B-cell
   and NK cell function and create an immunosuppressed
   environment^[131]32–[132]35. Thus, it is crucial to understand how
   these interactions vary between molecular subtypes and what their
   impact is on clinical outcomes.

   Cancer progression as well as the outcome of immune- and
   chemo-therapies is influenced by the immune-active or suppressed state
   of the TME^[133]36,[134]37. A higher presence of TILs in the TME can be
   indicative of a better prognosis^[135]38,[136]39. However, in our
   study, TNBC, which has the worst prognosis^[137]40, was found to have
   higher TILs in the TME. The pathway analysis of the genes
   differentially regulated in TNBC compared to non-TNBC suggested that
   the former have an immunosuppressed TME. This could likely be a reason
   why TNBC maintains an aggressive progression leading to its poor
   prognosis, despite having overall higher TILs in the TME. TNBC had a
   higher presence of B-cells and Tregs, which are shown to shape
   immunosuppressive TME^[138]41. It has been reported that tumors grow
   slower in B-cell deficient mice exhibiting stronger Th1 and CD8+ T-cell
   responses^[139]42. Additionally, the CD40-CD40L interactions between
   B-cells and tumor cells derives IL-10 production leading to
   immunosuppression in the TME^[140]43. Further, the depletion of B-cells
   in mice model increased effectiveness of the chemotherapeutic agent
   oxaliplatin^[141]44. In humans, high B-cell infiltration in bladder,
   prostate, breast, and renal cancer has been linked to worse
   outcomes^[142]45. It is suggested that the antigen presentation by
   B-cells induces tolerogenic pathways in CD8+ T-cells, leading to anergy
   and impaired cytotoxic function^[143]46–[144]48. The B-cells often
   differentiate into regulatory B-cells (Bregs) in TME, producing IL-10,
   IL-35, and TGFβ to suppress CD4+ Th1 and CD8+ T-cell
   functions^[145]49,[146]50. Along these lines, earlier studies have
   reported the highest percentage of regulatory B-cells in TNBC, compared
   to other subtypes^[147]51, and a reduced overall survival in BC
   patients whose tumors exhibit a higher correlative presence of Bregs
   and Tregs^[148]52.

   Interestingly, we observed a low abundance of the mast cells in TNBC.
   Their low count in the axillary lymph node is shown to be inversely
   proportional to the metastasis in BC patients^[149]53, and their higher
   presence in tumors is associated with resistance to anti-PD-1 therapy
   in the humanized mouse melanoma model^[150]54. Combining agonistic
   anti-OX40 immunotherapy with mast cell blockade results in robust
   anti-tumor immunity in pancreatic cancer^[151]55. Since we observed
   TNBC to have low mast cells, it makes them a good candidate for
   immunotherapy. All the above observation aligns with clinical data of
   TNBC being the most aggressive and lethal breast cancer subtype,
   particularly in Black women. Similarly, some clinical studies have also
   shown a more favorable response of TNBC to immune checkpoint therapy
   than non-TNBC^[152]56,[153]57. Since TNBC tumors have higher
   infiltration of Tregs and B-cells, a combined immune therapy targeting
   both PD-1 along with B-cell specific immune checkpoints such as TIM-1,
   CD40, or BTLA, could prove to be more effective in
   TNBC^[154]58,[155]59. Therefore, analyzing specific immune cell
   populations can assist in precision oncology and improve the survival
   of BC patients.

   The immune composition of TNBC varied significantly in Black and White
   patients and could have implications for disparate outcomes. Black TNBC
   had lower levels of Th1 and CD8+ T-cells in the tumor than White TNBC,
   along with an upregulation of immunodeficiency pathways and
   downregulation of immune-activating pathways. The Th1 and CD8+ T-cells
   promote immunity against BC cells, and higher infiltration of Th1 and
   CD8+ T-cells is associated with better overall survival in
   BC^[156]60–[157]62. On the other hand, the race-wise immune landscape
   comparison of non-TNBC tumors showed Black patients to have higher NK
   and Treg cells with lower DC cells. The NK cells can directly attack
   and kill the cancer cells, while Treg cells can inhibit the activity of
   NK cells by directly interacting with them or by secreting IL-10 and
   TGFβ, hindering the anti-tumor immune response^[158]63,[159]64.
   Decreased NK cell activity has been associated with higher metastasis
   in BC^[160]65. Activated DCs play a crucial role in initiating immune
   responses by antigen presentation to CD4+ and CD8+ T cells, which leads
   to the maturation and activation of tumor-specific cytotoxic T
   lymphocytes^[161]66. Lower DC infiltration in Black non-TNBC than White
   non-TNBC may thus cause activation of effector T-cells. The pathway
   analysis of the differentially expressed genes showed the
   immunodeficiency and PD-L1 pathway to be upregulated and the complement
   pathway downregulated in Black non-TNBC relative to White non-TNBC. The
   complement pathway stimulates the effector function of CD4+ Th1 cells
   and CD8+ T-cells while reducing the Treg population in
   tumors^[162]67–[163]69. This suggests that Black non-TNBC patients have
   an immunosuppressed microenvironment compared to White non-TNBC
   patients.

   We found metastatic tumors, both TNBC and non-TNBC, to have
   significantly higher infiltration of TAMs, which have been associated
   with immune suppression and tumor progression^[164]70,[165]71.
   Conversely, T-cell abundance, particularly CD8 + T-cells, was lower in
   metastatic non-TNBC patients, which is often associated with poorer
   prognosis in cancer patients^[166]21,[167]72. In White patients,
   metastasis was associated with lower TILs, including T-cells, DCs, and
   NK cells. We have recently shown that Black patients exhibit higher
   serum cortisol levels, and cortisol induces M2 polarization of
   macrophages to cause immunosuppression^[168]73. Other factors,
   including genetics and higher prevalence of obesity among Black women,
   could also contribute to the suppressive immune landscape and its
   relation to metastasis and disease prognosis^[169]74–[170]77.

   In conclusion, this study highlights important differences in the
   immune landscape between TNBC and non-TNBC, as well as significant
   racial disparities in immune infiltration and immune activation
   pathways. Despite higher immune infiltration, TNBC exhibits a more
   immunosuppressive environment characterized by the presence of B-cells
   and Tregs. Race-specific differences, particularly in the abundance of
   key immune cell types such as T-cells, NK cells, and DCs, may
   contribute to disparate outcomes in Black and White patients (Fig.
   [171]7). A potential limitation of our study; however, is that all the
   clinical samples were procured from a single institution and thus do
   not have broad geographic distribution. Therefore, a future
   multi-institutional collaboration could not only strengthen these
   observations but also demonstrate if geographic and other
   region-specific factors (such as cultural, food choices, etc.) have an
   impact on the immune landscape. Moreover, although race is a social
   construct and socioeconomic factors play a significant role in cancer
   health disparities, future studies could incorporate ancestry profiling
   to examine if there are lineage-specific differences as well.
   Nonetheless, our findings highlight the importance of considering both
   tumor subtype and patient race when designing immune-based therapies.
   Further, our data suggest that targeted strategies to modulate immune
   cell populations, such as enhancing DC and NK cell infiltration, may be
   particularly beneficial in improving the response to treatment in
   specific racial populations.

Fig. 7. Schematic summary depicting the subtype and race-specific differences
in BC immune landscape.

   [172]Fig. 7
   [173]Open in a new tab

   TNBC exhibit overall higher immune cell infiltration, specifically
   having a greater abundance of B-cells, Tregs, Th1, and CD8+ T-cells,
   compared to the non-TNBC. Pathway analysis of differentially expressed
   genes show activation of immunosuppressive pathways despite greater
   immune cell infiltration. TNBC from White patients have a higher
   presence of Th1 and CD8+ T-cells in the TME than those from Black
   patients, while non-TNBC from Black patients exhibit a higher
   accumulation of NK and Treg cells but lower DC, compared to that in
   White patients. Black patients show reduced immune activation
   regardless of molecular subtype. Higher TAM and lower T-cell
   infiltration in all BC patients, and lower TILs in White patients, are
   indicative of the metastatic disease.

Methods

Biospecimens

   The study was conducted following the determination of “exempt” status
   by the Institutional Review Board (IRB). De-identified Formalin-fixed
   paraffin-embedded (FFPE) BC tissue blocks were obtained from the
   Department of Pathology, University of South Alabama. The tissue
   samples consisted of histologically confirmed non-TNBC (n = 55) and
   TNBC (n = 26) cases and were selected following H&E staining-based
   determination by a certified pathologist for the presence of more than
   30% cancer cell area within the tumor. The description of basic
   characteristic features of the BC patients involved in the study is
   provided in Table [174]1. The racial classifications of the patients
   are based on self-reported data.

Table 1.

   Characteristics of breast cancer patients
   Patient cohort
   Total patients (n = 81) White (n = 40) Black (n = 41)
   Characteristics
    Age at diagnosis, range (mean) 25–85 (58.1) 25–85 (59.7) 32–83 (56.5)
   Race
    White 40 N/A N/A
    Black 41 N/A N/A
   Subtype
    Non-TNBC
    HR^+ HER2^+/− 48 (59.3%) 25 (62.5%) 23 (56.1%)
    HR^− HER2^+ 7 (8.6%) 3 (7.5%) 4 (9.8%)
    TNBC
    HR^− HER2^− 26 (32.1%) 12 (30%) 14 (34.1%)
   [175]Open in a new tab

Reagents

   The details of the reagents, kits, and other resources used in the
   study, along with their sources and catalog numbers, are provided in
   supplementary Table [176]S4.

RNA isolation

   FFPE tissues were sliced into 5 µm thick sections using a microtome and
   used for RNA isolation. The sections were collected in 1.5 ml
   microcentrifuge tubes, to which a deparaffination solution (Cat. No.
   19093, Qiagen, MD, USA) was added. Samples were heated to 56 °C for
   3 min to dissolve the paraffin, and the RNA was isolated using the
   RNeasy FFPE kit (Cat. No. 73504, Qiagen, MD, USA) according to the
   manufacturer’s protocol. The quantitative and qualitative analysis of
   RNA was performed using a Nanodrop 1000 (Thermo Scientific, MA, USA).

Immune profiling and gene expression data

   The immune profiling of the BC tissue was performed on the nanoString
   nCounter® pro analysis system using the PanCancer Immune Profiling
   Panel (nanoString Technologies, WA, USA). The RNA (150 ng in 5 µL
   volume) was mixed with 5 µL of hybridization buffer, 3 µL of reporter
   probe, and 2 µL of capture probe per sample and subjected to RNA-probe
   hybridization reaction for 18 h at 65 °C in a thermocycler. The
   hybridized samples were immobilized in the sample cartridge for imaging
   and counting using the nCounter Pro prep station, followed by cartridge
   scanning on the nCounter Pro Digital Analyzer (nanoString, WA, USA) to
   obtain the count data in RCC file format.

In silico data analysis

   The data analysis was performed using Rosalind online software. After
   performing the data quality check and normalization, the immune cell
   abundance score and differentially expressed genes for different
   comparisons were obtained from the Rosalind software. The unsupervised
   hierarchical clustering and heat maps for immune abundance in tissue
   samples were created using Clustergrammer online software^[177]78. The
   heat maps for differentially expressed genes were created by GraphPad
   Prism using the data obtained from analysis by Rosalind software. The
   KEGG pathway analysis of the up-regulated and down-regulated genes was
   performed using Enrichr^[178]79–[179]81 online software.

Statistical analysis

   All the statistical analysis was performed using the GraphPad Prism
   software (GraphPad Software, CA). Statistical comparisons between two
   groups were performed by an unpaired, non-parametric Mann-Whitney test.
   The Pearson correlation coefficient was used to determine the
   association between immune cell type infiltration. A p-value of lower
   than 0.05 was considered statistically significant.

Supplementary information

   [180]Supplementary information^ (344.2KB, docx)

Acknowledgements