Breast cancer is a heterogeneous disease with varying biological and clinical characteristics. It is the most common cancer among women worldwide, accounting for 25% of all cancer cases (1). According to GLOBOCAN 2020, the incidence and mortality of breast cancer reported worldwide were 34,65,951 new cases and 11,21,413 deaths, respectively; in India, 1,204,532 new cases and 436,417 deaths were recorded in 2020 (2).

Immunohistochemical analysis of breast tumors is the gold-standard method used in clinics to classify them based on the hormone receptor expression for improved therapeutic decisions. Based on this, breast cancer can be broadly grouped into five types, namely: i) Progesterone receptor (PR)-positive, estrogen receptor (ER)-positive and human epidermal growth factor 2 (Her2)-negative (luminal A); ii) ER-positive, PR-positive/negative and Her2-positive (luminal B); iii) Her2-overexpressing, ER- and PR-negative; iv) ER-, PR- and Her2-negative (basal-like or triple-negative), and v) normal-like (expression status similar to luminal A and resemble normal breast profile) (3–5). Additionally, molecular breast cancer analysis identified a distinctive phenotype with low claudin expression, immune receptor, and EMT markers expression (6). Cancer types with the claudin^low phenotype are highly metastatic and associated with poor prognosis (7). Her2-overexpressing cancer also displays high metastasis and poor prognosis (8). Among ER-positive subtypes, luminal B is associated with a significantly worse prognosis than luminal A (9,10). Patients with basal subtypes of cancer with BRCA1 mutations have a poor prognosis (9).

Based on specific gene expression patterns, breast cancers are categorized into five intrinsic or molecular subtypes. Among the intrinsic subtypes, basal-like triple-negative breast cancer (TNBC) accounts for 12–20% of breast cancers (11). TNBC has drawn specific attention due to the lack of expression of all three receptors (ER, PR, and Her2). Thus, it cannot be treated using anti-estrogen hormonal therapies or trastuzumab (12). Morphologically, TNBC is characterized by hyperdense masses without calcification, usually occurring in women <50 years of age. Histological features include significant lymphocyte infiltration, central necrosis, pushing tumor borders, and fibrosis (13). Cytokeratins, fascin, epidermal growth factor receptor (EGFR), caveolin, and vimentin are usually expressed in basal-like TNBC (14,15). TNBC is challenging to treat, as it is quite complex due to poor cell differentiation, molecular heterogeneity, and rapid metastasis, often leading to chemoresistance and recurrence of the disease (16). Fast relapse and invasions are common features of TNBC tumors and show poor prognosis (17). Recent advances in omics technologies have provided insight into the molecular mechanisms underlying TNBC (18). The present review focuses on the different subtypes of TNBC and therapeutic approaches currently employed in the treatment of TNBC.

Histologically, most TNBC is categorized as no special type (IDC-NST) (17). Most IDC is characterized by pleomorphic cells with prominent nucleoli. The cells are organized into diffuse sheets, cords, nests with ductal differentiation. The rest of the tumors are categorized into 47 specific subtypes, such as invasive lobular carcinoma (relatively common), metaplastic carcinoma, medullary carcinoma, mucinous carcinoma, adenoid cystic carcinoma, secretory carcinoma, acinic cell carcinoma, neuroendocrine tumors, as well as the rarest glycogen-rich clear cell carcinoma (19,20).

Among these specific subtypes (Fig. 1), medullary breast carcinoma occurs in <1% of patients and shows distinctive features, such as high lymphoplasmacytic infiltration, overexpression of BCLG (a pro-apoptotic gene); it bears more losses of heterozygosity than other subtypes and is immunomodulatory (21,22). It is associated with better outcomes compared with other TNBC subtypes (22). Metaplastic carcinoma presents unique pathologic features, where the glandular component may be partially or completely replaced by a non-glandular component(s), and based on their differentiation status further divided into i) Squamous type, tumor with keratinization and squamous differentiation; ii) matrix-producing type, tumor with more cells in the periphery; iii) mixed type, tumor showing both squamous differentiation and large high-grade cells with pleomorphic nuclei; and iv) spindle-cell type, tumor with storiform-like spindle cells. These metaplastic tumors harbor mutations in the PIK3CA, Wnt (Wingless-Type MMTV Integration Site Family) signaling pathway genes and display a unique copy number alteration pattern (23–25). Adenoid cystic carcinoma (ACC) is characterized by the presence of dividing epithelial cells and myoepithelial cells producing mucinous membrane. ACC occurs in 0.1% of patients with basal-like features (26,27) and expresses markers, such as cytokeratin 5, −5/6, −14 and −17 (28). Secretory carcinoma is characterized by microcystic, solid and tubular architecture and presence of vacuolated tumor cells producing intracellular and extracellular secretions. It occurs in <1% of the patients and is referred to as juvenile carcinoma, as it is common in adolescents and often reported to have favorable outcomes. It is also characterized by ETV6-NTRK3 fusion (29–31). The rarest among all the subtypes is glycogen-rich clear cell carcinoma, in which the tumors appear in sheets and cells are polygonal in shape, with a clear cytoplasm and the presence of glycogen (32). In these sheets, there are areas of lymphocytic infiltration and plasma cells.

Among the histological subtypes, adenoid cystic carcinoma has a median recurrence of only 2 months, and metaplastic carcinoma has ~9.9 months (33), compared to IDC-NST and matrix-producing metaplastic carcinoma, which are less aggressive, with 34 and 31.4 months of median time to recurrence, respectively (20).

Although the histological assessments were pointing to the presence of WBCs in and around the TNBC subtypes, focus on the presence of WBCs has led to the identification of TILS and TAMs, which are the parameters defining prognosis and therapy of TNBC. The TNBCs might be immunogenic due to mutations that lead to aberrant protein expression on the cell membrane (34). Tumor-infiltrating lymphocytes (TILs) are white blood cells that migrate towards the tumor from the bloodstream via the newly formed blood vessels (angiogenesis), which cancer cells use for their nutritional and oxygen requirements (35). They consist of a mixture of B cells, macrophages, natural killer cells and are dominated by T cells (35). TILs are present in ~20% of TNBC tumors and carry a pivotal prognostic and predictive value (36). The presence of TILs indicates a good prognosis (37). High number of TILs indicate that there is an equilibrium between the immune status and cancer (38). The ratio of cancer cells: TILs is tilted towards TILs after surgical removal of a tumor, resulting in an improved prognosis in TNBC (38). A high mutation load and clonal heterogeneity are associated with a low number of TILs, which may provide an escape route to tumor cells from immune surveillance (39). However, in addition to TILs, the tumor microenvironment components also influence the outcome of patients with TNBC (39). Relapsing patients with TNBC have been shown to have low levels of TILs and a high number of CD163⁺ tumor-associated macrophages (TAMs) compared with that of patients without relapse (39). High levels of CD8⁺ T cells may reflect improved sensitivity to chemotherapy, whereas high levels of TAMs correlate with poor patient outcomes (36). Nevertheless, a previous study in TNBC has reported paradoxical findings, with high levels of CD8⁺ T cells in the tumor stroma leading to the low infiltration of the tumor epithelium, thereby indicating a poor outcome (40). Therefore, immunohistological assessment for TILS or TAMS will help develop immunotherapies detailed in section 7.

Profiling based on gene expression has led to improved insight into tumor heterogeneity at the molecular level and has generated an impartial classification (Fig. 1). The PAM50 microarray set of 50 genes is used to identify breast cancer intrinsic subtypes (41). A set of 374 TNBC samples taken from 14 microarray datasets was analyzed to characterize TNBC subtypes using PAM50. The results from this analysis categorized most of the TNBC as basal-like (80.6%). The rest of the tumours were classified as Her2-positive(0.2%), normal-like (14.6%), luminal B (3.5%) and luminal A (1.1%) (Table I) (41).

Lehmann et al (42) performed gene expression profiling of 2,188 genes from 587 patients with TNBC and classified TNBC into six new groups, namely, basal-like 1 (BL1), basal-like 2 (BL2), immunomodulatory (IM), luminal androgen receptor (LAR), mesenchymal stem cell-like (MSL) and mesenchymal (M). The rest was classified as an unstable type (UNS/UNC). Each subtype had its characteristic feature. Basal-like was the most common type of TNBC (BL1, 22%; BL2, 12%) and was characterized by high Ki67 and DNA damage response levels. The IM subtype (18%) had basal-like characteristics with activation of IFNα and IFNγ signaling and high cytotoxic T-lymphocyte associated protein 4 gene expression. Mesenchymal subtypes (M, 21%; MSL, 10%), along with cell differentiation pathways, showed deregulation of EGFR, calcium signaling, MAPK, and PI3K signaling. In the LAR subtype (9%), an ~10-fold increase in androgen receptor (AR) expression was seen, compared with other subtypes. Activation of various pathways, such as steroid synthesis and FOXA1 and ERBB signaling, were observed in this subtype (Table I) (42,43).

Burstein et al (43) used a non-negative matrix factorization method to derive a panel consisting of 80 core genes that divided TNBC into four subtypes, luminal-AR (LAR), mesenchymal (MES), basal-like immune-suppressed (BLIS), and basal-like immune-activated (BLIA). BLIA has the best disease-free survival outcome compared to other subtypes (44). Based on DNA copy number, these subtypes can be placed into two groups, LAR or others (Table I) (31).

Liu et al (45) performed mRNA and long non-coding RNA (lncRNA) expression analysis in 165 TNBC tumor samples at Fudan University Shanghai Cancer Centre. The tumor samples were categorized into four subtypes (IM, LAR, MES, and BLIS subtypes), consistent with the classification by Burstein et al (43). The IM subtype comprised of genes related to immune functions such as CCR2, CXCL13, CXCL11, CD1C, CXCL10, and CCL5, along with ENST00000443397 long ncRNA. In contrast, the LAR subtype had enrichment of hormone regulation signaling and ENST00000447908 lncRNA (45). The MES subtype expressed lncRNA NR_003221 together with genes and pathways that promoted epithelial-to-mesenchymal (EMT) transition. Pathways and molecules such as DNA repair, replication, and mitosis, lncRNA TCONS_00000027 were enriched in the BLIS subtype (45,46).

Genomic/transcriptomic data from a set of 997 primary tumors were extracted, and an integrated analysis was performed by Curtis et al (47). A set of 995 tumors from the Molecular Taxonomy of Breast Cancer International Consortium (METABRIC) cohort was used as a validation set that divided TNBC into ten groups, named Integrated Clusters (IntClust) 1–10 (47). Basal-like breast cancer mostly fell in IntClust 4 and 10 (~80%). IntClust 4 is known to have greater TIL counts, while IntClust 10 subtype can display genomic instability and chromosomal aberrations (Table I) (47–49).

Through whole-exome and whole-genome data, it is evident that most of the genetic alterations in TNBC are copy number alterations and somatic mutations (40). The BRCA1 and BRCA2 tumor suppressor genes are required for the maintenance of genomic stability. These genes play a role in DNA repair and replication error control (50,51). A total of 10% of patients with TNBC are known to harbor germline mutations in BRCA1 or BRCA2 (12,26,27). The lifetime risk of breast cancer becomes 60–70% in the presence of such mutations (52). Gene alterations leading to homologous recombination (HR) defects other than germline BRCA mutations are termed ‘BRCAness’ (53). Moreover, ~35% of TNBC tumors show abnormalities in the HR pathway, making them sensitive to poly (ADP-ribose) polymerase (PARP) inhibitors and DNA-damaging agents (54).

Other common mutations observed in TNBC patients include those in TP53 (50–60%) and PIK3CA (~10%) (18,42). An analysis from the Catalogue of Somatic Mutations in Cancer (COSMIC) database revealed that the top genes mutated in TNBC, apart from BRCA1/2, TP53, and PIK3CA, were RB1, PTEN, NOTCH1 and BRAF (Fig. 2A). Among the point mutations observed, 34% of them were nonsense substitutions (where a base change leads to a stop codon in the DNA sequence), 21% were synonymous mutations (where a change in a base in the exon of a coding gene does not change the structure of the protein) (Fig. 2B). The rest of the mutations were missense mutations, frameshift insertion/deletions, and in-frame insertions/deletions. In the metastatic disease setting, genes from HR repair showed a larger frequency of biallelic loss-of-function mutations than in early TNBC (55).

Integrated analysis of The Cancer Genome Atlas (56) has demonstrated deletions in PTEN, DUSP4, and INPP48 involved in the PI3K-AKT pathway. Gene amplifications were seen in MYC, PIK3CA, KRAS, BRAF, FGFR, MET, and EGFR. Mutations in genes, such as ERBB2, AKT1, ATR, MAP3K1, CDKN2A, ATM, and NOTCH2 (18,42,51), were also observed. Based on the mutation signatures obtained from whole-genome sequencing of 560 tumors, TNBC could be classified into four mutation subtypes, namely, APOBEC-based signatures, HR deficiency-based signature (signature 3), Clock-like signatures (signatures 1 and 5), and mixed (no prominent signature) (56,57). These mutations suggest that DNA repair, the PI3K/AKT pathway, cell cycle checkpoints, and Notch signaling are possible druggable pathways in TNBC (58).

Recently, much focus has been put on bringing liquid biopsies, such as circulating tumor cells (CTC) and circulating tumor DNA (ctDNA), into the clinical setting for diagnostic and prognostic use (59). CTCs are nucleated cancer cells present in the bloodstream that can be detected using techniques, such as reverse transcription-quantitative PCR, flow cytometry, and immunohistochemistry (60). Tumor cells that undergo necrosis or apoptosis release DNA fragments into the plasma are referred to as ctDNA (61). In breast cancer, ctDNA and CTCs have been studied as potential biomarkers for prognosis (60). Stover et al (62) performed studies in metastatic breast cancer patients receiving chemotherapy and identified an association between CTCs and ctDNA and tumor burden, indicating that these could be used to measure early-treatment response in patients. A retrospective study in 164 patients with metastatic TNBC revealed that >10% of patients with ctDNA had worse disease-free survival (62). A study by Bidard et al (63), with metastatic breast cancer, revealed that patients with CTC levels >5 per 7.5 ml were associated with lower progression-free survival (PFS) and OS compared with patients who had CTC levels <5 per 7.5 ml. Cristofanilli et al (64) reported that CTC counts could be utilized to classify metastatic patients into two groups. Patients with CTCs levels >5 per 7.5 ml were categorized as aggressive stage IV and those <5 per 7.5 ml as indolent stage IV (64). ctDNA has been associated with chemotherapy in studies by Riva et al (65), in which ctDNA-positive patients before and after chemotherapy experienced poor OS and disease-free survival (DFS). Additionally, Radovich et al (66) reported that patients with early-stage TNBC and positive ctDNA after chemotherapy had a higher risk of disease relapse. Therefore, liquid biopsies are being developed as a non-invasive method to study recurrence, treatment response, and survival in the clinical setting.

TNBC treatment involves a combination of surgery, radiotherapy, and chemotherapy. New methods, such as targeted therapy and immunotherapy, have been developed to improve patient survival and prognosis. Lumpectomy and mastectomy are the surgical procedures performed for TNBC patients and are usually followed by radiotherapy and chemotherapy (67). Neoadjuvant therapy is given before the surgery, which may help shrink the tumor size and avoid mastectomy (Fig. 3) (61). Taxanes and anthracyclines form the current standard of care for TNBC in both the neoadjuvant and the adjuvant settings. Epirubicin and doxorubicin are the most common anthracyclines (anticancer antibiotics known to disrupt DNA replication and mitochondrial functions to activate apoptosis) (68,69). Taxanes block angiogenesis by inhibiting epidermal growth factor receptor signaling (70).

Paclitaxel and docetaxel are familiar examples of taxanes used in the first line of therapy (71). TNBC shows a 40% pathological complete response (pCR) for taxane and anthracycline-based therapy in the neoadjuvant setting (72–74). Adjuvant therapy guidelines are usually identical for all the subtypes of breast cancer and TNBC. Chemotherapy in the adjuvant setting is recommended for tumors >0.5 cm in size, as they exhibit increased aggressiveness, with a faster growth rate and metastasis (75). Anthracycline chemotherapy (cyclophosphamide and 5-fluorouracil) in patients with metastatic TNBC exhibited a response to survival within 22 months (69). However, acute toxicity is a major concern with anthracycline-based chemotherapy (76). Metastatic patients who develop resistance to anthracycline have shown sensitivity to capecitabine, gemcitabine and vinorelbine (77–79). The combination of docetaxel with capecitabine has improved the OS of patients with metastatic TNBC (78).

Carboplatin and cisplatin are platinum salts that are used in the treatment of TNBC. These generate DNA lesions, and apoptosis occurs in cells unable to repair these breaks (80). For TNBC, carboplatin as a neoadjuvant addition increases the response rate from 37 to 52.1% (81). A phase-II study of 86 patients evaluating the efficacy of platinum monotherapy demonstrated a 32% overall response rate (ORR) for cisplatin and 19% for carboplatin in early TNBC. Patients with BRCA1/2 mutations showed an improved response compared with patients without BRCA1/2 mutations (82). Moreover, phase-II trials showed an improved ORR of 72% in metastatic patients with BRCA mutation with neoadjuvant cisplatin monotherapy (83,84). Recently, the PEARLY trial (NCT02441933) has explored combination therapy of taxanes and carboplatin in the neoadjuvant setting (85). Carboplatin with docetaxel or paclitaxel combination has demonstrated promising efficacy in patients with TNBC and brain metastasis (86). Although TNBC is sensitive to chemotherapy, early relapse is a major concern (75). Therefore, optimizing a tailored standard regime to address chemotherapy issues, such as toxicity, and relapse has led to customizing personalized therapy based on tumor type.

Therapies targeted to TNBC are being developed based on the expression of specific pathways and genes. Targeted therapy focuses on customizing cancer therapy to an individual patient's tumor (87,88). TNBC being heterogenous, targeting alterations specific to the tumor would be the most effective treatment option. A study using genomics and transcriptomics has led to identifying molecular markers that could be effectively targeted in TNBC (89). PARP inhibitors, PI3K/AKT inhibitors, and anti-androgen therapy are under clinical investigation (Fig. 3) (58).

Sequencing of all the RNA species in a given cell using RNA-seq identified several RNA species, including mRNA. The two major classes of non-coding RNA studied in TNBC development and treatment are miRNA and Long non-coding RNA.

MicroRNA (miRNA/miR) is a small non-coding RNA, usually 20–22 nucleotides in length, regulating gene expression. miRNA is known to bind to the 3′untranslated region of mRNA. This binding either degrades mRNA or represses translation (154). miRNA is a key player in tumorigenesis, stemness, and drug resistance in TNBC (155–157). For instance, tumour suppressor miRNAs, involved in tumour development, miR-190a, miR-136-5p, miR-126-5p, miR-135b-5p and miR-182-5p are downregulated in TNBC (158). miR-22 is downregulation in TNBC, is associated with migration and metastasis. miR-22 exerts its effect through eukaryotic elongation factor 2 kinase (eEF2K) expression, which activates PI3K signaling pathway (159). Also, oncosuppressor, miR-200b, activate target genes like SRY-box transcription factor 2 (SOX2), CD133, and zinc finger E-box binding homeobox 1 (ZEB1), aiding in migration and invasion and stemness (157,160). High expression of miR-95 in TNBC indicates radiotherapy resistance that occurs by targeting sphingosine-1-phosphate signaling (161). Downregulated miR-449 upregulates CDK2, CCNE2 causing doxorubicin resistancein TNBC (162,163) (Table III). Multiple studies also show that miRNAs are expressed in different stages of TNBC Multiple studies also show that miRNAs are expressed in different stages of TNBC (164–166). These studies give hope for miRNA-based therapies, as the use of miRNA mimics or inhibitor oligonucleotides could serve as a therapeutic approach for TNBC (167). A study conducted by Shu et al (168) used miR-21 combined with aptamer targeting EGFR, blocking tumor growth in murine models. Yin et al (169) designed an RNA aptamer bound to CD133 with a sequence complementary to miR-21 carried by a three-way junction motif scaffold that reduced cell migration in TNBC cells (169). Non-coding RNA is being pursued as one of the TNBC therapy.

lncRNA (long non-coding RNA), ~200 nucleotides in length, regulates gene expression at the epigenetic, transcription, post-transcription levels, and post-translation modification (16). The long intergenic non-coding RNA for kinase activation activates HIF-1α by phosphorylating it via leucine-rich repeat kinase 2to promote glycolysis and tumorigenesis in TNBC (170). Yang et al (171) demonstrated the involvement of POU domain class 3 transcription factor 3 (POU3F3) in inhibiting apoptosis and promoting proliferation in TNBC (171). Nuclear paraspeckle assembly transcript 1 (NEAT1) plays a role in TNBC metastasis (172–174). Some lncRNAs (HOTAIR, LncRNA-ATB, LincRNA-ROR) are known to be co-expressed with transcription factors involved in EMT and proliferation (175). Vaidya et al (176) demonstrated that nanoparticle-mediated transfer of RNA interference molecules targeting differentiation antagonizing non-protein coding RNA, a lncRNA that is enriched in TNBC, showed some efficacy in a murine xenograft model of TNBC (Table III). These studies have shed light on the use of antisense oligonucleotides against oncogenic lncRNA as a potential approach to TNBC therapy.

TNBC is associated with poor prognosis compared to other breast cancer subtypes, and its treatment remains challenging. New technology and tools have provided insight into the molecular mechanism of the disease. This knowledge has led to the identification of druggable targets and the development of biomarker-driven therapy. The FDA-approved drugs for TNBC to date include PARP inhibitors for patients with BRCA1/2 mutations and atezolizumab for PD-L1⁺ tumors. Emerging targeted therapies have given hope for the treatment of TNBC. The inclusion of immunotherapy has shown promising results. Additionally, attempts to identify combinations that work effectively against TNBC are ongoing. A combination of the molecular profiles, including non-coding RNA and histology, has improved the prognosis and guided the treatment for TNBC.