Introduction

Breast cancer is the most common tumor among female patients with Li–Fraumeni syndrome (LFS) [1], a rare autosomal dominant hereditary syndrome characterized by germline TP53 mutations. Breast cancer affects women with LFS at a much younger age compared to the general population in the United States (median age at diagnosis 32 vs 61 years) [2, 3]. It is estimated that 3–8% of women diagnosed with breast cancer under the age of 30 carry a germline TP53 mutation [4]. The vast majority of TP53 germline mutations identified in LFS are highly penetrant [5], with a cumulative incidence of breast cancer in women of 85% by age 60 [6].

The immunophenotypic profile of LFS-associated breast cancers has been recently described, showing these tumors to be significantly associated with HER2 amplification and overexpression (53–83%) [2, 7, 8]. Approximately 50% of cases co-express both estrogen receptor (ER) and HER2 [2, 7]. However, compared with other hereditary breast cancers [9, 10], relatively little is known about the histopathologic characteristics of breast cancer in LFS patients. In this study, we sought to describe the morphologic features of invasive and in situ carcinoma of the breast in patients with known germline TP53 mutations.

Materials and methods

Eighteen cases of invasive mammary carcinoma and 9 cases of ductal carcinoma in situ (DCIS) from 25 patients with confirmed germline TP53 mutations were identified through the LFS Consortium [11]. Two patients had two cancers each, which were considered separate primaries. One patient had bilateral cancer, with DCIS on the left breast and invasive ductal carcinoma (IDC) on the right side. The second patient had DCIS and IDC on the same breast but in different quadrants with different receptor profiles, and therefore considered as two separate primaries. The cases included in this study constituted a subset of those with available slides and tissue blocks from a cohort previously reported by the LFS Consortium [2]. Clinical data including age and stage at diagnosis and germline mutation status were previously collected under Dana Farber Cancer Institute IRB protocol #10-458.

Histologic analysis

Up to 3 representative H&E-stained sections from each case were reviewed. DCIS cases were evaluated for architecture, nuclear grade, presence of necrosis (central vs punctate), calcifications, periductal stromal/lymphocytic response, and apocrine features. Invasive carcinomas were classified by type (ductal, lobular, mixed ductal and lobular, or other special types), tubule score, nuclear grade, mitotic score, modified Scarff–Bloom–Richardson (mSBR) grade, presence of extensive necrosis (defined as >1 4x high-power field), pushing border (if slides from excision were available), syncytial growth pattern, apocrine features and stromal tumor infiltrating lymphocytes (TILs). Syncytial growth pattern was assigned when anastomosing sheets of tumor cells with indistinct cell borders were present, as in the initial descriptions of medullary carcinoma by Moore [12] and Ridolfi [13]. Apocrine features were defined as tumor cells with ample granular eosinophilic cytoplasm with enlarged, central nuclei and prominent nucleoli [1]. Carcinomas of mixed type or special types were classified according to WHO criteria [1]. Scoring of stromal TILs was performed using the recommendations of the International TILs Working Group [14].

Tissue microarray construction

A tissue microarray (TMA) was constructed in the Dana Farber/Harvard Cancer Center Tissue Microarray Core Facility as previously described [2]. Briefly, three 0.6 mm cores from invasive carcinomas and 4–6 0.6 mm cores from DCIS cases were taken from areas marked by a breast pathologist (DAD) and placed into a recipient block using a manual arrayer (Beecher Instruments).

Immunohistochemical studies

Immunohistochemistry (IHC) was performed on 4 µm tissue sections (TMA sections and whole slide sections if available) following pressure cooking antigen retrieval (Target Retrieval Solution pH 6.0, DAKO) using antibodies for ER (SP1, 1:40 dilution, Thermo Scientific), progesterone receptor (PR) (PgR 636, 1:50 dilution, DAKO), HER2 (SP3, 1:50 dilution, Thermo Scientific), and p53 (DO-7, 1:500 dilution, DAKO). EnVision plus system-HRP (DAKO) was used as the detection system for all antibodies. External positive and negative controls were included in each run.

Immunohistochemistry (IHC) scoring

Cases were scored as positive for ER and PR when there was ≥1% nuclear expression in tumor cells, according to 2010 ASCO/CAP guidelines [15]. IHC for HER2 was scored as positive 3+ (strong complete membrane immunoreactivity in >10% of tumor cells), equivocal 2+ (weak to moderate complete membrane staining in >10% of tumor cells), negative 1+ (faint, weak partial membrane staining in >10% of tumor cells), and negative 0 (no immunoreactivity or ≤10% tumor cells with faint incomplete staining) according to 2018 ASCO/CAP guidelines [16]. The antibody clone DO-7 used in this study recognizes both mutant and wild-type p53 protein. P53 expression was scored as percent positive nuclei in tumor cells.

Fluorescence in situ hybridization

Fluorescence in situ hybridization (FISH) was performed as previously described [2]. Hybridization signals were scored in at least 20 tumor cells for each case. Results were interpreted according to 2018 ASCO/CAP guidelines [16].

Results

Patient population

The median age at diagnosis was 31.5 and 34 years for invasive carcinoma and DCIS, respectively (Table 1). All patients were women who carried a germline TP53 mutation. Specific type of mutation data was available for 23 of 25 patients. The majority were missense mutations (13/23), followed by nonsense mutations (6/23), deletions (2/23), and splice-site mutations (2/23).

Table 1 Cohort characteristics.

Two DCIS cases also showed microinvasion. Additional clinical information on stage at diagnosis is summarized in Table 1.

Invasive carcinoma

Morphology

Seventeen cases (94%) of invasive carcinoma were of ductal type (Fig. 1). The remaining case showed mixed ductal and lobular features. The majority of cases were mSBR grade 3 (61%, Fig. 2a). Fourteen cases had a tubule score of 3 (78%), and 13 tumors were of high nuclear grade (72%). A mitotic score of 2 or 3 was seen in 39% of cases. Only one case (6%) had apocrine features. None of the cases showed extensive tumor necrosis, pushing border, or syncytial growth pattern. Ten cases (56%) were associated with <10% stromal TILs. Only two cases (11%) had >50% stromal TILs.

Fig. 1: Morphologic features of invasive carcinomas.
figure 1

The bar chart shows the distribution of histologic features of 18 invasive carcinomas.

Fig. 2: Example of invasive carcinoma.
figure 2

Case 22: High-grade invasive ductal carcinoma (a) with p53 overexpression (b; mutant pattern) in a patient with a germline TP53 missense mutation.

Hormone receptor and HER2 Status

Seventeen cases had tissue available for IHC and/or FISH testing (Table 2). A total of nine cases were positive for HER2 (53%), all showing ER co-expression (Table 3). Mean ER expression in ER+/HER2+ cases was 58% (range 5–95%) compared to 92% in ER+/HER2− cases (range 80–95%) (p = 0.023, t-test). There were no ER-/PR + cases. Only two cases showed ER expression with negative PR results. Two cases were HER2 negative by IHC (1+) but positive by FISH. All remaining cases were HER2 concordant by IHC and FISH.

Table 2 Hormone receptors, HER2 immunohistochemistry (IHC), and fluorescence in situ hybridization (FISH) results.
Table 3 Hormone receptors and HER2 results in ductal carcinoma in situ (DCIS) and invasive carcinoma.

Ductal carcinoma in situ

Morphology

For the morphologic evaluation of DCIS, the in situ component of invasive tumors was also included in the analysis (n = 24, Fig. 3). DCIS cases showed a variety of architectural patterns; however, the most common predominant pattern was solid (15/24, 63%), followed by cribriform architecture (8/24, 33%). Twenty-two cases (92%) were of high nuclear grade (Fig. 4a). Central necrosis was present in 71% of cases (17/24). Associated calcifications and periductal fibrosis/lymphocytic response were seen in 58% (14/24) and 63% (15/24) of cases, respectively. Six cases (25%) showed apocrine features.

Fig. 3: Morphologic features of ductal carcinoma in situ (DCIS).
figure 3

The bar chart shows the distribution of histologic features of 9 DCIS cases.

Fig. 4: Example of ductal carcinoma in situ (DCIS).
figure 4

Case 3: High-grade DCIS (a) in a patient with a germline TP53 deletion showing absent p53 expression (b; null pattern).

Hormone receptor and HER2 status

Eight cases had tissue available for IHC and/or FISH testing. DCIS was most frequently HER2+ (6/8, 75%), including two ER+/HER2+ cases. There were only two ER+/HER2− cases.

p53 immunohistochemistry

Of cases with available tissue for p53 IHC analysis, the majority of cases with missense mutations (9/13) showed overexpression of p53 protein in >50% of tumor cells (Table 4; Fig. 2b). Four cases with missense mutations showed p53 expression in 5–50% of tumor cells. All cases with nonsense, deletion, and splice-site mutations (n = 10) showed little to no expression (<1%, null pattern; Fig. 4b). The type of germline mutation did not correlate with ER or HER2 status (Table 5). In addition, no association was found between type of mutation and histomorphology (data not shown).

Table 4 P53 immunohistochemistry (IHC) in tumor tissue and correlation with germline mutation.
Table 5 Hormone receptors and HER2 results according to type of TP53 germline mutation.

Discussion

In this report, we describe in detail the morphologic features of invasive and in situ carcinomas in LFS. In our cohort of 27 cases from 25 patients with germline TP53 mutations, both invasive and in situ carcinomas showed features similar to breast tumors with somatic TP53 mutations and distinct from other germline-associated cancers [9, 17,18,19,20]. Invasive carcinomas in our cohort were most often mSBR high grade, similar to invasive breast cancer with somatic TP53 mutations [21] and DCIS was most often high nuclear grade, solid type, with central necrosis, similar to DCIS with somatic TP53 mutations [22, 23]. While both BRCA1- and BRCA2-associated tumors show high mSBR grade they also tend to have pushing borders [10, 18, 24] and BRCA1-associated tumors are also associated with dense lymphocytic infiltrate and syncytial growth pattern. Pushing borders, dense lymphocytic infiltrate, and syncytial growth pattern are features rarely seen in the LFS-associated breast cancers in our study.

Our findings are similar to those recently reported by Packwood et al. [25]. In their study, which included 36 cases of invasive carcinoma and 9 containing DCIS only they found that the majority of invasive carcinomas were IDC of no special type, high mSBR grade (50%, 18/36), and HER2+ (55%, 20/36). In addition, they found that cancers in TP53 carriers were more likely to be associated with a densely sclerotic stroma compared to noncarriers, and showed high levels of αvβ6 integrin, α-SMA and pSMAD2/3 expression on IHC, suggesting that the dense stromal phenotype may be driven by upregulated TGFβ signaling. The type of stroma was not a variable that we were able to accurately evaluate in our study cohort, and this interesting finding should be validated in further studies.

LFS-associated breast cancer has been shown to be enriched for HER2-positive disease (53–83%) and in about half of all cases is both ER-positive and HER2-positive [2, 7, 8]. In a prior report including all the cases from the current study, 84% of LFS invasive breast cancers were positive for ER and/or PR and 81% were high grade. Sixty-three percent of invasive and 73% of in situ carcinomas were positive for HER2 (IHC 3+ or FISH amplified). Of the invasive tumors, 53% were positive for both ER and HER2 and DCIS was positive for ER and HER2 in 27% of the cases. The percentage of HER2 positivity seen in the current study is slightly lower than that previously reported by the LFS Consortium (60% vs 66%) [2], likely reflecting the particular mix of cases with tissue available for additional studies. It is worth noting the rates of HER2 positivity in invasive carcinoma are much higher than reported for breast cancer overall (53% vs 15%) [26]. In addition, all these HER2+ LFS-associated invasive carcinomas were also ER+, which is the rarest breast cancer subtype (only seen in 10% of breast cancers in the general population) [26]. In this study, ER+/HER2+ invasive carcinomas showed lower ER expression compared to ER+/HER2 negative cancers (mean ER expression 58% vs 92%), similar to non-LFS tumors [27]. The morphology of LFS-associated breast cancers is not entirely unique and shares features with unselected HER2+ disease, although with higher mSBR grade and less tendency to show apocrine features [28, 29].

Somatic TP53 mutations are frequent drivers in sporadic breast cancer, with an overall rate of TP53 mutation in breast cancer of about 28% [21] and higher rates in basal-like (65%) and HER2 enriched subtypes (53%) [21]. Prior studies have shown the distribution of TP53 mutations in germline and sporadic tumors to be similar, with about 75% of all mutations being missense mutations, primarily in the DNA-binding domain (exons 5–8), and a much lower fraction of other mutation types, including nonsense, frameshift, deletion, and splice-site mutations [30, 31]. Strong overexpression of p53 protein was seen in the single case in our current study for which the actual mutation sequence is not available, supporting a probable missense mutation; thus, the overall mutation distribution in the current study is 56% missense (15/27) and 44% non-missense (12/27). The somewhat higher proportion of non-missense mutations in the current cohort relative to prior studies may reflect improved current detection strategies, particularly for the detection of non-missense mutations and mutations in areas outside the DNA-binding domain. In the current study, tumors with high p53 protein expression were more commonly seen in individuals with TP53 germline missense mutations, whereas cases negative for p53 expression (or <1% nuclear expression) by IHC were seen in women carrying deletions or nonsense mutations. These patterns of immunoreactivity are similar to those previously reported for sporadic TP53 mutated breast cancer [22, 32]. Several tumors in our study from patients with germline missense mutations were associated with only moderate expression of p53 protein (5–40%), which may reflect tumor-specific regulatory mechanisms [33].

The morphologic features of LFS-associated breast cancers are not pathognomonic and are similar to those tumors with somatic TP53 mutations [21]. However, it is important to raise awareness among pathologists and clinicians of enrichment of ER+/HER2+ tumors in patients with germline TP53 mutations. The prevalence of germline TP53 mutations in an unselected cohort of breast cancer patients under the age of 50 with HER2+ disease has been previously reported at 1.4% [34]. Based on the clinical importance of correctly identifying germline TP53 carriers, consideration should be given to germline testing, with a panel including the TP53 gene, to young patients with HER2+ breast cancer, especially if there is a notable family history, regardless of morphology.

Limitations of our study include the retrospective nature, the limited number of slides available for pathology review, with only biopsy material available for some cases, and the age of stored paraffin blocks (some >40 years). LFS has been challenging to study, not only because the syndrome is rare and has high mortality, but also because of the complexity of the TP53 gene and its many roles in tumorigenesis. At the current time, there is little published data on the molecular pathogenesis of breast cancer in LFS. Somatic TP53 mutations have been reported in precursor lesions [22, 35], including atypical ductal hyperplasia [35, 36], and in DCIS [22]. These findings suggest that TP53 inactivation occurs in preneoplastic lesions; however, the early steps in breast carcinogenesis in the setting of germline TP53 mutations are poorly understood.

In conclusion, invasive carcinomas in LFS patients are most often high-grade but without the pushing borders, syncytial architecture and dense associated lymphocytic infiltrate characteristic of BRCA-associated cancers. More than half of TP53 germline mutated cancers are HER2 positive, and many also show co-expression of ER. The type of germline mutation present does not appear to correlate with ER and HER2 status. Further studies linking specific TP53 mutations or types of mutations with morphologic and molecular features may be helpful both for understanding disease pathogenesis and for the design of diagnostic and preventive strategies for LFS-affected individuals. In particular, studies of early lesions in LFS-associated breast cancer could be important for informing the design of trials of chemopreventive agents and potential environmental risk modifiers.