Abstract
Breast cancer during pregnancy is rare and is associated with relatively poor prognosis. No information is available on its biological features at the genomic level. Using a dataset of 54 pregnant and 113 non-pregnant breast cancer patients, we evaluated the pattern of hot spot somatic mutations and did transcriptomic profiling using Sequenom and Affymetrix respectively. We performed gene set enrichment analysis to evaluate the pathways associated with diagnosis during pregnancy. We also evaluated the expression of selected cancer-related genes in pregnant and non-pregnant patients and correlated the results with changes occurring in the normal breast using a pregnant murine model. We finally investigated aberrations associated with disease-free survival (DFS). No significant differences in mutations were observed. Of the total number of patients, 18.6% of pregnant and 23% of non-pregnant patients had a PIK3CA mutation. Around 30% of tumors were basal, with no differences in the distribution of breast cancer molecular subtypes between pregnant and non-pregnant patients. Two pathways were enriched in tumors diagnosed during pregnancy: the G protein-coupled receptor pathway and the serotonin receptor pathway (FDR <0.0001). Tumors diagnosed during pregnancy had higher expression of PD1 (PDCD1; P=0.015), PDL1 (CD274; P=0.014), and gene sets related to SRC (P=0.004), IGF1 (P=0.032), and β-catenin (P=0.019). Their expression increased almost linearly throughout gestation when evaluated on the normal breast using a pregnant mouse model underscoring the potential effect of the breast microenvironment on tumor phenotype. No genes were associated with DFS in a multivariate model, which could be due to low statistical power. Diagnosis during pregnancy impacts the breast cancer transcriptome including potential cancer targets.
Introduction
Breast cancer arising in young women exhibits distinct biological features and has been shown to pursue an aggressive clinical behavior (Cancello et al. 2010, Azim et al. 2012a, Collins et al. 2012, Narod 2012). Among the unique situations that young women with breast cancer may face is pregnancy. The fact that breast cancer is a hormonally driven tumor, and given the considerable structural changes that occur in the breast during pregnancy, raises concerns on the potential impact of pregnancy on breast cancer biology.
It is estimated that one out of 3000 pregnancies is complicated with breast cancer, and this incidence appears to be on the rise (Peccatori et al. 2013). While in recent years, there have been considerable efforts in defining safe and effective treatment strategies in managing breast cancer arising during pregnancy (Cardonick et al. 2010, Amant et al. 2012, Loibl et al. 2012), very little progress have been made regarding our understanding of the biology of these tumors.
Recently, several groups including ours have published on the prognosis of patients diagnosed with breast cancer during pregnancy (Azim et al. 2012b, Amant et al. 2013, Litton et al. 2013). Results of these studies have witnessed major controversy with studies showing worse outcomes, while others showing no difference in prognosis or even improved survival. These results further call for the importance of elucidating the effects of diagnosis during pregnancy on breast cancer biology.
In the current study, we investigated for the first time the effect of diagnosis during pregnancy on the biology of breast cancer arising in young women at the genomic level; particularly, the pattern of hot spot somatic mutations and gene expression profiles.
Materials and methods
We identified 65 patients who were diagnosed with breast cancer during pregnancy in the period from 1996 to 2010 at the European Institute of Oncology (IEO) in Milan, Italy. For each case, we identified two breast cancer patients who were not diagnosed during pregnancy, but who were matched according to age (±2 years), tumor size, nodal status, year of surgery (±2 years), and whether neoadjuvant chemotherapy was administered. Details regarding patient characteristics, pathological features, and treatments were published previously (Azim et al. 2012b). In brief, we did not find any obvious differences in the pathological features or the expression of estrogen receptor (ER), progesterone receptor (PgR), HER2, and Ki-67. However, at a median follow-up of 4 years, pregnant patients had significantly worse disease-free survival (DFS; Azim et al. 2012b).
All patients had available formalin-fixed paraffin-embedded (FFPE) tissue from the primary breast surgery. For the sake of this study, all samples were pathologically reviewed at IEO and three 1 mm tumor cores (>90% tumor cellularity) were sent to the Breast Cancer Translational Research Laboratory at Institut Jules Bordet for mutational and transcriptomic profiling. DNA and RNA extraction were performed using the ZYMO Research Pinpoint Slide DNA Isolation System and stored at −80 °C until use. Quality and quantity were evaluated using the NanoDrop.
All patients provided consent to donate biological samples for research purposes per IEO institutional policies. This study was approved by the Ethics Committee of Institut Jules Bordet in 2010 (number: 1782).
Mutational profiling
The samples were genotyped using the Sequenom MassARRAY Assay Design 3.1 with the default parameters. Multiplex PCR was done in 5 μl volume containing 5–10 ng DNA. Samples were considered to be of sufficient quality when genotyping could be performed for more than 75% of the mutations.
We queried the COSMIC database to identify a broad range of genes for hotspot somatic mutation profiling (http://cancer.sanger.ac.uk/cancergenome/projects/cosmic/). In total, we examined 84 hotspot somatic mutations on 19 cancer-related genes. Details regarding the genes and mutations that were evaluated are provided in Supplementary Table 1, see section on supplementary data given at the end of this article. We ultimately covered 94% (n=29) of the reported PIK3CA mutations.
Transcriptomic profiling
All samples were hybridized using Affymetrix Human Genome U219 array plates following standard Affymetrix protocols. This platform was used based on pilot studies performed in-house showing high reproducibility between transcriptomic information obtained from paired frozen and FFPE tumor samples. Detailed information on the used protocol is provided in Supplementary Materials and Methods, see section on supplementary data given at the end of this article.
We used two methods to determine breast cancer molecular subtypes: PAM50 (Parker et al. 2009) and 3-Gene (Haibe-Kains et al. 2012), as defined previously (Supplementary Materials and Methods). To compare the transcriptomic profiles of patients diagnosed with breast cancer during pregnancy and of those not diagnosed during pregnancy, we performed gene set enrichment analysis (GSEA) using the Gene Ontology classes as a reference gene set using the Bioconductor Package Piano (Varemo et al. 2013). The genes were sorted based on their differential expression between pregnant and non-pregnant patients using a conditional logistic regression model (http://hosho.ees.hokudai.ac.jp/∼kubo/Rdoc/library/Epi/html/clogistic.html). Results were cross-checked with those obtained using the hypergeometric probabilities delivered by the DAVID web tool (Huang da et al. 2009).
We then evaluated the expression of selected cancer-related genes such as BRCA1, BRCA2, PD1 (PDCD1), and PDL1 (CD274) in addition to previously published gene sets representing several important oncological biological processes in pregnant and non-pregnant patients. These gene sets included the following: RAS, SRC, MYC, Wnt/β-catenin (Bild et al. 2006), PI3K (Loi et al. 2010), insulin-like growth factor 1 (IGF1; Creighton et al. 2008), AKTmTOR (Majumder et al. 2004), MAPK (Creighton et al. 2006), PTEN (Saal et al. 2008); proliferation gene sets; GGI (Sotiriou et al. 2006) and GENE70 (van 't Veer et al. 2002), mammary stem cell and luminal progenitor gene sets (Lim et al. 2009); stroma gene sets; DCN (Farmer et al. 2009) and PLAU (Desmedt et al. 2008), and immune gene sets; and STAT1 (Desmedt et al. 2008) and IRM (Teschendorff et al. 2007).
Gene sets (or signature score) were computed by taking the sum of the products of the gene coefficient in the module (si) and the corresponding gene expression value (ei) according to the following formula:
Supplementary Table 2, see section on supplementary data given at the end of this article summarizes the size of the different gene sets and the number of genes that were available in the Affymetrix platform that we used in the current analysis. The raw gene expression data, together with patients' characteristics, are publically available on GEO http://www.ncbi.nlm.nih.gov/geo/, with accession number GSE53031.
In-silico analysis of publically available gene expression data derived from the normal breast using a pregnant mouse model
In order to evaluate the potential role of pregnancy in the normal breast tissue, we investigated the expression of genes or gene sets that were found to be differentially expressed between the pregnant and non-pregnant groups on a publically available murine data set derived from the normal breast of pregnant mice (GEO ID: GSE8191; Anderson et al. 2007). In the murine model, the breast tissue samples were obtained on days 1, 3, 7, 12, and 17 of pregnancy and also on day 19 (post-partum). Four replicates were available for each time point. Similar to the procedure that was applied in the human data set, the data were re-normalized using the robust multiarray analysis normalization, and the expression values of each set of probes corresponding to one Ensembl identifier were averaged (Bolstad et al. 2003).
Survival analysis
We performed updated DFS comparison between pregnant and non-pregnant breast cancer patients using a multivariate Cox regression model adjusted for age, tumor size, nodal status, histological grade, ER, HER2, and type of systemic therapy.
We performed an exploratory analysis to evaluate genes/gene sets or somatic mutations that were associated with DFS. We used the median expression value in the whole population (i.e. pregnant+non-pregnant) to define the cutoff of high and low gene or gene set expression. Analyses were performed using the SAS Software, version 9.2 (SAS Institute, Cary, NC, USA) and R Software, version 2.12.2 (available at www.r-project.org). All tests were two sided.
Results
Study population
The original data set included 195 breast cancer patients: 65 pregnant and 130 non-pregnant patients. For the current genomic analysis, we opted to exclude patients who received neoadjuvant therapy to avoid potential impact on the obtained results (n=21). Seven more patients were excluded due to poor RNA quality resulting in a total of 167 patients (85% of original data set): 54 pregnant and 113 non-pregnant patients who were included in the current analysis (Supplementary Fig. 1, see section on supplementary data given at the end of this article). Table 1 summarizes patient characteristics in comparison to the original data set. No major differences were observed.
Differences in characteristics between the whole study population and the subpopulation (∼85%) included in the genomic analysis
| Breast cancer patients diagnosed during pregnancy, n (%) | Breast cancer patients not diagnosed during pregnancy, n (%) | |||
|---|---|---|---|---|
| Whole study | Genomic analysis | Whole study | Genomic analysis | |
| Number | 65 | 54 | 130 | 113 |
| Tumor size ≥2 cm | 37 (56.9%) | 28 (51.9%) | 84 (64.6%) | 62 (54.9%) |
| Node positive | 35 (53.8%) | 28 (51.9%) | 70 (53.8%) | 61 (54%) |
| Histological grade III | 36 (55.3%) | 32 (59.2%) | 68 (52.3%) | 61 (54%) |
| ER-positive | 43 (66.1%) | 36 (66.7%) | 96 (73.8%) | 88 (77.8%) |
| PgR-positive | 42 (64.6%) | 35 (64.8%) | 85 (65.4%) | 79 (69.9%) |
| HER2-positive | 11 (16.9%) | 9 (16.7%) | 23 (17.7%) | 22 (19.5%) |
| Ki-67 ≥30% | 35 (53.9%) | 31 (57.4%) | 68 (52.3%) | 54 (47.7%) |
| No. of DFS events | 27 (41.5%) | 19 (35.2%) | 33 (25.3%) | 23 (20.4%) |
| No. of OS events | 13 (20%) | 10 (18.5%) | 19 (14.6%) | 11 (9.7%) |
ER, estrogen receptor; PgR, progesterone receptor; DFS, disease-free survival; OS, overall survival.
Mutational analysis
No significant differences in the pattern of hot spot somatic mutations evaluated in this study were observed between pregnant and non-pregnant patients (Table 2). Considering all eligible patients, 36 (21.6%) and 9 (5.4%) had a mutation in PIK3CA and TP53 respectively.
Differences in the prevalence of hot spot somatic mutations in tumors diagnosed in pregnant and non-pregnant breast cancer patients
| Patients diagnosed during pregnancy, n (%) | Patients not diagnosed during pregnancy, n (%) | P value | |
|---|---|---|---|
| Total number | 54 (100%) | 113 (100%) | 0.51 |
| PIK3CA | |||
| All | 10 (18.6%) | 26 (23%) | |
| Exon 9 mutations | 3 (5.6%) | 9 (8%) | |
| Exon 20 mutations | 4 (7.4%) | 12 (10.6%) | |
| Other PIK3CA mutations | 3 (5.6%) | 5 (4.4%) | |
| p53 | 2 (3.7%) | 7 (6.2%) | 0.51 |
| MAP3K1 | 1 (1.9%) | 2 (1.8%) | 0.97 |
| STK1 | 1 (1.9%) | 0 | 0.15 |
| RET | 0 | 1 (0.9%) | 0.49 |
Molecular subtyping
We found a high correlation between PAM50 and the 3-Gene classifiers in defining breast cancer molecular subtypes (r=0.857, P<0.0001). There was also a high correlation with immunohistochemistry (IHC)-defined subtypes based on the expression of ER, HER2, and a Ki-67 cutoff of 14% to distinguish luminal A and luminal B (both r>0.7, P<0.0001; Supplementary Table 3, see section on supplementary data given at the end of this article). We did not find any significant differences in breast cancer subtypes defined by either of the above classifiers between pregnant and non-pregnant breast cancer patients (Fig. 1).
Distribution of breast cancer molecular subtypes in patients diagnosed during pregnancy and patients not diagnosed during pregnancy using the PAM50 and 3-Gene molecular classifiers. No significant differences were observed using any of the two methods (P=0.14 and P=0.68 for PAM50 and 3-Gene respectively).
Citation: Endocrine-Related Cancer 21, 4; 10.1530/ERC-14-0111
Distribution of breast cancer molecular subtypes in patients diagnosed during pregnancy and patients not diagnosed during pregnancy using the PAM50 and 3-Gene molecular classifiers. No significant differences were observed using any of the two methods (P=0.14 and P=0.68 for PAM50 and 3-Gene respectively).
Citation: Endocrine-Related Cancer 21, 4; 10.1530/ERC-14-0111
Distribution of breast cancer molecular subtypes in patients diagnosed during pregnancy and patients not diagnosed during pregnancy using the PAM50 and 3-Gene molecular classifiers. No significant differences were observed using any of the two methods (P=0.14 and P=0.68 for PAM50 and 3-Gene respectively).
Citation: Endocrine-Related Cancer 21, 4; 10.1530/ERC-14-0111
Gene set enrichment analysis
After adjusting for multiple testing, two relevant pathways were significantly enriched in tumors diagnosed in pregnant vs non-pregnant breast cancer patients; the G protein-coupled receptor (GPCR) signaling pathway and the serotonin receptor signaling pathway (both P<0.0001; Fig. 2). The results obtained by the GSEA approach were also confirmed using the hypergeometric probabilities delivered by the DAVID web tool on the non-sorted list of significantly differentially expressed genes (Supplementary Fig. 2, see section on supplementary data given at the end of this article).
Gene set enrichment analysis (GSEA) showing high expression of the G protein-coupled receptor pathway and the serotonin receptor pathway in patients diagnosed with breast cancer during pregnancy (both P values after correction for multiple testing; FDR <0.0001).
Citation: Endocrine-Related Cancer 21, 4; 10.1530/ERC-14-0111
Gene set enrichment analysis (GSEA) showing high expression of the G protein-coupled receptor pathway and the serotonin receptor pathway in patients diagnosed with breast cancer during pregnancy (both P values after correction for multiple testing; FDR <0.0001).
Citation: Endocrine-Related Cancer 21, 4; 10.1530/ERC-14-0111
Gene set enrichment analysis (GSEA) showing high expression of the G protein-coupled receptor pathway and the serotonin receptor pathway in patients diagnosed with breast cancer during pregnancy (both P values after correction for multiple testing; FDR <0.0001).
Citation: Endocrine-Related Cancer 21, 4; 10.1530/ERC-14-0111
Differential expression of selected cancer-related genes and published gene sets on tumors diagnosed in pregnant vs non-pregnant breast cancer patients
In pregnant breast cancer patients, we found higher expression of PD1 (P=0.015), PDL1 (P=0.014), BRCA1 (P=0.018), and gene sets related to SRC (P=0.004), IGF1 (P=0.032), and β-catenin (P=0.019) (Fig. 3). Most of them showed a trend of higher expression in patients diagnosed later in the course of pregnancy, although this was not statistically significant. On the contrary, we found lower expression of AKTmTOR gene set (P=0.05) in pregnant breast cancer patients, with patients diagnosed during the third trimester of pregnancy having the lowest expression (Supplementary Fig. 3, see section on supplementary data given at the end of this article).
High expression of different genes and gene sets in pregnant compared with non-pregnant breast cancer patients. Differential expression according to period of gestation at breast cancer diagnosis is shown (i.e. first, second, and third trimesters). One patient had missing information on period of gestation at breast cancer diagnosis. The Y-axis represents the relevant expression value where each gene or gene set was treated as a continuous variable. (A) SRC gene set (pregnant vs non-pregnant, P=0.004; first trimester vs third trimester, P=0.28); (B) IGF1 gene set (pregnant vs non-pregnant, P=0.032; first trimester vs third trimester, P=0.61); (C) β-catenin gene set (pregnant vs non-pregnant, P=0.019; first trimester vs third trimester, P=0.45); (D) PD1 (pregnant vs non-pregnant, P=0.015; first trimester vs third trimester, P=0.1); (E) PDL1 (pregnant vs non-pregnant, P=0.014; first trimester vs third trimester, P=0.05); and (F) BRCA1 (pregnant vs non-pregnant, P=0.018; first trimester vs third trimester, P=0.74).
Citation: Endocrine-Related Cancer 21, 4; 10.1530/ERC-14-0111
High expression of different genes and gene sets in pregnant compared with non-pregnant breast cancer patients. Differential expression according to period of gestation at breast cancer diagnosis is shown (i.e. first, second, and third trimesters). One patient had missing information on period of gestation at breast cancer diagnosis. The Y-axis represents the relevant expression value where each gene or gene set was treated as a continuous variable. (A) SRC gene set (pregnant vs non-pregnant, P=0.004; first trimester vs third trimester, P=0.28); (B) IGF1 gene set (pregnant vs non-pregnant, P=0.032; first trimester vs third trimester, P=0.61); (C) β-catenin gene set (pregnant vs non-pregnant, P=0.019; first trimester vs third trimester, P=0.45); (D) PD1 (pregnant vs non-pregnant, P=0.015; first trimester vs third trimester, P=0.1); (E) PDL1 (pregnant vs non-pregnant, P=0.014; first trimester vs third trimester, P=0.05); and (F) BRCA1 (pregnant vs non-pregnant, P=0.018; first trimester vs third trimester, P=0.74).
Citation: Endocrine-Related Cancer 21, 4; 10.1530/ERC-14-0111
High expression of different genes and gene sets in pregnant compared with non-pregnant breast cancer patients. Differential expression according to period of gestation at breast cancer diagnosis is shown (i.e. first, second, and third trimesters). One patient had missing information on period of gestation at breast cancer diagnosis. The Y-axis represents the relevant expression value where each gene or gene set was treated as a continuous variable. (A) SRC gene set (pregnant vs non-pregnant, P=0.004; first trimester vs third trimester, P=0.28); (B) IGF1 gene set (pregnant vs non-pregnant, P=0.032; first trimester vs third trimester, P=0.61); (C) β-catenin gene set (pregnant vs non-pregnant, P=0.019; first trimester vs third trimester, P=0.45); (D) PD1 (pregnant vs non-pregnant, P=0.015; first trimester vs third trimester, P=0.1); (E) PDL1 (pregnant vs non-pregnant, P=0.014; first trimester vs third trimester, P=0.05); and (F) BRCA1 (pregnant vs non-pregnant, P=0.018; first trimester vs third trimester, P=0.74).
Citation: Endocrine-Related Cancer 21, 4; 10.1530/ERC-14-0111
No significant difference in expression of the other genes and gene sets was observed between patients diagnosed with breast cancer during pregnancy and non-pregnant patients considered in this study (Supplementary Fig. 4, see section on supplementary data given at the end of this article).
The expression of the differentially expressed genes and gene sets on the normal breast during the course of gestation in a pregnant mouse model
While evaluating the expression of the genes and gene sets that were highly expressed in tumors diagnosed during pregnancy in the normal breast using a pregnant mouse model, we found almost a linear increase in expression from D1 to D17 of pregnancy in all the genes and gene sets evaluated with the exception of BRCA1 (Fig. 4). The opposite trend was observed with AKTmTOR module where the expression was lower in patients diagnosed during pregnancy (P=0.05) with a linear decrease in expression in the normal breast from D1 to D17 in the pregnant mouse model (Supplementary Fig. 3).
Differential expression of genes and gene sets on the normal breast in a pregnant mouse model throughout the period of gestation. D1 represents conception, D17 end of pregnancy, and D19 post-partum. The Y-axis represents the relevant expression value where each gene or gene set was treated as a continuous variable. (A) SRC gene set (days 1 vs 17, P=0.035); (B) IGF1 gene set (days 1 vs 17, P=0.004); (C) BRCA1 (days 1 vs 17, P=0.11); (D) β-catenin gene set (days 1 vs 17, P=0.008); and (E) PD1 (days 1 vs 17, P=0.001).
Citation: Endocrine-Related Cancer 21, 4; 10.1530/ERC-14-0111
Differential expression of genes and gene sets on the normal breast in a pregnant mouse model throughout the period of gestation. D1 represents conception, D17 end of pregnancy, and D19 post-partum. The Y-axis represents the relevant expression value where each gene or gene set was treated as a continuous variable. (A) SRC gene set (days 1 vs 17, P=0.035); (B) IGF1 gene set (days 1 vs 17, P=0.004); (C) BRCA1 (days 1 vs 17, P=0.11); (D) β-catenin gene set (days 1 vs 17, P=0.008); and (E) PD1 (days 1 vs 17, P=0.001).
Citation: Endocrine-Related Cancer 21, 4; 10.1530/ERC-14-0111
Differential expression of genes and gene sets on the normal breast in a pregnant mouse model throughout the period of gestation. D1 represents conception, D17 end of pregnancy, and D19 post-partum. The Y-axis represents the relevant expression value where each gene or gene set was treated as a continuous variable. (A) SRC gene set (days 1 vs 17, P=0.035); (B) IGF1 gene set (days 1 vs 17, P=0.004); (C) BRCA1 (days 1 vs 17, P=0.11); (D) β-catenin gene set (days 1 vs 17, P=0.008); and (E) PD1 (days 1 vs 17, P=0.001).
Citation: Endocrine-Related Cancer 21, 4; 10.1530/ERC-14-0111
Survival analysis
At a median follow-up of 65 months, patients diagnosed during pregnancy continued to have a poorer DFS compared with the non-pregnant group (HR 2.5 (95% CI, 1.3–4.8), P=0.005), even after adjusting for classical prognostic factors (Supplementary Fig. 5, see section on supplementary data given at the end of this article).
No genes were found to be associated with DFS in the univariate model using false discovery rate (FDR) <0.05. Among the genes and gene sets that were highly expressed in the pregnant group, only IGF1 was significantly associated with DFS, both in the pregnant (P=0.01) and the non-pregnant groups (P=0.04) (Supplementary Fig. 6, see section on supplementary data given at the end of this article). However, this was not significant in a multivariate model adjusted for classical prognostic parameters (HR=1.0, P=0.4). No association was observed between PIK3CA mutation and DFS as well (Supplementary Fig. 7).
Discussion
In this study, we evaluated the impact of diagnosis during pregnancy on breast cancer biology. We found no obvious differences in classical pathological features, pattern of hot spot somatic mutations, or breast cancer molecular subtypes. However, tumors diagnosed during pregnancy were associated with different gene expression patterns and activated signaling pathways, namely the GPCR pathway and the serotonin receptor pathway.
GPCRs represent the largest family of cell surface molecules involved in signal transmission, which include ∼2–4% of genes in the genome. They regulate many cell functions including proliferation, angiogenesis, and cell survival (Dorsam & Gutkind 2007). CXCR4 and GPR30 are GPCRs that were shown to be of higher relevance in breast cancer; the former was found in preclinical models to increase the metastatic potential of breast cancer cells (Muller et al. 2001), while the latter was linked to the process of endocrine resistance (Thomas et al. 2005). The serotonin receptor pathway belongs to the GPCR family (Kvachnina et al. 2005). Serotonin is known to play a role in mammary gland development (Matsuda et al. 2004). It is synthesized within the breast epithelial cells during lactation and acts as a physiological regulator of involution following pregnancy, in part by favoring growth arrest and cell death (Marshall et al. 2010). The application of serotonin to primary human epithelial cells was shown to suppress cell growth (Pai et al. 2009). Conversely, serotonin was found to stimulate tumor growth when applied to invasive breast cancer cell lines, by promoting proliferation and epithelial–mesenchymal transition (Pai et al. 2009). Moreover, serotonin was found to induce the transcription of parathyroid hormone-related protein and the metastasis-associated transcription factor Runx2/Cbfal via the serotonin receptor, 5-HT2 (Hernandez et al. 2012). Hence, it is plausible that the activation of the serotonin receptor pathway in tumors arising during pregnancy may play a role in controlling the aggressive clinical behavior of these tumors.
We found that tumors diagnosed during pregnancy have higher IGF1 gene set expression. IFG1 is known to increase breast cancer development and breast cancer recurrence (Kim et al. 2007, Smith et al. 2011). In the normal breast, the IGF1 pathway is known to be critical for mammary gland development and progesterone regulation and is required for estrogen to cause ductal development (Macias & Hinck 2012). As the breast undergoes part of its development during pregnancy, we hypothesize that the increase in IGF1 expression that we observed in the normal breast over the course of gestation could be to support mammary gland development. However, a previous study has demonstrated that ‘previous’ pregnancy down-regulates IGF1 expression in the normal breast, suggesting that this could explain the long-term protective effect of pregnancy on breast cancer (D'Cruz et al. 2002). This is not in direct conflict with our results as pregnancy is known to exert a bi-directional effect on breast cancer development, promoting aggressive tumor growth on the short term with a long-term protective effect (Albrektsen et al. 2005). We speculate that pregnancy modifies the breast microenvironment, which consequently impacts breast cancer risk and biology with IGF1 emerging as one of the main pathways that appear to be implicated in the crosstalk between pregnancy and breast cancer.
The high expression of PD1 is also intriguing. PD1 is a key immune-checkpoint receptor expressed by activated T-cells and mediates immunosuppression by binding to its ligand, PDL1. During pregnancy, PD1–PDL1 interaction is activated to allow feto-maternal tolerance (D'Addio et al. 2011). On the other hand, in cancer, inhibition of the interaction between PD1 and its ligand PDL1 enhances T-cell responses in vitro and mediates antitumor activity (Dong et al. 2002, Iwai et al. 2002). In a recent study, an anti-PD1 was associated with an impressive clinical activity in patients with refractory solid tumors, mainly those with positive expression of PDL1 (Topalian et al. 2012). Our findings point out that it is likely that patients diagnosed during pregnancy express PDL1, which appears to be mediated by the normally higher PDL1 expression during pregnancy. It is plausible that the high expression of PD1/PDL1 on the tumor mediates immune suppression and hence results in aggressive tumor growth. However, this remains speculative requiring further confirmation.
In line with our findings, it was previously shown that BRCA1 is highly expressed in the breast during the first weeks of pregnancy possibly to counteract the state of proliferation and differentiation that occur in the breast to form mature alveoli (Marquis et al. 1995). It was later suggested that patients with germline BRCA mutation might have a higher incidence of pregnancy-associated breast cancer (Johannsson et al. 1998). Yet this was based on 21 patients mostly diagnosed in the year following pregnancy. Unfortunately, none of our patients underwent testing for germline BRCA mutation. We investigated the expression of a gene set that was derived from breast cancer patients with BRCA1 mutation (van 't Veer et al. 2002), yet no difference in expression was observed. Hence, we cannot draw solid conclusion on whether BRCA mutations are more common during pregnancy.
This is the first report on the pattern of selected hot spot somatic mutations, not only in the pregnant population, but also in young breast cancer patients in general. No obvious differences were observed. The overall incidence of PIK3CA mutations was ∼20%, which is rather in line with the known incidence of PIK3CA mutations in the general breast cancer population irrespective of age (Nik-Zainal et al. 2012, Stephens et al. 2012). Acknowledging that these tumors were probably initiated well before pregnancy, it is conceivable that no differences in mutations were observed between the two groups. However, we plan to perform whole-exome sequencing to refine our understanding regarding the pattern of mutations in these patients.
In contrast to our findings, two recent studies have shown that pregnant breast cancer patients do not have poorer prognosis (Amant et al. 2013, Litton et al. 2013). The discussion of this topic is beyond the scope of this manuscript; however, we believe that several factors could explain these contradicting results including the limitations of the case–control design, differences in identifying matching controls, lack of statistical power, and different treatment strategies. Nevertheless, the question related to the impact of pregnancy on breast cancer biology is valid and our results support that diagnosis during pregnancy modifies the transcriptome of these tumors.
Given the lack of statistical power, we cannot draw solid conclusions on the prognostic value of the different molecular aberrations associated with diagnosis during pregnancy. Hence, we cannot confirm that the poor outcome in the pregnant group is secondary to the unique gene expression pattern that was observed. Another limitation is that we did not have an adjacent normal tissue to evaluate the effect of pregnancy on the normal human breast. Thus, we evaluated the changes in the breast using a publically available mouse model instead, which obviously does not fully represent as to what occurs in the human pregnant breast.
In conclusion, this is a unique study interrogating the effect of pregnancy on breast cancer biology. Given the rarity of this disease, we believe that it could serve as a valuable resource for future research in this field. While we did not observe differences in selected somatic mutations, the altered breast microenvironment during pregnancy appeared to influence the tumor phenotype, as shown in our transcriptomic analysis. This could potentially impact the behavior of these tumors.
Supplementary data
This is linked to the online version of the paper at http://dx.doi.org/10.1530/ERC-14-0111.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
Funding
H A Azim Jr was supported by an ESMO translational research grant to conduct this Study. This research was supported by the fund from Les Amis De L'Institut Bordet and the ‘Fond de la Recherche Scientifique (FNRS)’ (ref: FC 55854).
Author contribution statement
H A Azim Jr, F A Peccatori, N Rotmensz, M Piccart, and C Sotiriou were responsible for study concept and design. H A Azim Jr, F A Peccatori, N Rotmensz, and P Dell'Orto were responsible for data collection. P Dell'Orto, G Pruneri, and G Viale were responsible for pathological evaluation. H A Azim Jr, S Brohée, S Majjaj, C Desmedt, S Loi, V Jose, M Ignatiadis, and C Sotiriou were responsible for gene expression profiling analysis. D Lambrechts was responsible for mutational profiling. S Brohée and N Rotmensz were responsible for the statistical analysis. All authors have contributed to data interpretation. H A Azim Jr, S Brohée, F A Peccatori, and C Sotiriou were responsible for manuscript writing. All authors approved the final version of the manuscript.
Acknowledgements
This work was presented as an oral presentation in the 9th European Breast Cancer Conference (EBCC), March 2014, Glasgow, UK.
References
Albrektsen G, Heuch I, Hansen S & Kvale G 2005 Breast cancer risk by age at birth, time since birth and time intervals between births: exploring interaction effects. British Journal of Cancer 92 167–175. (doi:10.1038/sj.bjc.6602302).
Amant F, Van Calsteren K, Halaska MJ, Gziri MM, Hui W, Lagae L, Willemsen MA, Kapusta L, Van Calster B & Wouters H et al. 2012 Long-term cognitive and cardiac outcomes after prenatal exposure to chemotherapy in children aged 18 months or older: an observational study. Lancet Oncology 13 256–264. (doi:10.1016/S1470-2045(11)70363-1).
Amant F, von Minckwitz G, Han SN, Bontenbal M, Ring AE, Giermek J, Wildiers H, Fehm T, Linn SC & Schlehe B et al. 2013 Prognosis of women with primary breast cancer diagnosed during pregnancy: results from an International Collaborative Study. Journal of Clinical Oncology 31 2532–2539. (doi:10.1200/JCO.2012.45.6335).
Anderson SM, Rudolph MC, McManaman JL & Neville MC 2007 Key stages in mammary gland development. Secretory activation in the mammary gland: it's not just about milk protein synthesis! Breast Cancer Research 9 204. (doi:10.1186/bcr1653).
Azim HA Jr, Michiels S, Bedard PL, Singhal SK, Criscitiello C, Ignatiadis M, Haibe-Kains B, Piccart MJ, Sotiriou C & Loi S 2012a Elucidating prognosis and biology of breast cancer arising in young women using gene expression profiling. Clinical Cancer Research 18 1341–1351. (doi:10.1158/1078-0432.CCR-11-2599).
Azim HA Jr, Botteri E, Renne G, Dell'orto P, Rotmensz N, Gentilini O, Sangalli C, Pruneri G, Di Nubila B & Locatelli M et al. 2012b The biological features and prognosis of breast cancer diagnosed during pregnancy: a case–control study. Acta Oncologica 51 653–661. (doi:10.3109/0284186X.2011.636069).
Bild AH, Yao G, Chang JT, Wang Q, Potti A, Chasse D, Joshi MB, Harpole D, Lancaster JM & Berchuck A et al. 2006 Oncogenic pathway signatures in human cancers as a guide to targeted therapies. Nature 439 353–357. (doi:10.1038/nature04296).
Bolstad BM, Irizarry RA, Astrand M & Speed TP 2003 A comparison of normalization methods for high density oligonucleotide array data based on variance and bias. Bioinformatics 19 185–193. (doi:10.1093/bioinformatics/19.2.185).
Cancello G, Maisonneuve P, Rotmensz N, Viale G, Mastropasqua MG, Pruneri G, Veronesi P, Torrisi R, Montagna E & Luini A et al. 2010 Prognosis and adjuvant treatment effects in selected breast cancer subtypes of very young women (<35 years) with operable breast cancer. Annals of Oncology 21 1974–1981. (doi:10.1093/annonc/mdq072).
Cardonick E, Dougherty R, Grana G, Gilmandyar D, Ghaffar S & Usmani A 2010 Breast cancer during pregnancy: maternal and fetal outcomes. Cancer Journal 16 76–82. (doi:10.1097/PPO.0b013e3181ce46f9).
Collins LC, Marotti JD, Gelber S, Cole K, Ruddy K, Kereakoglow S, Brachtel EF, Schapira L, Come SE & Winer EP et al. 2012 Pathologic features and molecular phenotype by patient age in a large cohort of young women with breast cancer. Breast Cancer Research and Treatment 131 1061–1066. (doi:10.1007/s10549-011-1872-9).
Creighton CJ, Hilger AM, Murthy S, Rae JM, Chinnaiyan AM & El-Ashry D 2006 Activation of mitogen-activated protein kinase in estrogen receptor α-positive breast cancer cells in vitro induces an in vivo molecular phenotype of estrogen receptor α-negative human breast tumors. Cancer Research 66 3903–3911. (doi:10.1158/0008-5472.CAN-05-4363).
Creighton CJ, Casa A, Lazard Z, Huang S, Tsimelzon A, Hilsenbeck SG, Osborne CK & Lee AV 2008 Insulin-like growth factor-I activates gene transcription programs strongly associated with poor breast cancer prognosis. Journal of Clinical Oncology 26 4078–4085. (doi:10.1200/JCO.2007.13.4429).
D'Addio F, Riella LV, Mfarrej BG, Chabtini L, Adams LT, Yeung M, Yagita H, Azuma M, Sayegh MH & Guleria I 2011 The link between the PDL1 costimulatory pathway and Th17 in fetomaternal tolerance. Journal of Immunology 187 4530–4541. (doi:10.4049/jimmunol.1002031).
D'Cruz CM, Moody SE, Master SR, Hartman JL, Keiper EA, Imielinski MB, Cox JD, Wang JY, Ha SI & Keister BA et al. 2002 Persistent parity-induced changes in growth factors, TGF-β3, and differentiation in the rodent mammary gland. Molecular Endocrinology 16 2034–2051. (doi:10.1210/me.2002-0073).
Desmedt C, Haibe-Kains B, Wirapati P, Buyse M, Larsimont D, Bontempi G, Delorenzi M, Piccart M & Sotiriou C 2008 Biological processes associated with breast cancer clinical outcome depend on the molecular subtypes. Clinical Cancer Research 14 5158–5165. (doi:10.1158/1078-0432.CCR-07-4756).
Dong H, Strome SE, Salomao DR, Tamura H, Hirano F, Flies DB, Roche PC, Lu J, Zhu G & Tamada K et al. 2002 Tumor-associated B7-H1 promotes T-cell apoptosis: a potential mechanism of immune evasion. Nature Medicine 8 793–800. (doi:10.1038/nm0902-1039c).
Dorsam RT & Gutkind JS 2007 G-protein-coupled receptors and cancer. Nature Reviews. Cancer 7 79–94. (doi:10.1038/nrc2069).
Farmer P, Bonnefoi H, Anderle P, Cameron D, Wirapati P, Becette V, Andre S, Piccart M, Campone M & Brain E et al. 2009 A stroma-related gene signature predicts resistance to neoadjuvant chemotherapy in breast cancer. Nature Medicine 15 68–74. (doi:10.1038/nm.1908).
Haibe-Kains B, Desmedt C, Loi S, Culhane AC, Bontempi G, Quackenbush J & Sotiriou C 2012 A three-gene model to robustly identify breast cancer molecular subtypes. Journal of the National Cancer Institute 104 311–325. (doi:10.1093/jnci/djr545).
Hernandez LL, Gregerson KA & Horseman ND 2012 Mammary gland serotonin regulates parathyroid hormone-related protein and other bone-related signals. American Journal of Physiology. Endocrinology and Metabolism 302 E1009–E1015. (doi:10.1152/ajpendo.00666.2011).
Huang da W, Sherman BT & Lempicki RA 2009 Bioinformatics enrichment tools: paths toward the comprehensive functional analysis of large gene lists. Nucleic Acids Research 37 1–13. (doi:10.1093/nar/gkn923).
Iwai Y, Ishida M, Tanaka Y, Okazaki T, Honjo T & Minato N 2002 Involvement of PD-L1 on tumor cells in the escape from host immune system and tumor immunotherapy by PD-L1 blockade. PNAS 99 12293–12297. (doi:10.1073/pnas.192461099).
Johannsson O, Loman N, Borg A & Olsson H 1998 Pregnancy-associated breast cancer in BRCA1 and BRCA2 germline mutation carriers. Lancet 352 1359–1360. (doi:10.1016/S0140-6736(05)60750-7).
Kim HJ, Litzenburger BC, Cui X, Delgado DA, Grabiner BC, Lin X, Lewis MT, Gottardis MM, Wong TW & Attar RM et al. 2007 Constitutively active type I insulin-like growth factor receptor causes transformation and xenograft growth of immortalized mammary epithelial cells and is accompanied by an epithelial-to-mesenchymal transition mediated by NF-κB and snail. Molecular and Cellular Biology 27 3165–3175. (doi:10.1128/MCB.01315-06).
Kvachnina E, Liu G, Dityatev A, Renner U, Dumuis A, Richter DW, Dityateva G, Schachner M, Voyno-Yasenetskaya TA & Ponimaskin EG 2005 5-HT7 receptor is coupled to Gα subunits of heterotrimeric G12-protein to regulate gene transcription and neuronal morphology. Journal of Neuroscience 25 7821–7830. (doi:10.1523/JNEUROSCI.1790-05.2005).
Lim E, Vaillant F, Wu D, Forrest NC, Pal B, Hart AH, Asselin-Labat ML, Gyorki DE, Ward T & Partanen A et al. 2009 Aberrant luminal progenitors as the candidate target population for basal tumor development in BRCA1 mutation carriers. Nature Medicine 15 907–913. (doi:10.1038/nm.2000).
Litton JK, Warneke CL, Hahn KM, Palla SL, Kuerer HM, Perkins GH, Mittendorf EA, Barnett C, Gonzalez-Angulo AM & Hortobagyi GN et al. 2013 Case control study of women treated with chemotherapy for breast cancer during pregnancy as compared with nonpregnant patients with breast cancer. Oncologist 18 369–376. (doi:10.1634/theoncologist.2012-0340).
Loi S, Haibe-Kains B, Majjaj S, Lallemand F, Durbecq V, Larsimont D, Gonzalez-Angulo AM, Pusztai L, Symmans WF & Bardelli A et al. 2010 PIK3CA mutations associated with gene signature of low mTORC1 signaling and better outcomes in estrogen receptor-positive breast cancer. PNAS 107 10208–10213. (doi:10.1073/pnas.0907011107).
Loibl S, Han SN, von Minckwitz G, Bontenbal M, Ring A, Giermek J, Fehm T, Van Calsteren K, Linn SC & Schlehe B et al. 2012 Treatment of breast cancer during pregnancy: an observational study. Lancet Oncology 13 887–896. (doi:10.1016/S1470-2045(12)70261-9).
Macias H & Hinck L 2012 Mammary gland development. Wiley Interdisciplinary Reviews. Developmental Biology 1 533–557. (doi:10.1002/wdev.35).
Majumder PK, Febbo PG, Bikoff R, Berger R, Xue Q, McMahon LM, Manola J, Brugarolas J, McDonnell TJ & Golub TR et al. 2004 mTOR inhibition reverses Akt-dependent prostate intraepithelial neoplasia through regulation of apoptotic and HIF-1-dependent pathways. Nature Medicine 10 594–601. (doi:10.1038/nm1052).
Marquis ST, Rajan JV, Wynshaw-Boris A, Xu J, Yin GY, Abel KJ, Weber BL & Chodosh LA 1995 The developmental pattern of Brca1 expression implies a role in differentiation of the breast and other tissues. Nature Genetics 11 17–26. (doi:10.1038/ng0995-17).
Marshall AM, Nommsen-Rivers LA, Hernandez LL, Dewey KG, Chantry CJ, Gregerson KA & Horseman ND 2010 Serotonin transport and metabolism in the mammary gland modulates secretory activation and involution. Journal of Clinical Endocrinology and Metabolism 95 837–846. (doi:10.1210/jc.2009-1575).
Matsuda M, Imaoka T, Vomachka AJ, Gudelsky GA, Hou Z, Mistry M, Bailey JP, Nieport KM, Walther DJ & Bader M et al. 2004 Serotonin regulates mammary gland development via an autocrine-paracrine loop. Developmental Cell 6 193–203. (doi:10.1016/S1534-5807(04)00022-X).
Muller A, Homey B, Soto H, Ge N, Catron D, Buchanan ME, McClanahan T, Murphy E, Yuan W & Wagner SN et al. 2001 Involvement of chemokine receptors in breast cancer metastasis. Nature 410 50–56. (doi:10.1038/35065016).
Narod SA 2012 Breast cancer in young women. Nature Reviews. Clinical Oncology 9 460–470. (doi:10.1038/nrclinonc.2012.102).
Nik-Zainal S, Alexandrov LB, Wedge DC, Van Loo P, Greenman CD, Raine K, Jones D, Hinton J, Marshall J & Stebbings LA et al. 2012 Mutational processes molding the genomes of 21 breast cancers. Cell 149 979–993. (doi:10.1016/j.cell.2012.04.024).
Pai VP, Marshall AM, Hernandez LL, Buckley AR & Horseman ND 2009 Altered serotonin physiology in human breast cancers favors paradoxical growth and cell survival. Breast Cancer Research 11 R81. (doi:10.1186/bcr2448).
Parker JS, Mullins M, Cheang MC, Leung S, Voduc D, Vickery T, Davies S, Fauron C, He X & Hu Z et al. 2009 Supervised risk predictor of breast cancer based on intrinsic subtypes. Journal of Clinical Oncology 27 1160–1167. (doi:10.1200/JCO.2008.18.1370).
Peccatori FA, Azim HA Jr, Orecchia R, Hoekstra HJ, Pavlidis N, Kesic V & Pentheroudakis G 2013 Cancer, pregnancy and fertility: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Annals of Oncology 24 (Suppl 6) vi160–vi170. (doi:10.1093/annonc/mdt199).
Saal LH, Gruvberger-Saal SK, Persson C, Lovgren K, Jumppanen M, Staaf J, Jonsson G, Pires MM, Maurer M & Holm K et al. 2008 Recurrent gross mutations of the PTEN tumor suppressor gene in breast cancers with deficient DSB repair. Nature Genetics 40 102–107. (doi:10.1038/ng.2007.39).
Smith J, Axelrod D, Singh B & Kleinberg D 2011 Prevention of breast cancer: the case for studying inhibition of IGF-1 actions. Annals of Oncology 22 (Suppl 1) i50–i52. (doi:10.1093/annonc/mdq666).
Sotiriou C, Wirapati P, Loi S, Harris A, Fox S, Smeds J, Nordgren H, Farmer P, Praz V & Haibe-Kains B et al. 2006 Gene expression profiling in breast cancer: understanding the molecular basis of histologic grade to improve prognosis. Journal of the National Cancer Institute 98 262–272. (doi:10.1093/jnci/djj052).
Stephens PJ, Tarpey PS, Davies H, Van Loo P, Greenman C, Wedge DC, Nik-Zainal S, Martin S, Varela I & Bignell GR et al. 2012 The landscape of cancer genes and mutational processes in breast cancer. Nature 486 400–404. (doi:10.1038/nature11017).
Teschendorff AE, Miremadi A, Pinder SE, Ellis IO & Caldas C 2007 An immune response gene expression module identifies a good prognosis subtype in estrogen receptor negative breast cancer. Genome Biology 8 R157. (doi:10.1186/gb-2007-8-8-r157).
Thomas P, Pang Y, Filardo EJ & Dong J 2005 Identity of an estrogen membrane receptor coupled to a G protein in human breast cancer cells. Endocrinology 146 624–632. (doi:10.1210/en.2004-1064).
Topalian SL, Hodi FS, Brahmer JR, Gettinger SN, Smith DC, McDermott DF, Powderly JD, Carvajal RD, Sosman JA & Atkins MB et al. 2012 Safety, activity, and immune correlates of anti-PD-1 antibody in cancer. New England Journal of Medicine 366 2443–2454. (doi:10.1056/NEJMoa1200690).
Varemo L, Nielsen J & Nookaew I 2013 Enriching the gene set analysis of genome-wide data by incorporating directionality of gene expression and combining statistical hypotheses and methods. Nucleic Acids Research 41 4378–4391. (doi:10.1093/nar/gkt111).
van 't Veer LJ, Dai H, van de Vijver MJ, He YD, Hart AA, Mao M, Peterse HL, van der Kooy K, Marton MJ & Witteveen AT et al. 2002 Gene expression profiling predicts clinical outcome of breast cancer. Nature 415 530–536. (doi:10.1038/415530a).
