Introduction
Breast cancer is among the two most commonly diagnosed cancers of pregnancy, along with cervical cancer [
1]. Up to 3.8% of all breast carcinomas are associated with pregnancy [
2]. These tumours account for 8% of breast cancers occurring in women younger than 45 years old and for as many as 18% of women younger than 30 years [
2]. One of the most used definitions of pregnancy-associated breast carcinoma (PABC) emphasizes the fact that these tumours develop during pregnancy itself or during the first year after delivery [
1,
3]. Accordingly, analysis of the literature shows that PABC develops in 60% of patients during pregnancy itself and 40% of cases occurred during the year following delivery [
4]. In addition, these tumours appear to exhibit peculiar profiles of evolution. Indeed, 14% to 28% of PABCs are of inflammatory types [
5]. These tumours were found to be significantly larger, with more extensive lymph node involvement, as compared with tumours developing in nonpregnant women [
6]. Finally, the SBR (Scarff, Bloom, Richardson; a histological index) grades of PABCs appear to be high [
6,
7].
Pregnancy induces the transfer of foetal cells that may persist for decades in the mother. Foetal transferred cells include differentiated cells as well as several progenitor cell types, such as haematopoietic [
8] and mesenchymal stem cells [
9]. These progenitors are capable of migration in maternal damaged tissues and can differentiate into various phenotypes [
10]. Several groups have suggested that microchimeric foetal cells were recruited to maternal lesional tissues such as kidney, liver, or brain tissues in order to help in the repair process, and may rescue genetically deficient maternal tissues [
11‐
13]. Similarly, we recently demonstrated the ability of foetal endothelial progenitor cells to form blood vessels in maternal inflamed skin [
14].
Microchimeric cells have also been detected in tumours. Indeed, in a series of kidney transplant recipients, we and others recently reported that epidermal skin carcinoma and, moreover, Kaposi's sarcomas may arise from microchimeric donor cells [
15]. Similarly, nontumoural cytokeratin-expressing foetal microchimeric cells have been identified in cervical cancer from women with a prior history of pregnancy [
16].
Because PABC appears to be a frequent condition with an aggressive course, and because foetal cells can participate in tissue remodelling, we conducted the present study to analyze the presence and phenotype of foetal cells in PABCs.
Discussion
This study was conducted to determine whether there is an interaction between a tumour type that is known to develop during pregnancy and foetal cells, which can migrate and differentiate in various maternal tissues. Our findings show that among women bearing a child, only those with breast carcinoma had foetal cells in their mammary tissue. Although the number of specimens analyzed was low, significant differences from benign control samples were identified; moreover, breast carcinomas nearly always had foetal cells within tumoural areas.
PABC is usually defined as occurring during pregnancy, lactation and up to 1 year after delivery [
1,
3]. In this study, in order to improve the stringency of detecting PABC and to avoid coincident cases, we restricted our sampling to tumours occurring up to 6 months after delivery. In our view, this allowed us also to avoid other unknown factors that may interfere with foetal cell trafficking over a prolonged period and to focus on a period characterized by higher levels of circulating foetal cells. Depending on country, breast carcinomas are the first or the second most common cancer occurring in pregnant women [
1]. It develops in about one in 3,000 pregnant women and therefore represents a troublesome but not infrequent disease [
19]. The profile of breast cancer during pregnancy has a number of distinct features. Women with PABC have larger tumours and with lymph node invasion occurring in 56% to 89%, as compared with 38% to 54% in breast carcinomas not associated with pregnancy [
3,
5,
6,
20‐
25]. Similarly, women with PABC were more likely to have tumours of high histological grade, to exhibit a high level of mitosis, and to be progesterone receptor negative than matched nonpregnant control women [
26,
27]. However, in two studies [
24,
28], when patients and nonpregnant control women were matched for histological grade and tumour stage, survival rates were equivalent. Therefore, the poorer scores of PABCs may result from a more advanced stage at diagnosis, because physiological changes in the breast delay diagnosis [
23,
29‐
32]. Alternatively, these higher scores may reflect an intrinsically worse profile for these breast carcinomas.
Analysis of the precise types of infiltrating foetal cells, which were present at very low levels, proved difficult. Although our team has accumulated experience in combining FISH and immunolabelling, each antibody in various tissues required careful evaluation, with multiple preliminary trials. Because this multistep, delicate technique was applied to few cells, only 40% of total foetal cells could be identified with precision, which is in accordance with our usual findings under such circumstances. Foetal cells in the vicinity of breast carcinomas were not circulating leucocytes nor endothelial cells, but seemed mainly of mesenchymal or epithelial origin, because these expressed vimentin or cytokeratin, respectively. Because the foetal cells that we identified were isolated and not organized as clusters, they were not part of the neoplastic clone, even if they expressed antigens that can be found within the adenocarcinoma.
Pregnancy induces the transfer of foetal haematopoietic, mesenchymal and endothelial stem cells to the mothers. These may persist decades after delivery in niches such as the mother's bone marrow. The persisting foetal stem cells may interact with maternal peripheral tissues. Indeed, several human and murine studies have demonstrated that foetal derived cells were present in thyroid, intestinal liver, neuronal, glial [
12] and kidney tubular cells [
11]. Interestingly, most studies have demonstrated that these foetal derived cells home to lesional tissues, where they adopt the phenotype of the affected tissues [
11,
13,
33]. Therefore, it appears that foetal cells are recruited in damaged organs, where they differentiate or fuse with host cells to adopt the phenotype of the involved tissue.
In the present study, we describe foetal derived cells expressing fibroblast, epithelial and even endothelial markers in breast carcinomas. These findings are in accordance with the above-mentioned previous studies and demonstrate that the stroma of malignant tumours developing during pregnancy nearly always recruits foetal derived cells. In mice as well as in humans, CD34 haemopoietic progenitors may develop into isolated epithelial cells without clusters in injured tissues, similar to our findings [
34,
35]. Mesenchymal stem cells can be recruited by cancer tissue, as was previously described [
36]. More recently, we reported the observation that endothelial progenitor cells of foetal origin could form blood vessels in maternal inflammatory skin during pregnancy [
14]. All of these data can therefore account for the presence of foetal fibroblasts, and endothelial and epithelial cells in breast carcinomas.
Although, Cha and coworkers [
16] in cervical carcinomas in women with a prior history of pregnancy reported the presence of CD45
+ leucocytes in 44.4% but also cytokeratin foetal derived cells in 24.3%. The present study is the first systematic evaluation of foetal cell invasion in carcinomas developing during or shortly after pregnancy. We previously conducted a study of skin epithelial tumours in skin cancers developing after kidney transplantation, a situation similar to pregnancy because it is characterized by the presence of a small number of circulating donor cells [
15]. In that context, microchimeric cells from the donors were frequently found in cutaneous malignant epithelial tumours, but they were not epithelial [
15]. In a unique and peculiar case, such donor cells were even able to give rise to a skin carcinoma, a situation that we did not identify in the present study. Interestingly, Kaposi's sarcoma in kidney transplant recipients occured more frequently in donor tissue [
37]. However, these tumours are not epithelial but derive partly from endothelial cells. Endothelial progenitors circulate and may therefore have been present in the transplanted kidney. Therefore, malignancies occurring during pregnancy appear to frequently recruit foetally derived cells and, in the case of breast, these are mainly cells stroma participating. Gadi and Nelson [
38] recently found that detection of foetal cells occurred more frequently in peripheral blood from women without breast carcinomas than in those presenting with breast cancer. These authors therefore suggest that circulating foetal microchimerism could 'protect' women from developing breast carcinoma [
38].
Pregnancy associated carcinomas are aggressive tumours. Our small series of 10 patients suggests that PABC is characterized by the presence of foetal cells in stroma. Stroma may influence tumor cell proliferation and may eventually be relevant to prognosis [
39]. This observation is potentially important enough to warrant further investigation. Indeed, foetal derived stromal cells may behave in a different way from adult maternal cells, because foetal mesenchymal cells exhibit a greater degree of plasticity, grow more rapidly and exhibit longer telomers than do their adult counterparts [
40]. In addition, the maternal immune system allows fetal cells bearing paternal nonshared antigens to avoid immune responses. Human leucocyte antigen (HLA)-G is among the molecules that are expressed by foetal cells to reduce maternal immune responses [
41]. HLA-G abnormal expression has also been suspected of playing a role in various malignancies, including breast cancer [
42]. One can therefore hypothesize that foetal cells in breast cancer tumours may influence tumour evolution. However, this question should be addressed through comparisons of foetal cell infiltration between PABC and similar tumours occurring years or decades postpartum.
Competing interests
The authors declare that they have no financial competing interests (political, personal, religious, ideological, academic, intellectual, commercial, or any other) to declare in relation to this manuscript.
Authors' contributions
SA, KK and SU contributed to study conception and design. GD, MO, RR and MCM contributed to generation of data. KK, SA, GD and RR were involved in the analysis and interpretation of data. KK, SA and GD were involved in drafting the manuscript or revising it critically for important intellectual content.