SEARCH

SEARCH BY CITATION

Keywords:

  • fetal chimerism;
  • allogeneic stem cells;
  • Y-chromosome;
  • breast cancer;
  • HER2;
  • pregnancy protection

Abstract

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Clinical observations suggest that pregnancy provides protection against cancer. The mechanisms involved, however, remain unclear. Fetal cells are known to enter the mother's circulation during pregnancy and establish microchimerism. We investigated if pregnancy-related embryonic/fetal stem cell integration plays a role in breast cancer. A high-sensitivity Y-chromosome assay was developed to trace male allogeneic cells (from male fetus) in females. Fixed-embedded samples (n = 206) from both normal and breast cancer patients were screened for microchimerism. The results were combined with matching clinicopathological and histological parameters and processed statistically. The results show that in our samples (182 informative) more than half of healthy women (56%) carried male cells in their breast tissue for decades (n = 68), while only one out of five in the cancer sample pool (21%) (n = 114) (odds ratio = 4.75, CI at 95% 2.34–9.69; p = 0.0001). The data support the notion that a biological link may exist between chimerism and tissue-integrity. The correlation, however, is non-linear, since male microchimerism in excess (“hyperchimerism”) is also involved in cancer. The data suggest a link between hyperchimerism and HER2-type cancers, while decreased chimerism (“hypochimerism”) associates with ER/PR-positive (luminal-type) breast cancers. Chimerism levels that correlate with protection appear to be non-random and share densities with the mammary progenitor components of the stem cell lineage in the breast. The results suggest that protection may involve stem/progenitor level interactions and implicate novel quantitative mechanisms in chimerism biology.

Abbreviations
CI

confidence interval

ER

estrogen receptor-alpha

HER2

human epidermal growth factor receptor 2

MIQE

minimum information for publication of quantitative real-time PCR experiments

OR

odds ratio

PR

progesterone receptor

SRY

sex-determining region Y

TSPY

testis specific protein, Y-encoded

Embryonic stem cells/regulators are frequently reported in adult cancers,[1-3] but the underlying mechanisms remain elusive. We found that a cohesin protein (APRIN/Pds5B/AS3),[4, 5] involved in embryonic stem cell exit and embryonic development[6, 7] was also silenced in cancers.[5, 8, 9] The finding established a link between embryonic cells and cancer and raised the possibility of cryptic embryonic/fetal cell involvement in tissue integrity. Remnants of autologous embryonic cells, however, typically generate teratomas/teratocarcinomas at an early age.[10] We investigated the alternative, therefore, that embryonic/fetal cells from heterologous, allogeneic sources may have a role in cancer.

A possible allogeneic source is the mutual exchange of fetal/maternal cells in pregnancy.[11, 12] Transfer of fetal cells into the mother is uniformly observed during pregnancy and involves multiple cell types including cells with multilineage potential.[13] Although most fetal cells disappear after birth, some populations survive in the mother. The established new symbiotic biology, called microchimerism, is present in women for decades after pregnancy.[13-15] Long-term survival of the fetal lineage, in turn, implies self-renewal and clonogenicity, the hallmarks of stemness. The observation strongly suggests that transplacental transfer of fetal stem cells occurs regularly and is well tolerated by the mother. Fetal cell integration has been reported in the brain, lungs, liver, heart, maternal neoangiogenesis, etc.[16] Moreover, these cells remain functional and involved in pathological processes in autoimmune and degenerative diseases.[16] Allogeneic stem/progenitor cells, therefore, appear to integrate into adult niches in the mother, but their biology and the niches are not well-known. Our data suggest that a potential niche for long-term integration is the mammary gland.

To investigate the phenomenon, we targeted male allogeneic cells that offer a unique marker, the Y-chromosome. We analyzed 206 clinical samples (182 informative) and found that the well-studied Y-linked DYS14 marker[17] was not sensitive in formalin-fixed samples. By using TSPY genes[18] we introduce a signal-integration approach to increase Y-detection. The results confirmed high incidence of male cells in normal breast tissues, but significantly less in most cancers, implying a correlation between allogeneic chimerism and tissue-integrity. The mechanism is unknown, but the role of pregnancy in breast cancer suppression is well documented.[19] Unexpectedly, we found that excessive male-chimerism also correlates with cancer. Altogether, the data suggest that transplacental transfer may contribute to maintain tissue integrity, but the mechanisms are complex. Chimerism correlates with both cancer protection and promotion, where the outcome appears to depend on novel quantitative aspects of chimerism biology.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Samples

Both normal and cancer tissues (breast- and prostate cancer) were purchased as formaldehyde-fixed paraffin-embedded (FFPE) sections (Cybrdi, Rockville, MD). The manufacturer also supplied matching pathological parameters, including cancer type, AJCC-staging; (available at: http://www.cancerstaging.org), TNM-staging (NCI, 2010; available at: http://www.cancer.gov/), estrogen receptor-alpha (ER), progesterone receptor (PR) and HER2 status of the corresponding samples. We selected normal and cancer samples to match by age. The average age for the normal samples was 48.6 years (standard deviation, SD = 10.08), comparable with that of the cancer group at 48.42 years (SD = 11.25) (Table 1). The age distribution patterns were also almost identical (see Fig. 2a; for raw data see Supporting Information Tables 1–3). In addition, the two cohorts were also comparable based on their hormonal status. The majority of subjects in both groups were premenopausal by standard criteria:[20] 69.1% in the normal and 63.2% in the cancer group (Table 1).

image

Figure 1. Y-chromosome detection by DYS14, SRY and symptotic assays. M, markers with base-pairs indicated; H2O, non-template control. (a) The OCT4-S and TSPY-S amplicons generated by end-point PCR from MCF7 and LNCaP-FGC cell lines (MCF7, breast cancer line from ATCC; maintained in high glucose DMEM/5 mM glutamine/5% FBS; Y-specific markers DYS14, SRY and TSPY assays verified Y-chromosome negativity; FISH tests for Y sequences were also negative; LNCaP-FGC, a prostate cancer line from ATCC with stable hormone response and fixed aneuploidy levels, maintained as above, but with 10% FBS; the cell line was verified for the Y-chromosome by DYS14, SRY and TSPY markers, all tests were positive). The assays and cell lines are indicated on top. (b) Pilot TSPY-S assays from prostate cancer FFPE tissues. Numbers on top indicate samples; T, TSPY-S test; S, SRY-test. (c) Pilot TSPY-S assays from normal breast FFPE tissues. Numbers on top indicate samples. C, control LNCaP assay.

Download figure to PowerPoint

image

Figure 2. Analyses of patient cohorts, PCR melting curves and male microchimerism parameters by quantitative symptotic assays. The scales at left display percentage where indicated (%), otherwise the scales are explained in the panel legends. Case numbers in the experimental groups are shown following the “n =” signs below the groups. (a) Comparison of the age distribution patterns. The distribution curves are arranged by increasing age representing normal (blue dots) and cancer patient samples (red dots). The two curves are shown in overlap. The scale at left indicates age at sample acquisition. The abscissa represents case numbers (not shown), the data are available in Supporting Information Tables 1–3. (b) Comparison of PCR melting curves. Denaturation parameters of amplicons from the standard DNA gradient PCR are shown (for the amplification diagram see Supporting Information Fig. 1). The scale at left indicates the manufacturer's melting derivative units (Bio-Rad), the temperature scale is shown at the abscissa. (c) Y-positive incidence in normal and breast cancer samples, as indicated at the bottom. Red, Y-positive; blue, Y-negative. (d) Dot-plot representation of Y-signal levels in normal (blue) and cancer samples (red), as indicated at the bottom. The scale at left shows normalized signal levels in genome equivalents. The abscissa represents case numbers (not shown). (e) Y-signal levels in the normal and cancer samples. The scale at left indicates normalized signal levels. For standard error of mean (SEM) values see Supporting Information Statistical File 1. (f) Lymph-node metastasis stage distribution in Y-positive versus negative cancers. Red, high-stage (N3 by the staging system; stage-3 in Supporting Information Statistical File 3); blue, low-stage (N1 and N2 by staging; stage-1 and 2 in Supporting Information Statistical File 3). (g) HER2 incidence distribution in Y-negative versus Y-positive cancers, as indicated at the bottom. Blue, HER2-negative; green, low-HER2 expression; red, HER2-positive (overexpression), as indicated on the top; the scale shows incidence in sample percentage. (h) Inversion of ER/PR-positive versus HER2-subtype incidences in the Y-positive cancer-pool. Blue, ER/PR-positive; red, HER2-subtype. Patient cohorts are indicated at the bottom (triple-negatives are not counted). (i) Positive correlation between Y-signal levels (male cell density) and HER2 expression. The dot-plot shows Y-signal levels in HER2-negative (blue) and HER2-positive samples (red), as indicated at the bottom. Normalized signal levels are shown in genome equivalents (scale at left). The abscissa represents case numbers (not shown). (j) Positive correlation between Y-signal levels and HER2 expression. Comparison of mean Y-signal levels between HER2-negative and HER2-positive samples as indicated at the bottom. The scale at left indicates normalized signal levels. For SEM values see Supporting Information Statistical File 6.

Download figure to PowerPoint

Table 1. Comparison of age and hormonal status between the normal and cancer groups
 NormalCancer
 TotalPremenopausalTotalPremenopausal
 NN%NN%
  1. Columns are indicated for the case numbers (N) and percentage values (%). In both the normal and cancer groups the Y-positive and Y-negative subsets are also shown. The average ages of the total and premenopausal subsets are calculated and shown at the bottom together with their standard deviations.

Y positive382771.7241666.7
Y negative302066.7905662.2
TOTAL684769.11147263.2
Age (mean)48.6 yr (SD = 10.08)43.9 yr (SD = 7.67)48.42 yr (SD = 11.25)41.82 yr (SD = 7.02)

Genomic DNA extraction

The samples were dissected manually and suspended in 100 µl Digestion Buffer (50 mM Tris HCl;100 mM EDTA;1% SDS;200 mg/L Proteinase-K). Incubation (rotating oven; 50°C; 48 hr) was followed by standard phenol/chloroform extraction and ethanol precipitation with 20 microgram glycogen enhancer. After centrifugation (18,000 rpm; 20 min; RT) the pellets were resuspended in 30 µl 1 mM Tris HCl pH 8, 0.1 mM EDTA, 1% Tween20. The manufacturer followed standard protocols to prevent contamination (initial sections were discarded, blades were regularly changed between blocks, etc.) and we also changed needles and tips between sample dissections.

Quantitative Y-chromosome detection

To design the TSPY-symptotic assay (TSPY-S assay), we used Clustalw (available at: www.ebi.ac.uk/clustalw/) and Oligo-Analyzer (available at: www.idtdna.com) software to select primers that can seed on multiple paralogous Y-restricted sequences (YSIS-F primer, 5'CCTCAGAATCATACACCCTCTGTG; YSIS-R, 5'CATGTATGATCTCCTTTCGCCTCC) and amplify the following target genes into a single amplicon (positions by the GRCH37 reference primary-assembly): TSPY1 (9.305 Mbp); TSPY2 (3.464 Mbp); TSPY3 (9.238 Mbp) TSPY4 (9.174 Mbp); TSPY6Pseudo(9.346 Mbp); TSPY7Pseudo (9.216 Mbp); TSPY8 (9.198 Mbp); TSPY9 (9.329 Mbp); and TSPY10 (9.370 Mbp). The OCT4-symptotic (OCT4-S assay) primer-pair (OCT4-SF primer, 5'GTGTTCAGCCAAACGACCATC; and OCT4-SR, 5'CACCCACTTCTGCAGCAAG) amplifies the following eight genes into a single amplicon (chromosome-contigs by the GRCH37-assembly): POU5F1 (OCT4, chr.6, NT_167248; diploid); POU5F1-Pseudo3 (chr.12, NT_009714; diploid); POU5F1-Pseudo4 (chr.1, NT_004487; diploid); POU5F1-Pseudo5 (chr.10, NT_030059; diploid). PCR protocol: 94°C-30 sec/5 cycles: 94°C-30 sec, 66°C-30 sec, 72°C-30 sec/45 cycles: 94°C-30 sec, 62°C-30 sec, 72°C-30 sec; extension: 72°C-3 min. The PCR products were analyzed in 3% gels and verified by sequencing (not shown).

Quantitative PCR experiments by MIQE guidelines

The qPCR-assays were performed following the recently modified “Minimal Information for Quantitative PCR Experiments” (MIQE-precis) guidelines.[21] TSPY-S and OCT4-S were performed in triplicates. After comparative PCR-efficiency analyses, we selected the iTaq SYBRGreen Supermix kit (BioRad, Hercules, CA) from three other vendor products. The Y-signal and OCT4 normalization triplicates were further normalized by MCF7 triplicate-standards across all experiments as inter-plate calibrators, together with non-template controls throughout. DNA quantifications were performed with all samples and only results with correct melting curves were accepted. The study was performed in UV-treated PCR-hoods and kept PCR plates sealed to prevent carry-over. We used facilities restricted to female associates to prevent Y-contamination. In addition, both DNA extractions and PCR reactions were performed by personnel blinded to patient identity and used only serial numbers.

Statistics

The quantitative data (sample triplicates, clinical staging and grading data) were analyzed by two Internet software programs, the StatPages and the GraphPad Prism package (available at: statpages.org/ctab2x2.htm and www/graphpad.com) (see Supporting Information Statistical Procedures). Calculations, results, standard deviations (SD), odds ratios (OR), confidence intervals (CI), p values and graphic displays are organized in Supporting Information Statistical Files.

Results

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Signal-enhancement by signal-integration: Symptotic PCR

Initially we targeted a widely used Y-marker, the DYS14 sequence in formalin-fixed paraffin-embedded (FFPE) tissue-core samples. These cores carry needle-biopsy equivalent DNA, but as the result of formaldehyde-fixation, the integrity of the DNA is typically low (average 80–120 bp).[22] The DYS14-amplicons, however, require at least 240 bp[17] or 200 bp intact sequences,[23] ruling out reliable DYS14 detection, particularly from FFPE samples where additional PCR-inhibitors were reported.[24]

Our approach took advantage of unique internal repeats on the Y-chromosome.[25] We noticed that the repeats can be used to augment the Y-signal and selected the TSPY-gene family with Y-restricted reiterations (Clustalw homology-based hierarchical clustering). An amplicon was designed with primer-seeding across nine genes in the TSPY-pseudogene family. The assay amplifies, therefore, multiple targets simultaneously, where all products “fall together” (symptotic) into a single amplicon, unlike multiplex-PCR where multiple targets generate multiple amplicons.[26] In addition, the extremely short amplicon (76 bp) was designed to match the degradation pattern in FFPE samples. We also selected a target gene-set for normalization, where representation of symptotic sequences were comparable with that of TSPY. In the OCT4 pseudogene family (POU5F1) seeding across four diploid chromosomes amplifies eight copies/cell. Signal integration by the new symptotic assays, therefore, had the potential to perform quantitative Y-chromosome detection in low-integrity templates.

Symptotic assays consolidate Y-marker detection from low-integrity DNA

We show that OCT4 symptotic assays (OCT4-S) amplified the normalization sequence from both female (MCF7) and male (LNCaP) cells, while symptotic TSPY-tests (TSPY-S) detected the Y-chromosome only in LNCaP (Fig. 1a). To evaluate formaldehyde-fixed templates, we tested FFPE prostate cancer samples. A representative result in Fig. 1b demonstrates a single specific TSPY band in each (lanes “T”). Amplifications in our SRY-project, however, indicated non-specific, multiple bands (lanes “S”) and the SRY-project was terminated. The regular DYS14-PCR failed to amplify the specific sequence.

Quantitative symptotic assays show extensive male-incidence in normal female breasts

During assay-development we used non-quantitative end-point amplifications on normal FFPE breast samples. The end-point PCRs readily detected Y-positive breast tissue samples and showed a single symptotic amplicon (Fig. 1c). To analyze the quantitative parameters of the qPCR approach, we performed sensitivity and linearity tests by using male genomic DNA gradients in female genomic background. The TSPY-S assays showed quantitative linear correlation from 0.1 pg to 0.41 ng of male genomic DNA (from normal prostate) in the presence of 10 ng female genomic DNA (see the amplification diagrams in Supporting Information Fig. 1 and the linearity analysis in Supporting Information Fig. 2). The melting curves in Figure 2b confirm the end-point PCR results: the single peak at 78.4°C melting temperature (Tm) is consistent with a single amplicon. The optimized quantitative protocols on the 68 informative samples detected Y-signals in 38 samples (55.88%) of our normal pool (Fig. 2c, normal) (for statistical analyses see below and Supporting Information Statistical File 1).

Decreased male microchimerism in ductal invasive breast cancer

We analyzed 142 cancer samples in a partially case-controlled study with focus on pathology (eliminated ductal carcinoma in situ and Paget's disease, six samples) and by using MIQE criteria (eliminated eight samples for normalization problems; eight for contamination; four were outliers; and two for other technical problems). We found that only 24 of our 114 informative cancer samples were positive for Y-chromosome presence (21%) (Fig. 2c, “cancer”). Odds ratio (OR) calculations indicate that male stem cell integration in the female breast increases the odds to remain cancer-free by 375% (OR = 4.75, CI at 95% 2.34–9.69; n = 114; p = 0.0001; see Supporting Information Statistical File 1).

A cohort of cancers carries significantly increased male cell load

Apparently, integration of male cells in the Y-positive cancer group did not correlate with protection from cancer. The result was puzzling and further studies showed that these cancers were not only Y-positive, but harbored male cells at significantly higher levels than in normal breast tissue (see dot-plot in Fig. 2d). The mean of the normalized Y-signal was an order of magnitude (>16-fold) higher than in the control (Fig. 2e) (n = 62; p = 0.0001; see Supporting Information Statistical File 2).

Y-positive cancers show correlations with the highest stage in nodal metastases

To investigate if male cell integration affected cancer pathology, we analyzed a set of clinicopathological markers (AJCC-stagings and TNM-staging). Primary-tumor status (T), regional lymph-node status (N) and anatomical staging showed differences, but did not reach statistical significance. However, within the node-positive subcohort, Y-positive cancers showed a trend for positive correlation with the highest stage (Fig. 2f). Although in these pilot studies the nodal metastasis case numbers were moderate, statistical analyses in Supporting Information Statistical File 3 indicated high odds ratio (OR = 6.4, CI 95% = 0.8–60.75; n = 43; Pearson's uncorrected p = 0.02).

Cancers with excessive male contingencies show correlation with HER2 positivity

To characterize Y-positive cancers at the molecular level we used molecular markers with critical roles in cancer biology and pathogenesis (estrogen receptor, ER; progesterone receptor, PR; and HER2/neu expression parameters). We found that in most Y-positive cancers ER and PR expressions tended to be low, but the statistical differences were not significant. In contrast, the incidence of HER2-positivity increased significantly (Fig. 2g) (in clinical terminology HER2-positives are level 3 or greater by histological evaluation, while 0 to 2+ are HER2-negative). In our 90 cases of Y-negative cancers (Supporting Information Table 2) only 21 showed HER2-positivity (23%), while it was positive in half of the Y-positive malignancies (n = 114; OR = 3.3, CI 95% = 1.2–9.3) (Supporting Information Table 3 and Supporting Information Statistical File 4). Moreover, by separating the absolute negatives from the low-expression negatives, as shown in Fig. 2g (HER2-neg. vs. HER2-low), we found that the absolute HER2-negative incidence changed little (blue values). The HER2-gain, therefore, represents a shift from low-level HER2 expression (HER2-low) to HER2 overexpression (HER2-pos) (see green-to-red transition). In fact, the odds of shifting from low-HER2 to HER2-overexpression are almost five times higher if the cancer integrates high level male chimerism (n = 69; OR = 4.57, CI at 95% is 1.14–19.67) (for two-way contingency table calculations, see Supporting Information Statistical File 4). The finding suggests that if chimeric cells are present, HER2 receptor expression by the cancer is a selective advantage and triggers powerful selection for overexpression. HER2 receptors are involved in stem/progenitor differentiation through their ligands, the epidermal growth factors (EGF). The finding raises the possibility that one of the mechanisms chimeric cells interact with the host is the EGF pathway.

HER2-subtype inversion in Y-positive cancers

Pathological and cancer markers in combination established three groups in our Y-positive cancer samples: hormone receptor (ER/PR) positive; ER/PR/HER2-negative (triple-negative) and HER2-positive types. In the general breast cancer population ER/PR-positive (luminal) incidence is close to 70%, while HER2-positive incidence is only 15%[27] (Fig. 2h, “General”). We show that male cell integration correlates with a dramatic change in subtype frequency where the HER2 pattern became inverted (see Supporting Information Statistical File 5). While ER/PR-positive incidence decreased, HER2-positve incidence increased to 50% (Fig. 2h, “Y-positive”). The statistical odds for HER2-positive cancers are 3.3-fold higher with male chimerism than without (n = 18; p = 0.01), while the odds are 3.1-fold higher for Y-negative malignancies to become ER/PR-positive than with Y-positive ones (n = 67; p = 0.02) (Supporting Information Statistical File 5).

Male cell chimerism shows direct quantitative links with HER2-overexpression in cancer

The findings suggested that the association between male cell integration and HER2 overexpression was not coincidental and raised the possibility of an underlying biological interaction. The results of our quantitative analyses in Figures 2i and 2j are consistent with a direct biological mechanism. We found high Y-signal levels in most HER2-positive cancers (see dot-plot in Fig. 2i). Analysis of the mean of normalized Y-signals (Fig. 2j) showed that it was four-times higher in the HER2–overexpressing samples than in the HER2-low/negative cancers (n = 22; p = 0.02) (Supporting Information Statistical File 6). The positive quantitative correlation demonstrates that HER2-levels increased concomitantly with the male cells load. The findings strongly suggest a biological link between the male cells and HER2-positivity. Although the interface and the mechanisms have yet to be established, our data suggest that the EGF pathway is involved (see above).

Chimeric cell densities that correlate with protection are at the same level as the mammary progenitors of the host

The data suggested that while under- and over-representation of male chimerism correlate with malignancies, at a particular level in between chimerism associates with cancer-protection. Microchimeric densities, therefore, appear to be critical and suggest that chimerism-biology may have a quantitative component. Calculations of male representation in genome-equivalents established that male cell density was between 196 and 300 cells in 10,000 (standard error range), average 248 × 10−4 in the normal breast (see Supporting Information Statistical File 2). To assess the biological significance of these levels, we compared the cellular densities of various cell types in the mammary gland. We found that the mammary progenitor component of the stem cell hierarchy was at levels comparable with the chimeric pool: between 120 and 400 cells[28] or 100 and 200 in 10,000 cells in humans[29] and 160/10,000 in the mouse.[30] The finding that protection level chimerism is almost identical with mammary progenitor densities raises the possibility of a common or shared niche and suggests a potential mechanism in cancer protection.

Discussion

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

It is becoming increasingly clear that the human organism is not entirely homogeneous genetically. A major source of heterogeneity comes from pregnancy where the placenta facilitates feto-maternal cell exchange. Integration of allogeneic cells, in turn, may result in long-term chimerism.[11, 31] Although tissue or organ transplantation and transfusion may also contribute to genetic heterogeneity in adults, pregnancy-related feto-maternal trafficking is by far the most common source of chimerism.[32] Therefore, and because we used commercial samples with limited case-histories, the epidemiology of transfer mechanisms was not the objective of the present studies. Our goal was to investigate the biological consequences of allogeneic chimerism as a marker. The data support that microchimerism is frequent in the normal breast, confirming that the breast is a male/female chimera in more than half of women.

The high frequency of chimerism may have significant evolutionary implications. Transplacental passage of allogeneic cells is a relatively new phenomenon in evolution. The placenta appeared 100 million years ago in eutherian reproduction, so mammals, e.g. mice were also reported to have microchimerism during pregnancy. The placenta of primates, however, took a unique turn. The growing fetal brain in primates increased oxygen consumption, while upright bipedalism restricted uterine circulation. To meet the high oxygen demand, the primate placenta became a deeply invasive hemochorial structure about 6 to 8 million years ago.[33] The striking extent of long-term chimerism suggests that by a biological serendipity the gates also opened for a new evolutionary experiment, stem cell exchange. Allogeneic stem cell integration, in turn, may have created a unique biological opportunity to refresh adult stem cell pools. It appears, however, that microchimerism is not uniformly beneficial. In cancer pathology chimerism correlates with both protection and promotion, where quantitative parameters appear to be critical.

Cancer may represent a unique biological environment for microchimerism as the undifferentiated paracrine milieu,[34] the neoplastic stromal microenvironment[35] and collateral angioneogenesis[36] may preferentially integrate allogeneic stem cells. Breast cancer is particularly intriguing, as delayed differentiation and periodic hormonal remodeling[37] may create unique entry points. Initial efforts showed correlations between circulating fetal cells and breast cancer protection[38, 39] and data on cancer-adjacent tissues further implied a negative correlation.[40] On the other hand, positive correlations between fetal cells and cancer have also been reported.[41-43] The data were puzzling and prompted us to perform statistically controlled quantitative studies.

We developed a highly sensitive symptotic assay that opened the possibility to analyze paraffin-embedded tissues that were previously inaccessible for chimerism studies. The new tests detected high incidence of male cells in normal mammary glands, decades after pregnancy. Our incidence numbers are in good correlation with similar data in normal tissues[40] and indicate the reliability of our assay. Fetal cells are known to circulate in the mother during pregnancy, but rapidly decline after parturition.[44, 45] Surviving chimeric mesenchymal stem cells were reported in bone marrow,[46] but the long-term chimeric integration mechanisms are not clear. The results suggest that a niche for allogeneic integration is the mammary gland.

The significance of this cryptic pool is further underlined by the lack of male cells in about 80% of cancers suggesting that the absence of allogeneic stem cells correlates with malignancy. The finding that the proliferating malignant tissue itself is depleted of chimerism is a new observation and is consistent with data on normal tissues adjacent to cancer.[40] The lack of fetal markers or “hypochimerism,” therefore, may be a valid term that signals cancer predisposition. Pregnancy-associated fetal microchimerism, in turn, may indicate protection.

In fact, pregnancy has been shown to confer protection from breast cancer, but the link is strong only with pregnancies at an early age.[19] Hormonal stem cell progression has been implicated,[47] but not completely understood. As the developing ductal system is not fully differentiated up to the early twenties,[19] we argue that the immature ducts may integrate allogeneic stem cells at this stage with beneficial effects on differentiation. In late pregnancies, however, this “window of opportunity” for integration closes, even as the mother still carries incompletely differentiated ducts. The observation that ∼90% of breast cancers initiate from ductal cells may result from this fact.[48] The fetal stem cell integration hypothesis is consistent, therefore, with the time frame, the clinical data and offers a testable model for the puzzling early restriction for cancer protection by pregnancy.

The most puzzling feature of microchimeric cells is, however, their robust biological effects in sharp contrast to their low cell numbers in tissues. It is difficult to explain how a few microchimeric cells surrounded by thousands of host cells can damage or benefit an entire tissue. We reasoned that if the interaction takes place at the stem cell level, very low cell numbers can orchestrate profound biological consequences. In cancer, for instance, a few host stem cells may carry genetic/epigenetic damage with cancer initiating potential. If fetal stem cells are able to compete or repress them in a niche, they may potentially eliminate or control cancer initiating cells. The niche competition theory implies, therefore, that microchimeric densities are critical. Moreover, if protection is mediated through the host's stem cells, protection-levels by chimeric cells may be restricted to densities close to the host's own stem cell levels. In fact, we found that allogeneic cells linked to cancer protection were at the same level as the host's own progenitors. Although speculative, the niche-competition idea offers a testable hypothesis and may explain the correlation between chimerism and tissue integrity.

Surprisingly, we found that the correlation with protection was non-linear and high chimeric presence (“hyperchimerism”) correlates with HER2-type malignancies.[49] Various levels of chimerism, therefore, associate with various biological outcomes, further supporting the quantitative nature of chimerism biology. The findings do not rule out, however, that qualitative differences in the composition of allogeneic pools may also contribute to the biological outcome.

Altogether, the extensive male representation in the breast and its quantitative correlations with both tissue integrity and cancer strongly suggest that chimerism is not an incidental epiphenomenon. Transplacental transfer of allogeneic stem cells may represent an emerging evolutionary strategy and may offer unique biological advantages. We showed, however, that in cancer it has both positive and negative consequences.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

The authors thank Diana Bianchi and Natasha Kreder for critical reading of the manuscript. Expert technical assistance by Emily Brown is also gratefully acknowledged. The project was supported by DOD Research Grants W81XWH08-1-0575 and W81XWH09-1-00411 (P.G.).

References

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
ijc28077-sup-0001-suppinfo.pdf1259KSupporting Information

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.