Service d'Anatomie Pathologique, Unité propre de recherche de l'enseignement supérieur (UPRES) N°3541, Université Paris XI, Hôpital Paul Brousse, Villejuif, France
Service d'Anatomie Pathologique, Unité propre de recherche de l'enseignement supérieur (UPRES) No. 3541, Université Paris XI, Hôpital Paul Brousse, 14 avenue Paul Vaillant Couturier, 94804 Villejuif, Cedex, France
Potential conflict of interest: Nothing to report.
Male microchimerism is frequent in the adult female liver and is attributed to fetal cells originating from previous male offspring. It has never been studied in pregnant women, female children, or fetuses. We examined its frequency and cellular nature in normal and diseased female livers from fetal life to adulthood. Forty-six liver samples from 29 women, 6 female children, and 11 female fetuses were screened for the Y chromosome via polymerase chain reaction (PCR) assay and fluorescent in situ hybridization (FISH). The X chromosome was used as an internal control. A third PCR assay was used for Y genotyping. The Y chromosome was detected in 5 of 6 children, 7 of 11 fetuses, 3 of 9 women with normal liver, 7 of 10 women with chronic hepatitis C, 5 of 6 women with acute liver disease during pregnancy with male offspring, and 2 of 4 nonpregnant women with fulminant hepatitis. In positive samples, the mean XY/XX ratio was 0.012 (±0.004). In women, male microchimerism was correlated with previous male offspring. Male hepatocytes, detected via FISH combined with anti-hepatocyte immunohistochemistry, were observed only in fetuses (4/9) and in postpartem women (4/6). Y genotypes were different from each other in 4 of 5 female livers. In conclusion, male liver microchimerism is frequent in normal and diseased female livers. The presence of male cells in the liver of female children and fetuses is probably due to the transplacental transmission of fetal cells preexisting in the mother and acquired either from previous pregnancy with male offspring or during the mother's own fetal life. (HEPATOLOGY 2005;42:35–43.)
Male cells have been detected in the liver in 30% to 70% of adult women by means of polymerase chain reaction (PCR)-based screening for the Y chromosome.1–3 The presence of male cells in adult female tissues is usually attributed to fetomaternal microchimerism originating from previous male offspring. Indeed, fetal cells are transfused into the maternal circulation during pregnancy4, 5 and can persist in maternal blood and tissues for decades after delivery.6–9 However, male cells have not yet been studied in the liver of female children or fetuses in whom such fetomaternal transmission is impossible.
The biological significance of these cells is unknown, although they have been incriminated in the pathogenesis of autoimmune diseases such as systemic scleroderma,10–14 Grave's disease,15 Hashimoto's thyroiditis,16 Goujerot-Sjögren syndrome,17 and primary biliary cirrhosis.18 In addition, these fetal cells may give rise to stem cells, which may participate in the repair of damaged tissues, in the same way as bone marrow–derived cells of donor origin in marrow transplant recipients.19–22 A recent study based on fluorescent in situ hybridization (FISH) detection of the Y chromosome suggested that male fetal cells could differentiate into mature thyroid follicles in the damaged thyroid.23 Johnson et al., using the same method, recently suggested the existence of male hepatocytes in the liver of a woman with chronic viral hepatitis C.24 The possibility of male hepatocytes in the adult female liver was clearly shown by the same group by combining hepatocyte-specific staining and FISH.25
In the present study, we screened normal and diseased female livers from fetal life to adulthood for male microchimerism. Our aim was to assess the role of liver lesions and the influence of current or recent pregnancy on the frequency and the type of male chimeric cells. To obtain good sensitivity and specificity, we screened for the Y chromosome with a combination of different PCR assays and FISH assays and we controlled Y chromosome polymorphism between different positive samples.
A total of 46 female liver specimens were studied (Table 1). Paraffin-embedded specimens were available in every case; frozen specimens were available in 36 of 46 cases. This project was approved by the local human Investigation Committee.
Table 1. Samples and Clinical Data
Age (Mean ± SD)
Fetuses (n = 11)
Children (n = 6)
8 mo (±3)
Adults (n = 29)
Acute Liver disease during pregnancy (n = 6)
31 y (±6)
Fulminant hepatitis due to HBV
Fulminant hepatitis of unknown origin
Chronic hepatitis C
Hyperhemesis gravidarum with lobular hepatitis
Chronic hepatitis C (n = 10)
52 y (±10)
Cirrhosis and hepatocellular carcinoma
Chronic hepatitis without cirrhosis
Fulminant hepatitis without pregnancy (n = 4)
45 y (±6)
Fulminant hepatitis due to HBV
Autoimmune fulminant hepatitis
Hepatectomy (n = 9)
43 y (±15)
Metastasis and normal histology
Hemangioma and normal histology
Liver graft with normal histology
Female Children and Fetuses.
Seventeen nonadult female subjects comprising 6 children (28 days to 23 months) and 11 fetuses—including 8 stillborn infants and 3 fetuses from therapeutic abortion (22 to 41 weeks)—were selected from the files of three local hospitals. The 6 female children were comprised of 2 girls who had undergone transplantation for biliary atresia, and 4 babies who died of sudden infant death syndrome between 0 and 1 year of life. The deceased babies and fetuses had undergone postmortem examination with their parents' informed consent. All data reported in the clinical charts of stillborn infants, fetuses, and babies - are detailed in Results.
Twenty-nine liver tissue specimens from adult women (mean age 44 yr; range 25-79 yr) were analyzed.
Six women who developed acute liver disease during a pregnancy were selected to investigate a condition associating fetal microchimerism and liver regeneration. Five among them had a pregnancy with a male child. One of them also had chronic hepatitis C. The liver specimens (one whole liver and five needle biopsy specimens) were obtained either just after termination of the current pregnancy or within weeks after delivery.
The 23 nonpregnant adult women comprised 9 women with a normal liver, 4 with fulminant hepatitis, and 10 with chronic hepatitis C. The women with normal liver histology were 5 liver graft donors who underwent routine pretransplantation surgical biopsies of the graft and 4 women undergoing liver surgery. Two of the 10 women with chronic hepatitis C had cirrhosis, 1 underwent liver surgery for hepatocellular carcinoma, and the other 7 had chronic hepatitis with slight to moderate activity.
Liver samples from 4 male subjects were used as positive controls for Y chromosome detection via PCR and FISH.
DNA was extracted from paraffin sections (8 sections 4 μm thick in each case) in 10 cases and from frozen samples in 36 cases using the QIAamp tissue kit (Qiagen GmBH, Hilden, Germany) according to the manufacturer's instructions. DNA concentration was estimated via spectrophotometry. A maximum of 1 μg of DNA was used as a template in each PCR experiment to minimize the PCR inhibitors.
PCR Detection and Quantification of the Y and X Chromosome
Two different PCR methods were used to detect and quantify Y chromosomes. One part of the Y chromosome-specific gene involved in azoospermia was detected via nested PCR assay that was previously described.2 The second method was a real-time detection PCR (RTD-PCR) method (Lightcycler; Roche Diagnostic, Meylan, France) of the Y chromosome, described elsewhere.26 Briefly, primer FP-Y (5′-AACTCACCTCCAACACATACTCCAC-3′), primer RP-Y (5′-TTCATGATGAAATCTGCTTTTTGTTT-3′), and a TaqMan probe P-Y (5′-CAGCCACCAGAATTATCTCCAAGCTCTCTGA-3′) synthesized according to the published sequence of the DFFRY gene27 were used together with the Light cycler FastStart DNA master hybridization probes kit (Roche Diagnostic) during 50 PCR cycles (95°C for 15 seconds and 60°C for 60 seconds).
To control the presence of PCR inhibitors and to normalize the quantification of the Y chromosome, quantification of the X chromosome was performed with RTD-PCR. The nucleotide sequences of the X chromosome–specific primers were designed from the nucleotide sequence of Xq25-26.3 (Genbank accession number AL161442; forward primer 5′-GGCTGAAATGACATGGGCT CTTA-3′, reverse primer 5′-GCTGCTTTCCCCGTAATTCAAGT-3′). PCR was performed with the Light cycler DNA master SYBR green I kit (Roche Diagnostic) and consisted of 50 cycles at 95°C for 15 seconds and 62°C for 45 seconds. For quantification of the X and Y chromosome, the same external standard curve was generated through the amplification of 10-fold dilutions of DNA extracted from 25 × 106 peripheral blood mononuclear cells (PBMC) of a male donor. Standard curves generated by the external male standard DNA showed that amplification of the X chromosome–specific sequence was linear between 500 and 500,000 male PBMC and that the amplification of Y was linear between 50 and 50,000 male PBMC (Fig. 1). The ratio between the X and Y quantifications from male DNA was around 1 (0.75-1.23) for all dilutions above 5 × 102 male cells. Normalized quantities of Y chromosomes in female liver samples were taken into account only if the quantity of the X chromosome in the same sample was 5 × 102 or more. In female liver with male microchimerism, the ratio between male and female cells was calculated as twice the ratio between quantification of Y- and X-specific sequences. The sensitivity threshold was determined by testing serial dilutions of male DNA in a constant dilution of female DNA previously shown to be Y chromosome–negative via RTD-PCR. More than 10 Y chromosomes were repeatdly detected among 105 female cells.
Y Chromosome Genotyping.
The two chromosome Y tetranucleotide repeats DYS38528 and DYS391 (Genbank accession number AF140637) are highly polymorphic markers used in individual identification and forensic analysis. DYS385 is duplicated on chromosome Y, and both copies show a variable number of tandem repeats between individuals. The haplotype pattern obtained via simultaneous analysis of DYS385 and DYS391 shows over 90% diversity in Europeans.29
Hemi-nested PCR amplifications were performed separately for the two microsatellites on eight female DNA samples found to be Y chromosome–positive via the other PCR methods. Primers DYS385.1B and DYS391RB were radioactively labeled by replacement of 5′P with gammaP33 using T4 polynucleotide kinase. The first round amplification was performed using primers DYS385.1 and DYS385.2 or primers DYS391 LB and DYS391.R.The second round amplification was performed using labeled primer DYS385.1B and primer DYS385.2 or labeled primer DYS391.RB and primer DYS391.LB. One millimole of each of the two PCR products (DYS385 and DYS391) for each sample were pooled together, heat–denatured, and then run in a 6% acrylamide denaturing gel. The gel was dried and exposed overnight to an autoradiography film.
FISH and Hepatocyte Immunostaining
Four-micrometer-thick sections were cut from paraffin-embedded liver samples. After deparaffinization, sections were treated with RNAse solution (100 mg/mL in 2× sodium saline citrate [SSC]) for 30 minutes at 37°C and then rinsed in 2× SSC. After microwave heating (750 W) for 4 × 5 minutes in 0.01 mol/L citrate buffer (pH 6.0), the sections were incubated in 1 mol/L sodium thiocyanate at 80°C for 45 minutes, rinsed in 2× SSC, digested with proteinase K solution (0.05 mg/mL in 2× SSC) for 10 minutes at 37°C, and rinsed again in 2× SSC. After dehydration, the sections were air-dried for 10 minutes. Hybridization was performed with X and Y chromosome–specific probes (alpha satellite centromeric probes specific for X [DXZ1] and Y [DYZ3] chromosomes [Qbiogen, Illkirch, France] in a HYBrite system [Vysis, Voisins-le-Bretonneux, France]) without formamide. Briefly, each section was incubated with a mixture of 1.5 μL of X and Y probes in 30 μL of Hybrisol. Probes were denatured on slides at the same time as the tissue DNA for 2 minutes at 72°C, and hybridization was performed for 20 hours at 37°C. The sections were then washed in 2× SSC for 5 minutes at 72°C and rinsed in phosphate-buffered detergent (Qbiogen) for 2 minutes. The X probe, labeled with biotin, was revealed with an avidin–Texas red complex; the Y probe, labeled with digoxigenin, was revealed with a fluorescein-conjugated anti-digoxigenin antibody (Fab fragments; Boehringer, Reins, France). The red signal generated by the X probe was amplified with an anti-avidin antibody and further avidin–Texas red complex treatment (Biotin Texas Red kit; Appligene-Oncor). The sections were finally counterstained with 4′,6-diamidino-2-phenylindole dihydrochloride.
The slides were analyzed with a Zeiss-Axiophot epifluorescence microscope equipped with a triple band-pass filter (Vysis) by a pathologist blinded to PCR results. Each section was analyzed by examining non-overlapping fields at a magnification of ×600 (10-100 fields according to the size of the specimen). X chromosomes were identified as intranuclear red spots; Y chromosomes were identified as intranuclear green spots. All cell types—hepatocytes, cholangiocytes, sinusoidal cells (including Kupffer cells, leukocytes, and endothelial cells), and inflammatory cells—were examined. At this step, hepatocytes were recognized via morphology, location within plates, and cytoplasmic autofluorescence. Only specimens containing one or two red spots in at least 50% of hepatocytes were retained. All fields identified under the microscope as bearing positive Y signals were captured with a cooled digital camera (Sensys 1400) and the QUIPS-smart capture system (both from Vysis). A second pathologist, also blinded to PCR results, validated the results on the captured pictures. The total number of Y chromosome–positive hepatocytes and non-hepatocyte cells was recorded in each case according to the number of fields.
Whenever Y chromosome–positive hepatocytes were detected, additional sections were tested via combined anti-hepatocyte immunohistochemistry (Hep-par1; DakoCytomation, Traypo, France) and FISH for X and Y chromosomes. Immunostaining was performed just after microwave treatment and before proteinase K digestion via the labeled stretavidin-biotin method. Diaminobenzidine specifically deposited on hepatocytes during anti-hepatocyte immunostaining generates green fluorescence with the FITC filter, allowing hepatocyte plates to be distinguished from sinusoids.
Because green fluorescence identifies both Y chromosome and hepatocyte cytoplasm, 5 of the cases with Y chromosome–positive hepatocytes (2 adult livers and 3 fetal livers) were tested in second intention using combined immunochemistry and FISH with modified colors to avoid any possible false positivity for the Y chromosome; the second Y probe (Yq12) was labeled with biotin and revealed with avidin–Texas red, and the second X probe (Xp11) was labeled with digoxigenin and revealed with a fluorescein-conjugated anti-digoxigenin antibody (probes kindly provided by S. Romana, Département de Cytogénétique Hôpital Necker-Enfants Malades, France).
FISH was positive for both Y and X chromosomes on paraffin sections from all male controls, and for Y chromosomes in 35% to 45% of hepatocytes because of the truncated nuclei (Fig. 1).
Male Microchimerism in Female Livers.
Real-time PCR revealed more than 5 × 102 X chromosomes in 35 of the 46 female samples (76%), which thus were qualified for Y chromosome detection via PCR. In these 35 samples, the mean quantity of XX cells per assay was 255,000 (±245,000). Y chromosomes were detected via RTD-PCR in 24 of these 35 (68%) samples. Nested PCR results confirmed the presence of Y chromosomes in 23 of the 24 RTD-PCR–positive samples tested and confirmed the absence of Y chromosomes in 11 of 11 RTD-PCR–negative samples. FISH results for Y chromosome identification were available for the 38 female samples (82%), which showed satisfactory labeling of X chromosomes. Among these 38 samples, FISH for Y was positive in 16 of 38 (42%). Overall, 43 of 46 (93%) female liver samples could be screened for Y chromosomes by means of PCR and/or FISH (30 cases with PCR and FISH, 5 cases with PCR alone, and 8 cases with FISH alone). FISH and reverse-transcriptase PCR results for Y chromosome detection are compared in Table 2. The results of the two methods correlated well in the 30 specimens for which both FISH and PCR were available (Fisher's exact test; P < .01).
Table 2. FISH and RTD-PCR Detection of Y Chromosomes in 46 Female Liver Samples
Among the 29 adult women, (Table 3), detection of the X chromosome via FISH or PCR assay was positive in 27. FISH and/or PCR detected Y chromosomes in 17 of 27 subjects (62%): 3 of 7 nonpregnant normal subjects, 7 of 10 nonpregnant women with hepatitis C, 5 of 6 pregnant women with acute liver disease, and 2 of 4 nonpregnant women with fulminant hepatitis. Among those 27 women, information on male siblings was available in 23: 15 of the 18 women with previous male offspring had male liver cells, while they were not detected in the 5 women who had not had male offspring (P < .001). The interval between the birth of the last male child and the liver biopsy or surgery ranged from 4 to 35 years.
Table 3. Detection of Y Chromosomes in the Liver of 27 Female Adults
Male DNA in Liver
Y Nested PCR
Number of Fields
Y FISH in Hepatocyte
Y FISH in Non-hepatocyte
Time After Delivery
Abbreviations: ALD, acute liver disease; N, negative; P, positive; NA, not available.
Among the 17 nonadult subjects (Table 4), detection of the X chromosome via FISH or PCR assay was positive in 16. The Y chromosomes were detected in the livers of 5 of 6 female babies and in 7 of 10 female fetuses. Information on previous pregnancy with male child of their mothers was available in 4 of 6 female babies and in 6 of 10 fetuses; no significant correlation existed between previous pregnancy with male child in the mother and the presence of male cells in the liver of their fetuses or children.
Table 4. Detection of Y Chromosomes in the Liver of 16 Nonadult Female Subjects
Y chromosome–positive hepatocytes were identified via FISH combined with immunohistochemistry in 8 of 38 cases, corresponding to 4 adult women having pregnancies with male offspring, and acute liver disease and in 4 of 9 female fetuses. Positive hepatocytes identified with combined immunohistochemistry were scarce and isolated; their number per field at an original magnification of ×600 ranged between 0.01 and 0.3. In hepatocytes containing one positive Y signal, no more than one X signal was observed. The results obtained in second intention in the 4 cases tested with red-stained Y chromosome were similar (Figs. 2–5).
Y chromosome–positive non-hepatocyte cells (i.e., sinusoidal and inflammatory cells) were identified via FISH in 14 of 38 cases. Four of these 14 subjects also had Y-positive hepatocytes. Positive non-hepatocyte cells were rare, ranging between 0.02 and 0.45 per field at an original magnification of ×600. Non-hepatocytic liver chimerism was observed in all patient subgroups: 2 of 7 normal livers, 2 of 6 livers with hepatitis C, 3 of 6 livers with acute hepatitis during pregnancy, 1 of 4 livers with fulminant hepatitis (nonpregnant women), 1 of 5 babies, and 4 of 10 fetal livers.
Y Chromosome Quantification via RTD-PCR in Female Liver.
X chromosome numbers determined via RTD-PCR exceeded 5 × 102 in 35 of 46 female samples, authorizing normalized quantification of Y chromosomes. Y chromosomes were detected via RTD-PCR in 25 (71%) of these 35 cases. Among these 25 cases, FISH was available in 19 subjects: the mean XY/XX ratio was 0.012 ± 0.004 (range 0.0002-0.036), and this ratio was higher in 12 FISH-positive than in 7 FISH-negative samples (0.016 ± 0.017 vs. 0.005 ± 0.005; P = .005). This ratio was not related to any of the other variables studied (pregnancy, age, adulthood, childhood or fetal life, chronic disease, and DNA extraction from paraffin-embedded or frozen tissues).
Y Chromosome Genotyping.
Markers DYS385 and DYS391 could be amplified in 5 of 8 female livers that were positive through PCR: 2 patients with fulminant hepatitis, and 3 fetuses. Y chromosome genotypes were different in 4 of 5 cases, ruling out contaminations by a common male DNA (Fig. 6).
Most previous studies of human fetomaternal chimerism and sex-mismatched transplant microchimerism have been based on a single screening method (FISH or PCR) for the Y chromosome in female tissues. In contrast, we used a combined approach based on FISH and two distinct PCR methods. The correlations between FISH and PCR and between the two PCR methods, together with the fact that each technique (FISH, nested PCR, and RTD-PCR) targeted different regions of the Y chromosome, strongly support the validity of the frequent male chimerism detected here. In addition, this validity is strengthened by the Y chromosome genotyping, which rules out contamination from a common male DNA.
Our study confirms that the presence of male cells in adult female livers is a frequent event, with an overall frequency of 62% in women in part preselected as having at least one male offspring and frequencies of 56% in nonpregnant women and 83% in pregnant or postpartem women. In a population of women selected according to the same criteria of having at least one male offspring, Tanaka et al., using a highly sensitive PCR method for Y chromosome screening, showed that 70% of liver samples were positive.1 In the adult population, we could also confirm the strong correlation between the presence of male cells in female liver and current or past pregnancy with male offspring.25 However, in the nonadult population, we did not observe such correlation in the few cases for which previous miscarriage of male offspring was known in the mothers of children or fetuses.
The combination of quantitative PCR and FISH allows us to ascertain that male liver microchimerism is the result of a small percentage of cells ranging from 0.03% to 5.8%. Most of these cells are non-hepatocyte cells, and it is noteworthy that hepatocyte chimerism was observed only during pregnancy or just after delivery. In the literature, male liver microchimerism was mainly explored in the setting of primary biliary cirrhosis, and in most studies, the frequency of male liver microchimerism was similar in women with primary biliary cirrhosis and in controls.2, 30 In this study, we tested acute and chronic liver diseases and failed to disclose any relationship between the presence and degree of male microchimerism and the existence and intensity of liver disease.
The most surprising finding in this study is the presence of male cells in the liver of female fetuses and children. The possibility of an embryological abnormality such as sex chromosome congenital chimerism (due to early fusion of twin embryos or tetragametic chimerism) or mosaicism as a result of separate nondysjunction events31 is unlikely, because (1) the frequency of these events is quite low when compared with the percentage of our female patients with intrahepatic male cells and (2) congenital chimerism is associated with a far higher percentage of male cells than that detected here. Male cells could also originate from cells transferred during blood transfusions with nonirradiated products, because long-term blood microchimerism after blood transfusions has been reported. This hypothesis cannot be totally ruled out for the children with biliary atresia but can be excluded for the female stillborn fetuses. Twin–twin transfusions between sex-mismatched twins may be another possible explanation for the presence of male cells within female tissues. In our patients, only one fetus was issued from a gemellar pregnancy, but the twin was also female and thus cannot be responsible for the presence of male cells.
Finally, the more plausible explanation for male chimerism in the liver of female fetuses and children is transplacental transmission of male cells present in the mother before pregnancy. Male cells present in the mother could originate from previous male pregnancies. Alternatively, they could be acquired transplacentally during the mother's own fetal life, and persist into adulthood. Thus adult women can harbor male fetal cells that were acquired during pregnancies with a male child or transplacentally during fetal life and that can survive for decades. Maternal cells have been detected in peripheral blood and inflammatory infiltrates present in muscle biopsy specimens from children with juvenile idiopathic inflammatory myopathy and in 20% of children with unrelated muscle diseases, demonstrating transplacental trafficking of maternal cells toward the fetal circulation, and the capacity of these maternal cells to survive after birth.32, 33 When pregnant, a woman would be able to transmit these early fetal cells transplacentally to her baby (male or female) in the same way that she can transmit her own cells. These fetal cells would be capable of surviving for decades in their new host. Further studies are clearly required to determine the lineage of such fetal cells in fetuses, children, and adults.
The second important finding in this study is that chimeric fetal cells appear to have stem cell–like properties, given their long-term persistence and ability to differentiate into hepatocytes. Very recently, such male microchimeric cells have been shown to express markers of epithelial, leukocyte, and hepatocyte differentiation within maternal organs, strongly suggesting the presence of fetal cells that may have multilineage capacity.25 However, in our study, we only observed hepatocyte microchimerism against a background of pregnancy (i.e., in fetuses and in pregnant women with acute liver disease). This finding suggests that the contribution of chimeric fetal stem cells to liver regeneration is low or nil in acute and chronic liver diseases outside of pregnancy, and that pregnancy could favor fetal cell transdifferentiation into hepatocytes. Recent studies in the setting of sex-mismatched bone marrow or liver transplantation have shown similar hepatocyte microchimerism in recipients; this outcome has been attributed to either bone marrow stem cell transdifferentiation19, 20, 22, 34 or fusion of hematopoietic stem cells with hepatocytes.35, 36 We obtained no evidence to support the latter possibility for Y chromosome–positive hepatocytes that never contained more than one X chromosome. In fetuses, chimeric fetal cells might conceivably play a role in liver development through cell-to-cell contacts or a paracrine action.
In conclusion, chimeric fetal cells may indeed be present in a large proportion of female subjects, be able to survive across generations, persist in non-autologous recipients, and differentiate into hepatocytes. Further studies are clearly required to determine the lineage of such fetal cells in fetuses, children, and adults.
We thank Bruno Péault and Daniel Azoulay for interesting discussions; Michel Beaugrand, Sophie Prévot, Martine Bucourt, Aurore Coulomb-Lherminé, and Caroline Rambaud for providing cases; Rachid Kara for technical help, and David Young for critically reviewing the manuscript.