Scientific advances in pregnancy-related research are challenging our notion of separate self. It is now well recognized that there is bidirectional traffic between mother and fetus during pregnancy. Fetal microchimerism is most readily detected by sampling plasma, and increases progressively over the course of gestation (1). That fetal cells persist in the maternal circulation decades thereafter (2) raises questions regarding the long-term consequences of pregnancy. Chronic graft-versus-host disease (GVHD) is a condition of human chimerism with similarities to some autoimmune diseases, including systemic sclerosis (SSc). The HLA relationship of donor and host is of central importance in the development of chronic GVHD, and HLA–class II genes are also known to be important in autoimmune disease. Considered together, these observations led to the hypothesis that microchimerism and the HLA genes of host and nonhost cells are involved in the pathogenesis of autoimmune disease (3). Autoimmune manifestations of chronic GVHD had been described more than 20 years ago; however, clinical similarities between chronic GVHD and autoimmune disease had not been previously considered in connection with cell transfer between fetus and mother during pregnancy. Studies of SSc, primary biliary cirrhosis, Sjögren's syndrome, pruritic eruption of pregnancy, myositis, and thyroid disease have both lent support and raised doubts about this hypothesis. Although all of these diseases have a female predilection, the effect of pregnancy on disease risk has been largely unexamined.
Pregnancy and risk of SSc
Whether or not parity affects SSc risk is of interest both because SSc has a peak incidence in women in the postreproductive years and because of recent knowledge that fetal cells persist long after pregnancy completion. Fertility and adverse pregnancy outcome prior to disease onset have previously been examined (4), but, prior to the current study by Pisa et al reported in this issue of Arthritis & Rheumatism (5), no study had investigated whether parity and/or gravidity is associated with the risk of SSc. Thus, the study by Pisa et al addresses an important gap in existing knowledge. The authors describe a carefully conducted case–control study comparing women with SSc with women with orthopedic disorders identified over a 30-month time period (5). Parous women (women with a prior birth) had a significantly decreased risk of SSc compared with nulliparous women (no prior birth). SSc risk was lowest in women with an earlier age at first pregnancy, and the risk decreased with an increasing number of children. Although the number of cases identified was modest, the study was well designed and all observations remained after evaluation for multiple potential confounding variables.
Pregnancy history before SSc onset was examined in another recent report (6). In contrast to the current case–control study, only patients with SSc were evaluated. One hundred consecutive women with SSc were identified, and the analysis compared women with diffuse SSc with those with limited SSc (6). Women with limited disease had significantly more children than did those with diffuse SSc. Women with limited SSc also had a shorter interval between first birth and SSc onset. The 2 reports differ but are not contradictory in that both the study design and study subjects were dissimilar. In the report by Launay et al (6), 28% of women had diffuse disease, whereas in the study by Pisa et al (5), 91% had diffuse disease (42 of 46 women compared with 28 of 100 in Launay et al). Thus, one possibility is that the effect of parity on disease susceptibility is different for diffuse SSc and for limited SSc. A differential effect is not implausible, in that the 2 disease subsets are distinct with respect to clinical manifestations, autoantibody response, and also immunogenetics (7).
The female predilection to SSc is expressed relative to that in men, and because neither study included observations in men, the risk of SSc in parous women relative to that in men is not known. Estimating from the findings in the study by Pisa et al, in which 70% of SSc patients were nevertheless parous (5), and using a female-to-male ratio of 4 to 1, parous women would have an increased risk compared with men and nulliparous women would have a further increased risk relative to men. Results of the current study do not necessarily conflict with the hypothesis that microchimerism resulting from pregnancy contributes to SSc; they do, however, suggest that pregnancy alone cannot explain the female predilection to SSc. A recent comprehensive review has summarized numerous aspects of biology and of exposures that could contribute to sex differences in disease (8). The microchimerism hypothesis proposes that HLA-similar (but not entirely identical) cells could result in disruption of the host immunoregulatory mechanisms (not that parity per se or multiparity increases SSc risk). Thus, microchimerism could be adverse within the context of other factors, including the particular HLA genes of mother and child, the HLA relationship between them, and environmental and/or infectious triggers. It is important to remember that persistent fetal microchimerism is a common phenomenon in normal healthy women. Thus, by far, the majority of fetal cells are expected to be HLA mismatched, and because persistent fetal microchimerism is common in healthy women, the overall effect of microchimerism could well be positive.
HLA genes of mother and child in SSc
The usual role attributed to HLA molecules in autoimmune disease is as a genetic risk factor. Long-term, persistent microchimerism from pregnancy indicates that HLA molecules could potentially contribute to disease in 3 different roles: the HLA genotype of the host, the HLA relationship of host to nonhost cells, and the HLA genotype of the nonhost cells. Considering the observations from transplantation, in which the HLA relationship of donor and host is of central importance, the hypothesis that microchimerism contributes to autoimmune disease incorporated the postulate that the HLA relationship of host and nonhost cells is important (3). To test this hypothesis, a study was conducted of SSc and control families (9). Women were included only if all children were willing to participate, because the hypothesis being tested was not that parity increases SSc risk, but that persistent fetal microchimerism with HLA-similar cells increases risk. Therefore, noninclusion of a child could result in misclassification. A 9-fold increased risk of SSc was found in women who had previously given birth to a child who was compatible for HLA–DRB1 genes that encode the basic HLA families of molecules DR1–DR14. No association was observed for HLA–class I genes. Interestingly, HLA–DRB1 is also the primary locus associated with susceptibility to SSc (7). If confirmed in an independent data set, this observation indicates that the DRB1 locus may play an important role in the regulation and/or pathogenicity of fetal microchimerism.
The specific HLA genotype of the microchimeric cells could also be a disease-contributory factor. Whether or not a woman had persistent fetal microchimerism within her T lymphocyte population was found to correlate with her own HLA genotype, but even more so with the HLA genotype of the child, in a recent study (10). Another interesting question that has not yet been addressed is whether women with SSc who do not themselves have SSc-associated HLA molecules harbor SSc-associated HLA molecules or peptides through microchimerism. In addition to fetal microchimerism, HLA peptides/molecules could derive from other sources, including from a twin, from a blood transfusion, or from maternal cells that have passed into the fetal circulation and persisted (11).
Fetal microchimerism in SSc
The initial study of microchimerism, which used a quantitative direct polymerase chain reaction (PCR) assay for male DNA, examined women with SSc who had given birth to at least 1 son (9). Women with sons were selected for the technical reason that it enabled use of a single assay that detected male DNA in a female host. The direct PCR assay had been validated and standardized previously for use in prenatal diagnosis and was conducted by testing DNA extracted from at least 10 aliquots of whole blood from a 16-cc sample. A significant quantitative difference in fetal microchimerism was found in women with SSc compared with healthy women. Results, expressed as the DNA equivalent number of male cells, indicated among healthy controls, a range of male-DNA cell equivalents of 0–2 (mean 0.38). In contrast, the range was 0–61 (mean 11.1) in women with SSc. Some women with SSc had levels of male DNA that were higher than what is found in most women who are currently pregnant with a normal male fetus, although patients had given birth to their sons decades previously. The study was prospective and blinded.
Another study tested DNA extracted from skin biopsy specimens and peripheral blood of SSc patients and controls using a nested PCR technique for male DNA (12). The strength of this study was inclusion of a disease-affected tissue. Nested PCR provides a qualitative result. Male DNA was found significantly more often in skin biopsy samples from SSc patients compared with controls, and male cells were seen in some skin biopsy samples using fluorescence in situ hybridization (FISH). Pregnancy history was obtained retrospectively for women who underwent skin biopsies and not provided for the majority of women for whom peripheral blood was tested (known for only 2 of 94 women). Although male DNA was used as the measure of microchimerism, women with SSc who had sons were studied for T lymphocyte microchimerism without comparison with control women with sons. The absence of appropriate controls was notable in view of the earlier description by Bianchi et al of fetal microchimerism among T lymphocytes in healthy women (2). Subsequent studies confirmed that simple detection of fetal microchimerism is common among T lymphocytes (and also in monocytes, natural killer cells, and B cells) in both women with SSc and healthy women (13).
Other studies of fetal microchimerism in SSc have provided variable results (13–17). Three studies examined fetal microchimerism in Japanese women (14–16). In a small study, no difference was found in the DNA extracted from peripheral blood mononuclear cells (PBMC) of SSc patients compared with controls, as assessed by nested PCR for male DNA in women with sons (14). A similar assay was used in a larger study in which a significant difference was found in women with SSc compared with controls, all of whom had sons (15). Whether DNA was extracted from whole blood or from PBMC in this latter study was not clear. The third study used a semiquantitative method to test DNA extracted from “white blood cells,” and reported no significant difference in the frequency of microchimerism, but levels of male DNA were greater in some SSc patients than controls (16). In another report, involving a largely Caucasian population, nested PCR for male DNA was used to test PBMC (13). The frequency of microchimerism was greater in SSc patients compared with controls, but the difference was only marginally significant. The most likely synthesis of the results of these studies is that the detection of microchimerism (i.e., using qualitative techniques) is common among healthy controls, and that patients are distinguished from controls primarily by quantitative differences in microchimerism. Some studies have included disease controls (15, 17), but no study included testing at more than 1 time point. Hence, there is a need for additional studies that are longitudinal, apply quantitative techniques, and include inflammatory and/or immunologic disease controls.
Available data are not sufficient to indicate that microchimerism has a role in disease pathogenesis. Microchimerism could be a secondary phenomenon, although the increased risk of SSc observed with prior birth of an HLA–DRB1–compatible child argues against this interpretation. The most persuasive argument for a role in disease pathogenesis has been provided in an experimental model reported by Christner et al (18). In an elegant series of experiments, the investigators created a model of SSc in mice with the use of injections of vinyl chloride, an agent that is associated with SSc. Increased levels of microchimerism were found, accompanied by pathologic changes in the skin and internal organs. Of further interest, marked splenomegaly was observed. Interestingly, a subsequent study of human SSc autopsy specimens used FISH with probes to the X and Y chromosomes to quantitate male cells in women who had sons and found the highest concentration of cells in the spleen (19).
Fetal microchimerism in other autoimmune diseases
In addition to SSc, fetal microchimerism has been investigated in pruritic eruption of pregnancy, Sjögren's syndrome, primary biliary cirrhosis, and thyroid disease. Results of studies have been variable. Pruritic eruption of pregnancy is a skin disorder that occurs during pregnancy. DNA was extracted from skin biopsy specimens from patients with SSc and compared with samples from healthy pregnant women, all of whom had male fetuses (20). Male DNA was found in the patients and not in the healthy controls. A study of women with Sjögren's syndrome examined DNA extracted from PBMC and from cells enriched for CD34 (21). Nested PCR was used for male DNA and no significant difference was found in patients compared with controls. Another limited study of peripheral blood in patients with Sjögren's syndrome also found no significant difference from controls (15).
Primary biliary cirrhosis sometimes occurs in conjunction with SSc and is another autoimmune disease with a peak incidence in the postreproductive years. Primary biliary cirrhosis strongly resembles chronic GVHD of the liver. Five studies have examined fetal microchimerism in this condition (22–26). The initial study used a quantitative assay to test DNA extracted from liver biopsy specimens (22). Male DNA was frequently detected in women with primary biliary cirrhosis but also in patients with other, nonautoimmune liver diseases. Quantitative results were somewhat greater among the patients compared with the controls, but the difference was not significant. Controls included patients with chronic hepatitis, α-antitrypsin deficiency, steatohepatitis with cirrhosis, hepatocellular carcinoma, polycystic liver disease, Wilson's disease, and alcohol-related liver disease.
Two other studies found no significant difference in DNA extracted from whole blood (23) or from PBMC and liver biopsy specimens (24) in patients with primary biliary cirrhosis compared with controls. In situ hybridization (ISH) was used to examine liver biopsy specimens for cells with a Y-chromosome signal in a small study; no Y chromosome–containing cells “within inflammatory cell infiltrates” were detected in 10 women with primary biliary cirrhosis, 5 of whom had sons (25). In a larger study using ISH, male cells were found in 42% of 19 primary biliary cirrhosis patients compared with none of the liver samples from 20 controls (26).
ISH or FISH may be more reliable than PCR techniques for paraffin-embedded tissues. Paraffin baths are often changed only weekly or biweekly. As a result, extracting DNA and using a very sensitive technique such as nested PCR for male DNA raises concern regarding contamination. We tested DNA extracted from 4 pediatric female, paraffin-embedded tissue samples using nested PCR and found that 3 of the 4 tissue samples had male DNA (Nelson JL et al: unpublished observations).
Fetal microchimerism is of special interest in thyroiditis because of the increased incidence of thyroiditis in the postpartum period. One study used nested PCR for male DNA to test the DNA extracted from thyroid specimens from women with and without sons (27). Positive results correlated with women who had sons, and an increased frequency of microchimerism was found in DNA extracted from thyroids affected by Hashimoto's disease compared with nodular goiter. Another study used carefully conducted FISH to identify male cells (28). Male cells were found significantly more often in diseased than in disease-free thyroid specimens. Both studies provide support for the idea that fetal microchimerism could affect the thyroid health of the mother. However, in the latter study, male cells were also found in patients with nonautoimmune disorders, including adenoma, multinodular goiter, and thyroid carcinoma. Therefore, results also raise questions as to whether fetal microchimerism is a secondary phenomenon or, alternatively, whether it could have a broader role in human health and disease.
Critical assessment of studies of microchimerism
There are both important technical issues and clinical issues to consider in the interpretation of studies of microchimerism. Most studies evaluate fetal microchimerism by testing for male cells or male DNA, presumably deriving from a male pregnancy. The specific Y-chromosome sequence is a variable. Some Y-chromosome sequences have cross-reactivity with autosomal sequences. Multiple-copy sequences are more sensitive than single-copy sequences, but single-copy sequences (e.g., SRY) are easily adapted to quantitative assays, whereas some multiple-copy sequences vary in copy number from individual to individual (e.g., DYZ1) and can give spurious results in quantitative assays. The number of aliquots/samples tested and the number of tests conducted can be expected to affect results. Contamination is a recognized risk of PCR-based techniques. Contamination is a concern in studies of paraffin-embedded tissues, since paraffin baths are not usually changed between samples. Studies of tissues may be more reliable when techniques such as FISH are used, avoiding tissue margins and counting only cells with 2 signals in a well-defined nucleus to avoid artifacts produced by overlapping cells.
Techniques that target male DNA or male cells in a woman provide only presumptive evidence for fetal cells. Tests for genetic polymorphisms unique to the child and/or microdissection of the nonhost cells can provide confirmatory data. Results may also vary according to whether PBMC, plasma, serum, or whole blood is tested (1). Clinical variables are of equal significance. Some studies lack data on pregnancy history and others do not provide this information in conjunction with the microchimerism results. Results may be expected to vary depending on whether a target organ derives from early or late in the disease. Ideally, studies should be blinded, pregnancy histories obtained prospectively, alternative sources of microchimerism queried, and disease therapy and duration specified.
Maternal microchimerism in health and disease
Mullinax proposed chimerism as the explanation for the development of systemic lupus erythematosus (SLE) in a patient who received in utero exchange transfusions (29). Another reason to consider maternal microchimerism in some cases of SLE is because a model of SLE in mice is created by introduction of parental cells into F1 progeny (30). IgG antinuclear antibodies are produced in large quantities and fatal immune complex glomerulonephritis develops, similar to spontaneously occurring SLE. Chronic GVHD sometimes exhibits features of SLE, including autoantibodies, neutropenia, and thrombocytopenia. Maternal microchimerism also merits investigation in neonatal lupus, a disorder that occurs in association with maternal autoantibodies and the maternal HLA genotype (31).
Maternal cells have long been known to engraft in immune-deficient progeny, but until recently, persistent maternal microchimerism was not examined in immune-competent progeny. Using converging lines of evidence with PCR-based tests for noninherited, nonshared HLA alleles and FISH for maternal cells in men, maternal microchimerism was found to persist into adult life in some SSc patients and also in healthy normal controls (32). Quantitative techniques will be necessary to determine whether maternal microchimerism plays a role in SSc, particularly in men and nulligravid women.
Addressing the related issue of HLA compatibility, another study reported the HLA compatibility of either the mother or the child as a risk factor for SSc (33). Unfortunately, it is not possible to draw conclusions from that study due to several methodologic issues. These included use of patients with another HLA-associated disease as controls, lack of assurance that all children were studied, and cross-tabulating different HLA–class II loci despite linkage disequilibrium of DRB1, DQA1, and DQB1 genes. In a study of men with SSc, the HLA compatibility of the mother did not differ from that in the control men and mothers (34).
Maternal microchimerism has recently been described in children with juvenile dermatomyositis. A strength of one of the studies was a study design in which unaffected siblings were used as controls (35). Among the patients, 13 of 15, compared with 5 of 35 unaffected siblings, had maternal microchimerism in peripheral blood samples. Maternal cells were found in the muscle tissue of 12 of 15 patients compared with 2 of 10 controls. In a concurrent report, 10 male patients with probable or definite juvenile idiopathic inflammatory myopathy and 10 male controls were investigated for maternal microchimerism (36). A significant difference was found in maternal microchimerism in CD4 and CD8 PBMC subsets and in muscle biopsy tissues when patients were compared with controls.
The long-term effects of maternal microchimerism might not be expected to be the same as those of fetal microchimerism. Long-lasting tolerance might be anticipated for maternal cells that passed into the fetus early in development. Studies of highly sensitized patients awaiting renal transplantation suggest that some individuals develop long-term tolerance to noninherited maternal HLA antigens (37). Factors such as timing during gestation, quantity and type of maternal cell exposure, and HLA genes could impact whether or not maternal cells that persist in the neonate and adult have pathogenic potential. Pregnancy complications and maternal health factors could impact cell traffic and/or its timing during pregnancy.
The recognition of bidirectional cell traffic during pregnancy raises important questions about the immunologic interaction of cells from the child and mother and the consequences thereafter. That fetal cells can persist long-term, probably for a life time, in the mother points to the need for epidemiologic studies that examine pregnancy in relationship to disease susceptibility. Because maternal cells can also persist into adult life, studies that examine the progeny of pregnancies complicated by preeclampsia, prematurity, and placenta abruptio for long-term consequences are also of interest. Microchimerism due to persistent fetal or maternal cells has been investigated in SSc, primary biliary cirrhosis, Sjögren's syndrome, thyroiditis, and myositis, with results that both lend support and raise doubts about its role in autoimmune disease.
The fact that microchimerism is common in healthy individuals suggests that there could be beneficial effects. Just as some HLA genes are associated with the risk of autoimmune disease and others are associated with protection, the particular HLA genes of nonhost cells and their relationship to the host could impact whether consequences are beneficial, neutral, or detrimental to the host. Women with rheumatoid arthritis, for example, often experience amelioration during pregnancy, an effect that occurs in association with fetal–maternal HLA disparity (38). Pregnancy provides modest protection against susceptibility to rheumatoid arthritis (39) and multiple sclerosis (40). If the results of the study by Pisa et al (5) are confirmed, SSc may be added to the list. Pregnancy also reduces the risk of breast cancer (41), suggesting broader beneficial effects may be considered.
Two very different interpretations could be proposed from currently available data. One is that microchimerism is an incidental byproduct of pregnancy without biologic significance. Another is that microchimerism is an active participant in human health with more far-reaching and diverse implications than previously anticipated. Unexpected cellular transformative potential has been demonstrated in other recent studies (42). Thus, another potential twist in the investigation of autoimmunity is if microchimeric cells also have transformative potential. For example, can microchimeric cells become hepatocytes, myocytes, or even neural cells, sometimes functioning to the detriment of the host as a target for “autoimmune” disease, or alternatively, providing a restorative function to the benefit of the host? Clearly, further studies are needed to elucidate the role of microchimerism from pregnancy as protector or insurgent in autoimmune disease, or both.