1 Department of Obstetrics and Gynaecology and 2 Department of Clinical Immunology, Aalborg Hospital, PO Box 561, DK-9100 Aalborg and 3 Fertility Clinic 4071, Rigshospitalet, Copenhagen University Hospital, DK-2100 Copenhagen, Denmark
4 To whom correspondence should be addressed at: Rolighedsvej 6, DK-9400 Norresundby, Denmark. e-mail: ckruse{at}dadlnet.dk
![]() |
Abstract |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Key words: abortion/HLA antigens/HLA class II/immunology/recurrent miscarriage
![]() |
Introduction |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
If immunological factors influence the risk of RM, the HLA system might play a role. Previous studies have focused on the theory of increased HLA compatibility between spouses as a cause of inappropriate immune recognition of the trophoblast and subsequent miscarriage (Coulam et al., 1987; Ho et al., 1990
), but the majority of studies could not demonstrate increased sharing of HLA antigens (Oksenberg et al., 1984
; Christiansen et al., 1989
; Orgad et al., 2000
), and in prospective studies increased HLA sharing exhibited no prognostic impact on future pregnancy outcome (Recurrent Miscarriage Immunotherapy Trialists Group, 1994
).
Numerous studies have reported that RM patients more often than fertile controls are positive for a series of autoantibodies, especially antiphospholipid antibodies (Branch, 1998). If systemic or localized autoimmunity is involved in the pathogenesis of unexplained RM, an association with certain HLA class II alleles would be expected, as almost all autoimmune disorders are associated with genes in these loci (Klein and Sato, 2000
). This theory is supported by the finding that HLA-DRB1*03 is associated with positivity for anticardiolipin antibodies and antinuclear antibodies in RM patients (Christiansen et al., 1998
). During the last decade, a series of studies by the authors group have focused on the association between HLA class II alleles and unexplained RM. In a casecontrol study using the restriction fragment length polymorphism (RFLP) technique, the frequencies of the HLA-DRB1*01 and -DRB1*03 allogenotypes were significantly increased among patients with at least four previous miscarriages, and it increased significantly with the number of previous fetal losses (Christiansen et al., 1994
). In a prospective study, RM patients with these allogenotypes had a higher miscarriage rate in their next pregnancy compared with other patients (Christiansen et al., 1993
), and in a family study, female relatives of RM patients had a significantly higher risk of miscarriage if they had a HLA-DRB1*01 or -DRB1*03 allogenotype compared with corresponding women without these types (Christiansen et al., 1995
). Finally, a meta-analysis including 18 previous studies exhibited a significantly increased odds ratio (OR) for being DRB1*01 antigen positive among RM patients of Caucasian ethnic origin compared with controls, whereas there was no difference between patients and controls with regard to the HLA-DRB1*03 allele (Christiansen et al., 1999
).
The aims of the present study were to confirm or reject the hypothesis of an association between RM and the HLA-DRB1*01 and -DRB1*03 alleles in a larger study, to separate the contribution of each of the alleles and to investigate interactions of importance with other HLA-DRB1, -DQA1 and -DQB1 alleles.
![]() |
Materials and methods |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Patients
The new patient group (study II) comprised 354 Caucasian women with RM, who had been admitted to the authors clinic from all over Denmark for investigation and treatment. All patients had a history of three or more miscarriages, confirmed by hospital records. None of the patients had uterine abnormalities found by hysteroscopy or hysterosalpingography, and all patients and their husbands had normal karyotypes. All women were regularly menstruating with a cycle length of <35 days and all had normal plasma thyroxin levels. Of these patients, 195 (55%) had had four or more previous miscarriages. Primary RM was found in 212 patients (60%) and 142 patients (40%) had secondary RM, with at least one birth after 28 weeks of gestation preceding RM. The patients had a median age of 32 years (range 2045) and had had a median number of four miscarriages (range 310).
The patient group of study I comprised 234 Caucasian women with RM who had undergone the same investigation programme as the new group.
In the combined patient group of 588 women, 292 (50%) had had four or more previous miscarriages, and 250 (43%) had secondary RM.
Controls
The control group of study II comprised 202 Caucasian individuals from couples with two or more children (median 2, range 24) and a history of no miscarriages. Most control couples (71%) were contacted in the maternity ward after their last delivery. The remainder were staff members and their spouses.
The control group of study I mainly comprised 360 healthy, unrelated, Caucasian Danish blood donors of both sexes.
PCR-SSP typing
All patients and controls in study II were genotyped using a previously described PCRsequence-specific primer (SSP) technique (Olerup et al., 1992, 1993). In short, DNA was isolated from nucleated blood cells. Sequence-specific amplification was carried out using 21 primers for the DRB1 locus and a primer for each of the DRB3, DRB4 and DRB5 loci, 21 primers for the DQA1 locus and eight for the DQB1 locus (Dynal Allsets, Oslo, Norway). The PCR was carried out using a DNA Thermal Cycler (Perkin-Elmer Gene Amp 9600). After agarose gel electrophoresis, the genotype for each patient was interpreted, using the SSP-tool program (Dynal Biotech, Oslo, Norway).
RFLP typing
Allelic TaqI DRB1, DQA1 and DQB1 RFLP typings were carried out in patients and controls in study I, as previously described (Christiansen et al., 1994).
Ethical aspects
Informed consent was obtained from all patients and controls, and the study protocol was approved by the local ethics committee. The study was considered in agreement with the Declaration of Helsinki regarding research in humans.
Statistical analysis
For comparisons between patients and controls in each study (study I and II), the OR was calculated by Haldanes modification of Woolfs method (Haldane, 1955). Whether OR values differed statistically significantly from unity was estimated by the
2 test. As studies I and II were different with regard to the control groups and to the laboratory methods, the combined ORs were calculated using a meta-analysis program [Review Manager (RevMan) Version 4.2 for Windows. Oxford, UK: The Cochrane Collaboration, 2002]. This program weighs the included studies not only according to the total number of individuals in each study, but also according to the number of patients positive for the HLA allele in question, a method which has been recommended (Svejgaard et al., 1974
). The P-value for heterogeneity between the included studies is also tested in the program. This should be >0.05 to allow a combined OR to be calculated.
In all tables of this study, the phenotype frequencies (frequency of individuals carrying the HLA allele) between patients and controls were compared, since this is generally recommended in studies of disease associations (Svejgaard and Ryder, 1994).
The test for increasing or decreasing trend was performed by Spearmans .
A P-value 0.05 was considered significant. P-values were corrected (Pc-values) according to previous recommendations (Svejgaard and Ryder, 1994
): as the two alleles HLA-DRB1*01 and -DRB1*03 were subjected to an a priori hypothesis, their P-values were multiplied by two. For all other HLA-DRB1 or -DQ alleles, the P-values were corrected by multiplication by the number of investigated alleles in each table. For analyses of subgroups of patients, these P-values were corrected further by multiplication by the number of investigated subgroups.
![]() |
Results |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
As expected, the distribution of HLA-DRB1 phenotypes was similar in male and female controls. As for HLA-DRB1*03, this phenotype was found in 21.5% of male and 20.5% of female controls, respectively, in the combined control group.
For study II alone, no differences between patients and controls were significant after correction for multiple comparisons (Tables IIII).
|
|
The frequency of patients with the HLA-DQA1*05 allele was increased, although the difference was not significant (OR = 1.3, 95% CI = 1.01.6, P = 0.07, Pc = 0.70; Table II).
|
The associations with the HLA-DRB1*03 allele were stronger in subgroups of patients with at least four miscarriages (OR = 1.8, 95% CI = 1.32.5, P = 0.0005, Pc = 0.004) and in patients with secondary RM (OR = 1.8, 95% CI = 1.32.5, P = 0.0007, Pc = 0.006) than the corresponding associations among all RM patients. The same trend was observed for the HLA-DRB1*10 allele, but not for the HLA-DRB1*01 and HLA-DRB1*04 alleles (Table IV).
|
The frequencies of the HLA-DRB1*03 allele in subgroups of patients with three, four or 5 previous miscarriages increased significantly with the number of miscarriages (P = 0.01, Pc = 0.02; Table V).
|
|
![]() |
Discussion |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
In the present study, two study groups were combined. The frequencies of some alleles differed between the old and new groups of patients and controls, respectively, (Table IIII). Whereas the patients in the two studies have been recruited under identical selection criteria, the control groups, although recruited from the same local area, had different origins: in study I, most controls were blood donors, and therefore selected to be healthy, but with an unknown reproductive history, whereas in study II the controls were couples with proven normal fecundity, and otherwise may be more representative of the normal population. However, couples with normal fecundity are generally healthy and, due to the low prevalence of RM in the general population, it can be expected that most blood donors have normal fecundity. The variation in the allele frequencies could be attributed to chance, since the test for heterogeneity for most alleles justified the combination of the two studies.
The present study confirmed our prior hypothesis that the HLA-DRB1*03 allele confers susceptibility to RM, since patients with this allele were found with a significantly increased frequency compared with controls before and after adjustment for multiple comparisons (Table I). The adjustment of this P-value was carried out according to the a priori hypothesis in our protocol that HLA-DRB1*01 and -DRB1*03 would be found with increased prevalence in the RM patients. This prior hypothesis was based on results from our previous casecontrol study (Christiansen et al., 1994), a prospective study (Christiansen et al., 1993
) and a family study (Christiansen et al., 1995
). In the latter two studies, the significance of the HLA-DRB1*01 and -DRB1*03 alleles was investigated together en bloc, and the main purpose of the present study was to separate the contribution of each of the alleles to the susceptibility to RM.
In subgroups of patients with severe RM (at least four miscarriages) or with secondary RM, the associations with HLA-DRB1*03 became highly significant (Table IV), and were still highly significant after correction of the P-value for the number of subgroup analyses. Since there was an overlap of 163 patients who had both severe and secondary RM, the stronger association between HLA-DRB1*03 and secondary RM could be caused by a higher number of miscarriages in this subgroup. However, it was calculated that the ORs were higher for patients with secondary RM compared with those with primary RM in the subgroup of patients with only three miscarriages (0.9 and 1.6, respectively) as well as in the subgroup with at least four miscarriages (1.4 and 2.0, respectively). The frequency of the HLA-DRB1*03 allele increased significantly with increasing number of previous miscarriages (Table V), which might indicate a causal relationship.
Surprisingly, the study could not confirm that HLA-DRB1*01 confers susceptibility to RM since, in the new sample of patients and controls (study II), the prevalence of this allele was lower in patients than in controls (Table I). In the control group of study II, the prevalence of individuals positive for HLA-DR1 was >6% higher than in study I. Some of this variation could be attributed to chance, since the test for heterogeneity between studies did not find them significantly different, or it might be explained by the different origin of the control groups, as previously discussed. On the other hand, study II gave further support for the theory that HLA-DRB1*10 or linked HLA-DQ alleles are risk alleles for RM. In study I, this allele was associated with RM exhibiting an OR of 3.2, which was, however, not statistically significantly increased due to the rarity of the allele. In study II, a similar OR was found, resulting in a combined OR of 3.4 that was significantly increased before adjustment for multiple comparisons. In our previous publication (Christiansen et al., 1994), we put forward the hypothesis that the DQB1*0501 allele rather than the HLA-DRB1*01 and -DRB1*10 alleles was the susceptibility allele for RM since this DQB1 allele is in strong linkage disequilibrium with both HLA-DRB1*01 and -DRB1*10 and, in our population, found on almost all haplotypes with these HLA-DRB1 alleles. The findings in the present study concerning HLA-DRB1*01 and HLA-DRB1*10 weaken this hypothesis, and the results shown in Table VI provide no unambiguous answer to the question of whether HLA-DQB1*0501 found together with HLA-DRB1*03 results in a genotype which confers a higher risk for RM than HLA-DRB1*03 on its own.
In Table VI the question was addressed of whether HLA-DRB1*03 in combination with HLA-DQB1*0501 or with HLA-DRB1*04 confers a higher susceptibility to RM than HLA-DRB1*03 alone. As mentioned above, it was hypothesized in a previous study (Christiansen et al., 1994) that HLA-DQB1*0501 is associated with RM. The finding of a higher OR (2.8) for the DRB1*03/DQB1*0501 combination than for HLA-DRB1*03 alone (1.8) in patients with four or more miscarriages might support this hypothesis. The frequency of HLA-DRB1*04 was increased in RM patients compared with controls (Table VI), and since HLA-DRB1*03 and -DRB1*04 have a synergistic effect on the predisposition to an autoimmune disease, insulin-dependent diabetes mellitus (Noble et al., 1996
), this might also apply to the susceptibility to RM. The present results indeed indicate that the HLA-DRB1*03/DRB1*04 genotype tended to confer a higher OR for RM than the HLA-DRB1*03 allele alone (Tables IV and VI).
The frequency of some alleles was significantly decreased in patients compared with controls (Tables IIII), although not after correction for multiple comparisons. This can be explained by the relatively higher frequency of other alleles in the patient group, and little emphasis has been put on this finding.
Strong linkage disequilibrium exists between HLA-DRB1, -DQA1 and -DQB1 alleles, meaning that some alleles in these loci are found together on the same haplotype much more frequently than expected by chance alone. An example is HLA-DRB1*03, which in Caucasian populations in practice always implies the presence of DQA1*0501 and DQB1*0201 on the same haplotype. Therefore, an association between an allele, e.g. HLA-DRB1*03, and a disorder could be secondary to an association between the DQA1*0501 or DBB1*0201 allele. A method has been recommended for calculating which allele among several candidates found in linkage disequilibrium with each other exhibits the strongest association with a disorder (Svejgaard and Ryder, 1994). These calculations were done for the HLA-DRB1*03, DQA1*0501, DQB1*0201 haplotype, as the demonstrated association between RM and HLA-DRB1*03 in principle could be secondary to associations with DQA1*0501, which is found with increased frequency (although not significantly) in patients, or with DQB1*0201. When the associations of the DQA1*0501 and DQB1*0201 alleles with RM are stratified in patients and controls according to positivity or negativity for HLA-DRB1*03, the ORs in HLA-DRB1*03-negative patients (1.1 and 0.7 for the DQA1*0501 and DQB1*0201 alleles, respectively) tended to be lower than the OR of 1.4 for HLA-DRB1*03, indicating that it is the HLA-DRB1*03 allele, rather than the DQA1 and DQB1 alleles found in linkage disequilibrium with it, that exhibits the primary association with RM. However, some reservations must be noted for these calculations, since no deviations were significant and certain combinations of alleles were not observed, either in patients or in controls.
The pathophysiolocigal explanation for the association between HLA-DRB1*03 and RM is still not clear. A possible model might be that HLA-DRB1*03 predisposes to an autoreactive immune response against one or more trophoblast antigens: HLA-DRB1*03 molecules on maternal antigen-presenting cells may present trophoblast-derived peptides to maternal autoreactive T cells. As another possible explanation, the association might be caused by a close linkage between the HLA-DRB1 genes and the genes for tumour necrosis factor (TNF), since it has been reported that HLA-DRB1*03 and -DRB1*04 are in linkage disequilibrium with alleles predisposing to in vitro hypersecretion of TNF- (Pociot et al., 1993
), which in murine and human in vitro studies has exhibited embryotoxic or trophoblast-inhibiting activity (Raghupathy, 1997
). A third explanation may be that HLA-DRB1*03 is in linkage disequilibrium with HLA-G alleles associated with RM (Hviid et al., 2002
). HLA-G is of particular interest with respect to the reproductive processes, as it is the dominant HLA antigen on the human trophoblast.
In conclusion, it was confirmed in the present study that HLA-DRB1*03 confers susceptibility to RM, and the association is probably not caused by the DQA1 and DQB1 alleles found in linkage disequilibrium with it. A genotypic combination of HLA-DRB1*03 with -DRB1*04 or with -DQB1*0501 might increase the susceptibility to RM, but this finding must be confirmed in further studies. The results of the present study could not confirm the hypothesis of an association between HLA-DRB1*01 and RM. Larger studies are needed if the main susceptibility gene sequences are to be identified with certainty. However, the results up until now indicate that RM is a multifactorial condition, and that HLA-DRB1*03, and perhaps other class II antigens, does not act as a sole factor to cause miscarriage, but increases the risk of fetal loss when found in conjunction with other immunological (Kruse et al., 2002) or haemophilic (Rey et al., 2003
) disturbances.
![]() |
Acknowledgements |
---|
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Bellingard V, Hedon B, Eliaou JF, Seignalet J, Clot J and Viala JL (1995) Immunogenetic study of couples with recurrent spontaneous abortions. Eur J Obstet Gynecol Reprod Biol 60,5360.[CrossRef][Medline]
Branch DW (1998) Antiphospholipid antibodies and reproductive outcome: the current state of affairs. J Reprod Immunol 38,7587.[CrossRef][Medline]
Christiansen OB (1996) A fresh look at the causes and treatments of recurrent miscarriage, especially its immunological aspects. Hum Reprod Update 2,271293.
Christiansen OB, Riisom K, Lauritsen JG and Grunnet N (1989) No increased histocompatibility antigen sharing in couples with idiopathic habitual abortions. Hum Reprod 4,160162.[Abstract]
Christiansen OB, Mathiesen O, Husth M, Lauritsen JG, Jersild C and Grunnet N (1993) Prognostic significance of maternal DR histocompatibility types in Danish women with recurrent miscarriages. Hum Reprod 8,18431847.[Abstract]
Christiansen OB, Rasmussen KL, Jersild C and Grunnet N (1994) HLA class II alleles confer susceptibility to recurrent fetal losses in Danish women. Tissue Antigens 44,225233.[Medline]
Christiansen OB, Andersen HH, Højbjerre M, Kruse TA, Lauritsen SL and Grunnet N (1995) Maternal class II allogenotypes are markers for the predisposition to fetal losses in families of women with unexplained recurrent fetal loss. J Immunogenet 22,323334.
Christiansen OB, Ulcova-Gallova Z, Mohapeloa H and Krauz V (1998) Studies on associations between human leukocyte antigen (HLA) class II alleles and antiphospholipid antibodies in Danish and Czech women with recurrent miscarriages. Hum Reprod 13,33263331.[Abstract]
Christiansen OB, Ring M, Rosgaard A, Grunnet N and Gluud C (1999) Association between HLA-DR1 and DR-3 antigens and unexplained repeated miscarriage. Hum Reprod Update 5,249255.
Coulam CB (2000) Understanding the immunobiology of pregnancy and applying it to treatment of recurrent pregnancy loss. Early Pregnancy 4,1929.[Medline]
Coulam CB, Moore SB and OFallon WM (1987) Association between major histocompatibility antigen (HLA) and reproductive performance. Am J Reprod Immunol Microbiol 14,177188.
Gerencer M and Kastelan A (1983) The role of HLA-D region in feto-maternal interactions. Transplant Proc 15,893895.
Haldane JBS (1955) The estimation and significance of the logarithm of a ratio of frequencies. Ann Hum Genet 20,309311.
Ho H, Gill TJ, Nsieh RP, Nsieh HJ and Lee TY (1990) Sharing of human leukocyte antigens in primary and secondary recurrent spontaneous abortions. Am J Obstet Gynecol 163,178188.[Medline]
Hviid TV, Hylenius S, Hoegh AM, Kruse C and Christiansen OB (2002) HLA-G polymorphism in couples with recurrent spontaneous abortions. Tissue Antigens 60,122132.[CrossRef][Medline]
Klein J and Sato A (2000) The HLA system. Second of two parts. N Engl J Med 343,782786.
Kruse C, Rosgaard A, Steffensen R, Varming K, Jensenius JC and Christiansen OB (2002) Low serum level of mannan-binding lectin is a determinant for pregnancy outcome in women with recurrent spontaneous abortion. Am J Obstet Gynecol 187,13131320.[CrossRef][Medline]
Laitinen T, Koskimies S and Westman P (1993) Foeto-maternal compatibility in HLA-DR, -DQ and -DP loci in Finnish couples suffering from recurrent spontaneous abortions. Eur J Immunogenet 20,249258.[Medline]
McIntyre JA, McConnachie PR, Taylor CG and Faulk WP (1984) Clinical, immunological and genetic definitions of primary and secondary recurrent spontaneous abortions. Fertil Steril 42,849855.[Medline]
Noble JA, Valdes AM, Cook M, Klitz W, Thomson G and Erlich HA (1996) The role of HLA class II genes in insulin-dependent diabetes mellitus: molecular analysis of 180 caucasian mutiplex families. Am J Hum Genetl 59,11341148.
Oksenberg JR, Persitz E, Amar A and Brautbar C (1984) Maternelpaternal histocompatibility: lack of association with habitual abortions. Fertil Steril 42,389395.[Medline]
Olerup O and Zetterquist H (1992) HLA-DR typing by PCR amplification with sequence-specific primers (PCRSSP) in 2 hours: an alternative to serological DR typing in clinical practice including donorrecipient matching in cadaveric transplantation. Tissue Antigens 39,225235.[Medline]
Olerup O, Aldener A and Fogdell A (1993) HLA-DQB1 and -DQA1 typing by PCR amplification with sequence-specific primers (PCRSSP) in 2 hours. Tissue Antigens 41,119134.[Medline]
Orgad S, Gazit E and Carp HJ (2000) Shared HLA antigens in couples with 5 or more, compared with 3 or 4 recurrent miscarriages. Early Pregnancy 4,144153.[Medline]
Pennesi G, Brioli G, Lulli P, Mariani B, Morellini M, Nicotra M and Trabace S (1998) HLA and complement factors alleles sharing in Italian couples with recurrent spontaneous abortions. Hum Immunol 59,382386.[CrossRef][Medline]
Pociot F, Briant L, Jongeneel CV, Molvig J, Worsaae H, Abbal M, Thomsen M, Nerup J and Cambon-Thomsen A (1993) Association of tumor necrosis factor (TNF) and class II major histocompatibility complex alleles with the secretion of TNF- and TNF-
by human mononuclear cells: a possible link to insulin-dependent diabetes mellitus. Eur J Immunol 23,224231.[Medline]
Raghupathy R (1997) Th-1 type immunity is incompatible with successful pregnancy. Immunol Today 18,478482.[CrossRef][Medline]
Recurrent Miscarriage Immunotherapy Trialists Group. (1994) Worldwide collaborative observational study and meta-analysis on allogeneic leukocyte immunotherapy for recurrent spontaneous abortion. Am J Reprod Immunol 32,5572.[Medline]
Rey E, Kahn SR, David M and Shrier I (2003) Thrombophilic disorders and fetal loss: a meta-analysis. Lancet 361,901908.[CrossRef][Medline]
Sasaki,T, Yamada H, Kato EH, Sudo S, Kishida T, Sasaki T, Nishigaki F and Fujimoto S (1997) Increased frequency of HLA-DR4 allele in women with unexplained recurrent spontaneous abortions, detected by the method of PCRSSP. J Reprod Immunol 32,273279.[CrossRef][Medline]
Sbracia M, Mastrone M, Scarpellini F and Grasso JA (1996) Influence of histocompatibility antigens in recurrent spontaneous abortion couples and on their reproductive performances. Am J Reprod Immunol 35,8592.[Medline]
Smith JB, Cowchock FS, Hankinson B and Iftekhar A (1989) Association of HLA-DR5 with recurrent spontaneous abortion in women treated with paternal leucocytes. Arthritis Rheum 32,15721576.[Medline]
Svejgaard A and Ryder LP (1994) HLA and disease associations: detecting the strongest association. Tissue Antigens 43,1827.[Medline]
Svejgaard A, Jersild C, Nielsen LS and Bodmer WF (1974) HLA antigens and disease. Statistical and genetical considerations. Tissue Antigens 4,95105.[Medline]
Takakuwa K, Adachi H, Hataya I, Ishii K, Tamura M and Tanaka K (2003) Molecular genetic studies of HLA-DRB1 alleles in patients with unexplained recurrent abortion in the Japanese population. Hum Reprod 18,728733.
Submitted on October 27, 2003; accepted on February 10, 2004.