Klinefelter's syndrome (47,XXY) in male systemic lupus erythematosus patients: Support for the notion of a gene-dose effect from the X chromosome

Abstract

Objective

Systemic lupus erythematosus (SLE) is a systemic autoimmune disease that predominantly affects women. Despite isolated reports of patients with coexisting Klinefelter's syndrome (47,XXY) and SLE, no association of Klinefelter's syndrome with SLE or any other autoimmune disease has been established. The present study was undertaken to investigate the prevalence of Klinefelter's syndrome in a large population of patients with SLE.

Methods

Sex chromosome genotyping was performed in 981 SLE patients, of whom 213 were men. A first group of 844 SLE patients from 378 multiplex families and a second group of 137 men with nonfamilial SLE were evaluated. In selected cases, chromosomes were enumerated by fluorescence in situ hybridization (FISH) and karyotyping in transformed B cell lines.

Results

Of 213 men with SLE, 5 had Klinefelter's syndrome (1 in 43). Four of them were heterozygous at X markers, and Klinefelter's syndrome was confirmed by FISH and karyotyping in the fifth. An overall rate of 47,XXY of 235 per 10,000 male SLE patients was found (95% confidence interval 77–539), a dramatic increase over the known prevalence of Klinefelter's syndrome in an unselected population (17 per 10,000 live male births). Asking men with SLE about fertility was highly sensitive (100%) for Klinefelter's syndrome. All 768 women with SLE were heterozygous at X.

Conclusion

The frequency of Klinefelter's syndrome (47,XXY), often subclinical, is increased in men with SLE by ∼14-fold compared with its prevalence in men without SLE. Diagnostic vigilance for 47,XXY in male patients with SLE is warranted. These data are the first to show an association of Klinefelter's syndrome with an autoimmune disease found predominantly in women. The risk of SLE in men with Klinefelter's syndrome is predicted to be similar to the risk in normal women with 46,XX and ∼14-fold higher than in men with 46,XY, consistent with the notion that SLE susceptibility is partly explained by an X chromosome gene-dose effect.

Systemic lupus erythematosus (SLE) is uncommon in men, being 10-fold more prevalent in women, a difference usually attributed to sex hormones (for review, see ref.1). Since 1969, there have been published reports of SLE in 30 men with Klinefelter's syndrome (2–6), with 2 others having anti–U1 RNP and “mixed connective tissue disease” (7, 8). Klinefelter's syndrome, resulting from a 47,XXY karyotype, is present in 17 of 10,000 live male births (95% confidence interval [95% CI] 14–20 per 10,000 as calculated from published data] [9–11]). Klinefelter's syndrome is characterized by abnormal sexual development at puberty due to sex hormone metabolic differences, with small testes, gynecoid body habitus, absent secondary sexual characteristics, gynecomastia, impotence, and sterility. In many men with 47,XXY, the condition remains undiagnosed until well after puberty.

In a study of 22 men with SLE, no instance of Klinefelter's syndrome was found (12), and in a study of 500 patients with Klinefelter's syndrome, none had or developed SLE after a decade of observation (13). These efforts failed to support the notion of an association between Klinefelter's syndrome and SLE, but the sample sizes were small for studies of uncommon diseases.

In our SLE genetic studies, ∼90% of the patients are women, as would be anticipated (14). In the present investigation, we studied patients with familial SLE and nonfamilial SLE to determine whether the prevalence of Klinefelter's syndrome is increased in men with SLE.

PATIENTS AND METHODS

Patient ascertainment and clinical data collection.

After providing written informed consent, patients who met the American College of Rheumatology classification criteria for SLE (15, 16) and their families (14) were enrolled. Participants who were members of multiplex lupus families (n = 1,258: 76 men with SLE, 768 women with SLE, and 414 men without SLE) were part of the lupus genetics studies conducted at Oklahoma Medical Research Foundation. Ninety-eight of the 137 male nonfamilial SLE cases were part of the PROFILE (The Genetic Profile Predicting the Phenotype) cohort (17), and the remaining 39 male nonfamilial SLE cases came from simplex families from the lupus genetics studies based in Oklahoma (Table 1). Data collection included a questionnaire, an interview, and medical record review. At the time of study enrollment 1 male SLE patient had known Klinefelter's syndrome, which had been diagnosed at age ∼17 years, coincident with the onset of SLE. Androgen therapy led to apparent improvement, with no evidence of an SLE recurrence in the subsequent 15 years (Scofield RH, Bruner GR, Harley JB: unpublished observations).

Table 1. Sex chromosome findings in the study subjects*

	Genotype (or karyotype)†
	46,XY	47,XXY	46,XX	45,XO
* PROFILE = The Genetic Profile Predicting the Phenotype. † In 3 of the men with systemic lupus erythematosus (SLE), 47,XXY was established by fluorescence in situ hybridization and karyotyping. All other sex chromosome designations were deduced from genotyping. ‡ Patient (see Figure 2) was found to have 47,XXY with a duplicated X chromosome.
Group 1 (familial SLE  [Oklahoma])
SLE men	74	2†	0	0
SLE women	0	0	768	0
Non-SLE men	414	0	0	0
Group 2 (nonfamilial SLE)
SLE men (Oklahoma)	38	1†‡	0	0
SLE men (PROFILE)	96	2	0	0
Total SLE men (groups 1 and 2)	208	5	0	0

Figure 2. The 47,XXY chromosomal karyotype of a male systemic lupus erythematosus patient. The diagnosis of Klinefelter's syndrome in this patient was made clinically when he was ∼17 years old. He was shown to be homozygous at all polymorphic X chromosome markers tested, and therefore must have Klinefelter's syndrome as a result of a meiosis II non-disjunction in his mother.

Genotyping and karyotyping.

Each SLE patient in the multiplex lupus group was genotyped with a panel of up to 16 microsatellite nucleotide repeats on the X chromosome, with 2 from the pseudoautosomal regions. Each male in the study was also typed for 2 Y chromosome markers, YS390 and YS389. Chromosomes from lymphoblastoid cell lines from 3 patients were analyzed by trypsin-Giemsa banding (18) and fluorescence in situ hybridization (FISH) analysis with X and Y centromere–specific DNA probes (Vysis, Downers Grove, IL).

The confirmatory cohort of 137 male patients with nonfamilial SLE was typed with 7 X and 2 Y chromosome markers. In both sets of microsatellite typing experiments, markers that spanned the length of the X chromosome were chosen. The male SLE patient with previously diagnosed Klinefelter's syndrome, from the latter cohort, was typed at 256 single nucleotide polymorphisms (SNPs) from the non-pseudoautosomal regions of the X chromosome, using the 10K GeneChip array (Affymetrix, Santa Clara, CA).

Statistical analysis.

Bayes theorem (P[B/A] = (P[A/B] × P[B])/P[A]) was used to estimate the frequency of SLE in Klinefelter's syndrome (where A = the frequency of Klinefelter's syndrome, B = the frequency of SLE in men, and P[B/A] indicates the probability of SLE within the group with Klinefelter's syndrome). Binomial 95% CIs were calculated with the raw data and then converted to numbers per 10,000, for ease of presentation.

RESULTS

In 378 families with 2 or more SLE patients (844 patients total), there were 76 men with SLE. Seventy-four of these men with SLE were monozygous at each of the X chromosome microsatellites tested (Table 1). One male SLE patient was heterozygous at 14 and the other at 12 of the 16 X chromosome markers. Both had allele assignments identical to those in their mothers (Figure 1). The 2 Y chromosome markers, and hence also the Y chromosomes, were present in both of the phenotypic males with heterozygous X chromosomes. A diagnosis of Klinefelter's syndrome or 47,XXY had not been previously considered for either of these men. Cytogenetic and FISH analyses confirmed that both patients had the Klinefelter's syndrome karyotype, 47,XXY. Genotyping studies in 414 men without SLE in these families showed no evidence of >1 X chromosome. All 768 women with SLE studied from the 378 families were heterozygous at multiple X chromosome microsatellite markers, providing no evidence of Turner's syndrome coexisting with SLE in this population.

Figure 1. Pedigree of a male systemic lupus erythematosus (SLE) patient (arrow) who was found to be heterozygous at an X chromosome marker. The marker shown (GATA144D04) is located at 44.7 Mb on the X chromosome. The SLE patient was also heterozygous at 13 of another 15 X chromosome markers (data not shown). He had alleles at this marker identical to those in his mother, and thus inherited both maternal X chromosomes. His father was not typed, but the father's identical twin brother was homozygous with an allele distinct from those in the patient and the patient's mother. Circles = females; squares = males; solid symbols = affected subjects; open symbols = unaffected subjects; symbols with diagonal lines = deceased subjects (not genotyped [NG]).

Thus, 2 of the 76 men with SLE in multiplex families had Klinefelter's syndrome, yielding a prevalence of 263 per 10,000 men with SLE (95% CI 32–918), which is >15 times higher than the rate of Klinefelter's syndrome in the general population and is well above the expected population rate of Klinefelter's syndrome of 17 per 10,000 (9–11). Thus, Klinefelter's syndrome occurred more frequently than expected in these men with SLE (P = 0.007 by Fisher's exact test).

In an effort to test and replicate our findings, we typed an independent group of 137 men with nonfamilial SLE. Of these, 135 were homozygous at the X and Y markers. However, in 1 of them, Klinefelter's syndrome had been diagnosed several years previously. Karyotype and FISH analysis of a lymphoblastoid cell line from this patient revealed 47,XXY (Figure 2), consistent with meiosis II non-disjunction resulting in a duplication of a maternal X chromosome, which is present in ∼18% of Klinefelter's syndrome patients (19). Ascertainment for study enrollment had been based on the patient's SLE, with the presence or absence of Klinefelter's syndrome having no consequence upon the decision to enroll him in the study. The other 2 men were homozygous at the Y markers and heterozygous at multiple X chromosome markers (Table 1). One had 2 alleles at 6 of 7 and the other had 2 alleles at 7 of 7 X chromosome microsatellite markers. Thus, the association of Klinefelter's syndrome and SLE was confirmed in 3 of 137 patients in this group (217 per 10,000 [95% CI 45–622]); the known rate of Klinefelter's syndrome, 17 per 10,000 live male births (9–11), is excluded from the confidence interval.

Given the estimates of the prevalence rates of Klinefelter's syndrome in the original and replication groups, a combined rate of 5 of 213 was obtained (235 men with 47,XXY for every 10,000 men with SLE [95% CI 77–539]), a >13-fold higher prevalence of Klinefelter's syndrome than is found in the general male population (9–11) (P < 0.001 by Fisher's exact test).

The prevalence of SLE is between 1 in 485 for African American women and 1 in 2,762 for European American women (20, 21). Of the SLE patients in the present study, 26.8% were African American and 55.9% European American. With this ethnic distribution, an SLE prevalence of 1 per 1,324 women would be predicted. Data on ethnic differences in SLE prevalence rates among males are not reliable because of the small samples studied (22, 23). Since lupus is 10 times more common in women than in men, we predicted an estimated prevalence of ∼1 in 13,240 for the SLE males studied. The rate of Klinefelter's syndrome among men with SLE (5 in 213) has been presented above, and the incidence of Klinefelter's syndrome in the general population is 17 in 10,000 (9–11). Therefore, since measures of the other 3 variables were available, using Bayes theorem we estimated the rate of SLE among Klinefelter's syndrome patients to be 1 in 960 (Figure 3). This number approximates the rate of SLE in women and is much (nearly 14-fold) higher than the rate of SLE in normal men with 46,XY (22, 23).

Figure 3. Prevalence of 47,XXY in males with systemic lupus erythematosus (SLE). We found that 1 in 43 men with SLE had Klinefelter's syndrome. The known rate of Klinefelter's syndrome in the general population is 17 in 10,000. Using Bayes theorem, we estimated that 1 of every 960 men with Klinefelter's syndrome will have SLE. This number of SLE men and Klinefelter's men would be found in a sample of ∼565,000 men ([960 × 10,000]/17) with the proportional racial composition of the present study population.

Upon study entry, the 135 men with SLE enrolled in this study through the Oklahoma genetics studies were asked “Are you infertile?” All 3 men subsequently found to have Klinefelter's syndrome, along with 6 other men, answered this question with a response other than “no.” Two stated “yes” and the third answered he “did not know.” Thus, this simple clinical question was 100% sensitive and 33% specific for identification of Klinefelter's syndrome in men with SLE (P = 0.00021 comparing men with Klinefelter's syndrome with the other men with SLE, by Fisher's exact test).

DISCUSSION

Klinefelter's syndrome and SLE, two very different conditions, occur with similar prevalence rates in the population. Herein we have shown that these diseases occur together more often than would be expected by chance alone (Figure 3).

There are reasons to suspect that Klinefelter's syndrome and SLE might be associated. First, numerous case reports document the coexistence of the two diseases (1–8, 12, 13, 24). Second, sex hormone similarities between men with Klinefelter's syndrome and women are associated with SLE (1, 4, 25, 26). The argument is bolstered by data from some animal models of SLE, demonstrating increased disease susceptibility with estrogen and disease protection with androgens (for review, see ref.1). Third, Klinefelter's syndrome is also associated with several sex-related conditions. For example, men with Klinefelter's syndrome die of breast cancer at a similar rate as women (27). Approximately 2 in 50 men with breast cancer have Klinefelter's syndrome (28), a rate similar to that found among men with SLE in the present study. There may be an increase in rheumatoid factor levels in the serum of patients with Klinefelter's syndrome (29). Testosterone treatment reverses immune activation abnormalities in Klinefelter's syndrome (30), and has also led to improvement of SLE as described in individual case reports (5, 31) and observed in one of our patients (Scofield RH, Bruner GR, Harley JB: unpublished observations). These observations may be related to the beneficial effect of the mild androgen dehydroepiandrosterone (prasterone) on mild SLE in women (32).

Investigation of X chromosome polymorphisms does not enable detection of 47,XXY caused by duplication of the X chromosome and produced by a maternal meiosis II non-disjunction, which we discovered in 1 patient upon followup of his clinical presentation. Any subclinical 47,XXY caused by this mechanism would not have been detected by X chromosome polymorphism screening, as was applied to some of our study subjects. The actual population prevalence of 47,XXY in our male SLE sample may thus be higher than the 5 in 213 we detected. However, the best estimate suggests that only 18% of patients with Klinefelter's syndrome have this mechanism of supernumerary X chromosomes (19). Therefore, there is a small but finite possibility of an additional man with undiscovered Klinefelter's syndrome among the male SLE patients studied by genotyping of the X chromosome only. Of course, typing of X chromosome markers is not the usual clinical modality for diagnosis for Klinefelter's syndrome, but karyotype and FISH, the usual clinical tests, require testing of cells, which were not available in some subjects.

In other work, we have typed several hundred X chromosome SNPs. We found heterozygosity for 40–60% of these markers in the men identified by microsatellite marker as having 47,XXY in the present study (Scofield RH, Harley JB: unpublished observations). This rate of heterozygosity is similar to that found in normal women. Thus, we are confident that the techniques used in this study have accurately identified men with Klinefelter's syndrome, as well as men with 46,XY.

Klinefelter's syndrome is a genetic abnormality in which androgen and estrogen levels are abnormal from at least the beginning of puberty (33). It may specifically predispose men to development of SLE, compared with other more common forms of hypogonadism, such as primary testicular failure. However, 5 of 35 men with SLE were found to have hypergonadotropic hypogonadism in one study, although the etiology of the hypogonadism was not otherwise delineated (34). Another showed a high rate of hypogonadism in men with rheumatic diseases (only 2 of whom had SLE). Of interest, of 13 men with rheumatic disease and untreated hypogonadism, 5 had Klinefelter's syndrome, 2 had Kallmann's syndrome, and 2 had idiopathic cryptorchism; thus, 9 of 13 had a congenital form of hypogonadism (6). There are case reports of Klinefelter's syndrome occurring in conjunction with other female-predominant autoimmune diseases (35, 36). However, the present data are the first to conclusively demonstrate an association of Klinefelter's syndrome with a female-predominant autoimmune disease.

We estimate that 1 SLE patient will be found among every 960 males with Klinefelter's syndrome. This is much closer to the 1 SLE patient in 1,324 women based on the ethnic distribution in our population than it is to the estimated prevalence of 1 in ∼14,000 for the SLE males. Thus, males with 47,XXY have a risk of lupus comparable with that in females with 46,XX, and not the ∼10-fold lower risk in males with 46,XY. This result is consistent with the notion of a gene-dose effect for lupus risk originating from the X chromosome, where XX (whether 46,XX or 47,XXY) confers a 10-fold higher risk than 46,XY.

Perhaps, future studies of large numbers of women with SLE will be conducted to estimate the relative risk of SLE associated with 45,XO (Turner's syndrome). If the gene-dose hypothesis is correct, then the rate of SLE among individuals with 45,XO should be similar to the rate in males with 46,XY. Our sample of 768 women with SLE is too small to determine whether this is the case. Turner's syndrome is ∼4-fold less prevalent than Klinefelter's syndrome (4 per 10,000) (11, 12). Even so, SLE in Turner's syndrome patients is virtually unreported (37, 38). In contrast, autoimmune thyroid disease has an increased prevalence among patients with Turner's syndrome, especially in patients with an Xq isochromosome (39). These differences suggest that SLE and autoimmune thyroid disease, both of which are female-dominated autoimmune diseases, have distinct X chromosome–dependent susceptibilities. In particular, our data imply that X chromosome monosomy, either congenital or acquired, may not be a risk factor for SLE, as has been suggested for primary biliary cirrhosis (40) or autoimmune thyroid disease (41).

The 46,XX and 47,XXY karyotypic risks for SLE appear to be similar. Consequently, some feature or features of Klinefelter's syndrome must be sufficient to confer the full risk of SLE in females, and, applying parsimony, it can be surmised that the many differences between normal females with 46,XX and males with 47,XXY are not sufficient to alter the risk for SLE. The decreased estrogen levels in individuals with 47,XXY compared with those with 46,XX, for example, do not alter the SLE risk. Thus, our data support the notion that differences in estrogen levels alone do not explain the much lower incidence and prevalence of SLE in men with 46,XY compared with women with 46,XX. Some might construe the lack of an association of oral contraceptive use with disease exacerbation in established SLE (42) as being consistent with this interpretation.

The increased androgen levels in individuals with 47,XXY compared with those with 46,XX do not protect against SLE. The other implication from the 47,XXY prevalence in males with SLE is that the Y chromosome does not appear to influence the overall risk of SLE in men. The increased prevalence of 47,XXY in the SLE population supports the idea that the difference between the prevalence of SLE in men with 46,XY and women with 46,XX is dominated by the X chromosome dose.

Based on the present findings, clinicians treating males with SLE should be aware that such patients have a greater likelihood of having 47,XXY. Increased awareness of Klinefelter's syndrome may improve diagnostic recognition. In the present study population, 3 of 3 male SLE patients with Klinefelter's syndrome answered the fertility question in a way that increased suspicion of infertility. There were only 6 such responses from other men. This question, which was asked and answered before this study was initiated, may be the most generally discriminating initial question to ask men with SLE when screening to identify 47,XXY. Certainly, in any male SLE patient whose fertility is questionable, an evaluation for the physical, and perhaps laboratory, features of Klinefelter's syndrome is warranted.

The finding of an increased prevalence of Klinefelter's syndrome in males with SLE has implications for understanding the genetics of SLE, by suggesting a gene-dose effect at the X chromosome. In addition, diagnosing the presence of 47,XXY among male SLE patients provides them access to potentially important medical management.

AUTHOR CONTRIBUTIONS

Dr. Scofield had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Scofield, Bruner, Kimberly, Li, Harley.

Acquisition of data. Scofield, Bruner, Namjou, Kimberly, Ramsey-Goldman, Petri, Reveille, Alarcón, Vilá, Reid, Harris, Li, Harley.

Analysis and interpretation of data. Scofield, Bruner, Kimberly, Ramsey-Goldman, Li, Kelly, Harley.

Manuscript preparation. Scofield, Bruner, Namjou, Kimberly, Ramsey-Goldman, Reveille, Li, Kelly, Harley.

Statistical analysis. Scofield, Kelly, Harley.

Acknowledgements

The authors are grateful for the help and cooperation of the SLE patients and their families and the referral of cases from support groups and physicians. The statistical advice of Dr. Barbara Neas, technical assistance of Michele Calvo, Parvathi Viswanathan, David Hutchings, Carisa Cooney, and Carrie Thornton, and critical reading of the manuscript by Dr. Michael Lockshin are all appreciated.