1857
Testicular Carcinoma and HLA Class II Genes Dirk J. A. Sonneveld, M.D., Ph.D.1 Martijn F. Lutke Holzik, M.D.1 Ilja M. Nolte, M.S.2 Dirk Th. Sleijfer, M.D., Ph.D.3 Winette T. A. van der Graaf, M.D., Ph.D.3 Marcel Bruinenberg, M.S.4 Rolf H. Sijmons, M.D., Ph.D.5 Harald J. Hoekstra, M.D., Ph.D.1 Gerard J. Te Meerman, M.S., Ph.D.2 1
Department of Surgical Oncology, University Medical Center Groningen, Groningen, The Netherlands.
2
Department of Medical Genetics, University of Groningen, Groningen, The Netherlands.
3
Department of Medical Oncology, University Medical Center Groningen, Groningen, The Netherlands.
4
Department of Medical Biology, University of Groningen, Groningen, The Netherlands.
BACKGROUND. The association with histocompatibility antigens (HLA), in particular Class II genes (DQB1, DRB1), has recently been suggested to be one of the genetic factors involved in testicular germ cell tumor (TGCT) development. The current study, which uses genotyping of microsatellite markers, was designed to replicate previous associations. METHODS. In 151 patients, along with controls comprising parents or spouses, the HLA region (particularly Class II) on chromosome 6p21 was genotyped for a set of 15 closely linked microsatellite markers. RESULTS. In both patients and controls, strong linkage disequilibrium was observed in the genotyped region, indicating that similar haplotypes are likely to be identical by descent. However, association analysis and the transmission disequilibrium test did not show significant results. Haplotype sharing statistics, a haplotype method that derives extra information from phase and single marker tests, did not show differences in haplotype sharing between patients and controls. CONCLUSION. The current genotyping study did not confirm the previously reported association between HLA Class II genes and TGCT. As the HLA alleles for which associations were reported are also prevalent in the Dutch populations, these associations are likely to be nonexistent or much weaker than previously reported. Cancer 2002;95:1857– 63. © 2002 American Cancer Society. DOI 10.1002/cncr.10903
5
Department of Clinical Genetics, University Medical Center Groningen, Groningen, The Netherlands.
KEYWORDS: testicular carcinoma, human leukocyte antigens, association, haplotype sharing, linkage disequilibrium.
T
Presented at the Annual Meeting of The American Society of Clinical Oncology, 2000 (poster presentation, abstract 787) and at the Fifth Germ Cell Tumor Conference, LEEDS, United Kingdom, September 13–16, 2001. Supported by Grant RUG 1999-2130 from the Dutch Cancer Society. The authors thank all families who donated their DNA to this project. Address for reprints: Gerard J. Te Meerman, M.S., Ph.D., Department of Medical Genetics, University of Groningen, A.Deusinglaan 4, 9713 AW Groningen, The Netherlands; Fax: 011-31-50-3633113; E-mail:
[email protected] Received February 7, 2002; revision received May 15, 2002; accepted May 30, 2002. © 2002 American Cancer Society
esticular germ cell tumors (TGCT) constitute the most common malignancy in men 20 – 40 years of age. The etiology of TGCT is still poorly understood. In addition to possible environmental predisposing factors, several observations point to a genetic susceptibility to the development of TGCT.1,2 First, familial and bilateral testicular carcinoma cases occur more frequently than expected by chance. The relative risk (RR) for brothers of TGCT patients ranges from 3 to 13.3–5 Second, a genetic susceptibility is reflected by an increased incidence of TGCT in persons with certain rare malformations of the urogenital system, some of which have a definite genetic component in the etiology.6,7 Furthermore, higher rates of urogenital developmental anomalies have been reported in families prone to TGCT.8 Third, the age distribution of TGCT may suggest a genetic origin of the disease. TGCT are usually diagnosed at a young age and the incidence declines after the age of 50 years. The young age at onset of testicular neoplasms indicates a role of important etiologic factors operating early in life, either in utero or shortly after birth. In addition to in utero exposure to maternal estrogens or exposure to infectious agents in early childhood, these early operating etiologic factors may also be genetic.6,9 –11 Finally, racial differences in the incidence of TGCT may point to a genetic component in the etiology of the disease. The
1858
CANCER November 1, 2002 / Volume 95 / Number 9
highest incidence is observed in Caucasians of Northern European descent. People of African descent have a universally low incidence of TGCT. In the United States, the incidence of TGCT in African Americans is only one fourth of that observed in Caucasians.1,6,12 Based on findings in numerous clinical and epidemiologic studies, a genetic susceptibility to TGCT is very likely. Several candidate genes have been proposed to play a role in the etiology of the disease.6,13–18 Recently, a high-resolution genotyping study in 55 Japanese TGCT patients showed a histocompatibility antigen (HLA)-DRB1 susceptibility allele (RR 3.26) and an HLA-DQB1 candidate protective allele (RR 0.26) for TGCT.19 The possible role of the HLA system in TGCT development could result from effects of HLA variation on immune response to carcinogenic factors, for example, viruses that may be etiologically associated with the development of TGCT.20 The importance of the HLA system in regulating susceptibility to, and tumor development in, a growing number of neoplastic conditions is becoming increasingly clear. An impaired immune system, genetically or acquired, favors carcinogenic factors.21 Linkage studies in TGCT families could be performed to map candidate genes for TGCT. Unfortunately, to date, the lack of a sufficient number of families with two or more affected men with TGCT has been a handicap to perform linkage studies with enough power to find effects of frequent alleles or genes with low marginal effects.14 An alternative approach to linkage studies in familial cases is to search for testicular carcinoma susceptibility genes among testicular carcinoma patients in founder populations by means of association analysis including haplotype methods. These approaches have more power than linkage analysis when high-frequency alleles are involved in the pathogenesis. However,unlike linkage analysis, they require markers that are very close to the causative genes. Patients in founder populations are expected to share a relatively high number of alleles from recent common ancestors, which is also expected to apply for mutations predisposing to testicular carcinoma. Therefore, founder populations are particularly suitable for finding genes predisposing to the development of testicular neoplasms through association-based methods. In a previous study,5 we showed the geographic clustering of testicular carcinoma in the northern part of The Netherlands, which is indicative of the importance of founder alleles. The current study, which includes testicular carcinoma patients and their relatives from this founder population, is the first extensive genotyping of the HLA region on chromosome 6p21 in a large number of TGCT patients using both standard methods and a new haplotype
TABLE 1 Patient Characteristics No. of patients Median age at diagnosis (range) Histologic type (%) Pure seminoma Nonseminomaa Familial testicular carcinoma (%) Affected first-degree relative Affected second-degree relative Bilateral testicular carcinoma (%) History of undescended Testis (%) a
151 29.4 (15.9–63.0) yrs 19 (13) 132 (87) 16 (11) 7 (5) 9 (6) 7 (5) 23 (15)
With or without a seminomatous component.
sharing method to examine the association between HLA Class II genes and TGCT.
MATERIALS AND METHODS Patients A total of 151 TGCT patients treated at the University Medical Center Groningen (UMCG) in The Netherlands during the period 1977–1998 were selected for the initial analysis. The majority of these patients descended from three provinces (Groningen, Friesland, Drenthe) in the northern part of The Netherlands, based on information collected about birthplace of the patients’ great-grandparents. Patient characteristics are listed in Table 1. The difference between the total number of nonseminomas (n ⫽ 132, 87%) and pure seminomas (n ⫽ 19, 13%) is due to different referral patterns for these histologic subtypes. Histologic diagnosis was established for all patients by the Department of Pathology, UMCG. For 108 patients, DNA from both parents was available for phase determination and for control (nontransmitted haplotype). For the remaining 43 patients, DNA from children (older than 18 years) and spouses was available for phase determination. In these cases, the haplotypes of the spouses were regarded as controls. All participants gave their informed consent and the Ethical Committee of the UMCG approved the study.
Genotyping High molecular weight genomic DNA was extracted from peripheral lymphocytes from 20 mL blood using standard protocols. After DNA extraction, a set of 15 polymorphic microsatellite markers in the HLA region on 6p21 was genotyped in all patients and controls. Markers were selected over a distance of approximately 8 cM. The marker order was determined by using sequence data of the major histocompatibility complex (MHC).22 Most of the 15 markers used in the current study are located on this 3.6-megabase MHC
Testicular Carcinoma/Sonneveld et al.
1859
TABLE 2 Marker Data Locus
Marker name
Position (bp)a
Forward primer (5ⴕ 3 3ⴕ)
Reverse primer (5ⴕ 3 3ⴕ)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
D6S1560 DNRNGCA RING3CA D6S2445 TAP1 D6S2444 D6S2443 G511525 D6S1666 D6S273 TNF␣ C3-4-13 D6S265 D6S478 D6S258
telomeric 355440 375387 461158 498634 601183 631213 648774 700043 1577323 1725592 2839280 3165149 3376450 centromeric
CTCCAGTCCCCACTGC AGGAATCTAGTGCTCTCTCC TGCTTATAGGGAGACTACCG AATATGATGGAAGAAGTAATCCAG GCTTTGATCTCCCCCCTC GAGCCAAGAACCCAGCATTC CCATACCAAAGTAAAACCCAG GGTAAAATTCCTGACTGGCC CTGAGTTGGGCAGCATTTG GCAACTTTTCTGTCAATCCA GCCTCTAGATTTCATCCAGCCACA GCATGACACTATAGTGGCTG ACGTTCGTACCCATTAACCT CCTCCATAATTGTGTGAGCC GCAAATCAAGAATGTAATTCCC
CCCAAGGCCACATAGC CTCTAGCAAAAGGAAGAGCC GATGGGAAGTTTCCAGAGTG GGATTACAGGTATAAGCCATTG GGACAATATTTTGCTCCTGAGG GGAAGGATTCTAAATAGGGGAG GAGGATGAAGGGAAATTAGAG GACAGCTCTTCTTAACCTGC ACCCAGCATTTTGGAGTTG ACCAAACTTCAAATTTTCGG CCTCTCTCCCCTGCAACACACA CATTGCACTCCAGTCTGGGC ATCGAGGTAAACAGCAGAAA CCAATCTTCTAACCCAAGCA CTTCCAATCCATAAGCATGG
a
Position (base pair) on sequence as published by the Sanger Centre.22
denaturation of 5 minutes at 95 °C, 35 cycles of 30 seconds at 95 °C, 30 seconds at 55 °C, and 1 minute at 72 °C. Post-PCR multiplexing was performed by combining 1–10 L (based on signal strength) of PCR products. Pooled fragments (2.3 L) were mixed with 2.5 L deionized formamide and 0.2 L ET-400R size standard (Amersham Pharmacia Biotech, Upsala, Sweden) and separated on a MegaBACE 1000 capillary sequencer (Amersham Pharmacia Biotech)according to the manufacturer’s protocol. Results were analyzed using genetic profiler v1.1 (Amarsham Pharmacia Biotech).
Statistical Methods FIGURE 1. Markers used in this study. They cover fully the two loci found by O¨ zdemir et al.19
sequence, particularly in the HLA Class II region. The position of the remaining markers on the immediate flanking regions was determined by using available marker information from data published in print or on the internet (see Table 2 and Fig. 1 for marker information).16,23,24 For each polymerase chain reaction (PCR), 0.25 units Taq DNA polymerase (Roche diagnostics, Mannheim, Germany) was used to amplify the fragments. The reaction volume was 10 L. Reaction mixtures contained 200 M of each dNTP, 1.5 mM MgCl2 , 10 mM tris-HCL, 50 mM KCl, and 0.25 M of each primer (with one primer 5⬘ labeled with a fluorochrome 6-FAM, HEX, or NED). Cycling was performed on a PTC-225 thermal cycler (MJ Research, Waltham, MA). Amplification consisted of an initial
After genotyping the 15 markers in all participants (each marker meets the Hardy-Weinberg criteria), the set of haplotypes present in patients and the set of nontransmitted haplotypes present in parents or spouse control haplotypes were determined. Differences between these two haplotype sets were analyzed using the haplotype sharing statistic (HSS), a new method for quantitative analysis of haplotype similarity. The validity of this method is demonstrated elsewhere by applying it extensively to simulated and empirical data.23,25–27 HSS assumes that mutations that predispose to TGCT will be present more often in patients than in controls. Some of the patients will have inherited a possible predisposing mutation to TGCT development from a common ancestor, especially in a founder population. Due to linkage disequilibrium, a small haplotype surrounding this mutation will also be identical by descent among these patients. The amount of identical DNA is dependent on the
1860
CANCER November 1, 2002 / Volume 95 / Number 9
number of recombinations that have taken place on either side of the mutation since it occurred. This number is particularly influenced by the number of meioses that have taken place and the recombination frequency of the region. The number of meioses between patients, selected from a population on the condition that a specific disease mutation is present on their haplotypes, is expected to be smaller than the number of meioses between a random sample of controls from that population. Therefore, the length of identical or shared DNA surrounding a predisposing mutation in patients is expected to be larger than the length of shared DNA surrounding that locus in controls. The difference in length of haplotype sharing surrounding a specific locus between patients and controls can be used as an indication for involvement of this locus in susceptibility to the disease. This difference is expected to be largest at the marker locus closest to the disease mutation. HSS defines the haplotype sharing between pairs of haplotypes as the number of intervals between loci that shows identical alleles on a row from a locus in both directions. This can be evaluated for all pairs of haplotypes for all marker loci. As a test of linkage disequilibrium, the statistical significance of the overlap of the observed haplotypes is evaluated. For this test, the observed alleles are redistributed randomly over their loci and the haplotype sharing in this randomized set is calculated. The mean haplotype sharing observed in the data is compared to the mean haplotype sharing in the randomized data in which there is linkage equilibrium among all marker loci. In addition to the new haplotype sharing test, we also performed single and two-locus association tests and a transmission disequilibrium test (TDT). For the one and two-locus association tests, the frequencies of the alleles and two locus haplotypes, respectively, are compared using a chi-square test, taking only those alleles into account that have an expected frequency of at least one copy. The TDT evaluates the transmission distortion of each allele versus all other alleles, using the test proposed by Spielman et al.28 For each locus, the result given will be the maximum distortion at that locus.
RESULTS Figure 2 shows the test of linkage disequilibrium. In both patients and controls, there is significant excess sharing (⫺log10 [P value] ⬎ 3) compared with random redistribution for all loci, indicating strong linkage disequilibrium in the entire genotyped region. Haplotype sharing in the presence of linkage disequilibrium indicates that similar haplotypes are likely to be identical by descent, i.e., inherited from a common ances-
FIGURE 2.
Deviation from multilocus linkage equilibrium (LE) in patients (solid line) and controls (dotted line), evaluated by haplotype sharing. Multilocus LE is simulated by permutation of the alleles over the haplotypes. The results are represented as the ⫺log10 of the significance of the difference in haplotype sharing between the observed and randomized haplotypes. Significance level is at ⫺log10 (P value) greater than 3.
FIGURE 3. Difference between haplotype sharing between patient and control haplotypes as calculated by a t test, expressed as ⫺log10 of the significance. The standard deviation is calculated by repeated sampling without replacement of 50% of the observed haplotypes. Significance level is at ⫺log10 (P value) greater than 3.
tor, suggesting that HSS has the power to detect differences between patients and controls. These differences in mean haplotype sharing between patients and controls are plotted in Figure 3. No significant differences (⫺log10 [P value] ⬍ 3) between patients and controls in haplotype sharing are observed over the entire region between marker loci 1 and 15. In addition, single and two-locus association tests (Fig. 4), as well as the TDT (Fig. 5), showed no significant association (⫺log10 [P value] ⬍ 3) between TGCT and marker loci in the HLA region on 6p21. Subanalysis of histologic groups (i.e., seminoma, nonseminoma) demonstrated no significant differences between patients and controls (data not shown). It must be noted, however, that the number of seminoma patients is
Testicular Carcinoma/Sonneveld et al.
FIGURE 4.
Association of the markers with testicular germ cell tumors. Results of single-locus (solid line) and two-locus (dotted line) association analysis. The distributions of alleles in patients and controls are compared using the chi-square test. The results are presented as ⫺log10 of the significance. Significance level is at ⫺log10 (P value) greater than 3.6.
FIGURE 5.
Results of the transmission disequilibrium test in 108 trios represented as ⫺log10 of the significance of the maximal transmission distortion at each locus. Significance level is at ⫺log10 (P value) greater than 3.2.
rather small due to the unique referral pattern of a cancer center. Our results particularly pertain to nonseminomas.
DISCUSSION It is likely that immune response differences based on HLA variation may play a role in carcinoma development and metastatic patterns. Several carcinomas have been reported to be HLA associated (e.g., Hodgkin lymphoma, Kaposi sarcoma, colorectal carcinoma, and Burkitt lymphoma).21 HLA variation has also been suggested to play a role in TGCT development. This may be caused by differences in effects of
1861
HLA variation on immune response to carcinogenic factors, such as viruses that may be etiologically associated with the development of TGCT.20 An impaired immune system, genetically or acquired, favors carcinogenic factors. The theory that TGCT may arise under conditions of reduced immune capacity is supported by the observation that patients with immune deficiencies following renal transplantation have a two to fivefold increased risk to develop a TGCT.29 In addition, there is evidence that the incidence of TGCT in patients with acquired immunodeficiency syndrome is higher than in the general male population.30 –33 Recently, a high-resolution genotyping study comprising 55 Japanese TGCT patients showed an HLA-DRB1 susceptibility allele (RR 3.26) and an HLADQB1 candidate protective allele (RR 0.26) for TGCT.19 This study suggested that one of the genetic factors involved in TGCT development may be associated with HLA, particularly the HLA Class II region. In ¨ zdemir et al.,19 Oliver34 found an associaddition to O ation between DR5 and seminoma and an increase of DR7 in patients with Stage IV disease (extralymphatic metastasis). Aiginger et al.35 pooled their data with those from Oliver et al.’s study (total ⫽ 233 patients) and found a significant increased frequency of the HLA antigens DR1 and DR5 in seminoma patients. The current extensive genotyping study of a large number of Dutch TGCT patients, however, fails to confirm the associations of HLA Class II genes with susceptibility to TGCT. A previous study showed geographic clustering of testicular carcinoma in the northern part of The Netherlands.36 The majority of patients participating in this study descend from this area. In addition to possible common environmental factors, this population is likely to share a relatively high frequency of mutations in genes involved in TGCT from recent common ancestors. However, for the HLA Class II region analyzed in this study, no difference between patients and controls is observed using both standard methods (association and TDT analyses) and HSS, even though HSS has the power due to the presence of strong linkage disequilibrium (Fig. 2). Strong linkage disequilibrium suggests that only a few different haplotypes are present in the data. Similar haplotypes are, therefore, likely to be identical by descent, i.e., inherited from a common ancestor. Because this similarity will be centered around the disease locus for patients and random over the entire region among controls, differences due to disease mutation in this region would have been revealed by HSS. The linkage disequilibrium observed in this study is in agreement with the results of former studies that have reported regions within the HLA region that show very
1862
CANCER November 1, 2002 / Volume 95 / Number 9
little recombination (e.g., between HLA-DR and DQ) and regions where recombination preferentially occurs.37 ¨ zdemir et al.19 was the first HLA The study by O genotyping study of TGCT patients. The HLA alleles for which associations were reported are also preva¨ zdemir et lent in the Dutch populations.38 However, O al. genotyped only 55 patients whereas the current study genotyped almost three times that number, resulting in more power, particularly for nonseminomas.19 Moreover, with HSS, we have additional power from phase information to detect differences due to genetic factors in the HLA region. Systematic analysis of patient and control haplotypes using a high-density genome screen can identify identity by descent within haplotypes of unrelated patients in a founder population. This is because haplotypes that are similar for many consecutive marker alleles are very likely to be inherited from common ancestors. The shared segment of the haplotypes that is most significantly overrepresented in patients compared with controls is likely to contain a predisposing gene. The previously suggested association between HLA and TGCT need not result from a functional role of the HLA system itself, but may also be due to an effect from a separate gene that is closely linked with HLA loci. The analysis of haplotypes where most genetic variation is associated with specific haplotypes enables conclusions about all genes present on the haplotypes. Therefore, based on the results in the current study, a role of HLA Class II genes in the development of TGCT seems much more limited than previously suggested by a few studies. In previous studies by Oliver et al.34 and Aiginger et al.,35 the association between HLA and TGCT was found by serotyping methods that are less precise and less efficient than genotyping methods. DNA typing is more sensitive and identifies more alleles. The reason for this is that the variation present in the HLA region is specific for the small number of founders. All variation that contributes to disease can be identified with haplotype association methods. DNA-based methods are more accurate and allow higher-definition HLA typing. High-resolution genotyping is the method of choice because the polymorphism of the peptide binding domain of MHC Class II molecules is more precisely determined by genotypes than by serotypes.39 In conclusion, although there are proven associations between HLA and several malignancies, this is not the case for TGCT. The current genotyping study did not confirm the previous reported association between HLA Class II genes and TGCT, despite a larger sample size, especially nonseminomas. As the HLA
alleles for which associations were reported are also prevalent in the Dutch populations, these associations are likely nonexistent or much weaker than reported. Further research focusing on other candidate loci should be performed to identify possible TGCT susceptibility genes.
REFERENCES 1. 2.
3.
4.
5.
6. 7.
8.
9.
10. 11.
12. 13.
14.
15. 16.
17.
18.
19.
Bosl GJ, Motzer RJ. Testicular germ-cell cancer. N Engl J Med. 1997;337:242–253. Van Basten JP, Schraffordt Koops H, Sleijfer DTh, Pras E, van Driel MF, Hoekstra HJ. Current concepts about testicular cancer. Eur J Surg Oncol. 1997;23:354 –360. Dieckmann KP, Pichlmeier U. The prevalence of familial testicular cancer: an analysis of two patient populations and a review of the literature. Cancer. 1997;80:1954 –1960. Dong C, Lonnstedt I, Hemminki K. Familial testicular cancer and second primary cancers in testicular cancer patients by histological type. Eur J Cancer. 2001;37:1878 –1885. Sonneveld DJA, Sleijfer DTh, Schraffordt Koops H, et al. Familial testicular cancer in a single-centre population. Eur J Cancer. 1999;35:1368 –1373. Heimdal K, Fossa SD. Genetic factors in malignant germ-cell tumors. World J Urol. 1994;12:178 –181. Savage MO, Lowe DG. Gonadal neoplasia and abnormal sexual differentiation. Clin Endocrinol (Oxf). 1990;32:519 – 533. Tollerud DJ, Blattner WA, Fraser MC, et al. Familial testicular cancer and urogenital developmental anomalies. Cancer. 1985;55:1849 –1854. Akre O, Lipworth L, Tretli S, et al. Epstein-Barr virus and cytomegalovirus in relation to testicular-cancer risk: a nested case-control study. Int J Cancer. 1999;82:1–5. Mills PK, Newell GR, Johnson DE. Familial patterns of testicular cancer. Urology. 1984;24:1–7. Newell GR, Mills PK, Johnson DE. Epidemiologic comparison of cancer of the testis and Hodgkin’s disease among young males. Cancer. 1984;54:1117–1123. Senturia YD. The epidemiology of testicular cancer. Br J Urol. 1987;60:285–291. Chaganti RS, Houldsworth J. Genetics and biology of adult human male germ cell tumors. Cancer Res. 2000;60:1475– 1482. Leahy MG, Tonks S, Moses JH, et al. Candidate regions for a testicular cancer susceptibility gene. Hum Mol Genet. 1995; 4:1551–1555. Murty VV, Chaganti RS. A genetic perspective of male germ cell tumors. Semin Oncol. 1998;25:133–144. Rapley EA, Crockford GP, Teare D, et al. Localization to Xq27 of a susceptibility gene for testicular germ-cell tumours. Nat Genet. 2000;24:197–200. Skotheim RI, Kraggerud SM, Fossa SD, et al. Familial/bilateral and sporadic testicular germ cell tumors show frequent genetic changes at loci with suggestive linkage evidence. Neoplasia. 2001;3:196 –203. Candidate regions for testicular cancer susceptibility genes. The International Testicular Cancer Linkage Consortium. APMIS. 1998;106:64 –70. ¨ zdemir E, Kakehi Y, Mishina M, et al. High-resolution O HLA-DRB1 and DQB1 genotyping in Japanese patients with testicular germ cell carcinoma. Br J Cancer. 1997;76:1348 – 1352.
Testicular Carcinoma/Sonneveld et al. 20. Bodmer WF. The HLA system: structure and function. J Clin Pathol. 1987;40:948 –958. 21. Bateman AC, Howell WM. Human leukocyte antigens and cancer: is it in our genes? J Pathol. 1999;188:231–236. 22. Complete sequence and gene map of a human major histocompatibility complex. The MHC sequencing consortium. Nature. 1999;401:921–923. 23. Boon M, Nolte IM, Bruinenberg M, et al. Mapping of a susceptibility gene for multiple sclerosis to the 51 kb interval between G511525 and D6S1666 using a new method of haplotype sharing analysis. Neurogenetics. 2001;3:221–230. 24. Foissac A, Crouau-Roy B, Faure S, Thomsen M, CambonThomsen A. Microsatellites in the HLA region: on overview. Tissue Antigens. 1997;49:197–214. 25. Beckmann L, Fischer C, Deck KG, Nolte IM, te Meerman GJ, Chang-Claude J. Exploring haplotype sharing methods in general and isolated populations to detect gene(s) of a complex genetic trait. Genet Epidemiol. 2001;21 (Suppl 1):S554–S559. 26. Dorum A, Moller P, Kamsteeg EJ, et al. A BRCA1 founder mutation, identified with haplotype analysis, allowing genotype/phenotype determination and predictive testing. Eur J Cancer. 1997;33:2390 –2392. 27. Levinson DF, Kirby A, Slepner S, Nolte I, Spijker GT, te Meerman GJ. Simulation studies of detection of a complex disease in a partially isolated population. Am J Med Genet. 2001;105:65–70. 28. Spielman RS, McGinnis RE, Ewens WJ. Transmission test for linkage disequilibrium: the insulin gene region and insulindependent diabetes mellitus (IDDM). Am J Hum Genet. 1993;52:506 –516. 29. Birkeland SA, Storm HH, Lamm LU, et al. Cancer risk after
30.
31. 32.
33.
34.
35. 36.
37. 38. 39.
1863
renal transplantation in the Nordic countries, 1964-1986. Int J Cancer. 1995;60:183–189. Bernardi D, Salvioni R, Vaccher E, et al. Testicular germ cell tumors and human immunodeficiency virus infection: a report of 26 cases. Italian Cooperative Group on AIDS and Tumors. J Clin Oncol. 1995;13:2705–2711. Calista D. Five cases of melanoma in HIV positive patients. Eur J Dermatol. 2001;11:446 – 449. Hentrich MU, Brack NG, Schmid P, Schuster T, Clemm C, Hartenstein RC. Testicular germ cell tumors in patients with human immunodeficiency virus infection. Cancer. 1996;77: 2109 –2116. Timmerman JM, Northfelt DW, Small EJ. Malignant germ cell tumors in men infected with the human immunodeficiency virus: natural history and results of therapy. J Clin Oncol. 1995;13:1391–1397. Oliver RT, Stephenson CA, Parkinson MC, et al. Germ cell tumours of the testicle as a model of MHC influence on human malignancy. Lancet. 1986;1:1506. Aiginger P, Kuzmits R, Kratzig C, et al. HLA antigens and germ-cell tumours. Lancet. 1987;1:276 –277. Sonneveld DJA, Schaapveld M, Sleijfer DTh, et al. Geographic clustering of testicular cancer incidence in the northern part of The Netherlands. Br J Cancer. 1999;81:1262–1267. Carrington M. Recombination within the human MHC. Immunol Rev. 1999;167:245–256. Roychoudhury AK, Nei M. Human polymorphic gens world distribution. New York: Oxford University Press, 1988:234–239. Corzo D, Salazar M, Granja CB, Yunis EJ. Advances in HLA genetics. Exp Clin Immunogenet. 1995;12:156 –170.
