In an effort to reassess the contribution of HLA-DRB1 polymorphisms to inter-individual variations of human longevity, we have compared their genotypic distributions between longevous and adult control groups in the French population. The longevous groups included two independent cohorts totalling 533 centenarians, and 163 nonagenarian siblings. Allelic distributions were significantly different between controls and longevous groups. Three individual alleles were mostly responsible for these differences: DR7, DR11 and DR13. Multivariate logistic analyses were performed in order to sort out interactions between gender- and age-specific genetic effects. DR7 frequency was elevated in longevous men, in centenarians as well as nonagenarian siblings [OR = 1.72 (1.2-2.5)]. DR11's influence on longevity displayed a significant interaction with sex, with an increase in women from longevous sibships [OR = 2.03 (1.4-3.0)]. DR13's frequency was increased in centenarians of both genders [OR = 1.46 (1.2-1.75)]. These results are discussed in the context of other pathophysiological effects of the implicated alleles. Our data support the direct involvement of three HLA-DR alleles in survival at very old ages. Two allele-specific effects on longevity appear to depend on gender and one on familial status for aggregation of this trait. The latter is an original finding for humans.
Evidence has been accumulating rapidly in support of the long-believed contention that there is a genetic component to inter-individual differences in human longevity (1,2). Genetic effects on human longevity were found at a few loci (3-5). The HLA locus, as the cradle of genetic polymorphism in humans, has been involved in associations with many diseases (6). MHC genotypes have strong influences on lifespan in mice (7,8). In the last 25 years, a fair number of studies have searched for the impact of polymorphic HLA genes on human longevity (9-16). These studies have yielded conflicting results so that, in spite of the impressive quantity of data generated on this issue, no clear consensus has been reached as to the role of genetic variation at the HLA locus in longevity. Major methodological problems account for the confusion in this field: (i) sample sizes are often insufficient to reach statistically significant conclusions considering the large number of alleles examined; (ii) age cut-offs for inclusion differ from study to study and may not be old enough for the purpose of finding genetic effects on survival; (iii) sex is not always taken into account; (iv) matching between the old and the control populations is not always checked; and (v) the loci investigated within the HLA region differ from study to study.
Drawing examples from three studies that reported no effect of HLA loci on survival, one examined 182 individuals aged 70-93 years with no mention of sex (10), another included 18 males and 41 females >80 years old, of whom three males and seven females were >90 years (11), and a third examined 228 subjects aged 58-86 years with a median age of 68 and no mention of sex (14). Proust et al. looked at HLA-A, -B, -C and -DR specificities in 55 male and 100 female nonagenarians from France with mean ages >93 (12). They found that certain HLA-C alleles were increased in either male or female nonagenarians and that a haplotype bearing DR3 was increased in male nonagenarians; no difference could be detected without taking sex into account. Takata et al. looked at HLA-A, -B, -C, -DR and -DQ specificities in 82 centenarians and 20 nonagenarians from Japan comprising 22 males and 80 females (15). They found a decrease of DR9 and an increase of DR1 alleles in longevous individuals. Lagaay et al. included 964 subjects >85 years old from Leiden, comprising 278 males and 686 females in what remains the largest-scale study to this date (16). These were compared with a control group, composed of 2444 young adults, for HLA-A, -B, -C, -DR and -DQ specificities. The only emerging results were a decrease in B40 and an increase in DR5 in the 85+ women. Altogether, it appears that effects on survival beyond age 90 of genetic variation at the HLA-DR locus were found whenever sex was taken into account, whereas they went undetected when sex was not considered. It may be relevant that in mice, where MHC effects on survival were first documented (7,17), MHC haplotypes have strong influences on fertility, fecundity, anatomy of reproductive organs and reproductive senescence (18). Therefore, there were strong grounds a priori to discriminate according to sex in the analysis.
In this study, we present data obtained by genotyping at the DNA level, gathered on two longevous populations selected according to different criteria: centenarians and long-lived siblings.
A total of 325 centenarians (VIE group, individuals in their 100th year or beyond) and 229 nonagenarian siblings (SIB group, see Materials and Methods for inclusion criteria) from the Chronos Collection were genotyped for HLA-DRB1 alleles. The following work bears on the 13 generic DRB1 alleles, hereafter simply named after the generic DR specificities, e.g. DR3. Genotypes from adult blood donors (aged 20-60) were gathered from seven regions of France, totalling 2950 controls. Allele distributionsfor the different populations are given in Table 1. Although fluctuations of certain allele frequencies do occur in several regions, there was no significant overall heterogeneity among the distributions in control populations (P > 0.4 for a [chi]2 test, df = 84). Therefore, genotypic data for controls were pooled in all subsequent calculations. In contrast, a comparison between centenarians and pooled controls revealed a significant heterogeneity in the HLA-DR allele distributions (P < 0.05, [chi]2 = 24 for df = 12). Four sets of longevous individuals were therefore distinguished: male or female centenarians (VIE), male or female nonagenarian siblings (SIB).
There are no sex-related differences in HLA-DR allele frequencies in control populations, a fact well established and directly verified here for the Paris and Strasbourg groups (P > 0.7 in a [chi]2 test comparing the distributions between genders in a group totalling 1302 individuals). The allele frequency distributions for five groups, comprising the controls and four longevous groups, are shown in Table 2. A [chi]2 test with the corresponding contingency table detected a significant disparity in the allelic distributions (P < 0.005, [chi]2 = 80.1 for df = 48). The three largest contributions to the [chi]2 come from DR11, DR7 and DR13 alleles, respectively. Indeed, striking differences appear at first sight. DR7 allele frequency was 14% in controls, as compared with 21.2% and 23.1% in VIE and SIB males, respectively, which is higher than in any regional control group. DR11 allele frequency, as compared with 12.8% in controls, was 23.4% in female SIB, which is also higher than in any region of France. Finally, DR13 allele frequency was 12.6% in controls as compared with 17.4% in male and female centenarians, once again a higher frequency than in any of the regions. Comparisons of DR7, DR11 and DR13 frequencies among controls and longevous groups are visualized in the histograms of Figure 1. These three alleles seemed to display diverse influences on longevity: whereas DR7 was increased in long-lived men, whether isolated centenarians or members of sibships, DR11 was increased only in long-lived women belonging to sibships and DR13 was increased in centenarians of both genders. There is, within the SIB group, a subset of individuals who have also become centenarians: 16 men and 50 women. The frequencies of DR7 in male SIB-centenarians and of DR11 in female SIB-centenarians are, respectively, 25% and 22%, which strengthens the group specificity of these alleles' influences on longevity.
Table 1
Table 2
Table 3
In order to estimate the statistical significance of the allelic associations, taking into account putative interactions between gender and specific alleles in the effect on survival, we applied a model of logistic regression. Computations were done for each allele, comparing controls with either VIE or SIB groups. Allele frequencies were assumed to be identical in male or female adult controls. Results are presented in Table 4 for the seven most common alleles, including the only three comparisons-out of 13-that reached statistical significance, namely for alleles DR7, DR11 and DR13. For DR7, the trend of sex-allele interaction due to higher frequencies in long-lived males, whether from the VIE or SIB group, did not reach significance. Male and female SIB were therefore pooled, yielding an OR of 1.50 (P < 0.01) in this group for DR7 carriers. There was a significant sex-allele interaction for DR11 in the SIB group (OR = 0.43 for men versus women, P < 0.01) and the OR for SIB women was 2.02 (P < 0.001; 95% CI: 1.38-2.9). DR13 frequency was equally elevated in male and female centenarians, yielding an OR of 1.44 (P < 0.01).
One year after obtaining the above data, our collection of centenarians had grown and we decided to evaluate the previously identified associations in a new cohort. Therefore, 208 more centenarians were genotyped, 185 women and 23 men. The allelic distributions were again significantly different between controls and this second cohort of centenarians (P < 0.05, [chi]2 = 22 for df = 12), whereas the distributions were similar in the two cohorts of centenarians. The male-specific increase in DR7 was again observed (23.9%) and almost reached significance, despite the small number of male centenarians. The increase in DR13 was confirmed too (17.8%, P < 0.05 for the allele-specific comparison between controls and this new cohort). Since the previously identified distribution pattern in centenarians was conserved, we felt entitled to merge the two cohorts into one population comprising 533 centenarians, 89 men and 444 women. Table 5 displays the overall allele frequencies. The logistic regression model applied to these data provides a better estimate of ORs for centenarians. The interaction with sex was now detected as significant for DR7, with an OR of 1.72 (95% CI: 1.19-2.5, P = 0.004) for male centenarians. DR13 displayed no sex interaction but was significantly increased with an OR of 1.46 (95% CI: 1.22-1.75, P < 0.0001) for centenarians. Interestingly, the overall allele distributions do not differ significantly between male and female centenarians (P > 0.3).
Table 4
Table 5
This study presents the largest sample of longevous individuals ever genotyped for HLA-DR alleles, totalling 533 centenarians and 229 nonagenarians. Our selection criteria are original, in that they are designed to allow differentiation between cases of isolated longevity in one generation, represented by individual centenarians, and cases where there is evidence of familial aggregation in the same generation, represented by longevous siblings. The inclusion threshold of 95 years for women and 90 years for men in sibships was selected to balance the gender differential in longevity-there are roughly as many women over 95 as there are men over 90 in France. Within the longevous populations, four groups were further distinguished, according to sex and familial status, since (i) there were a priori grounds to expect sex-related influences; and (ii) centenarians and nonagenarian sibling-pairs had been recruited following distinct criteria. For the control group, we gathered genotypic data generated at seven different immunology centres corresponding to seven regions of France, totalling a population of 2950 adults. In spite of occasional regional fluctuations in certain allele frequencies, there was no significant heterogeneity within the control populations. In contrast, a significant heterogeneity among the HLA-DR allele distributions appeared among the five populations including the controls and longevous groups. This provided a first indication that some HLA-DR alleles might affect longevity.
Therefore, in order to assess the contributions of individual alleles to longevity and their putative interaction with sex in this contribution, a log-linear model was implemented comparing either centenarians or nonagenarian siblings with adult controls for each allele. The allele-specific effects reached statistical significance for three alleles, namely DR7, DR11 and DR13, which made the largest contributions to the [chi]2 test for heterogeneity. The interaction with sex was significant for DR7 (elevated in longevous men, whether centenarians or siblings) and for DR11 (higher in women from sibships, unchanged in women centenarians). DR13 showed no interaction with sex, as its frequency was increased in centenarians of both genders.
We found that homozygotes were significantly more frequent in centenarians of both genders than in controls and that they also exceeded the proportion expected under the assumption of Hardy-Weinberg equilibrium: these results suggest the existence of a homozygous advantage at the HLA-DR locus, at least very late in life. Data have been apparently conflicting concerning the effect of heterozygosity at the HLA locus on survival. Several studies have reported a survival advantage for heterozygotes (9,13,19,20), while others have not confirmed this observation (10,11,14,16). All reports of a heterozygous advantage bear on HLA-B or HLA-A alleles. One exception is Dorak et al. (21), who found that a haplotype between a DQ gene in the class II region and the HSP70 gene in the class III region progressively disappears with age in males. We are not aware of published data mentioning an increased heterozygosity at the HLA-DR locus in association with ageing. On the contrary, Takata et al. noted an increase of HLA-DR homozygotes in their centenarians (15). It has been remarked that a bias for homozygosity may stem from imperfect matching between centenarians and controls, the former coming from a more rural background and a less admixed population. Although this may be the case, we note that the rate of homozygosity was not increased in the nonagenarian SIB group, for which the same argument would apply.
There is plentiful evidence for a direct role of HLA-DR alleles in differential susceptibility to a vast array of pathologies, in particular autoimmune diseases. The most classic examples include the association of DR4 and DR1 alleles with rheumatoid arthritis and of DR3 with systemic lupus erythematosus (22). The fact that these associations are found in a range of diverse ethnic groups, notwithstanding the various specific haplotypic combinations of polymorphic alleles extending over the whole region, strongly argues for a causal role of the implicated HLA-DR alleles. The penetrance of HLA-DR genotypes in disease phenotypes is commonly found to depend on sex (23). The influence of HLA-DR alleles on insulin-dependent diabetes mellitus (IDDM) is age-dependent (24). Furthermore, functional in vitro studies along with site-directed mutagenesis and sequencing have allowed in some instances the definition of the point mutations responsible for the different properties of certain alleles or subtypes (25,26). Among the numerous hypothetical mechanisms invoked to explain HLA class II disease associations, recent work has shown that signalling through HLA class II molecules mediates pathways leading to apoptosis (27).
In view of the implication of DR3 and DR4 alleles in autoimmune diseases, it is interesting that we did not find a significant effect of these alleles on longevity. DR3 has been proposed as a `male longevity factor', especially within the A1B8DR3 haplotype (12,28)-otherwise associated with autoimmune disorders. A clue to this apparent paradox may come from our finding that DR3 is increased in longevous males from sibships (13.4%) whereas it is decreased in male centenarians (6.7%) relative to controls (11.0%). Very advanced age (>99) and status of familial aggregation regarding longevity had not been discriminated for in previous studies.
We found a significant increase of DR7 in longevous men. DR7 appears to be a risk factor for CMV infection, either in renal transplant recipients (29) or in AIDS patients (30), and for chronic HBV infection (31). DR7 has been associated with the long-QT (LQT) syndrome, a polygenic electrocardiographic disorder characterized by syncope and fatal ventricular arrhythmias (32). While women are more affected than men by the LQT syndrome, the association was found only in men. Intriguingly, DR7 was positively and DR2 negatively associated with such disrelated diseases as LQT syndrome and HBV infection. HLA-DR7 has also been found as an additional risk for celiac disease in three independent populations from Spain, France and Ireland (33-35). Of note, though, celiac disease affects mainly women and, more importantly, the primary association comes from HLA-DQ rather than HLA-DR alleles (36). However, the fact that DR7 emerged in yet another population-from Germany-as the only class II allele significantly associated with Crohn's disease (37), a pathology of the digestive tract like celiac disease, adds support to a direct involvement of this allele. Finally, DR7 has been associated with a chronic B cell lymphocytosis in French women (38). HLA-DR7 has proven `protective' in two instances: against gold salt toxicity in rheumatoid arthritis treatment (39) and against disease progression in psoriatic arthritis (40).
We found a significant increase of HLA-DR11 in women belonging to long-lived sibships, but not in centenarians. Most relevant to this result is the finding of Lagaay et al. that the association with DR5 in longevous women was entirely borne by DR11, when DR5 was split into its subtypes DR11 and DR12 (16). Their population from Leiden included 681 women older than 85. HLA-DR11 has been associated with Graves' disease only in Caucasian American males (41) and with thyroid cancer in a Spanish population (42). The prevalence of thyroid disorders is higher in females and the latter study did not mention the sex of their patients. In this regard, it is interesting that centenarians were found to have lower levels of thyroid auto-antibodies than elderly subjects in the 70-85 years range (43), which could be a clue to underlying genetic factors. DR11 has been associated with alopaecia areata (44) and with hairy cell leukaemia (45), which affects mostly men above 40. The only suspected protective effect of HLA-DR11 is in IDDM, since a decrease of this allele has been reported in two independent populations of patients, from Sardinia and Belgium (46,47).
HLA-DR13 is the one allele that we found equally associated with longevity in both male and female centenarians. DR13 has been identified as protective against two major infectious diseases in West Africans: malaria parasitic and HBV infections (48,49). Functional data on peptide binding specificities suggest a molecular basis for these associations (50). DR13 confers a greater resistance to HIV infection in newborns from different ethnic groups (51). Negative associations of DR13 with cervical carcinoma have been reported in Hispanic and French women (52,53). DR13 was also negatively associated with mutiple sclerosis in Iceland (54). Therefore, this allele appears as protective in the context of both infectious and autoimmune diseases as well as neoplasia. Our findings are in agreement with these data, and suggest that HLA-DR13 imparts a survival advantage at very advanced ages.
Taken together, these data provide compelling evidence for direct influences on survival of three HLA-DR variants, taking sex-specific effects into account. Common features may be observed concerning the two variants that display sex-dependent effects: they are associated with sex-specific pathologies, in particular with lymphoproliferative diseases. Their associations with longevity may provide interesting examples of antagonistic pleiotropy (55), whereby the same allele has negative and positive effects on survival, the latter taking the edge very late in life. It is anecdotal that the oldest male centenarian in our cohort, who lived to 111 years, was homozygous for DR7.
Differences between the long-lived groups VIE and SIB are summarized in Table 6. Three alternative interpretations of these differences are possible: (i) they indicate different genetic components for familial and individual longevity; (ii) they merely reflect sampling fluctuations; or (iii) they arise from the age difference between the two groups. The third alternative would imply that an astonishingly strong selection on HLA-DR alleles occurs between ages 95 and 100. It is all the more unlikely that in the subset of SIB-centenarians, DR11 frequency was 22% in 50 women (versus 12% in female VIE) and DR3 frequency was 15.6% in 16 men (versus 6.7% in male VIE). The second alternative cannot be dismissed until an independent SIB cohort is examined. If the familial effect holds true, it may even be underestimated here for want of reliable information on the families of most centenarians in the VIE group. For example, our oldest centenarian, subject J.C. who lived to the ripe age of 122, was DR11/DR15. A genealogic study of her ancestry revealed a highly significant longevity trend up to the tenth generation and she had a brother who died at the age of 97 [(56), and unpublished data]. In fact, she would have qualified as a SIB-centenarian, long before she was included in our study. Our results may allow, for the first time, to distinguish between individual and familial longevity. In this respect, large-scale genealogic studies have demonstrated a correlation between fertility and longevity, and shown that longevity is more correlated among siblings of the same sex, especially female siblings (57,58).
Walford and other gerontologists view the whole MHC as a region involved in the control of mammalian longevity (59,60). This view is based on substantial work in mice, showing the differential longevity of congenic mice differing at the MHC locus (7,17). The MHC region displays unique features: it contains one of the highest density of genes in the genome, packing >70 functional genes within ~4 Mb (61); several of these genes present the highest degree of polymorphism; yet, extended haplotypes are so well conserved that they are used in anthropological history.
Going one step further, it may be seen that this exceptional region contains genes involved in several aspects of the life cycle: fertility, endocrine controls-and lifespan (62). Evolutionary considerations lead us to think that this clustering is not merely coincidental and that one may indeed expect epistasis between closely linked variants on life cycle traits, including lifespan (63). So the direct impact of some HLA-DR alleles on longevity does not preclude other such genetic effects to be unravelled in this region. On the contrary: the challenge, typical of HLA associations, will be to distinguish primary effects from secondary associations due to linkage disequilibrium (64,65). Other loci are currently under investigation in the HLA region. The involvement of HLA in longevity may well turn out to be exemplary, in complexity as well as instructiveness.
Centenarians were individuals in their 100th year recruited in the French population on the mere basis of chronological age: >99 years on the day of blood collection (3) (mean age: 100.9 ± 0.35, 101.25 ± 0.21, 101.2 ± 0.18 for male, female and overall groups, respectively). These figures are the average for the complete cohort of 533 centenarians: the first cohort of 325 centenarians comprises samples collected from 1991 to 1994, while samples from the second cohort of 208 centenarians were collected later. Longevous siblings were recruited according to our set criterion for familial longevity: at least two longevous members in the sibship, i.e. >90 years old for men, >95 years old for women (mean age: 93.4 ± 0.78, 95.7 ± 1.05 and 94.7 ± 0.69 for male, female and overall groups, respectively). The 229 nonagenarians that compose our SIB group belong to 127 different sibships. Control adults had donated blood at the different regional immunology centres throughout the country and genotypic data were kindly provided by these centres. HLA-DR allele distributions have been published for the Marseille (66) and Bretagne (67) populations.
Table 6
HLA-DR genotyping of the DNA for longevous groups was realized by reverse dot-blot with the Inno-Lipa DRB1 kit from Murex Diagnostics SA 68).
The hypothesis that allelic distributions were heterogeneous among the different populations of controls was tested by a [chi]2 test. The same test was used to globally compare allelic distributions between the adult controls, VIE and SIB groups. Then the comparison of each allele frequency in VIE versus controls and SIB versus controls was performed by multivariate logistic regression. In addition to the variable coding for the presence or absence of the allele under study, the gender was included in the multivariate logistic regression model since it was a priori considered to be relevant, as well as the interaction between the two variables. The significance of the interaction was first tested by examining the changes in the deviation when interaction was added in the logistic model. When the interaction was significant, we analysed the effect of the studied allele and computed ORs separately for each gender. When the interaction did not reach significance, the analysis was done globally on both genders. Analyses were made either by a maximum likelihood algorithm or by exact Monte-Carlo methods when necessary (LogXact software, Cambridge). Since multiple comparisons were made, the Bonferroni corrections to maintain the overall type I error threshold at 0.05 would require that P-values be multiplied by the number of comparisons. Considering that 13 alleles were studied and that we compared controls versus centenarians and versus SIBs, a conservative stand would be to conclude that a difference is significant for P-values < 0.05/26 = 0.002. In the second study with a new cohort, however, we were justified in examining a priori the alleles DR7 and DR13, so no Bonferroni correction would be required.
We thank D. Alcalay, J. Delobel, M. Drouet, J. Seignalet and M. Tongio for providing genotypic data on their control populations. We thank C. Billon, L. Cazes, F. Dufour and M. Legrand for technical assistance with cell cultures and DNA extraction. We are grateful to J. Dausset and C. Finch for encouragement and comments. R.I. was supported by a grant from the `Fondation des Amis des Sciences' and a `Poste Vert' fellowship from INSERM.
Human Molecular Genetics Pages
Introduction
Results
Control and long-lived populations defined
Overall distributions and allelic associations
Homozygosity in centenarians
Interaction between sex and HLA-DR alleles in longevity
Second cohort and conclusions
Discussion
Materials And Methods
Subjects
Genotyping
Statistical methods
Acknowledgements
References
Allele
Paris
n = 509Amiens
n = 43Limoges
n = 64Poitiers
n = 558Montpellier
n = 300Marseille
n = 108Bretagne
n = 150Strasbourg
n = 1218Total
n = 2950
DR1
12.7
11.6
12.7
11
10.8
7.4
12.7
11.5
11.5
DR3
11.7
16.3
11.9
10
10.2
13.3
11.7
10.8
11.0
DR4
11.7
15.1
8.7
13.4
16.8
11.7
19.7
12.5
13.3
DR7
12.8
8.1
14.3
14
16.8
13.3
15.3
13.9
14.0
DR8
2.65
1.2
6.3
2.2
1.8
3.9
3
3.1
2.8
DR9
0.3
0
3.2
0.7
0.6
1.1
1
0.8
0.7
DR10
0.9
1.2
0.8
1
0.6
1.5
0.3
0.7
0.8
DR11
14.2
16.3
11.1
13.4
13.8
13.8
4.3
13.3
13.1
DR12
0.7
2.4
2.4
1.8
1.8
0.5
0
2
1.6
DR13
13.5
12.8
12.7
15
9.6
13.8
9.7
12.5
12.8
DR14
4
1.2
4
3.2
4.2
5.8
3.7
3.9
3.8
DR15
12
5.8
8.7
10.7
12.0
10.8
18.3
13.2
12.4
DR16
3
7
3.2
3.4
1.2
3.2
0.3
1.7
2.3
Allele
Controls
n = 2950Female centenarians
n = 259Male centenarians
n = 66Female siblings
n = 77aMale siblings
n = 86b
DR1
11.5
9.1
14.4
9.1
8.0
DR3
11.0
10.8
8.3
11.0
13.4
DR4
13.3
13.1
12.9
10.7
10.4
DR7
14.0
15.0
21.2
15.6
23.1
DR8
2.8
3.3
1.5
0.6
4.1
DR9
0.7
1.0
0.8
0.0
1.7
DR10
0.8
0.8
0.8
1.6
1.2
DR11
13.1
11.6
6.8
23.4
11.7
DR12
1.6
1.3
0.0
0.6
0.6
DR13
12.8
17.4
17.4
13.6
11.0
DR14
3.8
3.7
1.5
1.0
3.2
DR15
12.4
9.5
10.6
9.2
8.4
DR16
2.3
3.5
3.8
3.5
3.2
Controls Paris
n = 509Controls Strasbourg
n = 1218Pooled controls
n = 1727Centenarians
n = 325Female centenarians
n = 259Male centenarians
n = 66
Homozygotes (observed)
56 (11%)
136 (11.2%)
192 (11.1%)
53 (16.3%)
41 (15.8%)
12 (18.2%)
Homozygotes (expected)
59.1 (11.6%)
138.8 (11.4%)
197.8 (11.5%)
39.2 (12.1%)
30.1 (11.6%)
9.1 (13.8%)
HLA-DRB1 allele
Centenarians versus controls Sex-genotype interaction, males versus females
OR, centenarians versus controls
Siblings versus controls Sex-genotype interaction, males versus females
OR, siblings versus controls
DR1
1.68 (0.93-3.05, P = 0.08)
0.87 (0.67-1.14, P = 0.3)
0.9 (0.4-1.96, P = 0.76)
0.7 (0.48-1.06, P = 0.33)
DR3
0.75 (0.37-1.5, P = 0.4)
0.93 (0.71-1.21, P = 0.6)
1.24 (0.62-2.46, P = 0.55)
1.13 (0.8-1.6, P = 0.5)
DR4
Not done
0.98 (0.78-1.26, P = 0.93)
0.95 (0.46-1.94, P = 0.88)
0.79 (0.55-1.13, P = 0.2)
DR7
1.52 (0.92-2.5, P = 0.1)
1.22 (0.97-1.52, P = 0.09)
1.64 (0.92-2.9, P = 0.09)
1.5a (1.14-2.0, P = 0.0044)
DR11
0.56 (0.27-1.18, P = 0.13)
0.78 (0.6-1.02, P = 0.07)
0.43 (0.23-0.8, P = 0.0075)
2.03b (1.4-3.0, P = 0.0003)
DR13
1.0 (0.6-1.7, P = 0.99)
1.43c (1.15-1.8, P = 0.0015)
0.79 (0.4-1.55, P = 0.49)
0.96 (0.68-1.34, P = 0.8)
DR15
1.14 (0.59-2.2, P = 0.7)
0.76 (0.58-1.0, P = 0.05)
0.94 (0.43-2.07, P = 0.9)
0.69 (0.47-1.02, P = 0.06)
Allele
Controls
Centenarians
Male centenarians
Female centenarians
n = 2950
Cohort I + II n = 535
Cohort I + II n = 89
Cohort I + II n = 446
DR1
11.5
10.5
12.4
10.1
DR3
11.0
9.5
6.7
10.1
DR4
13.3
12.0
11.2
12.1
DR7
14.0
15.3
21.9
14.0
DR8
2.8
3.6
1.7
3.9
DR9
0.7
0.7
0.6
0.7
DR10
0.8
0.8
0.6
0.8
DR11
13.1
11.8
10.7
12.0
DR12
1.6
1.6
0.6
1.8
DR13
12.8
17.7
16.3
17.9
DR14
3.8
3.0
2.3
3.1
DR15
12.4
10.6
11.2
10.4
DR16
2.3
3.2
3.9
3.0
Homozygotes
192/1727
80/535a
14/89
66/446
11.1%
15%
15.7%
14.8%
VIE
SIB
DR3 decreased in males (6.7% versus 11%)
DR3 increased in males (13.4% versus 11%)
DR11 not affected in females (11.8% versus 13.1%)
DR11 increased in females (23.4% versus 13.1%)
DR13 increased in both genders (17.7% versus 12.8%)
DR13 not affected in any gender (12.2% versus 12.8%)
Increased homozygosity (15% versus 11.1%)
No difference in homozygosity (11.5% versus 11.1%)
REFERENCES
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 6 Jan 1998
Copyright© Oxford University Press, 1998.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  M. Kimura, M. Barbieri, J. P. Gardner, J. Skurnick, X. Cao, N. van Riel, M. R. Rizzo, G. Paoliso, and A. Aviv Leukocytes of exceptionally old persons display ultra-short telomeres Am J Physiol Regulatory Integrative Comp Physiol, December 1, 2007; 293(6): R2210 - R2217. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Q. Tan, A. I. Yashin, E. M. Bladbjerg, M. P.M. de Maat, K. Andersen-Ranberg, B. Jeune, K. Christensen, and J. W. Vaupel A Case-Only Approach for Assessing Gene by Sex Interaction in Human Longevity J. Gerontol. A Biol. Sci. Med. Sci., April 1, 2002; 57(4): B129 - 133. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Giraud, G. Beaurain, A. M. Yamamoto, B. Eymard, C. Tranchant, P. Gajdos, and H.-J. Garchon Linkage of HLA to myasthenia gravis and genetic heterogeneity depending on anti-titin antibodies Neurology, November 13, 2001; 57(9): 1555 - 1560. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Barzilai and A. R. Shuldiner Searching for Human Longevity Genes: The Future History of Gerontology in the Post-genomic Era J. Gerontol. A Biol. Sci. Med. Sci., February 1, 2001; 56(2): 83M - 87. [Abstract] [Full Text]
|
|
|
|
 |
|
 G. DE BENEDICTIS, G. ROSE, G. CARRIERI, M. DE LUCA, E. FALCONE, G. PASSARINO, M. BONAFÉ, D. MONTI, G. BAGGIO, S. BERTOLINI, et al. Mitochondrial DNA inherited variants are associated with successful aging and longevity in humans FASEB J, September 1, 1999; 13(12): 1532 - 1536. [Abstract] [Full Text]
|
|
|
|
 |
|
 H. Hendel, S. Caillat-Zucman, H. Lebuanec, M. Carrington, S. O'Brien, J.-M. Andrieu, F. Schachter, D. Zagury, J. Rappaport, C. Winkler, et al. New Class I and II HLA Alleles Strongly Associated with Opposite Patterns of Progression to AIDS J. Immunol., June 1, 1999; 162(11): 6942 - 6946. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||