Human Molecular Genetics, 2002, Vol. 11, No. 14 1585-1597
© 2002 Oxford University Press
Genetic susceptibility to childhood common acute lymphoblastic leukaemia is associated with polymorphic peptide-binding pocket profiles in HLA-DPB1*0201
1Immunogenetics Laboratory, St Mary's Hospital, Manchester M13 0JH, UK, 2Leukaemia Research Fund Centre at the Institute of Cancer Research, London, UK, 3Public Health Sciences, University of Edinburgh, Edinburgh, UK and 4Academic Department of Paediatric Oncology, Christie Hospital and Royal Manchester Children's Hospitals, Manchester, UK
Received November 4, 2001; Accepted April 22, 2002
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSUKCCS INVESTIGATORSREFERENCES |
|---|
In a previous study, we obtained preliminary evidence in a small series of patients (n=63) suggesting that susceptibility to childhood common acute lymphoblastic leukaemia (c-ALL) was associated with an allele at the HLA-DPB1 locus, DPB1*0201. We have now tested this hypothesis by comparing the frequency of children with leukaemia (n=982) who typed for specific DPB1 alleles and two groups of non-leukaemic children, one consisting of children with solid tumours, excluding lymphomas (n=409), the other consisting of normal infants (n=864). We found that significantly more children with c-ALL and T-ALL, but not pro-B ALL or acute non-ALL typed for DPB1*0201 as compared with children with solid tumours [odds ratio (OR), 95% confidence interval (CI) for c-ALL: 1.76, 1.20–2.56; T-ALL: 1.93, 1.01–3.80] and normal infants (OR, 95% CI for c-ALL: 1.83, 1.34–2.48; T-ALL: 2.00, 1.10–3.82). In childhood c-ALL, significantly more children than those with solid tumours or normal infants typed for DPB1 alleles coding specific polymorphic amino acids lining the antigen-binding site of the DPß1*0201 allotypic protein, suggesting that susceptibility to childhood c-ALL may be influenced by DPß ABS amino acid polymorphisms shared by DPß1*0201 and other DPß1 allotypes. These results point to a mechanism of c-ALL susceptibility that involves the presentation of specific antigenic peptides, possibly derived from infectious agents, by DPß1*0201-related allotypic proteins, leading to the activation of helper T cells mediating proliferative stress on preleukaemic cells.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSUKCCS INVESTIGATORSREFERENCES |
|---|
Leukaemia is the most common childhood malignant disease in populations with a high socio-economic status, such as the UK, Western Europe and the USA, where it accounts for approximately 30% of childhood cancers (1–6). Childhood leukaemia consists predominantly (85% of cases) of acute lymphoblastic leukaemia (ALL), with non-ALL (comprising mainly acute myeloid leukaemia, AML) accounting for the rest (2,4). Precursor B-cell or ‘common’ ALL (c-ALL) constitutes about 65–70% of childhood ALL (7) and is characterized by a unique peak of cases arising between 3 and 5 years of age (8–11). At least 90% of childhood acute leukaemias carry non-random, clonotypic chromosome rearrangements, including reciprocal translocations that give rise to novel fusion genes involved in leukaemogenesis (7). The observation that the MLL–AF4 and TEL–AML1 fusion genes may be present in the blood at birth of children who subsequently developed ALL with these gene rearrangements suggests that the initiation of leukaemia can, at least in some cases, take place prenatally (12–14). However, clonal evolution, characterized by complex karyotypic changes, plus the modest concordance rate for ALL in identical twins strongly suggest that additional genetic events, occurring mainly in the postnatal period, are required for progression to full malignancy (15).
Whilst evidence linking environmental carcinogen exposures to the aetiology of childhood leukaemia has been far from consistent or conclusive, support for the role of infection has been increasing (16–18). Two main mechanisms have been proposed for an infectious causation: an as-yet unidentified oncogenic virus (17) or the indirect effect of an abnormal immune response to one or more common (bacterial or viral) infections (18). Since no specific virus has yet been implicated by serological (19,20) or molecular screening of candidates (21–23), a model based on insertional mutagenesis seems less likely than one based on an indirect, proliferative stress-induced effect of infection on the developing immune system.
Epidemiological evidence of an infectious aetiology for ALL is indirect and incomplete, but nevertheless substantial and persuasive. It includes data showing a transient increase in leukaemia rates following several examples of population mobility and mixing (24–26), as well as increased risks associated with certain demographic features (27), seasonal presentation (28,29) and space–time clustering (30,31). Other evidence implicates a paucity of infections in the first year of life as leading to an increased risk of childhood ALL (32–34), and potentially protective effects of immunization with certain vaccines (34–36), allergies (37) and breast feeding (38) in infancy. These observations lend credence to the idea that lack of early programming of the developing immune system, followed by later exposure to common infections, could lead to infection-related, immune-mediated stress on a preleukaemic clone that is sufficient to progress to leukaemia (18).
In the continuing absence of direct evidence for the identity of a causative infection in childhood ALL, we have adopted a ‘reverse immunogenetic’ approach (39). This involves the detection of HLA class II alleles whose protein ``footprints'' provide functional markers of immune responses to antigens that may influence susceptibility to childhood ALL. The HLA class II (DR, DQ and DP) genes encode highly polymorphic cell surface glycoproteins that play a key role in adaptive immune responses to infection, and are known to influence the risk of certain autoimmune diseases with suspected infectious aetiologies (40,41). The impact of different HLA class II allotypes on disease susceptibility is thought to be mediated through interactions between polymorphic peptide-binding pockets formed by the 
and ß chains of each HLA class II allotypic protein with anchor residues on processed antigenic peptide ligands, leading to the directed activation of T-cell-receptor bearing CD4+ helper T (Th)-cell subsets (40,42). By inference, evidence of an increased frequency of a specific HLA class II allele in childhood ALL could therefore indicate a role for differential immune response regulation in response to one or more infection-derived peptides involved in the causation of childhood ALL, and may ultimately enable predictions to be made about their identity (40).
Previous attempts to determine the role of HLA-associated susceptibility in childhood leukaemia were hampered by ascertainment bias and low-resolution HLA-typing techniques. Substantial improvements in the survival of children with leukaemia now allow for more complete biological sample acquisition during the remission period, and avoid the potential that deleted sequences on chromosome 6p in leukaemia cells (43,44) may lead to incorrect HLA typing Furthermore, more detailed and standardized diagnostic information (7,11) and the development of high-resolution HLA molecular typing (45) now make it feasible to accurately predict the functional contribution of HLA polymorphisms to the aetiology of childhood leukaemia. In a previous pilot study using HLA molecular genotyping, we reported an increased frequency of the HLA-DPB1*0201 allele in children with c-ALL (relative risk 
2) (46). However, the small number of patients studied (n=63), and the highly polymorphic nature of the DPB1 gene did not preclude a chance association with DPB1*0201. In this paper, we now confirm and extend our preliminary finding by analysing a much larger and different series of childhood leukaemias in comparison with non-leukaemic subjects as part of the UK Childhood Cancer Study (47). The UKCCS is an epidemiological case–control study of childhood cancer aetiology that commenced in Scotland 1991 and in England and Wales in 1992. One aim of the UKCCS is to test the hypothesis that childhood ALL may arise as an abnormal response to delayed infection (16). Our results, presented here, suggest that certain peptide-binding pockets present in DPß1*0201, and also in other DPß1 allotypic proteins, contribute to ALL susceptibility. We interpret these results to suggest that qualitative and/or quantitative differences in the helper (Th) response to infection-derived peptides may be potentially important in the aetiology of childhood ALL.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSUKCCS INVESTIGATORSREFERENCES |
|---|
The UKCCS was designed to accrue at least 1000 cases of childhood ALL in order to provide sufficient power for aetiological analysis (47). As part of the UKCCS study, we collected remission blood samples from 982 children with acute leukaemia (ALL and non-ALL), 409 children with solid tumours (excluding non-Hodgkin's lymphoma and Hodgkin's lymphoma), and 864 normal infants born in Manchester, comprising a case-comparison series not included in our previous study (46). We extracted genomic DNA from blood samples, typed them for DPB1 alleles using a polymerase chain reaction–sequence-specific oligonucleotide probe (PCR–SSOP) technique (see Materials and Methods) (45,46) and determined the percentage of subjects (i.e. phenotype frequencies) in the leukaemia and comparison groups typing for specific DPB1 alleles.
As shown in Table 1, the leukaemia series included 875 children with ALL and 107 with non-lymphoid leukaemia (‘non-ALL’, i.e. mainly AML). Based on data obtained by immunophenotyping the pretreatment leukaemia cells (47), 559 of the children with ALL had a confirmed diagnosis of c-ALL, 19 had pro-B ALL, and 69 had T-ALL. For 228 children diagnosed morphologically as ALL, either an ALL subtype could not be assigned (168 children) or the immunophenotyping results were incomplete (60 children). These cases were included in the analysis of the total leukaemia and ALL series. The pretreatment leukaemia cells from 95 children with ALL were found to be positive for TEL–AML1 by fluorescence in situ hybridization (FISH) and/or RT–PCR, and 298 leukaemias were hyperdiploid (i.e. with >50 chromosomes) by FISH (47).
|
Comparison of the leukaemia and solid tumour patient characteristics with those included in the total interviewed UKCCS patient series (47) revealed no major differences in age distribution, sex or deprivation category, indicating that the DPB1 typed patients were representative of the total patient series.
In order to test the prior, but unconfirmed, hypothesis of an association between DPB1*0201 and childhood ALL (46), we compared the percentage of leukaemic children who typed for the six most common DPB1 alleles (DPB1*0101, *0201, *0301, *0401, *0402 and *0501), and the combined frequency of the remaining alleles (‘Others') with the percentage of children with solid tumours and normal infants who typed for these alleles. The results of these comparisons are shown as % phenotype frequencies in Table 2. Measures of relative risk attributable to each allele tested are depicted as cross-product odds ratios (ORs) and 95% confidence intervals (95% CI) in Table 3. The results in Table 2 show that a higher percentage of children with c-ALL typed for DPB1*0201 (17.5%) than did children with solid tumours (10.8%) or normal infants (10.4%). This increased frequency in leukaemia was also seen in the total (i.e. ALL plus non-ALL) patient series (16.9%), and in the total ALL series (17.5%), but was less marked in the non-ALL leukaemia series (12.1%) as compared with the solid tumour and normal infant groups. Within the ALL patient group, the percentages of children with T-ALL and pro-B ALL who typed for DPB1*0201 (18.8% and 15.8% respectively) also exceeded the percentages of children with solid tumours and normal infants typing for this allele. Furthermore, the percentages of children with c-ALL whose pretreatment leukaemia cells carried rearrangements of TEL–AML1 or a hyperdiploid karyotype and who typed for DPB1*0201 (18.9% and 18.1% respectively) were greater than in the solid tumour and infant groups.
|
|
To calculate the significance of the increased percentage of leukaemic children typing for each of the alleles shown in Table 2 as compared with the non-leukaemic children, we computed ORs and 95% CI values. The results of this analysis, shown in Table 3, indicate that ORs significantly greater than 1.0 were obtained for DPB1*0201 in the total leukaemia patient series, the total ALL series, for children with c-ALL, pro-B ALL and T-ALL, but not for children with non-ALL, as compared with the children with solid tumours and normal infants. ORs significantly greater than 1.0 were also found in children in the total leukaemia and pro-B ALL groups who typed for DPB1*0401, and in children in the total ALL, c-ALL and pro-B ALL group who typed for DPB1*0402. In contrast, ORs significantly less than 1.0 were obtained for children who typed for DPB1*0101 in the total leukaemia series, the total ALL series and the children with c-ALL, but not in children with non-ALL.
Our previous results showing an association between c-ALL and DPB1*0201 (46) made no correction for multiple allelic comparisons, and we could therefore not exclude a chance observation. The larger series of children with leukaemia compared with those with solid tumours and normal infants included in the present study provided the analysis with sufficient statistical power to detect an OR similar to that reported in our previous study, and also enabled us to correct P-values obtained in Fisher's exact tests for each leukaemia–non-leukaemia DPB1 phenotype comparison to exclude multiple testing (type 1, Bonferroni) errors. Even with this conservative approach, the increased DPB1*0201 phenotype frequency in children with c-ALL remained significant after correction for multiple alleles and diagnostic groups. In other comparisons (data not shown), however, we could find no difference in DPB1*0201 phenotype frequency between boys and girls with c-ALL, or between children with c-ALL who were diagnosed between 2 and 4 years of age (N=290) when compared with those children with c-ALL who were diagnosed between 5 and 14 years (N=214). Furthermore, the increased DPB1*0201 phenotype frequency in children with c-ALL remained significantly greater than in the comparison groups even when the small percentage of children with c-ALL of non-white parents were removed from the analysis.
To assess whether susceptibility to ALL with the TEL–AML1 fusion gene or chromosomal hyperdiploidy was influenced by the DPB1 phenotype, we compared the frequencies of leukaemic children with these markers and children with solid tumours and normal infants. The results in Table 4 show that ORs significantly greater than 1.0 were obtained for children with TEL–AML1 and hyperdiploid leukaemia cells who typed for DPB1*0201, but ORs significantly less than 1.0 were obtained for children with these leukaemias who typed for DPB1*0101 as compared with normal infants but not as compared with children with solid tumours. The increased risk of hyperdiploid leukaemia in children typing for DPB1*0201 was significant after correction for multiple comparisons, but the risk of TEL–AML1 leukaemia was not significant, although the number of cases available for analysis was small. Of the 95 children with TEL–AML1+ ALL, 65 had a confirmed diagnosis of c-ALL and the percentage typing for DPB1*0201 was no greater than those in the total c-ALL series (Table 2). By contrast, only 2 of the 65 TEL–AML+ c-ALL children (3%) typed for DPB1*0101, as compared with 7% of the 227 children with hyperdiploid leukaemia and 10.4% of the normal infants (Table 2), suggesting that DPB1*0101 may afford additional protection against TEL–AML1+ c-ALL.
|
To determine the influence of DPB1*0201 and *0101 genotypes, respectively, on susceptibility and resistance to c-ALL, we calculated ORs for *0201 and *0101 homozygous and heterozygous genotypes using normal infants for comparison. As Table 5 shows, children with c-ALL who were inferred as being homozygous for DPB1*0201 (i.e. only one allele was detected) were at a lower risk than children who were heterozygous for this allele. By contrast, children who were heterozygous for alleles that included DPB1*0201 and DPB1*0401 or DPB1*0402 had ORs significantly greater than 1.0. Children who were homozygous for DPB1*0101 were significantly protected from c-ALL as compared with newborn infants, whereas those children with c-ALL who were heterozygous were not significantly protected by this allele.
|
The results of typing other, less common, DPB1 alleles, in childhood leukaemia suggested that HLA-DPB1-associated susceptibility to childhood c-ALL might not be confined to a single allele, DPB1*0201, but might also involve other alleles. However, the number of children with c-ALL who typed for rarer alleles, and the total number of DPB1 alleles detected, rendered an analysis of individual rare alleles using a conventional Fisher's exact test unreliable. To overcome this, we used the CLUMP Monte Carlo simulation test described by Sham and Curtis (48) to test the post hoc hypothesis that certain (rare) alleles are more common in children with c-ALL than in normal infants. We also used CLUMP to identify these alleles, and, based on the polymorphic amino acids lining the peptide-binding pockets of the antigen-binding sites (ABS) of these allotypic proteins that are also present in DPß1*0201, we compared ORs obtained with most of these alleles with that of DPB1*0201.
Using CLUMP (48) to compute the T1 statistic from a 2x38 contingency table of DPB1 allele frequencies in children with c-ALL and normal infants, we obtained a 
2 of 103.93 that was not exceeded in 2x104 simulations (P<0.000001), suggesting that there is indeed a significant difference in DPB1 phenotype frequency distribution between the c-ALL and infant groups. The maximized 2 T4 statistic obtained by CLUMP analysis tests the hypothesis that certain alleles are more frequent in the case than in the comparison group. The 2 of 42.41 (P= 0.0001) supported this conclusion, and suggested that, in addition to *0201, the following alleles may contribute to c-ALL susceptibility: DPB1*0202, *0402, *0601, *0801, *1001, *1401, *1601, *1701, *1901, *2001, *3301, *3501, and *4901.
CLUMP analysis suggested that susceptibility to c-ALL might not be confined to DPB1*0201, but does not identify the cause. One explanation for DPB1 multi-allelic susceptibility might be that it is related to functional properties shared by different allotypes, based on the presence of polymorphic amino acids lining the peptide-binding pockets of DPß1*0201 and these allotypes. Comparison of each putative peptide pocket profile of the 14 ‘CLUMPED’ DPB1 alleles (including DPB1*0201) with the DPß1*0201 allotypic protein reveals that 8 alleles (57%) code leucine–phenylanine–glutamine–glycine (single-letter amino acid code LFQG; HVR-A) at positions 8–11, 11 alleles (79%) code phenylalanine–valine (FV; HVR-B) at positions 35 and 36, 6 alleles (43%) code aspartic acid–glutamic acid–glutamic acid (DEE; HVR-C) at positions 55–57, 8 alleles (57%) code isoleucine–leucine–glutamic acid–glutamic acid–glutamic acid (ILEEE: HVR-D) at positions 65–69, 9 alleles (64%) code methionine (M; HVR-E) at position 76, and 4 alleles (29%) code glycine–glycine–proline–methionine (GGPM; HVR-F) at positions 84–87. By contrast, the numbers of ‘CLUMPED’ alleles that share pocket profiles with the protective allotype DPß1*0101 are as follows: VYQG8-11: 0 (0%); YA35,36: 1 (7%); AAE55-57: 1 (7%); ILEEK65-69: 3 (21%); V76: 4 (29%); DEAV84-86: 8 (57%). Thus, partial sharing of pocket profiles with DPß1*0201 and a deficit of pocket profiles shared with DPß1*0101 in the ‘CLUMPED’ alleles strongly suggest that susceptibility to childhood c-ALL may involve contributions from individual polymorphic peptide pocket profiles (40) common to the ABS of DPß1*0201 and the other DPß1 allotypic proteins (49), rather than by the DPß1 peptide-binding matrices that constitute the complete allotypic protein.
To investigate this further, we analysed the frequency of each of the polymorphic amino acids in five of the six DPß1 pocket profiles (HVRs A–E) in all children with leukaemia as compared with the solid tumour patients and normal infants. By calculating phenotype frequencies for each amino acid, we took account of the possibility that an individual pocket profile (i.e. LFQG8-11) might be phenotypically homozygous, despite the subject typing for that pocket profile being DPB1 heterozygous. Thus, subjects heterozygous for DPB1*0201/*0401, which is associated with increased susceptibility to c-ALL, are homozygous for LFQG8-11.
The results in Table 6 show that, with one exception (FA35,36 in hyperdiploid leukaemia), ORs significantly greater than 1.0 were confined to amino acids constituting the DPß1*0201 peptide-binding matrix, consisting of LFQG8-11, FV35,36, DEE55-57, ILEEE65-69 and M76. Even after correction for multiple testing, significant associations were found with the total leukaemia series, ALL, c-ALL, TEL–AML1+ ALL and hyperdiploid ALL. ORs significantly greater than 1.0 were also found in pro-B ALL, T-ALL and non-ALL, but P-values for these subtypes were not significant after correction for multiple testing. Additional support for the association with individual DPß1*0201 pocket profiles (and hence potentially functional interactions with anchor residues on relevant ligands) derives from the absence of associations (i.e. ORs<1.0) with non-DPß1*0201-associated polymorphic amino acids – specifically those amino acids present in the DPß1*0101 peptide-binding matrix.
|
Susceptibility to c-ALL thus seems to be influenced by – though not confined to – single DPß1*0201-associated pocket profiles. This suggests that different combinations of pocket profiles might have an important influence on susceptibility to c-ALL. Since the number of combinations of the six peptide-binding pockets shared by DPß1*0201 and other allotypes is potentially large, and the contribution of different combinations of these peptide pockets to ligand binding is difficult to model statistically, we ranked the majority of DPß1 allotypes proteins predicted by CLUMP analysis of c-ALL that exceeded the frequency in the infant group (except those absent from the infant group) by the number and identity of pocket profiles shared with the DPß1*0201 allotype (Table 7 ). Although not a quantitative relationship, this comparison suggests that DPß1 allotypes with P9, P4 and P2 pocket profiles containing the DPß1*0201 peptide pockets FV35,36, ILEEE69 and M76, respectively, confer the greatest level (in terms of OR) of susceptibility to c-ALL. None of the pocket profiles that contribute to c-ALL susceptibility are found in DPß1*0101, adding further support to a specific role for interactions between DPß1*0201-related pocket profiles and as-yet unidentified peptide ligands in the causation of c-ALL.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSUKCCS INVESTIGATORSREFERENCES |
|---|
In this study, we have presented the results of one of the largest comparative analyses of HLA class II alleles ever undertaken to assess the contribution of HLA-associated genetic susceptibility in childhood leukaemia. Based on the prior hypothesis of an association between childhood c-ALL and HLA-DPB1*0201 (46), power calculations predicted that the number of leukaemic and non-leukaemic children available for analysis in the present study would exclude a chance association. Since our results were obtained with a different series of leukaemia patients and non-leukaemia comparison groups from those included in our preliminary study, they provide confirmation that the prior evidence of an association between DPB1*0201 and susceptibility to childhood c-ALL was not a chance finding. Our results indicate that there is a 70–80% increase in the risk of common ALL in children typing for DPB1*0201, and they suggest – though they do not prove – that there may be a similarly increased risk in other types of childhood ALL.
In addition to confirming the previously documented association with DPB1*0201 (46), we report here for the first time that DPB1*0101 conferred significant protection against c-ALL. Furthermore, genotype analysis showed that DPB1*0101 homozygosity resulted in even greater protection against c-ALL. In contrast, there was no increase in susceptibility to c-ALL in patients with DPB1*0201 homozygosity, whereas there was an increase in susceptibility in DPB1*0201 heterozygous patients. One of the mechanisms thought to be responsible for the maintenance of HLA polymorphisms is infectious-pathogen-driven selection, leading to heterozygote advantage (49). The functional interpretation of heterozygote advantage is that it allows for a wider diversity of HLA-restricted epitopes to be presented to T cells, and consequently the development of a more effective protective immune response. Although direct evidence for heterozygote advantage in resistance to infectious disease remains limited, a role for heterozygous advantage in DR–DQ-associated resistance to persistent hepatitis B infection has been reported by Thursz et al. (50). There are also reports that antibody responses to hepatitis B (51) and measles vaccines (52) are stronger in HLA class II heterozygous than homozygous individuals. Our finding that DPB1 heterozygosity is associated with increased susceptibility to c-ALL accords with the Greaves hypothesis (15,16) that an active immune response that is protective against infection(s) is likely to be associated with an increased risk of c-ALL.
One of the important findings of this study that was not reported in our previous study (46) (since the patient group there was too small) is that susceptibility to c-ALL, and possibly to other subtypes of childhood ALL, is not confined to DPB1*0201. Evidence obtained by CLUMP analysis suggests that susceptibility is associated with several less common DPB1 alleles. One explanation for this multi-allelic susceptibility is that it is due to the presence in other alleles of sequences encoding polymorphic amino acids lining the ABS that are also present in the DPß1*0201 allotypic protein. Support for this conclusion was obtained from the analysis of the polymorphic amino acid frequencies encoded by DPB1 HVRs which showed susceptibility to be associated with putative DPß1 peptide pocket profiles. This interpretation of our results leads us to conclude that the association of HLA-DPB1 alleles with susceptibility to childhood c-ALL, and possibly with other ALL subtypes, may be related to the ligand-binding properties of individual DPß1 peptide pocket profiles. Although the peptide-binding characteristics of DPß allotypic proteins are not as well characterized as those of DRß allotypes, similarities between naturally processed self-peptides eluted from DP2 and DR proteins, and in the binding of synthetic peptides by DP2 and DR (53), suggest that the functions of DP pocket profiles are analogous to those established for DR. Evidence of rapid allelic diversification in DPB1 suggests that recent diversifying selection pressure has been applied to the DPß1 ABS (54), and points to an active functional role for this locus. Nonetheless, evidence for the role of DPB1 alleles in susceptibility to childhood ALL does not preclude additional contributions from alleles at other HLA class II loci (55,56), in a manner similar to that known to occur for DQ and DR alleles in insulin-dependent (type 1) diabetes mellitus (IDDM) (57). Indeed, Djoulah et al. (58) have used a multilocus model to determine the contribution of the DQ P4 and P9 and DR P4 pockets to IDDM susceptibility with high predictive value, which is similar to our single-locus approach to the role of peptide-binding pockets in c-ALL.
Our data suggest that a ligand that has binding affinity for DPß1*0201-associated ABS pocket profiles may be capable of contributing to the risk of c-ALL. The precise functional details require further analysis, but could be interpreted to suggest that a ligand derived from any protein that binds to a DPß1 peptide matrix with pocket profiles related to DPß1*0201 may carry epitopes leading to T-cell activation and proliferative stress on pre-c-ALL cells. It is thus possible to envisage that such T-cell epitopes could be derived from more than one type of infection. Evidence that children typing for DPB1*0101 were at a 20–30% lower risk of developing c-ALL further supports the conclusion that certain DPß1 allotypes bind peptide ligands that lead to T-cell responses that protect against c-ALL, whilst others participate in the development of c-ALL.
Since we interpret our findings as suggesting that it is the functional properties of DPß1 allotypes that contribute to childhood c-ALL susceptibility, we suggest that this may be mediated by a mechanism that involves the differential binding of peptide ligands by DPß1*0201 and related allotypes. We propose that this leads to the activation of a CD4+ helper T (Th)-cell population by the DPß1-ligand complex, which contributes to the proliferative stress exerted on a clone of premalignant c-ALL cells, or alternatively to the apoptotic suppression of normal haematopoietic cells, thereby facilitating the emergence of a clone of leukaemia cells. Such an interpretation makes it unlikely that childhood c-ALL is caused by a mutation in an as-yet unidentified ‘leukaemia gene’ that is linked to DPB1. Weak linkage disequilibrium (LD) between DPB1 and DQ/DR alleles (59) also makes it unlikely that the association that we have observed with DPB1*0201 is secondary to stronger effects of DQ/DR alleles in LD with DPB1*0201. We found no evidence for such an explanation in our previous study of DQB1 association with childhood c-ALL (60), although we do not exclude an independent contribution by DQ and DR allotypes to c-ALL susceptibility (55,56).
In addition to confirming the association of DPB1*0201 with susceptibility to c-ALL, our data suggest that this allele may also contribute to susceptibility to TEL–AML1+ and hyperdiploid c-ALL, to pro-B ALL and to T-ALL. If this is correct, it would imply that the functional contribution of the DPß1 allotypic protein is not confined to a single ALL subtype. However, the number of cases with each of these subtypes was too small to allow us to reach a definite conclusion on this point, and confirmation will require studies of additional series of patients.
The result of our previous study of DPB1 alleles in childhood c-ALL showed that the relative risk (odds ratio) associated with DPB1*0201 was low (2.0) (46). This is not unexpected in a multifactorial disease such as c-ALL, but it did indicate that confirmation of this level of risk would require a much larger patient and comparison group. The present series of 559 patients with confirmed c-ALL and 864 newborn infants, which yielded an odds ratio for DPB1*0201 of 1.83, had 98% power at a P-value of 0.05. Similar calculations revealed that the strength of the associations between DPB1*0201 and hyperdiploid and TEL–AML1+ c-ALL had 94% and 71% powers, respectively.
Studies of HLA associations with rare diseases such as childhood leukaemia can present problems with the selection of a relevant comparison group, which can lead to spurious associations. In the present study, ethical and logistic considerations precluded the collection of blood samples from the UKCCS case-matched control children. For case-comparison purposes, we therefore used two quite different series of non-leukaemic children. One series consisted of UKCCS case children with solid tumours (excluding lymphoma), who were ascertained nationally in the same way as the leukaemic children. The other group consisted of an unselected series of normal newborn infants born in Manchester, UK. In the absence of a national birth cohort in the UK, both comparison series represent the best non-leukaemic comparison groups of children that could be achieved. Theoretical considerations relating to the use of cancer controls in epidemiological studies of cancer have been reviewed (61–63), and suggest that where the hypothesis being tested concerns the specificity of exposure in relation to a specific type of cancer, as is the case in the present study, then the advantages outweigh the disadvantages. Smith et al. (62) discussed the value of multiple control groups and cited studies in which both cancer and non-cancer population groups were used as controls for specific cancers.
The normal full-term babies who were born in Manchester comprised a population-based sample. Collection of parental ethnic background information enabled us to exclude the non-white ethnic groups from the HLA analysis. However, had we not done this it would have made virtually no difference to the overall result, since the frequency of white and Asian newborn infants with DPB1*0201 was not significantly different. Furthermore, our results showed that the increased DPB1*0201 phenotype frequency was present in the total leukaemia series, the ALL group and the c-ALL patients, regardless of the ethnic background of the infants. In a previous study, it was reported that about 5% of children with ALL in the UK have Asian parents (64). Our expectation was that the presence of this proportion of children in our leukaemia series would make little if any difference to the result. This was confirmed when we removed non-white children from the c-ALL series from the analysis, resulting in a DPB1*0201 phenotype frequency in the remaining white children that was virtually unchanged (17.0% in the white-only, compared with 17.5% in the total c-ALL series; 10.2% in the white-only solid tumours, compared with 10.8% in the total series), giving an OR of 1.80 with a 95% CI of 1.18–2.75. Our results thus suggest that, whilst the sources of comparison groups are not ideal, population stratification does not explain the association that we found between DPB1*0201 and susceptibility to childhood c-ALL. Previous concerns about spurious associations due to population stratification (65) are likely to be a problem only where disease and allele frequencies differ among subpopulations (66). Wacholder et al. (67) has estimated that the effect of ignoring ethnicity on the frequency of N-acetyltransferase genotypes in non-Hispanic US cancer patients of diverse European origins only biased results by 1%. Our results suggest that ethnic stratification is unlikely to explain the increased frequency of children with ALL typing of DPB1*0201 and related alleles.
The modular architecture of the DPß1 allotypic proteins can lead to the same peptide pocket profiles occurring in different DPß1 allotypes. However, different allotypes have different combinations of pocket profiles, resulting in a peptide-binding matrix that is uniquely encoded by that allele. Predictions based on peptide binding by DRß1 allotypes (40) suggest that the same pocket profile occurring in different allotypes has the same binding affinity for a given ligand. We found that several DPB1 alleles are associated with susceptibility to c-ALL, and that, with only one exception, these alleles share peptide pockets with DPß1*0201 more frequently than with the protective allele DPß1*0101. Thus, of the 14 CLUMPED alleles, 8, 11, 5, 8, 9 and 5 code the same polymorphic acids as DPB1*0201 at positions 8–11, 35, 36, 55–57, 65–69, 76 and 84–86, respectively. By contrast, only 0, 1, 1, 3, 4 and 8 of these 14 alleles code the same polymorphic amino acids at these positions, as the protective allele, DPB1*0101, with alleles coding DPß1*0101 amino acids exceeding in number those of DPB1*0201 at positions 84–86 (i.e. DEAV84–86>GGPM84–86).
If the functional interaction of DPß1 allotypic proteins with antigenic peptides is similar in c-ALL to that in other DPB1- and DRB1-associated diseases, this would predict a role for the DPß1*0201 ABS in eliciting a CD4+ Th response to epitopes on a range of antigenic peptides. At present, the identity of these peptides in c-ALL is unknown, but the peptide with the highest known binding affinity for DPß1*0201 is hepatitis B virus surface antigen, HbsAg14–33 (53) – and the hepatitis B virus has already been identified in children with leukaemia (68,69). Evidence that DPB1*0201 is protective against the IgE response to the house dust mite allergen Der p1 (70) might explain why there is thought to be a deficit of allergy among children with leukaemia (37,71), although this requires further investigation. Sensitivity to chronic beryllium disease (CBD) (72,73) and cobalt-induced hard metal disease (74) are specifically associated with DPB1*0201 and more importantly with glutamic acid (E69) in the DPß1 P4 peptide-binding pocket. It has been suggested that beryllium forms hapten complexes with endogenous peptides that can lead to metal sensitization and lung disease (73). Although heavy metal sensitization restricted through DPB1 seems an unlikely mechanism of susceptibility to c-ALL, it is possible that the selective presentation of certain other peptides in the context of DPß1*0201 and related allotypes could lead to proliferative stress on pre-c-ALL cells. The TEL–AML1 fusion peptide can probably be excluded as the source of this DPß1*0201-bound ligand, since we found no evidence of DPß1*0201 anchor residues (53) in the TEL–AML1 breakpoint sequence. In contrast, TEL–AML1 peptide presentation restricted by DP5 and DP17 has been shown to induce CD4+ T-cell clonal proliferation (75). Of these two alleles, only DPB1*1701 is associated with susceptibility to c-ALL. The association of DPB1*0101 with protection from TEL–AML+ c-ALL suggests a role mediated by the DPß1*0101 ABS, although this requires confirmation by direct peptide binding and T-cell functional studies (76).
Although direct cellular studies have yet to be carried out, our results could be interpreted to suggest that the increased risk of c-ALL associated with DPB1*0201 and the decreased risk associated with DPB1*0101 might indicate differences in the polarity of the Th responses in response to infection-derived peptides that lead to, or protect against, c-ALL. Evidence that Th1 responses primarily mediate cellular immunity, while Th2 responses support humoral immunity (77), points to a possible mechanism for Th1 or Th2 cytokines in the proliferative stress that may lead to c-ALL. Although evidence suggests that Th1/Th2 response polarization is MHC-linked (78), this has not yet been directly demonstrated for DPB1 allotypes, so a role for differential Th reponses remains a matter for conjecture – though amenable to investigation. Nonetheless, the apparently inverse relationship between leukaemia and allergy (38,71) and the protective effect of DPB1*0201 against house dust mite IgE responses (70) suggest, albeit indirectly, that Th1 responses may be more important than Th2 responses in the immune-mediated proliferative stress exerted in pre-c-ALL.
The low relative risk (OR) for DPB1*0201 in c-ALL of approximately 1.8 in our study strongly suggests that other genetic and environmental factors are likely to be involved in the aetiology of childhood c-ALL. Nevertheless, identification of an HLA-DPB1 component in this multifactorial aetiological pathway suggests that further investigation of the HLA class II–peptide–T-cell receptor trimolecular complex in c-ALL might aid the identification of the antigenic peptides (40,76), the nature of the reactive Th population and the cellular events involved in the proliferative stress leading to c-ALL.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSUKCCS INVESTIGATORSREFERENCES |
|---|
Patients
All UKCCS patient data (diagnoses, ages and ethnic background) used in this paper have been checked and validated by the UKCCS Data Centre at the LRF Epidemiology Unit, University of Leeds. Diagnostic immunophenotyping of pretreatment leukaemias was carried out as previously described (47) according to criteria used in the MRC UKALL XI trial (11,79). ALL was subdivided on the basis of immunophenotyping into pro-B (early pre-B or null) ALL (CD10-CD19+T-); common (precursor B-cell) ALL (CD10+ CD19+TdT+T-) and T-cell ALL (CD2/CD7+CD19-DR-). The non-ALL cases included in this study predominantly included acute myeloid leukaemias, but also a small number of patients with juvenile chronic myeloid leukaemia, acute myelomonocytic leukaemia and myelodysplasia. Molecular and conventional cytogenetic analysis of pretreatment leukaemic bone marrow was carried out as previously described (47). Leukaemias with the TEL(ETV6)–AML1(CBFA2) fusion gene were identified using RT–PCR and/or FISH. Leukaemias with a hyperdiploid karyotype were identified using conventional cytogenetics, and by FISH (47).
Blood samples
Blood samples for HLA-DPB1 typing were obtained during the post-treatment period from 65% of the children diagnosed with leukaemia who fulfilled criteria for inclusion in the national UK Childhood Cancer Study (47) between 1992 and 1998. Blood samples from the two comparison groups used in the DPB1 phenotype frequency comparisons were obtained from children with solid tumours, but excluding lymphoma (Hodgkin's and non-Hodgkin's lymphoma) included in the UKCCS, and comprising 30% with brain tumours and 70% with other solid tumours, and from 864 full-term infants of white UK parents born at St Mary's Hospital, Manchester.
HLA-DPB1 molecular typing
Genomic DNA was extracted from case and comparison group blood samples and HLA-DPB1 molecular typing carried out as previously described (46,80,81). DPB1 typing involved a high-throughput PCR–SSOP method in which 96 DNA samples were amplified using a single pair of generic DPB1-specific primers. Arrays of 384 PCR products were blotted onto nylon membranes, and spot hybridized with 28 32P-labelled DPB1 SSOP. SSOP hybridization was detected using real-time autoradiography, and alleles assigned from published DPB1 ideograms.
Data analysis
DPB1 types for each study subject were stored in a Microsoft Access database and transferred to Microsoft Excel spreadsheets for statistical analysis. DPB1 allele frequencies were determined by the method of allele (gene) counting, and these were used to derive genotypes, from which DBP1 phenotype frequencies (i.e. numbers of subjects with a given DPB1 allele) were computed. Homozygosity was inferred in subjects who typed for only one DPB1 allele, rather than being directly detected in family studies. Case–control comparisons were carried out on phenotype frequency data using the RELRISK program in the Linkage Utility Package LINKUTIL (82) by calculating cross-product ORs and 95% confidence intervals by the method of Mantel and Haenszel (83). Phenotype frequency comparisons were also carried out by two-sided Fisher's exact tests using the 2by2 program (82), and P-values were corrected by multiplication by the number of disease subtypes and the number of DPB1 alleles detected. The CLUMP program, version 1.2 (48), was used to identify common and rare DPB1 alleles contributing to c-ALL susceptibility.
|
UKCCS INVESTIGATORS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSUKCCS INVESTIGATORSREFERENCES |
|---|
Management Committee
K.K. Cheng, Central region; N. Day, East Anglia region; R. Cartwright, A. Craft, North East region; J.M. Birch, O.B. Eden, North West Region; P.A. McKinney, Scotland; J. Peto, South East Region; V. Beral, E. Roman, South Midlands region; P. Elwood, South Wales region; F.E. Alexander, South West region; C.E.D. Chilvers, Trent region; R. Doll, Epidemiological Studies Unit, University of Oxford; G.M. Taylor, Immunogenetics Laboratory, University of Manchester; M. Greaves, Leukaemia Research Fund Centre, Institute of Cancer Research, London; D. Goodhead, MRC Radiation and Genome Stability Unit, Harwell; F.A. Fry, National Radiological Protection Board; G. Adams, UK Coordinating Committee for Cancer Research.
Regional Investigators
K.K. Cheng, E. Gilman, Central region; N. Day, J. Skinner, D. Williams East Anglia region; R. Cartwright, A. Craft, North Eastern region; J.M. Birch, O.B. Eden, North West Region; P.A. McKinney, Scotland; J. Deacon, J. Peto, South East Region; V. Beral, E. Roman, South Midlands region; P. Elwood, South Wales region; F.E. Alexander, M. Mott, South West region; C.E.D. Chilvers, K. Muir, Trent region.
Leukaemia Research Fund Data Management Processing Group
R. Cartwright, G. Law, J. Simpson, E. Roman.
A complete list of UKCCS investigators is given in (47).
|
ACKNOWLEDGEMENTS |
|---|
We are indebted to the children and families who took part in the UK Childhood Cancer Study for enabling us to carry out this work. We thank J. Simpson and Dr E. Roman at the Leukaemia Research Fund Centre (LRF) for Clinical Epidemiology, University of Leeds for providing detailed information on patient diagnoses and other characteristics. We thank Dr D. Curtis for the CLUMP program and advice on its use, R. Carter for help in documenting blood sample details, and midwives at St Mary's Hospital, Manchester for obtaining cord blood samples from newborn infants. We are grateful to M.D. Robinson, Dr C. Watson and Dr D.A. Gokhale for help in sample processing and setting up the molecular methods during the early part of the study. The Laboratory work was carried out with the support of the Kay Kendall Leukaemia Fund. Professor O.B. Eden is supported by the Cancer Research UK, and Professor M.F. Greaves by the Leukaemia Research Fund and the Kay Kendall Leukaemia Fund. The United Kingdom Childhood Cancer Study (UKCCS) was sponsored and administered by United Kingdom Coordinating Committee on Cancer Research. The UKCCS has been conducted by 12 teams of investigators (10 clinical and epidemiological and 2 biological) based in university departments, hospitals, research institutes and the Scottish health service. The work is coordinated by a Management Committee and in Scotland by a Steering Group. It is supported by the United Kingdom Children's Cancer Study Group, consisting of paediatric oncologists, and by the National Radiological Protection Board. Funding has been provided by a consortium of statutory bodies, cancer charities and industrial sponsors.
|
FOOTNOTES |
|---|
* To whom correspondence should be addressed. Tel: +44 161 2766472; Fax: +44 161 2766414; Email: gmtaylor{at}man.ac.uk
For a full list of Investigators, see the end of the paper.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSUKCCS INVESTIGATORSREFERENCES |
|---|
1 Breslow, N.E. and Langholz, B. (1983) Childhood cancer incidence: geographic and temporal variations. Int. J. Cancer, 32, 703–716.[Web of Science][Medline]
2 Parkin, D.M., Stiller, C.A., Draper, G.J. and Bieber, C.A. (1988) The international incidence of childhood cancer. Int. J. Cancer, 42, 511–520.[Web of Science][Medline]
3 Coeburg, J.W.W., van der Does-van den Berg, A., van Wering, E.R., van Steensel-Moll, H.A., Valkenburg, H.A., van 't Veer, M.B., Schmitz, P.I.M. and van Zanen, G.E. (1989) Childhood leukaemia in The Netherlands, 1973–1986: Temporary variation of the incidence of acute lymphocytic leukaemia in young children. Br. J. Cancer, 59, 100–105.[Web of Science][Medline]
4 Linet, M.S. and Devesa, S.S. (1991) Descriptive epidemiology of childhood leukaemia. Br. J. Cancer, 63, 424–429.[Web of Science][Medline]
5 Gurney, J.G., Severson, R.K., Davis, S. and Robison, L.L. (1995) Incidence of cancer in children in the United States. Cancer, 75, 2186–2195.[Web of Science][Medline]
6 Stiller, C.A., Allen, M.B. and Eatock, E.M. (1995) Childhood cancer in Britain: the National Registry of Childhood Tumours and incidence rates 1978–1987. Eur. J. Cancer, 31A, 2028–2034.
7
Pui, C.-H., Behm, F.G. and Crist, W.M. (1993) Clinical and biologic relevance of immunologic marker studies in childhood acute lymphoblastic leukaemia. Blood, 82, 343–362.
8 Greaves, M.F., Colman, S.M., Beard, M.E.J., Bradstcok, K., Cabrera, M.E., Chen, P.-M., Jacobs, P., Lam-Po-Tang, P.R.L., MacDougall, L.G., Williams, G.K.O. et al. (1993) Geographical distribution of acute lymphoblastic leukaemia subtypes:second report of the collaborative group study. Leukemia, 7, 27–34.[Web of Science][Medline]
9 McKinney, P.A., Alexander, F.E., Cartwright, R.A., Scott, C.S. and Staines, A. (1993) Acute lymphoblastic leukaemia in the UK by immunophenotype. Leukemia, 7, 1630–1634.
10 Buckley, J.D., Buckley, C.M., Ruccione, K., Sather, H.N., Waskerwitz, M.J., Woods, W.G. and Robison, L.L. (1994) Epidemiological characteristics of childhood acute lymphocytic leukemia. Analysis by immunophenotype. Leukemia, 8, 856–864.[Web of Science][Medline]
11 Hann, I.M., Richards, S.M., Eden, O.B. and Hill, F.G.H. (1998) Analysis of the immunophenotype of children treated on the Medical Research Council United Kingdom Acute Lymphoblastic Leukaemia Trial XI. Leukemia, 12, 1249–1255.[Web of Science][Medline]
12 Ford, A.M., Ridge, S.A., Cabrera, M.E., Mahmoud, H., Steel, C.M., Chan, L.C. and Greaves, M.F. (1993) In utero rearrangements in the trithorax-related oncogene in infant leukaemias. Nature, 363, 358–360.[Medline]
13
Gale, K.B., Ford, A.M., Repp, R., Borkhardt, A., Keller, C., Eden, O.B. and Greaves, M.F. (1997) Backtracking leukaemia to birth: identification of clonotypic gene fusion sequences in neonatal blood spots. Proc. Natl Acad. Sci. USA, 94, 13950–13954.
14 Wiemels, J.L., Cazzaniga, G., Daniotti, M., Eden, O.B., Addison, G.M., Masera, G., Saha, V., Biondi, A. and Greaves, M.F. (1999) Prenatal origin of acute lymphoblastic leukaemia in children. Lancet, 354, 1499–1503.[Web of Science][Medline]
15 Greaves, M.F. (1999) Molecular genetics, natural history and the demise of childhood leukaemia. Eur. J. Cancer, 35, 173–185.
16 Greaves, M.F. and Alexander, F.E. (1993) An infectious etiology for common acute lymphoblastic leukemia in childhood? Leukemia, 7, 349–360.[Web of Science][Medline]
17 Kinlen, L.J. (1995) Epidemiological evidence for an infective basis in childhood leukaemia. Br. J. Cancer, 71, 1–5.[Web of Science][Medline]
18 Greaves, M.F. (1997) Aetiology of acute leukaemia. Lancet, 349, 344–349.[Web of Science][Medline]
19
Sutton, R.N.P., Marston, S.D., Pullen, H.J.M., Darby, C.W., Evans D.I.K. and Emond, R.T.D. (1974) Antibodies to Epstein–Barr and other viruses in children with acute lymphoblastic leukaemia. Arch. Dis. Child., 49, 540–544.
20 Schlehofer, B., Blettner, M., Geletneky, K., Haaf, H.-G., Kaatsch, P., Michaelis, J., Mueller-Lantzsch, N., Niehoff, D., Winkelspecht, B., Wahrendorf, J. and Schlehofer, J.R. (1996) Sero-epidemiological analysis of the risk of virus infections for childhood leukaemia. Int. J. Cancer, 65, 584–590.[Web of Science][Medline]
21 MacKenzie, J., Perry, J., Ford, A.M., Jarrett, R.F., Greaves, M. (1999) JC and BK virus sequences are not detectable in leukaemic samples from children with common acute lymphoblastic leukaemia. Br. J. Cancer, 81, 898–899.[Web of Science][Medline]
22 Smith, M.A., Strickler, H.D., Granovsky, M., Reaman, G., Linet, M., Daniel, R. and Shah, K.V. (1999) Investigation of leukemia cells from children with common acute lymphoblastic leukemia for genomic sequences of the primate polyomaviruses JC virus, BK virus, and simian virus 40. Med. Paediatr. Oncol., 33, 441–443.[Web of Science][Medline]
23 MacKenzie, J., Gallagher, A., Clayton, R.A., Perry, J., Eden, O.B., Ford, A.M., Greaves, M.F. and Jarrett, R.F. (2001) Screening for herpesvirus genomes in common acute lymphoblastic leukemia. Leukemia, 15, 415–421.[Web of Science][Medline]
24 Kinlen, L.J., Clarke, K. and Hudson, C. (1990) Evidence from population mixing in British New Towns 1946–1985 of an infective basis for childhood leukaemia. Lancet, 336, 577–582.[Web of Science][Medline]
25 Kinlen, L.J. (1997) High contact paternal occupations, infection and childhood leukaemia: five studies of unusual population-mixing of adults. Br. J. Cancer, 76, 1539–1545.[Web of Science][Medline]
26
Kinlen, L.J. and John, S.M. (1994) Wartime evacuation and mortality from childhood leukaemia in England and Wales in 1945–1949. Br. med. J., 309, 1197–1202.
27 Alexander, F.E., McKinney, P.A., Ricketts, T.J. and Cartwright, R.A. (1990) Community lifestyle characteristics and risk of acute lymphoblastic leukaemia in children. Lancet, 336, 1461–1465.[Web of Science][Medline]
28 Badrinath, P., Day, N.E. and Stockton, D. (1997) Seasonality in the diagnosis of acute lymphocyctic leukaemia. Br. J. Cancer, 75, 1711–1713.[Web of Science][Medline]
29 Westerbeek, R.M., Blair, V., Eden, O.B., Kelsey, A.M., Stevens, R.F., Will, A.M., Taylor, G.M. and Birch, J.M. (1998) Seasonal variations in the onset of childhood leukaemia and lymphoma. Br. J. Cancer, 78, 119–124.[Web of Science][Medline]
30 Alexander, F.E., Boyle, P., Carli, P.-M., Coebergh, J.W., Draper, G.J., Ekbom, A., Levi, F., McKinney, P.A., McWhirter, W., Magnani, C. et al. (1998) Spatial temporal patterns in childhood leukaemia: further evidence for an infectious origin. Br. J. Cancer, 77, 812–817.[Web of Science][Medline]
31 Birch, J.M., Alexander, F.E., Blair, V., Eden, O.B., Taylor, G.M. and McNally, R.J.Q. (2000) Space–time clustering patterns in childhood leukaemia support a role for infection. Br. J. Cancer, 82, 1571–1576.[Web of Science][Medline]
32
Van Steensel-Moll, H.A., Valkenburg, H.A. and van Zanen, G.E. (1986) Childhood leukemia and infectious diseases in the first year of life: a register-based case–control study. Am. J. Epidemiol., 124, 590–594.
33 Neglia, J.P., Linet, M.S., Shu, X.O., Severson, R.K., Potter, J.D., Mertens, A.C., Wen, W., Kersey, J.H. and Robison, L.L. (2000) Patterns of infection and day care utilization and risk of childhood acute lymphoblastic leukaemia. Br. J. Cancer, 82, 234–240.[Web of Science][Medline]
34 McKinney, P.A., Juszczak, E., Findlay, E., Smith, K. and Thomson, C.S. (1999) Pre- and perinatal risk factors for childhood leukaemia and other malignancies: a Scottish case control study. Br. J. Cancer, 80, 1844–1851.[Web of Science][Medline]
35 Groves, F.D., Gridley, G., Wacholder, S., Shu, X.O., Robison, L.L., Neglia, J.P. and Linet, M.S. (1999) Infant vaccinations and risk of childhood acute lymphoblastic leukaemia in the USA. Br. J. Cancer, 81, 175–178.[Web of Science][Medline]
36 Auvinen, A., Hakulinen, T. and Groves, F. (2000) Haemophilus influenzae type B vaccination and risk of childhood leukaemia in a vaccine trial in Finland. Br. J. Cancer, 83, 956–958.[Web of Science][Medline]
37 Schüz, J., Kaletsch, U., Meinert, R., Kaatsch, P. and Michaelis, J. (1999) Association of childhood leukaemia with factors related to the immune system. Br. J. Cancer, 80, 585–590.[Web of Science][Medline]
38
Shu, X.O., Linet, M.S., Steinbuch, M., Wen, W.Q., Buckley, J.D., Neglia, J.P., Potter, J.D., Reaman, G.H., Robison, L.L. (1999) Breast-feeding and risk of childhood acute leukaemia. J. Natl. Cancer Inst., 91, 1765–1772.
39 Davenport, M.P. and Hill, A.V. (1996) Reverse immunogenetics: from HLA-disease associations to vaccine candidates. Mol. Med. Today, 2, 38–45.[Web of Science][Medline]
40 Hammer, J., Stumiolo, T. and Sinigaglia, F. (1997) HLA class II peptide binding specificity and autoimmunity. Adv. Immunol., 66, 67–100. [Web of Science][Medline]
41 Nishimura, Y., Oiso, M., Fujisao, S., Kanai, T., Kira, J.-I., Chen, Y.-Z. and Matsushita, S. (1998) Peptide-based molecular analyses of HLA class II-associated susceptibility to autoimmune diseases. Int. Rev. Immunol., 17, 229–262.[Medline]
42 Germain, R. (1995) Binding domain regulation of MHC class II molecule assembly, trafficking, fate and function. Semin. Immunol., 7, 361–372.[Medline]
43 Baccichet, A., Qualman, S.K., Sinnett, D. (1997) Allelic loss in childhood acute lymphoblastic leukemia. Leukemia Res., 21, 817–823.[Web of Science][Medline]
44
Takeuchi, S., Bartran, C.R., Wada, M., Reiter, A., Hatta, Y., Seriu, T., Lee, E., Miller, C.W., Miyoshi, I., Koeffler, H.P. (1995) Allelotype analysis of childhood acute lymphoblastic leukemia. Cancer Res., 55, 5377–5382.
45 Bignon, J.D., Femandez-Vina, M.A., Cheneau, M.L., Fauchet, R., Schrueder, G.M. Th., Clayton, J., Marsh, S.G.E. and Charron, D. (1997) HLA DNA class II typing by PCR-SSOP: 12th International Histocompatibility Workshop experience. In Charron, D. (ed), HLA. Proceedings of the Twelfth International Histocompatibility Workshop and Conference. Vol. 1, pp. 21–25.
46 Taylor, G.M., Robinson, M.D., Binchy, A., Birch, J.M., Stevens, R.F., Jones, P.M., Carr, T., Dearden, S. and Gokhale, D.A. (1999) Preliminary evidence of an association between HLA-DPB1*0201 and childhood common acute lymphoblastic leukaemia supports an infectious aetiology. Leukemia, 9, 440–443.
47 UK Childhood Cancer Study Investigators (2000) The United Kingdom Childhood Cancer Study: objectives, materials and methods. Br. J. Cancer, 82, 1073–1102.[Web of Science][Medline]
48 Sham, P.C. and Curtis, D. (1995) Monte Carlo tests for associations between disease and alleles at highly polymorphic loci. Ann. Hum. Genet., 59, 97–105.[Web of Science][Medline]
49 Hughes, A.L. and Nei, M. (1988) Pattern of nucleotide substitution at major histocompatibility complex class I loci reveals overdominant selection. Nature, 335, 167–170.[Medline]
50 Thursz, M.R., Thomas, H.C., Greenwood, B.M. and Hill, A.V.S. (1997) Heterozygote advantage for HLA class II-type in hepatitis B virus infection. Nat. Genetics, 17, 11–12.[Web of Science][Medline]
51 McDermott, A.B., Cohen, S.B.A., Zuckerman, J.N. and Madrigal, J.A. (1999) Human leukocyte antigens influence the immune response to a pre-S/S hepatitis B vaccine. Vaccine, 17, 330–339.[Web of Science][Medline]
52 Hayney, M.S., Poland, G.A., Jacobson, R.M., Rabe, D., Schaid, D.J., Jacobsen, S.J. and Lipsky, J.J. (1998) Relationship of HLA-DQA1 alleles and humoral antibody following measles vaccination. Int. J. Infect. Dis., 2, 143–146.[Medline]
53 Chicz, R.M., Graziano, D.F., Trucco, M., Strominger, J.L. and Gorga, J.C. (1997) HLA-DP2. Self peptide sequence and binding properties. J. Immunol., 159, 4935–4942.[Abstract]
54 Gyllensten, U., Bergstrom, T., Josefson, A., Sundval, M. and Erlich, H.A. (1996) Rapid allelic diversification and intensified selection at antigen recognition sites of the Mhc class II DPB1 locus during hominoid evolution. Tissue Antigens, 47, 212–221.[Web of Science][Medline]
55 Dorak, M.T., Owen, G., Galbraith, I., Henderson, N., Webb, D., Mills, K.I., Darke, C. and Burnett, A.K. (1995) Nature of HLA-associated predisposition to childhood acute lymphoblastic leukemia. Leukemia, 9, 875–878.[Web of Science][Medline]
56
Dorak, M.T., Lawson, T., Machulla, H.K., Darke, C., Mills, K.I. and Burnett, A.K. (1999) Unravelling an HLA-DR association in childhood acute leukemia. Blood, 94, 694–700.
57 Yasunaga, S., Kimura, A., Hamaguchi, K., Ronningen, K.S. and Sasazuki, T. (1996) Different contribution of HLA-DR and -DQ genes in susceptibility and resistance to insulin-dependent diabetes mellitus (IDDM). Tissue Antigens, 47, 37–48.[Web of Science][Medline]
58 Djoulah, S., Busson, M., Sasazuki, T., Maillere, B., Yasunaga, S., Kimura, A., Charron, D. and Hors, J. (1999) A new predictive model for insulin-dependent diabetes mellitus susceptibility based on combinations of molecular HLA-DRB1 and HLA-DQB1 pockets. Tissue Antigens, 54, 341–348.[Web of Science][Medline]
59 Begovich, A.B., McClure, G.R., Suraj, V.C., Helmuth, R.C., Fildes, N., Bugawan, T.L., Erlich, H.A. and Klitz, W. (1992) Polymorphism, recombination, and linkage disequilibrium within the HLA class II region. J. Immunol., 148, 249–258.[Abstract]
60 Dearden, S.P., Taylor, G.M., Gokhale, D.A., Robinson, M.D., Thompson, W., Ollier, W., Binchy, A., Birch, J.M., Stevens, R.F., Carr, T. and Bardsley, W.G. (1996) Molecular analysis HLA-DQB1 alleles in childhood common acute lymphoblastic leukaemic. Br. J. Cancer, 73, 603–609.[Web of Science][Medline]
61
Linet, M.S. and Brookmeyer, R. (1987) Use of cancer controls in case–control cancer studies. Am. J. Epidemiol., 125, 1–11.
62
Smith, A.H., Pearce, N.E. and Callas, P.W. (1988) Cancer case–control studies with other cancers as controls. Int. J. Epidemiol., 17, 298–306.
63
Lasky, T. and Stolley, P.D. (1994) Selection of cases and controls. Epidemiol. Rev., 16, 6–17.
64 Stiller, C.A., McKinney, P.A., Bunch, K.J., Bailey, C.C. and Lewis, I.J. (1991) Childhood cancer and ethnic group in Britain: a United Kingdom Children's Cancer Study Group (UKCCSG) study. Br. J. Cancer, 64, 543–548.[Web of Science][Medline]
65
Lander, E.S. and Schork, N.J. (1994) Genetic dissection of complex traits. Science, 265, 2037–2048.
66 Pritchard, J.K. and Donnelly, P. (2001) Case–control studies of association in structured or admixed populations. Theor. Pop. Biol., 60, 227–237.[Web of Science][Medline]
67
Wacholder, S., Rothman, N. and Caporaso, N. (2000) Population stratification in epidemiologic studies of common genetic variants and cancer: quantification of bias. J. Natl Cancer Inst., 92, 1151–1158.
68 Masera, G., Locasciulli, A., Jean, G., Jankovic, M., Rossi, M.R., Recchia, M. and Uderzo, C. (1981) Hepatitis B and childhood acute lymphoblastic leukaemia. J. Pediatr., 99, 98–100.[Web of Science][Medline]
69 Pontisso, P., Locasciulla, A., Schiavon, E., Cattoretti, G., Sciro, R., Stenico, D. and Alberti, A. (1987) Detection of hepatitis B virus DNA sequences in bone marrow of children with leukaemia. Cancer, 59, 292–296.[Web of Science][Medline]
70 Hu, C.-Y., Hsu, P.-N., Lin, R.-H., Hsieh, K.-H., Chua, K.-Y. (2000) HLA-DPB1*0201 allele is negatively associated with immunoglobulin E responsiveness specific for house dust mite allergens in Taiwan. Clin. Exp. Allergy, 30, 538–545.[Web of Science][Medline]
71 Wen, W., Shu, X.O., Linet, M.S., Neglia, J.P., Potter, J.D., Trigg, M.E. and Robison, L.L. (2000) Allergic disorders and the risk of childhood acute lymphoblastic leukemia (United States). Cancer Causes Control, 11, 303–307.[Web of Science][Medline]
72
Richeldi, L., Sorrentino, R. and Saltini, C. (1993) HLA-DPB1 glutamate 69: a genetic marker of beryllium disease. Science, 262,, 242–244.
73
Wang, Z., White, P.S., Petrovic, M., Tatum, O.L., Newman, L.S., Maier, L.A. and Marrone, B.L. (1999) Differential susceptibilities to chronic beryllium disease contributed by different Glu69 HLA-DPB1 and DPA-1 alleles. J. Immunol., 163, 1647–1653.
74 Potolicchio, I., Festucci, A., Hausler, P. and Sorrentino, R. (1999) HLA-DP molecules bind cobalt: a possible explanation for genetic association with hard metal disease. Eur. J. Immunol., 29, 2140–2147.[Web of Science][Medline]
75 Yun, C., Senju, S., Fujita, H., Tsuji, Y., Irie, A., Matsushia, S. and Nishimura, Y. (1999) Augmentation of immune response by altered peptide ligands of the antigenic peptide in a human CD4+ T-cell clone reacting to TEL/AML1 fusion protein. Tissue Antigens, 54, 153–161.[Web of Science][Medline]
76 Sturniolo, T., Bono, E., Ding, J., Raddrizzani, L., Tuereci, O., Sahin, U., Braxenthaler, M., Gallazzi, F., Protti, M.P., Sinigaglia, F. and Hammer, J. (1999) Generation of tissue specific and promiscuous HLA ligand databases using DNA microarrays and virtual HLA class II matrices. Natl Biotechnol., 17, 556–561.
77 Abbas, A.K., Murphy, K.M. and Sher, A. (1996) Functional diversity of helper T lymphocytes. Nature, 383, 787–793.[Medline]
78 Schountz, T., Kasselman, J.P., Martinson, F.A., Brown, L. and Murray, J.S. (1996) MHC genotype controls the capacity of ligand density to switch T helper (Th)-1/Th-2 priming in vivo. J. Immunol., 157, 3893–3901.[Abstract]
79 Eden, O.B., Harrison, G., Richards, S., Lilleyman, J.S., Bailey, C.C., Chessells, J.M., Hann, I.M., Hill, F.G.H. and Gibson, B.E.S. (2000) Long-term follow-up of the United Kingdom Medical Research Council protocols for childhood acute lymphoblastic leukaemia, 1980–1997. Leukemia, 14, 2307–2320.[Web of Science][Medline]
80 Taylor, G.M., Dearden, S., Payne, N., Ayres, M., Gokhale, D.A., Birch, J.M., Blair, V., Stevens, R.F., Will, A.M. and Eden, O.B. (1998) Evidence that an HLA-DQA1-DQB1 haplotype influences susceptibility to childhood common acute lymphoblastic leukaemia in boys provides further support for an infection-related aetiology. Br. J. Cancer, 78, 561–565.[Web of Science][Medline]
81 Bugawan, T. L., Begovich, A.B. and Erlich, H.A. (1990) Rapid HLA-DPB typing using enzymatically amplified DNA and nonradioactive sequence-specific oligonucleotide probes. Immunogenetics, 32, 231–241.[Web of Science][Medline]
82 Ott, J. (2000) Linkage utility programmes. http://linkage.rockefeller.edu/ott/linkutil.htm.
83 Mantel, N. and Haenszel, W. (1959) Statistical aspects of the analysis of data from retrospective studies of disease. J. Natl Cancer Inst., 22, 719–748.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  A. P. Chokkalingam and P. A. Buffler Genetic susceptibility to childhood leukaemia Radiat Prot Dosimetry, December 1, 2008; 132(2): 119 - 129. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. T. Dorak, R. J. Q. McNally, and L. Parker RE: "CHILDHOOD ACUTE LYMPHOBLASTIC LEUKEMIA AND INFECTIONS IN THE FIRST YEAR OF LIFE: A REPORT FROM THE UNITED KINGDOM CHILDHOOD CANCER STUDY" Am. J. Epidemiol., August 1, 2007; 166(3): 364 - 365. [Full Text] [PDF]
|
|
|
|
|
|
E Roman, J Simpson, P Ansell, S Kinsey, C. Mitchell, P. McKinney, J. Birch, M Greaves, T Eden, and on behalf of the United Kingdom Childhood Cancer S Childhood Acute Lymphoblastic Leukemia and Infections in the First Year of Life: A Report from the United Kingdom Childhood Cancer Study Am. J. Epidemiol., March 1, 2007; 165(5): 496 - 504. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. J. Q. McNally and T. O. B. Eden Re: Childhood Leukemia Incidence in Britain, 1974-2000: Time Trends and Possible Relation to Influenza Epidemics J Natl Cancer Inst, December 6, 2006; 98(23): 1746 - 1746. [Full Text] [PDF]
|
|
|
|
 |
|
 X. Ma, M. B Does, C. Metayer, C. Russo, A. Wong, and P. A Buffler Vaccination history and risk of childhood leukaemia Int. J. Epidemiol., October 1, 2005; 34(5): 1100 - 1109. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C Gilham, J Peto, J Simpson, E Roman, T O B Eden, M F Greaves, F E Alexander, and for the UKCCS Investigators Day care in infancy and risk of childhood acute lymphoblastic leukaemia: findings from UK case-control study BMJ, June 4, 2005; 330(7503): 1294. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. F. Greaves, A. T. Maia, J. L. Wiemels, and A. M. Ford Leukemia in twins: lessons in natural history Blood, October 1, 2003; 102(7): 2321 - 2333. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||