Tissue Antigens ISSN 0001-2815
BRIEF COMMUNICATION
Genetic response to an environmental pathogenic agent: HLA-DQ and onchocerciasis in northwestern Ecuador F. De Angelis, A. Garzoli, A. Battistini, A. Iorio & G. F. De Stefano Department of Biology, University of Rome Tor Vergata, Rome, Italy
Key words adaptation; Ecuador; human leukocyte antigens; onchocerciasis Correspondence Flavio De Angelis Department of Biology University of Rome Tor Vergata Via della Ricerca Scientifica 1 00133 Rome Italy Tel: +39 0672 594 310 Fax: +39 062 023 500 e-mail:
[email protected] Received 22 February 2011; revised 7 October 2011; accepted 13 November 2011 doi: 10.1111/j.1399-0039.2011.01811.x
Abstract The aim of this study is to explore human leukocyte antigen (HLA)-DQ variability in two populations (Cayapas Amerindians and Afro-Ecuadorians) who live near one another along the Cayapa River and who are exposed to the same environmental stresses, such as infection by Onchocerca volvulus. HLA-DQA1 and HLA-DQB1 of 149 unrelated individuals (74 Cayapas and 75 Afro-Ecuadorians) have been analyzed. HLA high-resolution molecular typing was performed by sequencebased typing, sequence-specific oligonucleotides hybridization and sequence-specific primer (SSP) amplification. The comparison between affected (cases) and unaffected people (controls) in both populations shows the key role of several HLA-DQA1 alleles in susceptibility and protection against onchocerciasis. In both populations, there is strong evidence related to the protective role of DQA1*0401 against onchocerciasis. Alleles HLA-DQA1*0102 and *0103 seem to represent risk factors in Afro-Ecuadorians, while HLA-DQA1*0301 is only a suggestive susceptibility allele in Cayapas. These findings represent new positive/negative associations with onchocerciasis in South America, whereas previous findings pertained only to African populations.
The notion that selection during epidemics or long periods of exposure to infectious diseases might play a major role in development of allelic diversity of human populations is not new (1). Clearly, an analysis of the human genome with respect to susceptibility to infections has already provided important new insights into the mechanisms of human diversity. Recent studies have shown that host genetics is an important determinant of the intensity of infection and morbidity due to helminthes (2): in fact, several studies about the basis of susceptibility to helminth infection have been reported by human leukocyte antigen (HLA) association studies (3–5). Onchocerciasis is caused by nematode Onchocerca volvulus and causes blindness and debilitating skin lesions (6–8). The latest counts by World Health Organization Programmes about onchocerciasis estimated that in 2009 around 37 million individuals were affected (9). In the Americas, onchocerciasis was introduced by the slave trade (10, 11) and occurs in 13 discrete foci in six countries: Brazil, Colombia, Ecuador, Guatemala, Mexico and Venezuela. Differences in the clinical presentation of the disease have been described between African and American populations (12). In the Ecuadorian focus of Esmeraldas, in the northwestern rainforest, the native Cayapas and the Afro-Ecuadorians, whose origin traces back to the 17th century transatlantic slave trade, live very
© 2011 John Wiley & Sons A/S · Tissue Antigens 79, 123–129
close to each other, yet there is very little intermarriage between them (13–16). Both populations are coresidential in a marginal tropical forest area known to be hyperendemic for onchocerciasis (17–19): several studies have verified that the prevalence of infection among the indigenous population and Afro-Ecuadorians is virtually identical. Nevertheless, it was possible to highlight some important population differences in the reaction to infection, most notably in the distribution and location of onchocercal nodules on the body. Among Cayapas, the nodules are mostly located in the nuchal and iliac crest region, while among Afro-Ecuadorians the distribution is mainly sacrococcygeal. Among the Cayapas, the microfilariae related to eye pathology are located almost exclusively in the anterior chamber, while among Afro-Ecuadorians they are commonly found in the posterior chamber (20). Therefore it seems reasonable to assume that clinical response to infection by O. volvulus differs significantly between populations. These differences may be explained by particular genetic factors, as has been shown to occur with other parasitic diseases such as schistosomiasis and elephantiasis (21). Several researchers have indicated the influence of HLA class II alleles in nematode infections (22, 23). A correlation between allelic variants of HLA-DQA1 and HLA-DQB1 and changes in clinical manifestations of the disease caused by
123
F. De Angelis et al.
HLA-DQ alleles in onchocerciasis from Ecuador
O. volvulus was highlighted (24–26). This evidence clearly indicates that HLA class II variants may influence infection by O. volvulus and may reflect genetic protective immunity or susceptibility. This research highlights the presence of susceptibility/protective alleles in people scattered in the onchocerciasis hyperendemic focus of Esmeraldas. It is the first HLA-onchocerciasis association screening in the Americas. Further, this study exploits the unique opportunity to test two populations of different ethnic backgrounds colocated in a single environment, which may be very meaningful for detecting immunogenetic strength in response to onchocerciasis. Recruiting was carried out in Rio Cayapas area by one of us (GFDS) along with medical personnel of Quito Vozandes Hospital (Quito, Ecuador) as the secondary care level. This area represented the major onchocerciasis focus in Esmeraldas and it consisted of several positive communities with an average infection rate of 51.1% (27). DNA was extracted from whole-blood samples (28) of 149 unrelated individuals of both sexes, between 7 and 66 years of age, who gave their informed consent for this project. Among these, 74 are Cayapas (40 affected and 34 not affected by onchocerciasis) and 75 are Afro-Ecuadorians (30 affected and 45 not affected), with none presumed to be first-degree relatives based on interviews and surname comparisons. Individuals were classified as affected (cases) or not affected (controls) according to the presence of nodules on the body surface after the examination by Vozandes Hospital medical doctors, who collaborated in the field study according to the hospital’s Ethics Committee. Cases and controls have been matched for age, sex and residence in each community.
HLA typing
High-resolution sequence-based typing was used for the determination of the second and the third exons of DQA1 and DQB1 loci: these exons respectively encode for the α1 and
β1 regions of the HLA-DQ antigen cleft. The primers for amplifications and sequencing are listed in Table 1: they were used according to references (29, 30). Conditions for the polymerase chain reaction (PCR) assay by custom-designed primers were as follow: 10 min at 95◦ C, 35 cycles of 30 s at 95◦ C, 1 min at 53.5◦ C, 30 s at 72◦ C and a final extension of 7 min at 72◦ C. The sequencing reactions used BigDye Terminator v.1.1 Cycle Sequencing Kit (Applied Biosystems, Carlsbad, CA), according to manufacturer’s instructions and amplification forward primers. All the sequences were compared to the IMGT/HLA Database v.3.5.0 (31) (http://www.anthonynolan.org.uk/) in order to determine the allelic assignments at the two HLA class II loci. Further, reverse sequence-specific oligonucleotide hybridization technique (Inno-Lipa Multiplex and DQB1 Update kits; Innogenetics, Gent, Belgium) and PCR using sequencespecific primers (Dynal, New York, NY) (32) were also used respectively to determine HLA-DQB1 and HLA-DQA1 genotypes. Both techniques were used according to the manufacturer’s instructions for every sample.
Statistics
Overall distributions of HLA-DQ alleles were determined by the direct counting method. Deviation from Hardy–Weinberg equilibrium was examined with Arlequin software v.3.5.1.2 (33) with the default setting, where the exact P -value is calculated based on the Markov chain method (34). GraphPad Instat v.3.00 was employed to calculate alleles’ odds ratios (ORs) and their P -values using Fisher exact tests (GraphPad Software, San Diego, CA). Bonferroni correction was used to calculate corrected P -values (Pc ). Commonly, Bonferroni correction is performed when the marker has more than two alleles. Although the significance level of an elementary test is not changed just because other tests have been carried out on the same data, a correction for multiple testing is required
Table 1 Primer used to amplification and direct sequencinga Primer sequence 5 –3 TGTAAAACGACGGCCAG TCGCCGCTGCAAGGTCG GCTCACTCTCCTCTGCAA GCTCACTCTCCTCTGCAG TTTTCCTGTCTGTTACTGCCC TCAATATCCCCTTACGCCACT CATCTTCACTCATCAGCTGACC GTAGAGTTGGAGCGTTTAATCAG GTAGAGTTGTAGCGTTTAATCAT GTAGAGTTGGAGCGTTTAATCAC AGGTTCCTGAGGTCACAGTGTTT CTTGACAGACAAGAAAGCATC
Amplification
References
Code
DQB1 exon 2 fwd DQB1 exon 2 rev DQB1 exon 2 rev DQB1 exon 2 rev DQB1 exon 3 fwd DQB1 exon 3 rev DQA1 exon 2 fwd DQA1 exon 2 rev DQA1 exon 2 rev DQA1 exon 2 rev DQA1 exon 3 fwd DQA1 exon 3 rev
Van Dijk et al. (29) Van Dijk et al. (29) Van Dijk et al. (29) Van Dijk et al. (29) Custom designed Custom designed Cordovado et al. (30) Cordovado et al. (30) Cordovado et al. (30) Cordovado et al. (30) Cordovado et al. (30) Cordovado et al. (30)
2Bf 2Br1 2Br2 2Br3 3Bf 3Br 2Af 2Ar1 2Ar2 2Ar3 3Af 3Ar
fwd, forward primer; rev, reverse primer. a The amplification couples are as follows: 2Bf-(2Br1 or 2Br2 or 2Br3) for DQB1 exon 2; 3Bf-3Br for DQB1 exon 3; 2Af-(2Ar1 or 2Ar2 or 2Ar3) for DQA1 exon 2; 3Af-3Ar for DQA1 exon 3. Thermal conditions are specified in the references or in the text.
124
© 2011 John Wiley & Sons A/S · Tissue Antigens 79, 123–129
F. De Angelis et al.
when a particular result is highlighted to be significant in a set of similar tests, and it is suspected to be due just to chance. In this case, the obtained P -values should be corrected based on the number of tests by Bonferroni correction, that is, equal to the number of tested alleles. This correction made more stringent the statistical significance of the results. The calculated ORs measure the association between an allele and a disease, and it is defined as the ratio of the odds of allele presence among people with the disease to the odds of allele presence in people without the disease. The OR varies between 0 and ∞ with a threshold of 1 to indicate independence between variables, so an OR less than 1 indicates protection against a disease, while an OR more than 1 means that the allele plays a susceptibility role in the disease. Genotype frequencies significantly deviate from predictions of the Hardy–Weinberg equilibrium in every locus, which may suggest some selective pressure on HLA genes. Allelic distributions at HLA-DQA1 and HLA-DQB1 in both populations are shown in Tables 2 and 3, respectively. The haplotype analysis was performed by Phase v.2.1 (35), but it does not add any information because there is no statistically different haplotype between cases and controls (unpublished data). HLA-DQA1 and HLA-DQB1 genetic diversity among Afro-Ecuadorians
The population sample of the Afro-Ecuadorian community comprises 75 unrelated individuals. Table 2 shows the HLADQA1 distribution of the data sets. This points out an expected allele diversity that encompass 16 variants at locus HLADQA1, with the leading frequency of two alleles, HLADQA1*0102 and *0401, with 32% (n = 48) and 29% (n = 43), respectively, while the other allelic variants are somewhat less represented, with the sole exception of *0103 with 11%. Similarly, 15 alleles are seen at locus HLA-DQB1(Table 3); the leading presence of DQB1*0301 (26%) is due to 39 alleles, followed by *0402 with 32 alleles (21%), and *0611 reaches 10%. The subdivision of the Afro-Ecuadorian population according to onchocerciasis status uncovers several associations. In fact it is very interesting to note that DQA1*0102 may play a susceptibility role reaching statistical significance (Pc = 0.003) along with DQA1*0103 (Pc = 0.002). The OR calculated for *0102 is 4.00 [confidence interval (CI) 1.94–8.24; P < 0.01], and 13.39 is the OR for *0103 (CI 2.92–61.49; P < 0.01): both may indicate a specific susceptibility role of these alleles in onchocerciasis. On the contrary, DQA1*0401 showed an OR of 0.12 (CI 0.04–0.34; P < 0.0001) (Pc = 0.002) (Table 2), suggesting the protective role of this allele in onchocerciasis. No DQB1 alleles were significantly different between cases and controls. A trend for protectiveness was seen for DQB1*0301, with an estimated OR of 0.50 (CI 0.22–1.10; Pc = ns), but it did not reach statistical significance possibly because of small sample size. Notably, this allele represent a protective variant against
© 2011 John Wiley & Sons A/S · Tissue Antigens 79, 123–129
HLA-DQ alleles in onchocerciasis from Ecuador
onchocerciasis in African samples as highlighted by Meyer and colleagues (24). HLA-DQA1 and HLA-DQB1 genetic diversity among Cayapas
The population sample of Cayapas comprises 74 unrelated individuals, and Table 2 points out their HLA-DQA1 allelic pool. This encompasses 11 allelic variants with the foremost frequency of three alleles: DQA1*0301, *0302 and *0401 (30%, 18% and 24%, respectively); with the rest in somewhat low frequencies (Table 2). As for locus HLA-DQB1, Cayapas show nine alleles (Table 3), with the leading frequencies of alleles being *0301 (14%), *0302 (31%), *0303 (24%) and *0402 (22%). The *03 group encompasses the majority of the variation with 75.7% of the entire allelic pool. The population was subdivided into two classes according to presence/absence of onchocerciasis features, and some interesting findings can be outlined by observation of the differences. In fact, the allele HLA-DQA1*0401 seems to represent a significant protective allele: the OR for onchocerciasis associated with the DQA1*0401 allele is 0.29 (CI 0.13–0.65; Pc < 0.05); individuals with this allele have significantly lower odds of being affected by the disease. Conversely, the allele DQA1*0301 seems to be a susceptibility HLA-DQA1 variant, with a leading frequency in affected individuals but without the statistical significance (Pc > 0.05) (Table 2). For locus HLA-DQB1 (Table 3), no allele reaches statistical significance. This typing analysis also shows how variable the HLA region is, and this feature is underlined by the multiallelic structure of the HLA genes, which may not only be maintained by genetic drift but also by natural selection. In fact, as in all New World populations, few alleles may reach high frequencies, and others are maintained in somewhat low frequencies (36). Thus each population may be different because the high-frequency allele may be dissimilar in each population; however, the high frequencies found for DQA1*0401 in both Esmeraldas samples suggest that this distribution may not be due solely to genetic drift or gene flow. The genetic separation between the Esmeraldas populations (13, 15, 37) shows that migration did not play a significant role in shaping the genetic background of these human groups. In addition, the frequencies of DQA1*0401 in Central and South American populations vary considerably but are much lower than the frequencies in the Esmeraldas populations (38). Despite the general heterogeneous scenario, the critical DQA1*0401 frequencies separation between onchocerciasis and healthy people in the Esmeraldas communities is interesting. A putative interpretation may be that an effective selective pressure positively selected this DQA1 allele in both Esmeraldas populations. The Mantel–Henszel OR = 0.2 (Pc < 0.0001) (Table 2) calculated in both Afro-Ecuadorians and Cayapas populations may be consistent with the positive selective pressure of
125
126
0 30 14 1 0 0 0 4 0 5 0 1 0 0 5 0 60
0.00 50.00 23.33 1.67 0.00 0.00 0.00 6.67 0.00 8.33 0.00 1.67 0.00 0.00 8.33 0.00 100.00
2 18 2 0 7 1 2 1 1 38 4 1 4 2 6 1 90
Pvalue 3 2 0 0 0 0 30 12 0 11 5 3 6 5 3 0 80
3.75 2.5 0 0 0 0 37.5 15 0 13.75 6.25 3.75 7.5 6.25 3.75 0 100.00
3 2 0 0 0 0 14 14 0 24 1 4 0 3 1 2 68
OR (CI 95%)
— 2.04 8.67 — — — — — — 0.20 — — — — — — —
PCrude Mantel– value Pc (11) Haenszel OR
4.41 — ns ns 2.94 — ns ns 0.00 — ns ns 0.00 — ns ns 0.00 — ns ns 0.00 — ns ns 20.59 — ns ns 20.59 — ns ns 0.00 — ns ns 35.29 0.29 (0.13–0.65) 0.003 0.03 1.47 — ns ns 5.88 — ns ns 0.00 — ns ns 4.41 — ns ns 1.47 — ns ns 2.94 — ns ns 100.00 — — —
N N % % Pc (16) cases cases controls controls
2.22 — ns ns 20.00 4.00 (1.94–8.24) 0.0002 0.003 2.22 13.39 (2.92–61.49) 0.0001 0.002 0.00 — ns ns 7.78 — ns ns 1.11 — ns ns 2.22 — ns ns 1.11 — ns ns 1.11 — ns ns 42.22 0.12 (0.04–0.34)
