The immune system is largely regulated by cytokines and various receptor interactions. Killer immunoglobulin-like receptors (KIRs) participate in regulating the activity of natural killer cells and certain T-cell subsets, thereby affecting innate and adaptive immunity. This cellular regulation can influence the course of infections, autoimmunity, cancers and other diseases.1, 2, 3, 4, 5

The KIR cluster contains 15 genes and 2 pseudogenes,6 which are segregated within the centromeric and telomeric portions of the cluster. The centromeric and telomeric regions are flanked by framework genes KIR3DL3 and KIR3DL2, respectively. These regions are also separated by framework genes KIR3DP1 and KIR2DL4. The framework genes are highly conserved throughout human populations. The individual KIR genes inside these regions code for either activating or inhibitory receptors whose interaction with corresponding MHC ligands on target cells can, respectively, promote or inhibit cell function. Individual KIR gene associations have been made with specific disease resistance and susceptibility within a population.1, 2, 3, 4, 7 In addition, population differences in gene and genotypes profiles suggest that KIR clusters may contribute to distinct disease outcomes between ethnic populations.8, 9, 10 Considerably, less is understood about the disease interactions of KIR haplotypes. KIR genes have been designated into inheritable A or B haplotypes. Haplotype A is defined by the presence of either KIR2DL3 or KIR2DL1 in the centromeric portion and KIR3DL1 or KIR2DS4 in the telomeric region.6 In contrast, the B haplotype is defined by KIR2DS2 or KIR2DL2 in the centromeric region and KIR3DS1, KIR2DL5A, KIR2DL5B, KIR2DS5 or KIR2DS1 in the telomeric segment. Centromeric and telomeric influences can also provide another level of complexity.11, 12

Ethnically defined disease burdens have been largely attributed to environmental factors.13, 14 However, epidemiological data on reduced rates of progression to chronic hepatitis C infection in North American Aboriginal people despite similar environmental factors15, 16, 17, 18 suggest that underlying biological events participate in population-specific disease outcomes. In an effort to address genetic factors that might contribute to these differences in health outcomes, we evaluated the KIR genes of Caucasian and Aboriginal populations residing in the Canadian province of Manitoba.



This study was approved by and conducted according to the policies of the University of Manitoba Ethics Board. It was also performed in discussion with the Assembly of Manitoba Chiefs, Health Information Research Committee. Canadian Aboriginal people consist of First Nation, Inuit and Métis. In this study, Caucasian (n=93) and Aboriginal (n=86) adults were recruited through local advertisement. Caucasian individuals defined as being of European descent acted as a population control. First Nation individuals were predominately from the Oji-Cree First Nation (70%). The remaining Aboriginal individuals were primarily Oji-Cree, Ojibway, Cree or Métis. All individuals were negative for serologic evidence of HCV exposure or chronic HBV infection, and were not receiving medications for other chronic infections. The percentage of male participants was 39.3% in the Caucasian cohort and 45.5% in the Aboriginal cohort (ns, χ2, Pearson's correction). The median age (range) was also not significantly different; 43 (22–75) and 36 (20–61) in the Caucasian and Aboriginal cohorts, respectively.

Genetic analysis

After the signing of consent, whole blood was donated. Genomic DNA was extracted from isolated buffy coat cells following the QIAGEN protease digestion and absorption onto QIAamp silica-gel (Qiagen, Mississauga, ON, Canada). The product was eluted on a column to enhance the purity and concentration of the extracted DNA. The presence of KIR genes was determined using the Dynal Biotech (Brown Deer, WI, USA) Pel-Freez SSP UniTray KIR genotyping kit according to the manufacturer's instructions. Gene products were visualized with UV light on a transilluminator and photographed for subsequent analysis. In this manner, the presence or absence of the following KIR genes was assessed: KIR2DL1, KIR2DL2, KIR2DL3, KIR2DL5A, KIR2DL5B, KIR2DS1, KIR2DS2, KIR2DS3, KIR2DS4, KIR2DS5, KIR3DL1, KIR3DS1 and KIR2DP1, as well as framework genes KIR2DL4, KIR3DL2, KIR3DL3 and KIR3DP1.

Data analysis

KIR gene frequencies for a population were calculated as 1−(√1−f), where f equals the observed KIR gene frequency. Differences in KIR cluster parameters between Caucasian and Aboriginal cohorts were analyzed by χ2, Pearson's exact test. A P<0.01 was considered as statistically significant. Linkage disequilibrium was assessed as D′, a statistical calculation of linkage between two single nucleotide polymorphisms (SNPs) (SNP & Variation Suite 7; Golden Helix Incorporated, Bozeman, MT, USA). An analysis of the association between two SNPs, r2, was also performed. Haplotype designations were made as previously detailed.19, 20 (1) Framework genes were recognized as homozygous. (2) For the centromeric portion of the cluster, as KIR2DL3 and KIR2DL2 are thought to be alleles of the same locus, if only one was present, then that allele was assumed to be homozygous and the centromeric portion was assumed to be an AA or BB haplotype, respectively. If both KIR2DL3 and KIR2DL2 were present, the cluster was considered to be an AB haplotype. (3) For the telomeric portion of the cluster, as KIR3DL1 and KIR3DS1 are considered alleles of the same locus, if only one was present, then homozygosity was assumed and the telomeric portion of the cluster was identified as an AA or BB haplotype, respectively. If both were present, the cluster was considered to be an AB haplotype. (4) Independent of haplotype, KIR2DS4 was considered homozygous if other KIR2DS genes were not present. However, if another KIR2DS gene was apparent in the presence of an ABtelomeric haplotype, the KIR2DS4 was considered to be present on the haplotype A and absent on the haplotype B.


Indigenous people have a differential presence of KIR B haplotype genes

Caucasian and Aboriginal populations differ in the prevalence of certain infectious diseases, autoimmunities and cancers. To better understand the impact of underlying biology in disease demographics, KIR gene profiles were assessed in Caucasian (n=93) and Aboriginal (n=86) individuals. As expected, the framework genes KIR2DL4, KIR3LD2, KIR3DL3 and KIR3DP1, and pseudogene KIR2DP1 were present with few exceptions. In addition, KIR2DL1, KIR2DL3, KIR2DS4 and KIR3DL1 genes within the A haplotype were present in over 84% of all individuals, a finding paralleled in other ethnic groups (Table 1; see Supplementary Table 1 for gf differences between Manitoban populations and details regarding the presence of KIR2DL5A and KIR2DL5B in these cohorts).19, 20, 21, 22

Table 1 Aboriginal gene profiles reflect that of other indigenous people of the Americas

Conversely, marked variation was apparent in the genes associated with the B haplotype. Compared with Caucasian samples, the Aboriginal population exhibited a decreased presence of B haplotypecentromeric genes KIR2DL2, KIR2DS2, as well as KIR2DS3 (Table 1). This aligned with the contrasting homogeneity of the A and B haplotypes for the centromeric section (see Supplementary Table 2). KIR2DL3 and KIR2DL2 are alleles of the shared locus within A and B haplotypes, respectively. The KIR2DL3 allele was present in a high proportion of the Caucasian (90.0%) and Aboriginal (95.3%) samples (Table 1). However, relative to the Caucasian cohort this allele was more often present in a homogenous manner in the Aboriginal cohort (P<0.05; see Supplementary Table 2) corresponding to the lower presence of KIR2DL2 in this population (Table 1). Concomitantly, the opposite was observed with the telomeric KIR3DL1 and KIR3DS1 alleles. Here, KIR3DL1 displayed greater homogeneity in the Caucasian cohort (see Supplementary Table 2; P<0.01). This was supported by the reduced presence of the KIR3DS1 allele in the Caucasian, relative to the Aboriginal, population (Table 1). Concomitantly, there is an increased presence of B haplotypetelomeric genes in the Aboriginal cohort.

To evaluate the potential distribution of the Aboriginal KIR phenotype, the Manitoban KIR gene profiles were analyzed against previously published studies (Table 1). The gene pattern from the Caucasian cohort paralleled Caucasians from other continents;19, 23 whereas, the Aboriginal gene profile bore similarity to patterns from other indigenous peoples of the Americas.19, 20 The resemblance was particularly acute with respect to the genes from the B haplotype. As with the Aboriginal population, the reduced presence of the Bcentromeric genes, coupled with the prominence of the Btelomeric genes was also observed in other American indigenous peoples.

Manitoban Aboriginal genotypes parallel those of other American indigenous peoples

The Caucasian and Aboriginal KIR genotypes were numbered according to the global prevalence as per the meta-analysis published by Yawata et al.8 Reflecting this global prevalence, genotypes 1 and 2 were foremost in our Caucasian cohort (Table 2; Supplementary Figure 1 for full description of genotypes and six genotypes designated N01–N06 not reported by Yawata et al). The dominant genotype in the Aboriginal population, however, was genotype 4 (32.6%). Genotypes 2 and 3 consisting of a quarter of the Caucasian genotypes were rare in the Aboriginal population (2.3%).

Table 2 Indigenous people of the Americas display a unique genotype prevalence

The distinctiveness of the Manitoban Aboriginal KIR genotype profile was further apparent when compared with other indigenous populations (Table 2). Manitoban Caucasian genotypes were analogous to Caucasian profiles from other continents, and were also similar to East and West Asian profiles.8 Conversely, genotypes 4, 1 and 8 accounted for 72% of the Aboriginals genotypes. This hierarchy was comparable to the Mexican Purepecha and Tarahumara, and Argentinean Chiriguanos.19, 20 Interestingly, all the indigenous populations, including the Huichol and Wichis, had an increased presence of genotypes 4 and/or 8 that possess genes belonging within the Btelomeric segment. This occurred simultaneously to a muted presence of genotypes 2 and 3 with limited connections to the Btelomeric segment. Indeed, the genotypes with activating genes (KIR3DS1, KIR2DS1 and KIR2DS5) belonging to the haplotype Btelomeric, in a minority of Caucasian samples (36.6%), emerged with a high frequency in Aboriginal counterparts (67.4%, P<0.05; Supplementary Figure 1). A similar finding was observed with the Yucpa, Bari and Warao Amerindians from Venezuela, where 72.1, 65.2 and 70.0% of the genotypes, respectively, possessed the activating genes belonging to the haplotype Btelomeric.24

Caucasian and Aboriginal haplotype analysis

Despite these differences between ethnic cohorts, there appeared a high degree of linkage (D′=1.00) for genes within the centromeric portion for both the Caucasian (P=1.5 × 10−18) and Aboriginal (P=3.0 × 10−19) populations (Table 3). The exception was the KIR2DS3 gene which interacted with other centromeric genes with a D′=0.780 (P=2.5 × 10−8) in the Caucasian cohort. This interaction was further diminished in the Aboriginal cohort (D′=0.272, P>0.05). The uniqueness of the KIR2DS3 interaction may reflect the lower percentage of KIR2DS3 (39%) within the Caucasian cohort and its further reduced presence (7%) in the Aboriginal genome. The telomeric segment also demonstrated strong intra-regional interactions for the Caucasian (D′=0.972, P=6.2 × 10−16) and Aboriginal (D′=1.00, P=6.2 × 10−18) populations. These linkages suggest that genes within either the centromeric region (except for KIR2DS3) or the telomeric region are generally inherited as a regional unit.

Table 3 Centromeric and telomeric regional interactions

In contrast, between centromeric and telomeric segments linkage was lacking within Caucasian samples (D′=0.170, P>0.05), but was still notable within the Aboriginal population (D′=0.678, P=5.0 × 10−4). KIR2DS3 again displayed a differential interaction with genes of the telomeric region.

Due to these interactions, A and B haplotypes were analyzed relative to the centromeric and telomeric positions.11 Compared with the Caucasian cohort, there was a reduced prevalence of the AB-AA haplotype in the Aboriginal cohort (18.3 vs 2.3%, P<0.001). Within the Aboriginal population Bcentromeric haplotypes were present only in combination with a Btelomeric haplotype counterpart, with the exception of genotype 2 (see Supplementary Figure 1). Conversely, Btelomeric haplotypes occurred in 53.9% of the Caucasian and 72.1% of the Aboriginal populations (P<0.001; Table 4). The greater presence of the Btelomeric haplotypes in the Aboriginal cohort was also evident in that only 3.3% Caucasian population possessed the rare BBtelomeric haplotype, in contrast to 16.3% of the Aboriginal cohort (P<0.01).

Table 4 Aboriginal individuals have an increased prevalence of Btelomeric haplotypes

Current KIR genetic profiles may reflect historical immune selection

The similarity in the KIR cluster within indigenous people of the Americas likely represents a historical selection bias. One mechanism to introduce such a bias would be a common ancestry. Genetic studies have theorized that American indigenous peoples migrated from East Asia.25, 26 This particular migration relationship was not apparent in this KIR analysis, as the KIR genetic phenotypes of indigenous groups were not similar to East and West Asian populations (Table 2).8 Alternatively, the immune systems of indigenous populations may have evolved in convergent manner due to the long-term infectious selection pressure on them by indigenous infectious agents. In support of this theory, archeological evidence indicates that before the arrival of the Europeans; indigenous peoples were more likely subject to bacterial and parasitic infections, such as tuberculosis and pneumonia, as opposed to viral infections.27

Lastly, the similarity in KIR genotypes may be the result of an abrupt immune selection following the sudden introduction of novel infectious agents. During the European colonization, viral infections including smallpox, measles and influenza introduced to indigenous peoples resulted in widespread deaths ranging between 25 and 90% of these populations throughout the Americas.27, 28, 29, 30 In this scenario, a greater genetic tendency for active/pro-inflammatory immunity could have provided an immunological edge enabling individuals to survive the devastation of ‘old world’ viral infections. Finally, it may be a combination of these factors that contribute to the KIR cluster similarities within indigenous peoples of the Americas.

Implications for current disease resistance and susceptibility

Whatever the mechanism, historically the immune profile of indigenous peoples must have provided a survival advantage suggesting that the prevalence of Btelomeric genes had a particular benefit. Conversely, survival may have been curtailed by Bcentromeric genes, potentially associated with severer forms of autoimmune disease and diabetes.31, 32

This immunological legacy would also be evident in existing disease patterns, although interactions with recent environmental influences may render these gene profiles less beneficial. Interestingly, the absence of KIR2DS3 could have implications for current disease manifestations as KIR2DS3 appears protective against certain types of immune-mediated disease.33, 34 However, it is unknown if the presence of the activatory Btelomeric genes increases susceptibility to autoimmune diseases prevalent in indigenous peoples throughout the Americas.35, 36, 37, 38 Nonetheless, North American Aboriginal ancestry is also associated with a reduced progression to chronic HCV infection.15, 16, 17, 18 While data on HCV sequelae in other indigenous peoples of the Americas are lacking, the possibility that inherited KIR gene profiles could contribute to these disease patterns is strengthened by the observation that similarities exist between the KIR gene profiles of this Aboriginal cohort and those reported in association with spontaneous clearance of HCV.1, 39, 40


The contribution of KIR genotypes and haplotypes to disease resistance and susceptibility on an individual basis, much less on a population basis, has yet to be deciphered. This assessment of the KIR cluster indicates that Aboriginal peoples have a stronger activation phenotype compared to Caucasians. Other genetic and functional studies support the finding that Aboriginal and other indigenous people have an enhanced pro-inflammatory microenvironment.41, 42, 43 The uniqueness of this phenotype within these populations suggests that it provided a historical survival advantage and that it currently participates in the resistance to chronic HCV infections observed in North American Aboriginals. In light of current environmental influences, however, this KIR genetic phenotype may also detrimentally contribute to the greater prevalence of chronic immune-mediated disease within the indigenous peoples of the Americas.15, 16, 17, 18