Frequency of the CCR5-∆32 Mutation in the Atlantic Island Populations of Madeira, the Azores, Cabo Verde, and São Tomé e Príncipe

There is evidence that he CCR5-A32 mutation confers protection against HIV-1 infection tohomozygous individuals. It is believed that this mutation spread through Europe with the Vikings and that it has been subjected to positive selection, leading to a high frequency in Europe (-10%). We carried out the present study to determine the 32-bp deletion allele and genotype frequencies of the CCR5 gene (CCR5-A32 ) in the Atlantic island populations of Madeira, the Azores, Cabo Verde, and Säo Tomé e Principe. These Atlantic archipelagos were all colonized by the Portuguese in the 15th and 16th centuries, but the latter two received most of their settlers from the West African coast. The frequency of the CCR5-A32 mutation varies between 0% in Säo Tomé e Principe and 16.5% in the Azores. The Azores Islands have one of the highest frequencies of homozygotes found in Europe (4.8%). There are significant differences (P < 0.05) between some of these populations, for example, between Säo Tomé e Principe and Cabo Verde, and even within populations (e.g., Portugal, Madeira, nd the Azores). The CCR5-A32 mutation consists of a 32-nucleotide deletion in the CCR5 gene located on chromosome 3p21 (Dean et al. 1996). This gene encodes a protein that serves as an entry port for HIV-1, and this mutation causes a truncation i that protein. Individuals who are homozygous for this mutation are highly resistant o HIV-1 infection (Samson et al. 1996; Dean et al. 1996; Libert et al. 1998), whereas heterozygotes have a slower rate of progression to disease (Grimaldi et al. 2002; Samson et al. 1996; Dean et al. 1996). The CCR5-A32 mutation also seems to confer protection against several diseases, namely, arterial occlusive disease (Ghilardi et al. 2004), rheumatoid arthritis (Pokorny et al. 2005), breast cancer (Degerli et al. 2005), and hepatitis C virus (Wald et al. 2004). The CCR5-A32 mutation seems to have had its origin in northern Europe, but there is no agreement about its age, which varies from about 700 years ago (with a range of 275-1,875 years) (Stephens et al. 1998) to 5,075 years ago (with 'Human Genetics Laboratory, University of Madeira, 9000-390 Funchal, Portugal. Human Biology, December 2006, v.78, no. 6, pp. 697-703. Copyright © 2006 Wayne State University Press, Detroit, Michigan 48201-1309

a range of 3,150 to 7,800 years) (Sabeti et al. 2005). Lucotte (2001) put forward the hypothesis of a Viking origin for the mutation, suggesting that the mutation already existed in Scandinavia before the Vikings dispersed 1,000-1,200 years ago. A study in Europe shows a high frequency of the CCR5-A32 mutation in areas usually associated with Viking culture, namely, Iceland (14.7%) and Sweden (14.2%) (Lucotte 2001), with decreasing frequencies running toward southern Europe (Libert et al. 1998). A study of sub-Saharan Africans shows a very low frequency (0.16%) of the mutation for this region (Lucotte 1997), and for North Africa the values are lower (-1.5%) than those in Southern Europe (Lucotte 2001), suggesting a European influence in North Africa.
In Europe the gene frequency of the CCR5-A32 mutation is about 10%, resulting in a heterozygous frequency of 18% and a homozygous frequency of 1% (Martinson et al. 1997). The mutation's relatively high frequency in Europe has been explained by positive selection of this mutation in those affected by bubonic plague (Stephens et al. 1998), but Galvani and Slatkin (2003) demonstrated that smallpox is a better candidate for positive selection for the CCR5-A32 mutation, thus explaining the cline prevailing today in Europe. HIV and poxvirus (responsible for smallpox) enter leukocytes by using chemokine receptors (Lalani et al. 1999), and it is biologically reasonable that the obliteration of the CCR5 chemokine receptor could confer resistance to both HIV and smallpox. Nevertheless, clinical characterizations of HIV and the plague-causing bacillus Yersinia pestis are quite distinct, and Elvin et al. (2004) and Mecsas et al. (2004) showed that deficiency in the chemokine receptor CCR5 in humans is unlikely to protect against plague.
In the present study we tested samples from the Atlantic islands of Madeira, the Azores, Cabo Verde (Cape Verde), and Säo Tomé e Principe, which were all settled by the Portuguese in the 15th and 16th centuries. In particular, Cabo Verde and Säo Tomé e Principe were settled with sub-Saharan slaves brought from the African coast of Guinea and the Gulf of Guinea, respectively, constituting the majority of its population (>90%; Russell-Wood 1998). We subdivided the population of Cabo Verde into two groups (Cabo Verde North and Cabo Verde South) because recent studies have shown that there are significant differences in the gene pools of these two groups Spfnola et al. 2002;Fernandes et al. 2003;Gonçalves et al. 2003). The aim of this study was to observe the frequency of the CCR5-A32 mutation in the Atlantic islands of Madeira, the Azores, Cabo Verde, and Säo Tomé e Principe and to compare the results with frequencies from Portugal (Western Europe) and Guinea-Bissau (West African coast).

Material and Methods
Blood samples were obtained with informed consent from unrelated individuals originating from the archipelagos of Madeira (n = 119), the Azores ( n = 124), Cabo Verde South ( n = 127), Cabo Verde North ( n = 97), and Säo Tomé e Principe (n = 100). Two populations were further analyzed for comparisons: one from mainland Portugal, further divided into three subregions (south, n = 106; north, n = 92; central, n = 91); and a sub-Saharan population from Guinea-Bissau (n = 50), located on the West African coast.
DNA extraction was done using conventional phenol-chloroform extraction methods. The CCR5-A32 gene fragment was amplified according to the method of Martinson et al. (1997), and DNA fragments were visualized with silver staining. We assessed Hardy-Weinberg equilibrium of samples and basic genetic parameters such as genotype and allele frequencies for each population using the software program GenePop (Raymond and Rousset 1995). Frequency differences were calculated according to Fisher's exact test using Arlequin (Schneider et al. 1997). Probability values of P<0.05 were considered statistically significant.

Results
All populations are in Hardy-Weinberg equilibrium except the southern Portugal population. This population presented a high frequency of homozygotes for the CCR5-A32 mutation, but the number of heterozygotes was lower than expected, causing the disequilibrium. Other differences in the genotype frequencies for the CCR5-A32 mutation among the populations are listed in Table 1 . The allele frequency of the CCR5-A32 mutation varies between 0% in Säo Tomé e Principe and 16.5% in the Azores (Figure 1).
Interestingly, although expected, the population of Cabo Verde had a relatively high frequency for the mutant allele. In comparison, Guinea-Bissau, the most probable place of origin of slaves brought to Cabo Verde, lacks the CCR5-A32 mutation completely. This clearly shows that the CCR5-A32 mutation from Cabo Verde was introduced through admixture by male Caucasian settlers from Europe. In central Portugal the frequency of the mutation is also lower than expected. As verified in a previous study (Fernandes and Brehm 2002), central and northern Portugal show significant genetic differences (P<0.05). The Azores and Madeira also show significant differences with central Portugal. In the present study we did not find significant differences within the Cabo Verde archipelago nor between Cabo Verde and central and southern Portugal. The present study revealed significant differences between the Cabo Verde and Säo Tomé e Principe archipelagos. This is perhaps not surprising because the admixture among Caucasian settlers and sub-Saharan slaves was more pronounced in Cabo Verde than in Säo Tomé e Principe, giving rise to a typical Creole population Spinola et al. 2002;Fernandes et al. 2003;Gonçalves et al. 2003;Trovoada et al. 2004).  Figure 1. Allele frequencies of the CCR5-A32 mutation from the Atlantic islands of Madeira, the Azores, Cabo Verde (subdivided into north and south groups), and Sào Tomé e Principe plus mainland Portugal (north, central, and south) and Guinea-Bissau.

Discussion
It seems clear that the frequency of the CCR5-A32 mutation in the population from the Azores Islands is most probably the result of a strong founder effect. The structure of the Azorean population suggests that the founder effect and genetic drift were due to a great influence from northern European populations, a notion that is strengthened by studies of Y-chromosome (Montiel et al. 2005), mtDNA (Santos et al. 2003), andHLA (Spinola et al. 2005) markers. Positive selection must also have occurred, and history describes a smallpox epidemic in 1844-1845 that was devastating in the islands (Amaral 1960).
The population from southern Portugal is in Hardy-Weinberg disequilibrium, although Fernandes and Brehm (2002), in a study performed with exactly the same samples but using neutral autosomal STR markers, found the contrary. This is further evidence that positive selection is acting, probably as a result of the entry of smallpox carried by North African populations (Amaral 1960;Hopkins 2002) that settled in the region on several occasions. Unlike other studies Spinola et al. 2002;Fernandes et al. 2003), we found no significant differences between Cabo Verde South and Cabo Verde North. This result can be explained by the fact that the CCR5-A32 mutation was carried mainly by Caucasian male settlers from Europe and there is a large Caucasian influence in both groups of islands: 25% in the southern group and 42% in the northern group, according to Gonçalves et al. (2003). Although the explanation for a relative high frequency of the CCR5-A32 mutation in Cabo Verde lies primarily in the strong influence of Caucasians, a second reason could be the occurrence of positive selection for the mutation. Cabo Verde was an important slave-trading center, with slaves coming mainly from Guinea. History reports the introduction of smallpox in Brazil by African slaves in 1555 (Hopkins 2002), and the population of Cabo Verde must have been exposed to the virus at this stage of early colonization.
The Azores and southern Portugal populations have a higher frequency of homozygotes (4.8% and 2.8%, respectively) than other European populations (1%) (Martinson et al. 1997). Because these homozygous individuals seem to be resistant to AIDS, this is a good perspective on these regions in Portugal. The high percentage of the CCR5-A32 mutation in Cabo Verde (compared to other African populations) should always be taken into consideration when studying epidemiological models of the spread of AIDS in Cabo Verde.