Thus far, innate resistance to malaria infections in humans has been attributed to blood group polymorphisms. Duffy blood group polymorphisms are important in areas where P. vivax predominates, because this molecule acts as a receptor for this protozoan (but not for the other human malaria parasites) on the surface of RBCs . Field observations from West Africa and Ethiopia have indeed established a strong correlation between absence or low endemicity of P. vivax malaria and the high prevalence of the Duffy negative allele [20, 21]. Little is known on the frequency of RBC polymorphisms that confer either partial or complete resistance against malaria. The data obtained in the present study emphasize the importance of the evaluation of Duffy blood group genotypes in malaria endemic areas of the Brazilian Amazon region.
The Brazilian population has a highly heterogeneous ethnic composition, a result of the hybridization of the numerous native indigenous populations and immigrants from Europe, Africa and Asia. The immigration flow was not uniform in the different regions of the country [22–24]. The differential distribution of Duffy antigenic determinants among ethnic groups is an aspect characteristic of this blood system. Hence, this has been used as a marker for the ethnic composition as well as an indicator in the populational evolution. In the current study, significant differences in the allelic frequencies of the FY gene were identified when compared with previous studies of patients infected with human plasmodium in Brazilian Amazon regions [17, 25–27]. However, the FYA, FYB and FYB-33 alleles showed differential distributions compared to the population of the southeastern region of Brazil . The genotypic compositions obtained in this study demonstrated significant variations compared to previous studies performed in the states of Pará and Rio Grande do Sul [28, 29] and in the cities of São Paulo , Campinas  and Ribeirão Preto  all of which are in the state of São Paulo, Southeastern region of Brazil.
The currently obtained results showed that the FYA/FYB genotype was the most common, followed by the homozygotes for the FYB and FYA alleles and by the heterozygotes FYA/FYB-33 and FYB/FYB-33 (Table 1). The FYB-33/FYB-33 genotype, which is classically seen in individuals unaffected by P. vivax infection, was identified in 3.2% of the general population and in 5.5% of blood donors. These numbers differ from previous reports on a group of malaria patients uninfected by P. vivax in the state of Rondônia , where the frequency of this genotype was 12% and also for a mixed sample from the Northern and Southeastern regions of Brazil . Some populations had higher genotypic frequencies for FYA/FYB and FYB/FYB-33 than for a caucasian-like and negroid population from the Southeastern of Brazil . On the other hand, the frequency of the FYB-33/FYB-33 genotype was lower. In fact, with the exception of negroid ethnic groups, this genotype is extremely rare [12, 30, 32, 33].
The FYA allele is common in European and Oriental peoples but it is rare in African Negroes. Additionally, the FYB is more common in populations classified as white than in asiatic and negroid populations [34, 35]. The frequency of the FYA and FYB alleles in the Brazilian Amazon population was higher than those in the Southeastern region  probably due to the massive influence of Portuguese colonization in the North region of Brazil as well as the presence of Amerindians , as recent molecular analysis has corroborated that Amerindians have an Asian origin . Indeed, in the North region of Brazil, studies carried out with different tribes of Amerindians have shown that the FYA allele is the most common [31, 38]. As expected, a lower frequency of the FYB-33 allele was observed in the North region (P = 0.0001). Although a high prevalence of this allele was demonstrated in isolated communities in the states of Amapá and Pará, North region of Brazil , the introduction of a negroid ethnic component in the Amazon region is recent [40, 41]. In respect to the FY
allele, there does not seem to be a differential ethnic distribution as, the frequency detected here is similar to those described in other Caucasian-like populations in Brazil  and also in Europe .
The different FY allelic frequencies in individuals from North compared to Southeastern Brazil, may be due to the contribution of the three major ethnic groups (Europeans, in particular Portuguese; Blacks and Amerindians) in the formation of both populations. The Amerindian contribution was higher in the north of the country whereas European migration took place more in the South and Southeastern regions [43–45]. In spite of the current knowledge of the relationship between structure/function and tissue location of DARC, the functional significance of each of the alleles and the different genotypic combinations require further elucidation. The variation in the ethnic composition of the urban and rural populations of the Brazilian Amazon region and of distinct regions in Brazil [26, 46], may influence the allelic and genotypic distributions reflecting in matrixes of genetic mechanisms favorable to the susceptibility to infectious and parasitic diseases, in particular to malaria.
The frequency of the allele with the FY GATA mutation (FYB-33) was greater in blood donors than in patients infected with malaria (P = 0.0033), suggesting that there is a reduction in the infection rate of carriers of the FYB-33 allele. These results were recently reinforced by data from of individuals infected by P. falciparum when compared to those infected by P. vivax in Brazilian Amazon regions . The FYB-33 allele is a variant of the FYB allele resulting from a T → C point mutation in the gene promoter region (neucleotide-33), which abolishes its expression . The same occurs in the FYA allele, determining the FYA
allele . The presence of the FYB-33 allele results in a 50% reduction in the Duffy protein expression on the erythrocyte surface [15, 47]. This process demonstrates the action of the dose-related effect of the gene [15, 48], which may limit the invasion process of red blood cells by the parasite, although susceptibility to P. vivax may occur in heterozygotic Duffy-negative individuals [13, 14]. Thus the FY/FYB-33 genotypic combination, with either Fy(a+b-) or Fy(b+b-), seems to convey a reduction in the susceptibility to malaria. In vitro studies  support this hypothesis as RBCs that express these phenotypes have a significant reduction in cytoadherence of the parasite when compared to RBCs that express Fy(b+b+). Recently, in Papua New Guinea, a significant reduction in infection by P. vivax was observed in Duffy negative individuals heterozygotic for the FYA
In this study a low frequency of the negative Duffy genotype (FYB-33/FYB-33) was detected among uninfected subjects. As has previously been demonstrated, the absence of the Fy antigen in many ethnic Negro groups and their descendents, does not seem to exert any deleterious effect, however it does bestow a natural resistance against infection by P. vivax [2, 9, 14, 35, 50, 51]. These individuals are homozygotes for the FYB mutation in the GATA, that completely abolishes the Fy expression in erythrocytes but not in cells of other tissues . Recent reports have provided evidence on the transmission among Duffy-negative patients in Africa  and in Brazil . These data suggest the possibility that P. vivax is utilizing alternative receptors, apart from Fy, for binding in the erythrocyte invasion process. Whether FYB carrying the GATA box mutation is the primordial gene that encodes the Duffy system antigens or whether the GATA box mutation has evolved to escape malaria infection per se is controversial. However, the presence of this allele (FYB-33) may be a demonstration of the selection advantage by mutation in the population. Thus, in the Brazilian Amazon, where P. vivax predominates, the frequency of the FYB-33 allele is higher than expected given the ethnic population, both in terms of heterozygotes and homozygotes. This might, over time, cause an increase in the number of Duffy-negative individuals in the population and, as a consequence, a reduction in the rate of infection by P. vivax in this region.
In the group of patients infected by P. vivax, different allelic combinations with FY
were detected, including five (1.6%) FYA/FY
individuals, four (1.3%) FYB/FY
and two (0.6%) FY
(Table 2). The frequency of this allele among patients and blood donors did not show statistical differences (P = 0.3873). Studies carried out in blood samples from European and American Caucasians and Afro-American individuals, demonstrated that the C265T (FY
) mutation altered the Duffy protein expression on the RBC surface differently, depending on the ethnic group, using at least two mechanisms. One mechanism involves silent transcription in one of the FYB alleles and the other affects the translation and/or the stability of the protein [47, 54]. These studies showed that heterozygote individuals for the FY
) allele have 50% less of the Duffy protein on the surface of the membrane of the erythrocyte. On the other hand, individuals homozygotic for the FY
allele have about one tenth of the antigen expression in the erythrocytes . However, the information shown by the present study suggest that these polymorphisms do not seem to be associated with susceptibility to malaria.
As illustrated in Table 2, a higher frequency of the FYA/FYB genotype in patients compared to blood donors (P = 0.0404) was detected. In principle, this seems to indicate that the condition of heterozygosis, resulting from the expression of the FYA and FYB genes, favors infection by P. vivax. In fact, as has already been demonstrated by us, based on a phenotyping study, Fy(a+b+) individuals may be more susceptible to infection by this species of Plasmodium . Although no qualitative or quantitative measurements of the Duffy glycoprotein expressions were made in this study, we saw a larger number of malaria episodes among patients with the heterozygote genotype than the homozygote genotype. The basis of this observation has not been determined yet, but the phenotypic  and genotypic data presented from the same Brazilian Amazon areas associated with pertinent publications, point to the possibility that heterozygote individuals modulate the susceptibility to malaria by P. vivax by means of quantitative and/or qualitative variations that affect the Duffy antigen expression on erythrocytes. In the first aspect, in vitro studies demonstrate differences in the levels of the expression of FY glycoprotein on the surface of the reticulocytes of Caucasian-like and Afro-American individuals with the Fy(a+b+) erythrocytic phenotype . These authors verified that individuals with FYA and FYB alleles expressed a lower quantity of DARC than heterozygotes. Hence, it is possible that heterozygote individuals have a greater repertory of receptors for the possible variations that occur in the parasite protein binder to the human erythrocytes.
The recent demonstration of polymorphisms in the Duffy binding protein in isolates of P. vivax from the Brazilian Amazon  seems to support the observed results. An association, between human receptor polymorphisms and variations in the parasite binders of Plasmodium falciparum that modulate susceptibility to malaria, was also demonstrated [19, 56–60]. Hence, apart from the different levels of the expression, the specific conformation of the Fya and Fyb antigens may determine differences in the susceptibility to infection. Nevertheless, one of the possible consequences of differential susceptibility to vivax malaria could be modifications in allelic frequencies of FYA and FYB in populations exposed to P. vivax, the most prevalent species in the Brazilian Amazon region. If in fact this process occurs as a consequence of infection, it seems to be necessary, but not sufficient to eliminate heterozygote individuals, even if they are affected by repetitive episodes of malaria by P. vivax, as the infection is rarely severe and occurs in the entire age range with different frequencies for different ages and regions of the Brazilian Amazon without lethality [61, 62].
Based on the significant associations described herein, our data differ from previous studies  carried out in the western Brazilian Amazon region as well as recent studies of patients from the Brazilian Amazon as a whole . In respect to the first studies, these differences may be a result of the small sample sizes initially evaluated; the control group of the current study was composed of blood donors who reported never having experienced malaria against samples from individuals who had P. falciparum previously evaluated and the various ethnic groups in the different regions studied today (Eastern and Western Amazon). In relation to the recent studies, different to the findings observed in this study Brazilian Amazon population, the authors detected a significant association among infections by P. vivax and the FYB/FYB genotype, also conflicting with results of Wooley et al  in North American individuals (Caucasians and Afro-Americans). These divergences probably occurred due to the smaller sample size involved, and the fact that the control group consisted patients infected by P. falciparum.