Temporal changes in Plasmodium falciparum reticulocyte binding protein homolog 2b (PfRh2b) in Senegal and The Gambia

Archived P. falciparum DNA samples from infected blood samples of consenting individuals were sourced from previous studies conducted between 2007 and 2013 at the Service de Lutte Anti-Parasitaire (SLAP) clinic in Thiès (70 km from Dakar, the capital city of Senegal) and from 1984 to 2013 in Western Gambia. Overall, malaria prevalence is moderate in The Gambia with high seasonal transmission [31, 32]. In contrast, Thiès is characterized by a perennial hypo-endemic transmission. These studies had received ethical approval from the Institutional Review Boards of the Harvard School of Public Health, the Ethics Committee of the Ministry of Health in Senegal and the Joint Gambian Government/MRCG Ethics Committee. A total of 1380 (849 from Thiès and 531 from Western Gambia) P. falciparum malaria infected blood samples were analysed. Among the Thiès samples, 580 that has been previously genotyped using a molecular barcode of 24 SNPs by Daniels et al. [33] were used for determining the association between alleles of PfRh2b and the specific barcode of parasite.

Semi-nested PCR method, as described previously [10], was used to amplify an PfRh2b gene fragment including the deletion in all extracted DNA samples. A positive control (P. falciparum laboratory cloned line 3D7) and negative control (reagent grade water) were included in all PCR amplifications. The size of the PCR products was estimated using Gene Ruler 100 bp DNA ladder marker (Quick Load^®, 100pb DNA Ladder).

For continuous variables median and interquartile range were calculated, while for categorical variables, the proportion or prevalence of the outcome with 95% CI was calculated.

For each year, to take into account the prevalence in mixed genotype isolates, the prevalence of PfRh2b full-length and PfRh2b deletion was calculated as follows:

(number of PfRh2b allele isolates + (0.5 * number of mixed isolates))/total number of isolates [10]. The Chi square linear trends was used to determine if the differences in frequency over the years was statistically significant. Differences between groups were assessed using Mann–Whitney U-test. Wright’s fixation index (Fst) was also calculated to assess the extent of genetic differentiation of PfRh2b polymorphism in Thiès and Western Gambia over time.

In Thiès, the prevalence of the deletion decreased significantly from 66.54% in 2007 to 38.1% in 2013 (P < 0.0001). This decline was not homogeneous with the presence of a peak in 2012, where PfRh2bdel form was present in 43.62% of infections (Fig. 1a).

In Western Gambia, temporal variation in prevalence in PfRh2b deletion was observed between 1984 and 2013. The prevalence of PfRh2b deletion increased between 1984 and 2005 from (58.04%) to (69.33%) (P = 0.03). From 2005 to 2007 there was a decline of the deletion to (47.47%) (P = 0.004). Between 2007 and 2012, the prevalence of this deletion increased significantly from 47.47 to 83.02% (P = 0.00005) and finally declined significantly to 57.94% in 2013 (P = 0.001) (Fig. 1b). Since mixed infections are uncommon in endemic populations, the frequency of infections with both deleted and full-length parasites (mixed) was determined. The result was highest in 2007 (0.10) and lowest in 2009 (0.02) in Thiès. In Western Gambia, mixed infections were most common in the earliest population from 1984 (0.23) and lowest in 2008 with (0.04) (Additional file 1: Table S1).

The presence of the deletion at high frequency in the general population and the acquisition of antibodies in an age-dependent manner against the c-terminal region of PfRh2b [10], have raised the interest to determine the prevalence of the deletion by age-group to see if the deletion would be more frequent in adults since they are long exposed. Thus, among the 1380 Plasmodium isolates assayed, 934 samples were available for age data and were analysed (651 from Thiès and 283 from Western Gambia). A median age was used to divide each population into two numerically equal groups.

In Thiès, since the study population mainly consists of adults (62%), the median age of the 651 patients was 18 years. Thus, the PfRh2b deletion form was less common in children compared to older patients (Fig. 2a). Using the Mann–Whitney U-test, a significant difference in the presence of PfRh2b deletion was found (P = 0.037), suggesting that there is an association between age and the presence of this deletion in Thiès.

In contrast, in Western Gambia, the population study is consisting of children (under 16 years of age); the median age of the 283 patients was 5 years. The PfRh2b deletion was more common in younger children compared to older children (Fig. 2b), but the difference is not significant by the Mann–Whitney U-test (P = 0.316).

Genetic diversity of the genes in natural parasites populations is a real obstacle for the validation of potential vaccine candidate. In this study, the degree of divergence to which PfRh2b gene is subject to selection was estimated by calculating Fst from the allelic frequencies of this locus according to years.

In Senegal the Fst value obtained from 2007 to 2013 was high (0.09), suggesting high degree of allelic divergence in this region over the time. In contrast, the Fst value obtained in The Gambia from 1984 to 2013 was low (0.057), suggesting less genetic differentiation of these alleles in this area over the time (Table 1).

It has been observed by barcoding in Thiès that the parasite, having undergone various interventions and pressures, has adapted by evolving towards to a type of parasite [28]. It is also assumed that parasite, initially having the PfRh2b full-length, has adapted by presenting the deletion following the different pressures. To test the hypothesis that the parasite with PfRh2b deletion may be linked by specific cluster, the prevalence of PfRh2b deletion according to molecular barcode was determined. Among the samples that were analysed for PfRh2b polymorphism, 580 from Thiès had been previously genotyped using the 24-SNP barcode [33] and the results were classified into two groups: a group named “cluster” which includes the samples having similarities with at least one other isolate in their nucleotide sequences and a group called “unique” grouping parasites with unique SNP barcode. 41.73% (242/580) of samples belonged to the cluster-group and 58.27% (338/580) were in the unique group (Table 2).

No significant difference in the frequency of PfRh2b deletion between the cluster-group and the unique group was found (P = 0.5014). Among the 338 isolates having unique barcodes, 166 (49.11%) had the deletion and 172 (50.89%) had the full-length fragment of PfRh2b gene. Therefore, both alleles PfRh2b deletion and PfRh2b full-length, had a similar distribution within the unique genotype group. Furthermore, all the 242 isolates belonging to the cluster-group were also tested to assess the frequency of PfRh2b alleles. In this group, 62 distinct subgroups were identified. The results showed that 20 haplotype clusters had only the deletion variant of PfRh2b gene, 24 subgroups had only the full-length fragment and the remaining 18 subgroups were parasites having the deletion and full-length (Table 2).

The number of samples is differently distributed in each of the 20 haplotype clusters with the deletion, and this distribution varies also over time (Table 3). Evolution over time of the 20 haplotype clusters containing only the PfRh2b deletion shows a decrease of the prevalence between 2008 and 2010 as well as a slight increase between 2010 and 2013 (Fig. 3). Since the numbers in the majority of the cluster group with the PfRh2b deletion are limited, haplotypes with n > 5 were examined further (haplotype 4 and 16) (Fig. 4). Haplotype cluster 4 with the deletion was found in 2008 (n = 12) and 2009 (n = 3), but not in the other years. Meanwhile, haplotype cluster 16 with the deletion was present in 2011 (n = 1), and 2012 (n = 6), but not in the other years (Fig. 5). The analysis of the results over time suggests that the frequency of PfRh2b deletion is related to the presence of some haplotype clusters in this population.