Host age and Plasmodium falciparum multiclonality are associated with gametocyte prevalence: a 1-year prospective cohort study

Ethical clearance for this study was provided by the Institutional Review Board of the National Institute of Allergy and Infectious Diseases and the Ethics Committee of the Faculty of Medicine, Pharmacy, and Odontostomatology, University of Bamako. The study is registered with Clinicaltrials.gov (Identifier No. NCT01829737). Written informed consent was obtained from all participants aged ≥ 18 years, or the parents or guardians of individuals aged < 18 years.

In June 2013, a prospective cohort study was initiated in Kenieroba, Mali, to assess the population-level dynamics of asexual and sexual parasite carriage over 1 year. Malaria transmission in this area is seasonal, with most cases occurring during the June-December rainy season [14, 15]. The remainder of the year is mostly dry, with relatively few malaria cases reported. A cohort of 500 individuals aged 1–65 years, selected to have a similar age distribution as that of the entire village population, was recruited for the study. Enrollment was based on age and residency within Kenieroba for the 1-year study period. The average retention rate for the study period was 84%. The demographic characteristics of the cohort, and information on clinical malaria and total parasite (asexual and sexual parasites) prevalence during the 1-year study period have been previously reported [14].

Finger-prick blood samples were collected as dried blood spots (DBS) on Whatman 3MM filter paper and also as whole blood in RNAprotect^® Cell Reagent (Qiagen) for nucleic acid stabilization. Filter paper samples were stored under dry conditions at room temperature, and used for total parasite detection [14]. RNAprotect^® samples were stored at − 80 °C until nucleic acid extraction and molecular analyses were performed. For participants who reported malaria symptoms (i.e., history of fever, headache, body ache, and/or malaise) at any time during the study, thin and thick blood smears were prepared, Giemsa-stained, and read by experienced microscopists for parasitological diagnosis. Confirmed cases were referred to the Kenieroba Health Centre for standard-of-care anti-malarial treatment after samples collection. It should be noted that microscopic detection of parasites was only performed on samples collected from such symptomatic individuals for parasitological diagnosis of malaria cases, and not for general detection of P. falciparum infection in asymptomatic individuals in this study.

DBS samples were screened for P. falciparum infection by species-specific nested-PCR as previously described [14]. Using whole blood samples collected in RNAprotect^® Cell Reagent, P. falciparum gametocyte prevalence was measured by RT-PCR targeting expression of the female gametocyte specific Pfs25 gene. Total RNA extraction was performed using the RNeasy Plus 96 Kit (Qiagen), with an additional on-column DNase I treatment to eliminate any remaining contaminating genomic DNA (gDNA). RNA was eluted in 100 µL of RNase-free H₂0 and stored at − 80 °C until molecular assays were performed. Prior to the RT-PCR assay for gametocyte detection, all extracted RNA samples were tested for gDNA contamination with a 40-cycle Pfs25-targeted PCR, which was analysed with the LabChip GX and the HT-DNA 5 K LabChip capillary electrophoresis setup in accordance with the manufacturer’s protocol (Caliper Life Sciences). After confirming absence of gDNA, RNA samples were used as templates in a 1-step Pfs25-targeted RT-PCR assay for gametocyte detection. Primers and probe used in this assay were previously described by Wampfler et al. [16]. The RT-PCR assays were performed on CFX Connect™ Real-Time PCR Detection System (Bio-Rad), with the One Step PrimeScript™ RT-PCR Kit (Clontech). Each reaction included 7.5 µL of reaction buffer, 0.3 µL each of Ex Taq and RT enzymes, 400 nmols of each primer, 200 nmols of probe, and 2 µL of extracted RNA for a total reaction volume of 15 µL. The following thermocycler protocol was used: 1 cycle of 42 °C for 5 min and 95 °C for 10 s, 45 cycles of 95 °C for 5 s and 58 °C for 30 s. Assays were performed in duplicates for each sample. The threshold for positivity was set at 40 cycles based on testing with quadruplicate serial dilutions of a plasmid standard carrying one copy of the Pfs25 target sequence [16]: the lower detection limit of this assay occurred at a one plasmid copy per µL concentration with an average threshold cycle (Ct) of 40 cycles (StDev = 0.19). This limit of detection is equivalent to 0.05 gametocytes/µL [16]. Therefore, samples with one or both replicates at Ct of < 40 cycles were considered positive. No-template negative controls were included in each run.

Multiclonality was assessed among Pf-positive samples at four time-points (June, November, February, and April), which were selected to reflect periods of varying transmission intensities during the year. Multiclonality was measured using a SNP DNA barcode assay as previously described [14, 17]. Plasmodium falciparum multiclonality was reported for each sample in two ways, as described elsewhere [14]. In brief, polymorphic proportion (PmP) was defined as the proportion (%) of polymorphic reactions among the total number of successful reactions (reactions that yielded monomorphic or polymorphic nucleotide calls) from the 24 reactions that comprise the 24-SNP barcode assay, and complexity of infection (COI) was the estimated number of unique P. falciparum genotypes in each infection calculated by the COIL program [14].

Cross-sectional Pf gametocyte prevalence was calculated at each time-point in two ways; (1) the number of gametocyte-positive blood samples out of the total number of blood samples collected at the respective time-point, termed “population gametocyte prevalence” in this manuscript, and (2) the number of gametocyte-positive samples out of P. falciparum-positive (by nPCR) blood samples (called “proportion of gametocyte positive infections”). The latter measure was used for all statistical analysis. The proportion of gametocyte positive infections between wet (June to December) and dry (January to May) season time-points were compared by Mann–Whitney test. To analyse the effect of age on proportion of gametocyte positive infections, the cohort was categorized into seven age groups such that each group had a similar number of individuals (60–91 in each group). Proportion of gametocyte positive infections in each age group was then determined at each of the 12 time-points (one data point per month). Age-effect (12 proportion data for each age group) was analysed by One-way ANOVA test followed by Tukey’s multiple comparison test. Similarly, the proportion of gametocyte positive infections in each gender was calculated for each of the 12 time-points, and the gender-effect (n = 12 for each gender) was evaluated using Mann–Whitney test. Multiclonality was compared between gametocyte-positive and -negative samples using Mann–Whitney test for PmP, and Chi square test for COI.

During the 1-year study period, a total of 469 clinical malaria cases were recorded among 232 individuals; viz. 46.4% (232 out of 500) of all study participants experienced at least one clinical malaria episode during the year [14]. The highest monthly incidence of 116 malaria episodes was measured in October, and the lowest incidence of no malaria episodes in May 2014 (Fig. 1). Nearly all Plasmodium infections were caused by P. falciparum (Plasmodium malariae was the only other species detected, at ≤ 5% prevalence throughout the year) [14]; therefore, only P. falciparum parasites (total P. falciparum positivity and P. falciparum gametocyte positivity) were evaluated in this study, irrespective of potential, but rare infections by other Plasmodium spp.

Population P. falciparum prevalence (by nested PCR) was previously determined at 2-week intervals within this cohort over the 1-year study period [14]. Here, gametocyte positivity was assessed at one of two time-points for each month by Pfs25-targeted RT-PCR among P. falciparum-positive samples (i.e., samples showing parasitaemia as detected by nested-PCR, regardless of their clinical status). Cross-sectional P. falciparum gametocyte prevalence in the entire cohort (population gametocyte prevalence) showed a seasonal profile, ranging from a minimum of 8.3% in May 2014 to a maximum of 30.1% in October 2013, similar to the seasonal effect seen in population total P. falciparum prevalence (Fig. 1).

Figure 2 shows gametocyte prevalence among all P. falciparum-positive individuals (proportion of gametocyte positive infections) at one-month intervals over 1 year. The proportion of gametocyte positive infections ranged from 51.3% (61/119, July 2013) to 88.9% (32/36, May 2014), with no significant effect of seasonality on proportion (wet vs. dry season, p = 0.343, Mann–Whitney test). A previous assessment of the impact of several host factors, including age, gender, and red blood cell polymorphisms (i.e., ABO/Rh types, haemoglobin phenotype, and G6PD deficiency and α-thalassaemia genotypes) identified significant age and gender effects on persistent total P. falciparum positivity, while all other factors tested showed no significant association [14]. Since more than half of P. falciparum-positive individuals were identified as gametocyte-positive throughout the year, when cross-sectional population gametocyte prevalence was analysed, similar age and gender effects were observed on gametocyte positivity as was observed on total P. falciparum-positivity ([14], Additional file 1). Therefore, subsequent analyses were performed using proportion of gametocyte positive infections, not the population gametocyte prevalence.

To assess age effect, the cohort was categorized into seven age groups and the proportion was measured for each age group at each of the 12 months during the year (Fig. 3a). Proportion of gametocyte positive infections was similarly high (74–82%, median of 12 time-points in each age group) in the five age groups ≤ 17 years. Older individuals had relatively lower proportions (medians of 58 and 43% in 18–35 and > 35 age groups, respectively), and the > 35 age group showed significantly lower proportion than age groups ≤ 17 years (p = 0.003, One-way ANOVA; p < 0.05, Tukey’s multiple comparison test). Similarly, the proportion of gametocyte positive infections in each gender group was determined at 12 time-points (Fig. 3b), and there was no significant difference between males and females (p = 0.755, Mann–Whitney test). In both analyses, there was no appreciable effect of seasonality on proportion (Fig. 3). Furthermore, none of the RBC polymorphisms listed above were associated with gametocyte positive rate (number of gametocyte positive samples among all P. falciparum positive samples tested throughout a year in each individual) in a multiple linear regression analysis after adjusting for age and gender (Additional file 2). Out of the 1176 P. falciparum-positive samples, 57 RNA samples were collected from participants with symptomatic malaria at the time of sampling. Of those, 54 (94%) samples were identified as gametocyte positive. Hence, the chance of gametocyte positivity in symptomatic individuals was significantly higher than that in asymptomatic individuals (786/1119, 70%; p < 0.0001 by a Fisher’s exact test).

It was previously observed that there was a significant positive correlation between persistent P. falciparum positivity over 1 year and multiclonality at multiple time-points throughout the year [14]. Additionally, multiclonality, as measured by either PmP or COI, remained stable between wet and dry season time-points. Here, the relationship between multiclonality of P. falciparum infections and gametocyte positivity was investigated. Both PmP and COI were compared between gametocyte-positive and gametocyte-negative infections at 2 wet season and 2 dry season time-points. There was a general trend of higher PmP and COI in gametocyte-positive compared to gametocyte-negative infections throughout the year, although these differences were not statistically significant at individual time-points except in November (Additional file 3). When data from all four time-points were combined, PmP and COI were significantly higher in gametocyte-positive than gametocyte-negative infections (Fig. 4a, p = 0.009, Mann–Whitney test; Fig. 4b, p = 0.033, Chi square test).