Background
The significant reduction in malaria mortality and morbidity seen in the last 10 years is the result of combined efficient control measures such as early diagnosis, effective treatment with artemisinin-based combination therapy (ACT), active surveillance and vector control. Nevertheless, in 2016, 216 million cases were still reported worldwide [
1] and major concerns exist about the extent to which the emergence and spread of insecticide [
2] and artemisinin resistance [
3] may affect worldwide malaria control and elimination efforts. New strategies may be needed to sustain recent gains and accelerate malaria elimination initiatives. These new strategies include the development and deployment of transmission-blocking strategies that aim to reduce malaria incidence by targeting the infection reservoir involved in maintaining parasite transmission from humans to
Anopheles mosquitoes.
Transmission to mosquitoes is mediated by the presence of gametocytes in peripheral blood of a human host. Gametocytes are sexually dimorphic and both sexes are required to ensure the development of the parasite inside the mosquito. Although the likelihood of mosquito infection is largely dictated by gametocyte density [
4], the gametocyte sex ratio may also play a significant role in ensuring fertilization [
5‐
7]. Since one female gametocyte (FG) produces only one gamete, while a male gametocyte (MG) produces eight gametes [
8,
9], the gametocyte sex ratio is usually female-biased in the proportion of 3–5 females: 1 male [
10,
11] with indications from rodent malarias and natural
Plasmodium falciparum infections that sex ratios may be adjusted in the presence of other parasite clones [
6,
12], in relation to gametocyte density [
7,
13] during infections and in response to environmental factors, such as anaemia [
10]. Recent findings further suggest that anti-malarial drugs may have differential effects on MG and FG [
14,
15]. Understanding gametocyte sex ratios is thus of interest to understand
Plasmodium biology, better predict transmission potential during natural malaria infections and estimate the likelihood of onward transmission to mosquitoes after treatment with anti-malarial drugs.
Gametocytes usually circulate in blood at low levels as only 0.2–1% of asexual parasites commit to sexual development at every cycle of red blood cell invasion [
16]. Several studies have observed infected mosquitoes after feeding on blood containing gametocyte densities as low as 0.25–0.3 gametocytes/microlitre of blood, well below the threshold for detection by routine microscopy [
7,
17‐
21]. As a consequence of the abundant presence of submicroscopic densities of gametocytes in clinical and asymptomatic infections [
22‐
24], they represent a silent infectious reservoir in the population. In the last 20 years, sensitive molecular techniques based on sexual stage-specific mRNA transcripts have been developed to improve the detection and quantification of both gametocyte sexes.
Pfs25 mRNA has been widely used as a mature gametocyte marker [
19‐
21,
25] and was recently confirmed to be female-specific or at least considerably female-biased [
26,
27]. Based on RNA-seq analysis,
PfMGET was recently presented as a novel male-specific gametocyte marker [
26,
28]. The use of intron-spanning primers allows for sensitive detection of MG in samples of naturally infected parasite carriers [
28]. Thus far a combination of separate
Pfs25 and
PfMGET qRT-PCR assays has been used to estimate gametocyte sex ratios in natural and controlled infections [
7,
15,
28‐
30]. However, estimating sex ratios by using two separate qRT-PCR assays may affect assay precision and throughput. Here, a novel target for FG is proposed,
CCp4, that was previously identified as a gametocyte-specific transcript [
31] and allows for intron-spanning primer design. This manuscript describes a one-step multiplex qRT-PCR assay for robust assessments of gametocyte sex ratios at densities below the microscopic threshold for gametocyte detection.
Methods
Selection of male and female marker transcript
The selection of the male marker
PfMGET was described previously [
28]. The female marker
CCp4 was identified by integrating transcriptomics data and validated as gametocyte-specific [
31].
Preparation of gametocyte material
Sex-sorted gametocytes were generated as described previously [
26,
28]. In brief, cultures of the PfDynGFP/P47mCherry line [
26] were treated with
N-acetyl glucosamine and stage V gametocytes were FACS sorted for their fluorescence signal (MG are sorted as GFP-positive/mCherry-negative, FG as mCherry-positive/GFP-negative) and afterwards counted with a Bürker-Türk counting chamber. For both MG and FG tenfold dilution series were prepared in whole-blood in the range of 10
6/mL to 10
1/mL and stored in RNAProtect to serve as standard curves for gametocyte quantification.
Preparation of ring stage parasites to assess transcript stage specificity
Asexual parasites of the NF54 strain were synchronized by the selection of late trophozoites and schizonts as described [
28]. In brief, a 63% Percoll density gradient was followed by a 5% sorbitol treatment, killing the remaining schizonts after 5 h and ensuring tight synchronization. NF54 ring stage parasites were harvested 10–20 h after the Percoll treatment and stored in lysis buffer (5.25 M GuSCN; 50 mM Tris–HCl pH6.4; 20 mM EDTA; 1.3% Triton X-100) for later analysis.
To obtain pure asexual stage reference material without plausible contamination by gametocytes, ring stage parasites of the gametocyte-deficient F12 clone were used that have a loss-of-function mutation in the gene encoding the gametocyte master transcription factor AP2-G [
32]. In addition, ring stage parasites were generated from the recently described AP2-G knock-down line 3D7/AP2-G-GFP-DDglmS that does not express AP2-G when grown in the presence of 2.5 mM (
d)-+-glucosamine (GlcN) (Sigma Aldrich) [
33]. Under these conditions both the
ap2-
g-
gfp-
dd transcript and AP2-G-GFP-DD protein are degraded, resulting in no gametocyte production. Parasites were synchronized twice 16 h apart to obtain an 8-h growth window. 30 mL parasite culture at 3–4% parasitaemia and 5% haematocrit was harvested at 8–16 h post invasion. Parasites were released from infected RBCs by saponin lysis and total RNA was directly isolated using Ribozol (Amresco) according to the manufacturer’s manual.
Samples from naturally infected gametocyte carriers
Samples from two previously published clinical trials in light microscopy-positive gametocyte carriers from Kenya [
28] and Mali [
15] were used to directly compare gametocyte density estimates using the
PfMGET/
CCp4 multiplex assay with the previous qRT-PCR assays targeting
PfMGET (male) and
Pfs25 (female) transcripts separately. The multiplex assay was performed on Kenyan samples collected prior to treatment (day 0; n = 31), and after treatment on day 2 (n = 16), day 7 (n = 46), and day 14 with dihydroartemisinin–piperaquine (DP) alone or with primaquine (n = 28; total n = 121). The samples from Mali that were included in the current study were collected on day 7 after treatment with either DP (n = 15) or DP + methylene blue (15 mg/kg given daily for the first 3 days of treatment; DP + MB, n = 19) or sulfadoxine–pyrimethamine and amodiaquine (SP–AQ, n = 19) or SP–AQ with a single dose of primaquine (0.25 mg/kg given together with the first dose of SP–AQ, SP–AQ + PQ, n = 19) [
15]. In both studies considerable variation in gametocyte densities and sex ratios was previously reported.
Nucleic acid extraction and target amplification
Total nucleic acids were extracted from 50 µL whole blood in five volumes RNAProtect with the MagNAPure LC automated extractor (Roche) using the MagNAPure LC Total Nucleic Acid Isolation Kit—High Performance, with the exception of F12- and 3D7/AP2-G-GFP-DDglmS-derived material (bulk Ribozol (Amresco) extraction according to the manufacturer’s manual). Total nucleic acids were eluted in 50 µL of MagNAPure elution buffer, of which 5 µL was used in the multiplex assay. For the multiplex assay, we used the Luna
® Universal Probe One-Step RT-qPCR Kit (NEB). Gene IDs of the male and female markers and respective primer and probe sequences can be found in Table
1, for additional primer sequences see Additional file
1. Probe and primer concentrations were optimized (see Additional file
1) to obtain efficient amplification of both targets. The optimal conditions are summarized in Table
2. Negative controls were run to ensure there were no unspecific signals detected from no-template controls, or from total nucleic acids without reverse transcription (intron-spanning primers do not bind to genomic DNA); and melt curves were visually inspected.
Table 1
Primer and probe sequences for qRT-PCR assays, with references for earlier reported methods and primers
PF3D7_1031000 | GAAATCCCGTTTCATACGCTTG AGTTTTAACAGGATTGCTTGTATCTAA | – | – |
PF3D7_1031000 | | 6FAM-ccgtttcatacgcttgtaa-MGB | FAM |
PF3D7_0903800 CCp4 (MPX) | CACATGAATATGAGAATAAAATTG* TAGGCGAACATGTGGAAAG | AGCAACAACGGTATGTGCCTTAAAACG | Texas Red |
PF3D7_0903800 CCp4 (qRT-PCR) | CACATGAATATGAGAATAAAATTG* TAGGCGAACATGTGGAAAG | – | – |
PF3D7_1469900 PfMGET (MPX) | CGGTCCAAATATAAAATCCTG* TGTGTAACGTATGATTCATTTTC | CAGCTCCAGCATTAAAAACAC | FAM |
PF3D7_1469900 | CGGTCCAAATATAAAATCCTG* GTGTTTTTAATGCTGGAGCTG | – | – |
Table 2
Multiplex conditions for male–female assay
Female primers CCp4 | 900 nM | 55 °C 15 min | RT-step |
Female probe—Texas Red | 200 nM | 95 °C 1 min | |
Male primers PfMGET | 225 nM | 95 °C 10 s | 45 cycles |
Male probe—FAM | 200 nM | 60 °C 1 min | |
Input total nucleic acid | 5 µL | | |
Serial dilutions of all ring stage materials were used to detect ring stage transcripts (SBP-1) and early gametocyte transcripts (Pfg27, Pfs16, Pfg14-744, Pfg14-748) as well as the mature gametocyte transcripts CCp4, PfMGET and Pfs25. Total nucleic acids were used for the intron-containing genes SBP-1, Pfg14-744, Pfg14-748, CCp4 and PfMGET while Pfs16, Pfg27 and Pfs25 required RQ1 DNase I treatment (Promega). cDNA was prepared with the High Capacity cDNA Reverse Transcription Kit (Applied BioSystems) and 2 μL of cDNA was run in the GoTaq qPCR Master Mix (Promega). All primers were used at 900 nM, except for SBP1, PfMGET and Pfs25 which were run at 225 nM primers.
Synthetic RNA standard curve material
Linear dsDNA templates for the target regions of
CCp4 and
PfMGET were synthesized by BaseClear B.V. the Netherlands (for sequences, see Additional file
2) and purified by agarose gel electrophoresis. Bulk RNA was transcribed with the MEGAShortscript T7 high yield transcription kit (Invitrogen) at 10–50 nM dsDNA input according to the manufacturer’s instructions. Transcription samples were DNaseI treated with the TURBO DNA-
free kit (Invitrogen) and subsequently purified over an RNeasy mini spin column (Qiagen). RNA standards were calibrated against sex-sorted gametocyte standards, both standards were prepared using tenfold serial dilutions. Copy numbers in initial samples were quantified on a Qubit 2 (ThermoFisher), absolute RNA amounts were calculated into copy numbers with the specific sequence weight.
Gametocyte transcript stability testing
Synchronized mature gametocytes (NF54, 1.8% parasitaemia) were diluted in whole EDTA-blood starting at concentrations of 105 gametocytes/mL. These samples were either stabilized by adding five volumes of RNAProtect (Qiagen) or left unstabilized (no protective buffer added) for further treatments: One to three aliquots were kept at room temperature (22–25 °C) for 0 h, 1 h, 2 h, 4 h, 6 h, 8 h or 24 h before freezing them at − 80 °C and subsequent processing. A subset of samples (n = 3) in RNAProtect (added at 0 h) were additionally freeze–thawed five times (37 °C/− 80 °C cycling for at least 1 h each) before extraction of nucleic acids. The stability of the transcripts after extraction was also tested by freeze–thaw cycles at which the samples were left at room temperature (22–25 °C) or 37 °C for 1 h, interspersed by at least 1 h at − 20 °C.
Statistical analysis
Graphs and statistical analyses were made with GraphPad Prism (version 5.0.3) or R statistical software (version 3.4.0). The concordance between separate qRT-PCR assays and the multiplex approach was assessed by estimating the slope and 95% confidence interval (95% CI) in linear regression. No statistical comparisons were made on the CT values of different marker genes since this was beyond the scope of the current manuscript and meaningful comparisons would require a larger number of replicates. Where transcript abundance differences were estimated, this was based on the assumption of doubling of transcript after each cycle (1 Ct difference), regardless of reaction efficiencies.
Discussion
In the current study, a multiplex assay for the rapid quantification of female and male gametocytes is reported. The assay utilizes a new female gametocyte marker
CCp4 in conjunction with the reported male gametocyte marker
PfMGET [
28]. The use of intron-spanning primers allows simultaneous quantification of male and female-specific transcript levels in total nucleic acids without prior DNase I treatment. The presented analysis concludes low but non-negligible gametocyte transcripts in gametocytes of the opposite sex and gametocyte-free ring-stage asexual parasites. The stability of
CCp4 and
PfMGET transcripts was similar under suboptimal storage conditions; gametocytes can be reliably detected and quantified at densities 0.1–1 gametocyte/μL.
CCp4 is a member of the LCCL-domain containing adhesion protein family and orthologous to LAP6 in
Plasmodium berghei, where this gene is reported to be translationally repressed with protein expression occurring at the ookinete stage only [
37]. Earlier, a DOZI (development of zygote inhibited) knock out indicated that
LAP6 transcripts (then called PB000955.03.0) are accumulated in an mRNA storage complex [
38]. In
P. falciparum, the CCp4 protein is predominantly expressed at the gametocyte stage, with only minor expression in male gametocytes [
39] and no evidence for translational repression. Gametocyte-specific
CCp4 transcripts were reported in an integrated analysis of eight
Plasmodium transcriptomes [
31]. The initial validation in qRT-PCR experiments confirmed
CCp4 expression to be at least 1000-fold upregulated in gametocytes of different
P. falciparum strains compared to asexual blood stages. The 1000-fold higher mRNA levels in FG over MG reported here confirm and exceed the previous estimates by RNAseq (38-fold higher in females by RPKM values) [
26].
Unlike the commonly used female marker
Pfs25,
CCp4 allows the design of intron-spanning primers. The current assay utilizes intron-spanning primers of both male and female reporter genes with two marker-specific probes. Importantly, the multiplex gametocyte assay can be performed on total nucleic acids without DNase I treatment, which may affect gametocyte detection at low densities [
28,
36]. The multiplex male–female assay is thus faster than separate assays. Sensitivity for detecting female gametocyte was lower than for the single qRT-PCR targeting
Pfs25, which is at least in part explained by a lower estimated number of
CCp4 transcripts per female gametocyte as compared to
Pfs25. The LOD for female and male gametocytes in the multiplex assay is 0.1/μL, well below the limit of microscopic detection and in the same range as other molecular sex-specific assays [
28,
34,
35]. More sensitive total gametocyte assays have been reported [
40] but the current multiplex LOD allows for the detection of infections that are likely to be transmissible to mosquitoes. An increasing likelihood of mosquito infections is consistently observed at gametocyte densities above 1–5 gametocyte per microlitre [
7,
29,
41]. The current assay reliably quantifies gametocytes at these densities. The lower sensitivity to detect female gametocytes as compared to
Pfs25 may be a concern in studies where very low overall gametocyte densities are observed [
29,
30] but the operational attractiveness of a multiplex assay that does not require DNase treatment is considerable for many other studies.
Previous work indicated that the stability of transcripts is a relevant concern when estimating gametocyte prevalence or density [
36,
41]. When assessing gametocyte sex ratio, transcript stability is a particular concern since differences in the stability between target transcripts may affect bias estimates. In a limited set of experiments there were no indications for differences in the stability of
PfMGET and
CCp4 transcripts, provided samples are transferred to RNA-protective buffer within 1–2 h of blood collection. Freeze–thaw cycles resulted in limited RNA loss once blood samples are in this protective buffer. A similar stability of male or female signal is of particular relevance for studies conducted in low resource settings where there may be challenges in ensuring optimal storage conditions. Repeated or prolonged freeze–thaw cycles may thus affect overall gametocyte detection or quantification [
41,
42] but current results indicate they would not disproportionally affect MG or FG quantification and thus sex-ratio estimates.
The presented multiplex assay is a fast route to accurate P. falciparum sex ratio determination, saving about 25% of the time—with similar material costs—compared to two separate assays of which one requires a DNase treatment step. Medium sample through-put in 96-well format is the recommended application, providing accurate gametocyte quantification and sex ratio determination for blood and culture samples.
Previous studies concluded no or negligible
Pfs25 transcript numbers in asexual parasites. In the current set of experiments, we detected gametocyte transcripts in different preparations of ring-stage asexual parasite material. Whilst low-level contamination of gametocytes in supposedly pure asexual parasites may have contributed to the detection of
CCp4,
Pfs25 and
PfMGET transcripts in asexual parasite material from the NF54 strain, the detection of these transcripts in the gametocyte-deficient F12 line and under knock-down conditions for AP2-G in a more recent gametocyte-less line 3D7/AP2-G-GFP-DDglmS [
33] provides convincing evidence for low-level expression of
Pfs25,
CCp4 and
PfMGET in asexual ring stage parasites. The current findings of detectable transcript expression for all gametocyte markers in high densities of asexual blood stages despite different strategies to avoid gametocyte contamination have implications for past and future gametocytaemia estimates. Whilst the difference in transcript abundance between gametocytes and asexual parasites is sufficiently pronounced to conclude a marginal impact on gametocyte density estimates, gametocyte prevalence estimates may be inflated in some populations. Given the high parasitaemia of some acute malaria infections (with densities typically above 10,000 parasites/µL [
43] as opposed to asymptomatic infections where densities commonly lie below 10 parasites/µL [
44]), earlier studies recruiting clinical malaria cases may have overestimated gametocyte prevalence by molecular assays. In studies with asymptomatic parasite carriers and in low-endemic settings, where lower asexual parasite densities dominate, this overestimation will be less pronounced and often negligible. A previously reported rapid decline in gametocyte prevalence based on
Pf25 mRNA detection in the first 3 days following treatment of high-density asexual infections [
45] may thus be (partially) explained by the detection of
Pfs25 transcripts arising from asexual parasites, whilst the gradual decline in gametocyte prevalence following treatment of lower-density asymptomatic asexual parasite carriers [
15,
46] or gametocyte transcript kinetics in the period following asexual parasite clearance may better reflect gametocyte clearance and gametocyte half-life [
47]. With a better appreciation of caveats in gametocyte detection, the molecular tools for gametocyte detection are of value for studies aiming to quantify the human infectious reservoir for malaria, the kinetics of gametocyte production and the impact of interventions on gametocyte carriage. As a consequence of the detection of low level transcripts of gametocyte markers in rings, it is advised to report gametocyte prevalence in samples with parasite densities > 1000 parasites/µL together with a qualifying remark on the reliability of gametocyte prevalence and quantification. The presented results suggest that gametocyte prevalence determined in samples below assay-specific cut-off values indeed can be trusted (Table
3). Stating the limitations of molecularly determined gametocyte prevalence for densities, if required, will re-confirm the validity of molecular gametocyte detection.
With a cautious interpretation of low gametocyte density estimates in samples with high concurrent asexual parasite densities, molecular gametocyte diagnostics such as the multiplex assay presented in this manuscript are valuable tools to obtain sensitive and robust estimates of gametocyte prevalence and density. With these tools, gametocyte densities and sex ratios can be assessed across the gametocyte density range that is likely to contribute to onward transmission to mosquitoes [
7,
19], which in many settings is well below the threshold for detection by microscopy.
Authors’ contributions
Work was conceptualized and overseen by TB. KL, LMK, KC and TB planned and conducted experiments, additional samples were prepared and provided by EC, AM, PSa, HD, MVB. CA and LMK wrote the initial manuscript, editing was done by TB, KC, WS, CD, AD, TV, EC, IF, PSc and KL. All authors read and approved the final manuscript.