The transmission of Plasmodium falciparum remains intense in many areas of sub-Saharan Africa. As a consequence, a considerable fraction of infected individuals in many settings may have multiclonal infections. Multiclonal infections may arise either from superinfection, whereby a single individual is infected with different parasites from multiple mosquito bites, or from cotransmission where a single mosquito transmits multiple parasite clones at one time. Repeated sampling to determine the complexity of infection (number of clones present during infection) indicates that a substantial proportion of clones may remain undetected when relying on single time point samples.1–3 The ability to detect all clones present during an infection is affected by fluctuating parasite densities, parasite sequestration patterns, and technical challenges such as competition in the polymerase chain reaction (PCR) reaction mix. Understanding the detectability of clones is relevant to appreciate to what extent single time point observations may be flawed. This is particularly important for drug efficacy studies when using genotypes on the day of treatment to determine whether an infection post treatment is recrudescence (treatment failure) or a new infection.4 Few studies have sampled naturally infected individuals for clonal complexity repeatedly over a short time window. In this study, we determined P. falciparum clones on 3 consecutive days in incident infections and chronic infections in Burkina Faso. Our hypothesis was that repeated sampling would increase the number of clones detected and that the detectability of clones would be lowest in chronic infections where parasite densities may be generally lower.
We addressed these hypotheses in a cohort of children aged 5–10 years, from Balonghin, Burkina Faso, an area exposed to intense malaria transmission. Informed consent was provided by the parent or guardian of each child. Ethical approval for the study was granted by the Ethical Review Committee of the Ministry of Health of Burkina Faso (Deliberation number 2015-3-033) and the Ethics Committee of the London School of Hygiene and Tropical Medicine (#9008). Full details of study cohorts are reported elsewhere.5 Briefly, the incident infection cohort included children (N = 42) who were parasite-free at the start of the transmission season by molecular methods and were monitored weekly by 18S nested PCR (nPCR)6 for incident infections. Once detected, infections were monitored daily. The chronic infection cohort included children (N = 54) who were parasite positive on at least two consecutive visits by nPCR that were spaced 4 weeks apart. Once this condition was met, daily sampling was performed as for incident infections. All individuals received treatment with the first-line antimalarial drug artemether-lumefantrine as soon as symptoms developed (measured axillary temperature ≥ 37.5°C or reported fever in last 24 hours). For the current study, only the first 3 days of intensive follow-up were considered. On those days, P. falciparum densities were measured by 18S quantitative PCR7 and clones were detected by PCR targeting the polymorphic marker gene merozoite surface protein 2 (MSP2) and capillary electrophoresis. Following automated sizing by PeakScanner® software (version 2) using three basepair bins (Applied Biosystems, Thermos Fisher Scientific, Inc., UK), we applied a cutoff of 1,000 relative fluorescent units (RFUs) to conservatively distinguish true peaks from background signal and stutter peaks. All major peaks and any additional alleles with a minimum height of 10% of the major peak were scored. Complexity of infection (COI) was defined as the total number of clones detected on days 1–3 combined. The proportion of these clones that were detected either on day 1 of sampling, on combined days 1 and 2, or on combined days 1 and 3 was calculated for incident and chronic infections separately. The MSP2 allelic diversity was estimated by GeneAlEx version 6.58 as equivalent to heterozygosity estimate. The association between parasite density and proportion of detected clones was assessed by a generalized linear model (GLM) with a logit link and the binomial family; parasite density was added to these models after log10 transformation. Analyses were performed in Stata 16.0 (Statacorp, College Station, TX).
When combining the 3 days of sampling, the total number of distinct parasite clones detected was 158 for the incident infection cohort and 335 for the chronic infection cohort. Among these clones, there were 105 (66.5%) distinct 3D7-specific alleles ranging in size from 207–460 bp and 53 (33.5%) FC27-specific alleles ranging in size from 209–424 bp for the incident cohort. For the chronic cohort, there were 217 (64.8%) 3D7-specific alleles ranging in size from 212–427 bp and 118 (35.2%) FC27-specific alleles ranging in size from 220–425 bp. The estimated haploid diversity for MSP2 was 0.981 in incident infections (with 88 detected alleles) and 0.976 in chronic infections (with 120 detected alleles). The median number of clones per infected individual was 3.0 (interquartile range [IQR] 2.0–5.0) for incident infections compared with 7.0 (IQR 4.0–8.0) in chronic infections (Mann–Whitney test, P < 0.0001), and this was underestimated with a single day of sampling (Table 1). All infections in the chronic infection cohort and 76.2% (32/42) of the incident infections were multiclonal (Figure 1A). Although this high COI may reflect the accumulation of clones in chronic infections over time, the data indicate that incident infections are also commonly multiclonal. Our intensive monitoring, with confirmed parasite negativity before incident infections, therefore, suggests that children experiencing incident infections were either bitten by multiple infected mosquitoes over a very short time period, which is plausible given the high mosquito exposure in some households in the study area,9 or they experienced genetically complex inocula from a single mosquito.10 Sampling on individual days considerably underestimated COI: approximately two-thirds of all present clones (day 1–3 combined) were detected on day 1 of sampling in both incident (69.0% 109/158) and chronic (68.1% 228/335) infections (Figure 1B). Interestingly, this level of detectability with 1 day of sampling was also observed in a similar study using the same genotyping method in a lower endemicity setting, Papua New Guinea.2 When samples were included from days 1 and 2, 82.3% (130/158) and 90.7% (304/335) of all clones were detected in incident and chronic infections, respectively; when including samples from days 1 and 3, these proportions were 86.7% (137/158) and 77.3% (259/335). Sampling on days 1 and 2 detected a similar proportion of all clones compared with sampling on days 1 and 3, with additional unique clones detected on day 3 that were not detected on day 2. Parasite density was a median of 676 parasites/µL (IQR 84–3,248) in incident and 545 parasites/µL (IQR 156–1,316) in chronic infections, and was strongly associated with the fraction of clones detected on the first day of sampling. With increasing total parasite density, the proportion of clones detected on day 1 increased (GLM model coefficient 0.70, 95% CI 0.41–0.98, P < 0.001). After adjusting for parasite density, the detectability of clones was not different between incident and chronic infections in our cohort (GLM model coefficient −0.41, 95% CI −1.03 to 0.22, P = 0.20).
Clonal complexity in incident and chronic infections
|Incident infections (N = 42)||Chronic infections (N = 54)|
|Number of clones detected (n/N [%])|
|Day 1||109/158 (69.0%)||228/335 (68.1%)|
|Day 2||21/158 (13.3%)||76/335 (22.7%)|
|Day 3||28/158 (17.7%)||31335 (9.3%)|
|Day 1 + Day 2||130/158 (82.3%)||304/335 (90.7%)|
|Day 1 + Day 3||137/158 (86.7%)||259/335 (77.3%)|
|Average complexity of infection (median clones per person [IQR])|
|Day 1||2.0 (1–3)||5.0 (2–6)|
|Day 1 + Day 2||2.0 (1–4)||6.0 (4–7)|
|Days 1–3 combined||3.0 (2–5)||7.0 (4–8)|
IQR = interquartile range.
In conclusion, our findings demonstrate that a large proportion of clones are missed by sampling a single time point. An additional sampling time point 24 hours later (i.e., days 1 and 2) or 48 hours later (i.e., days 1 and 3) greatly improved the detectability of clones, but either combination of 2 day sampling still missed between 9.3% and 22.7% of clones detected with 3 days of sampling. Since additional unique clones were detected on the third day of sampling, we conclude that parasite sequestration is unlikely to be the only explanation for missing parasite clones. Minority clones are difficult to detect11 and this complicates, for instance, reliably identifying antimalarial treatment failures if the parasite clone that survives treatment is present at low abundance at the initiation of the study. Application of novel, highly sensitive genotyping protocols, for example, amplicon deep sequencing, could help detect minority clones12 especially when repeated sampling is not possible.13 Our study was conducted in children residing in an area of high transmission intensity; this resulted in high COI estimates among study participants. Frequent sampling allowed us to detect infections early and, uniquely, investigate clonal complexity early-on in incident infections. The downside of this approach is that parasite densities were relatively low (resulting from early detection of infection by PCR) and this will have influenced detectability of clones. We also consider it plausible that sampling beyond day 3 would have increased the number of detected clones beyond our current findings. This would not affect any of the study conclusions but further highlight the complexity of accurately assessing infection clonal composition with sparse sampling approaches.
Farnert A , Snounou G , Rooth I , Bjorkman A , 1997. Daily dynamics of Plasmodium falciparum subpopulations in asymptomatic children in a holoendemic area. Am J Trop Med Hyg 56: 538–547.
Koepfli C , Ross A , Kiniboro B , Smith TA , Zimmerman PA , Siba P , Mueller I , Felger I , 2011. Multiplicity and diversity of Plasmodium vivax infections in a highly endemic region in Papua New Guinea. PLoS Negl Trop Dis 5: e1424.
Koepfli C , Schoepflin S , Bretscher M , Lin E , Kiniboro B , Zimmerman PA , Siba P , Smith TA , Mueller I , Felger I , 2011. How much remains undetected? Probability of molecular detection of human Plasmodia in the field. PLoS One 6: e19010.
Martensson A , Ngasala B , Ursing J , Isabel Veiga M , Wiklund L , Membi C , Montgomery SM , Premji Z , Farnert A , Bjorkman A , 2007. Influence of consecutive-day blood sampling on polymerase chain reaction-adjusted parasitological cure rates in an antimalarial-drug trial conducted in Tanzania. J Infect Dis 195: 597–601.
- Search Google Scholar
- Export Citation
Martensson A Ngasala B Ursing J Isabel Veiga M Wiklund L Membi C Montgomery SM Premji Z Farnert A Bjorkman A 2007. Influence of consecutive-day blood sampling on polymerase chain reaction-adjusted parasitological cure rates in an antimalarial-drug trial conducted in Tanzania. J Infect Dis 195: 597– 601.
Barry A et al.2021. Higher gametocyte production and mosquito infectivity in chronic compared to incident Plasmodium falciparum infections. Nat Commun 12: 2443.
Snounou G , Pinheiro L , Gonçalves A , Fonseca L , Dias F , Brown KN , do Rosario VE , 1993. The importance of sensitive detection of malaria parasites in the human and insect hosts in epidemiological studies, as shown by the analysis of field samples from Guinea Bissau. Trans R Soc Trop Med Hyg 87: 649–653.
- Search Google Scholar
- Export Citation
Snounou G Pinheiro L Gonçalves A Fonseca L Dias F Brown KN do Rosario VE 1993. The importance of sensitive detection of malaria parasites in the human and insect hosts in epidemiological studies, as shown by the analysis of field samples from Guinea Bissau. Trans R Soc Trop Med Hyg 87: 649– 653.
Hermsen CC , Telgt DS , Linders EH , van de Locht LA , Eling WM , Mensink EJ , Sauerwein RW , 2001. Detection of Plasmodium falciparum malaria parasites in vivo by real-time quantitative PCR. Mol Biochem Parasitol 118: 247–251.
Peakall R , Smouse PE , 2012. GenAlEx 6.5: genetic analysis in Excel. Population genetic software for teaching and research–an update. Bioinformatics 28: 2537–2539.
Guelbeogo WM et al.2018. Variation in natural exposure to anopheles mosquitoes and its effects on malaria transmission. eLife 7: e32625.
Das S , Muleba M , Stevenson JC , Pringle JC , Norris DE , 2017. Beyond the entomological inoculation rate: characterizing multiple blood feeding behavior and Plasmodium falciparum multiplicity of infection in Anopheles mosquitoes in northern Zambia. Parasit Vectors 10: 45.
- Search Google Scholar
- Export Citation
Das S Muleba M Stevenson JC Pringle JC Norris DE 2017. Beyond the entomological inoculation rate: characterizing multiple blood feeding behavior and. Parasit Vectors Plasmodium falciparummultiplicity of infection in Anophelesmosquitoes in northern Zambia 10: 45.
Bretscher MT , Valsangiacomo F , Owusu-Agyei S , Penny MA , Felger I , Smith T , 2010. Detectability of Plasmodium falciparum clones. Malar J 9: 234.
Lerch A , Koepfli C , Hofmann NE , Kattenberg JH , Rosanas-Urgell A , Betuela I , Mueller I , Felger I , 2019. Longitudinal tracking and quantification of individual Plasmodium falciparum clones in complex infections. Sci Rep 9: 3333.
Gruenberg M , Lerch A , Beck HP , Felger I , 2019. Amplicon deep sequencing improves Plasmodium falciparum genotyping in clinical trials of antimalarial drugs. Sci Rep 9: 17790.