Plasmodium falciparum Merozoite Surface Proteins Polymorphisms and Treatment Outcomes among Patients with Uncomplicated Malaria in Mwanza, Tanzania

Background The severity of malaria infection depends on the host, parasite and environmental factors. Merozoite surface protein (msp) diversity determines transmission dynamics, P. falciparum immunity evasion, and pathogenesis or virulence. There is limited updated information on P. falciparum msp polymorphisms and their impact on artemether-lumefantrine treatment outcomes in Tanzania. Therefore, this study is aimed at examining msp genetic diversity and multiplicity of infection (MOI) among P. falciparum malaria patients. The influence of MOI on peripheral parasite clearance and adequate clinical and parasitological response (ACPR) was also assessed. Methods Parasite DNA was extracted from dried blood spots according to the manufacture's protocol. Primary and nested PCR were performed. The PCR products for both the block 2 region of msp1 and the block 3 regions of msp2 genes and their specific allelic families were visualized on a 2.5% agarose gel. Results The majority of the isolates, 58/102 (58.8%) for msp1 and 69/115 (60.1%) for msp2, harboured more than one parasite genotypes. For the msp1 gene, K1 was the predominant allele observed (75.64%), whereas RO33 occurred at the lowest frequency (43.6%). For the msp2 gene, the 3D7 allele was observed at a higher frequency (81.7%) than the FC27 allele (76.9%). The MOIs were 2.44 for msp1 and 2.27 for msp2 (p = 0.669). A significant correlation between age and multiplicity of infection (MOI) for msp1 or MOI for msp2 was not established in this study (rho = 0.074, p = 0.521 and rho = −0.129, p = 0.261, respectively). Similarly, there was no positive correlation between parasite density at day 1 and MOI for both msp1 (rho = 0.113, p = 0.244) and msp2 (rho = 0.043, p = 0.712). The association between MOI and ACPR was not observed for either msp1 or mps2 (p = 0.776 and 0.296, respectively). Conclusions This study reports high polyclonal infections, MOI and allelic frequencies for both msp1 and msp2. There was a lack of correlation between MOI and ACPR. However, a borderline significant correlation was observed between day 2 parasitaemia and MOI.


Introduction
Sub-Saharan Africa has the highest burden of Plasmodium falciparum malaria, with eleven countries accounting for 70% of all cases and 94% of the recorded deaths [1]. Malaria still remains a public health challenge in Tanzania despite the reported decrease in prevalence of about 10% over the past 10 years. Te severity of malaria infection depends on the host, parasite and environmental factors [2]. Parasite factors such as merozoite surface protein polymorphism (msp), Plasmodium falciparum surface protein (Pfs47), and apical membrane antigen (AMA1) may infuence treatment outcomes.
Merozoite surface protein diversity determines transmission dynamics, P. falciparum immunity, evasion and virulence. Te merozoite surface proteins are regarded as a potential target for malaria vaccine development [3,4] and antigenic polymorphisms have been associated with reduced vaccine efcacy [5,6].
Diversity in merozoite surface protein (msp) genes is also employed in the characterization of P. falciparum strains. Among the blood stage surface antigens, merozoite surface protein 1 (msp1) and merozoite surface protein 2 (msp2) are the most commonly used markers for the identifcation of genetically distinct P. falciparum parasite populations [7]. Te msp1 gene, located on chromosome 9 [8] is a major P. falciparum surface protein encoding a 185-215 kDa protein which is cleaved into several polypeptides during merozoite maturation and red cell invasion [7]. Te msp1 gene may be divided into 17 blocks of diverse sequences fanked by conserved regions. Block 2 (a region near the N-terminal of the gene) is the most polymorphic part of the msp1 gene [9] and is grouped into three major allelic families, namely MAD20, K1 and RO33 based on the variable nucleotide sequence and copy number of repeats of block 2 [8].
Te msp2 gene is located on chromosome 2 and is encoding the merozoite surface protein 2 which is a glycoprotein with an approximately 30 kDa [10,11]. It is composed of fve blocks whereby the central block (block 3) is the most polymorphic. Te msp2 alleles are grouped into two allelic families, namely, FC27 and 3D7/IC1 [12,13]. Fragment size polymorphisms in MAD20, K1 and RO33 (for msp1) and FC27 and 3D7 (for msp2) are used as molecular markers in studying P. falciparum malaria transmission dynamics and virulence [14]. Genotyping of msp1 and msp2 is also employed in diferentiating recrudescence from reinfection in antimalarial efcacy studies.
Multiplicity of infection (MOI), also referred to as complexity of infection (COI), is defned as the average number of distinct parasite genotypes concurrently infecting a patient [15] or the number of diferent P. falciparum strains coinfecting a single host [16]. MOI is used as an indicator for malaria transmission and immune status (in areas with stable malaria transmission) [16]. Parasite genotype and MOI are suggested to modulate infection outcomes and are determined by recombination during the sexual life cycle and injection of multiple clones during mosquito bite, respectively [17].
Individuals in areas of high transmission experience multiple mosquito bites associated with multiple clones per bite, unlike in low transmission areas where most of the mosquito bites are associated with monoclonal strains [3,18]. Te low genetic diversity observed in low transmission areas is associated with a strong linkage disequilibrium (LD) and a defned structure of parasite populations, unlike in high transmission areas where there is a weak LD and nondefned population structures [19,20]. Recent studies in Kenya and Myanmar reported no change in P. falciparum diversity and population structure after many years of intensive use of insecticide treated nets (ITNs) and insecticide residual spraying (IRS) despite a decline in malaria transmission due to the above interventions [19,21]. In general, the mechanisms controlling parasite genetic diversity are many and complex [18].
Te frequencies of the msp1 and msp2 alleles have been extensively reported globally. However, the infuence of msp polymorphisms and MOI on P. falciparum treatment outcomes has been reported in very few areas [22] with contradicting fndings. Studying MOI and the frequency of multiclonal infections is important in understanding malaria transmission intensity [15] in order to enable the establishment or modifcation of malaria control strategies. It is also imperative to determine the association between the msp polymorphisms and treatment outcomes. Tere is limited updated information on P. falciparum msp polymorphisms and their impact on transmission and treatment outcomes in Tanzania. Terefore, this study aimed at examining msp genetic diversity and MOI in a meso-endemic region and their association with peripheral parasite clearance and adequate clinical and parasitological response (ACPR) among P. falciparum malaria patients treated with artemether-lumefantrine.

Study Area and Patient Recruitment.
Tis study was carried out at Karume Health Centre in Igombe, a semiurban and malaria meso-endemic area in Ilemela District, Mwanza region. Te samples were collected during the rainy season of the year. Data for the study were prospectively collected from patients with uncomplicated Plasmodium falciparum (confrmed on the malaria rapid diagnostic test (MRDT)) malaria as part of the efcacy study involving artemether-lumefantrine and dihydro-artemisinin piperaquine published earlier [23]. Only patients on artemetherlumefantrine treatment were studied for parasite genetic diversity. Te inclusion and exclusion criteria are described in the previous study [23]. Details on the patient's clinical examination and drug administration were in accordance with the Tanzania guideline for the management of malaria.

Follow-Up and Sample Collection.
Patients were requested to return to the clinic for follow-up on days 1, 2, 3, 7, 14, 21, 28 and 35. Blood from fnger pricks was collected on flter paper (FTA ® Whatman paper) during the visits. Te flter papers were then dried at room temperature and kept in sealed plastic bags. Tick and thin blood smears were stained by Giemsa (on each day of the visit) according to the standard protocol [24]. Parasite identifcation and counting were done by two independent experienced microscopists.

DNA Extraction and
Genotyping for msp1, msp2, and Allelic Families. Parasite DNA was extracted from the dried blood spots (DBS) using the Life Sciences genomic DNA kit for dried spots according to the manufacture's protocol. Primers for genotyping of the block 2 for msp1, block 3 for msp2, specifc allelic families for msp1 (MAD20, K1, and R033) and specifc allelic families for msp2 (FC27 and 3D7) were in accordance to Somé et al. [25]. Primary and nested PCR was done using the method described previously [25]. Te PCR products for both the msp1 and msp2 and their specifc allelic families were visualized on a 2.5% agarose gel (containing red safe dye) under a UV illuminator. Te band sizes were recorded. Recrudescence and reinfection were distinguished according to the WHO guideline [26].

Treatment Outcomes.
Te WHO 2015 protocol [27] was used to classify treatment outcomes at day 28 as early treatment failure (ETF), late clinical failure (LCF), late parasitological failure (LPF), and adequate clinical and parasitological response (ACPR). Treatment failures were also classifed as recrudescence or reinfection after PCR correction.

Statistical Analysis.
Statistical analyses were performed using STATA version 13.1 (Texas, USA). MOI was defned as the number of distinct parasite genotypes coexisting within a single infection [10] and was calculated as the maximum number of PCR fragments for block 2 (msp1) and 3 (msp2) regions visualized for each sample [22]. Te mean MOI was estimated by dividing the total number of distinct msp1 or msp2 genotypes detected by the number of positive samples for the same markers [28]. Te percentage of polyclonal infections in the study samples was computed based on the proportion of isolates with multiple genotypes per marker. Categorical data were compared using the chi-square test or Fisher exact test. Spearman's rank correlation coefcient was used to fnd out the relatedness of continuous variables. p value less than 0.05 was considered statistically signifcant.

Allelic Diversity of P. falciparum msp1 and msp2 and MOI
among Patients with Uncomplicated Malaria. Alleles of msp1 and msp2 were classifed according to the PCR amplifed fragments. K1 was the predominant allele observed for msp1 (75.6%) and yielded 9 fragments (160-300 bp) as shown in Figure 1 and Table 1. Te msp1 allele with the lowest frequency was RO33 (43.6%) and yielded 4 fragments (120-200 bp) as shown in Figure 1 and Table 1. For msp2, the 3D7 allele was observed at a higher frequency (81.7%) than FC27 (76.9%). Te number and sizes of fragments are shown in Figure 1 and Table 1.
High MOI (above 2) was observed for both msp1 and msp2, but the diference between the two groups was not statistically signifcant (Table 2). Te majority of the patients harboured polyclonal infections for both merozoite surface proteins although the diference between the two groups was not statistically signifcant. Te overall mean MOI (msp1 and msp2 combined) was also high (2.36) ( Table 2).

Association/Correlation between Age, Parasite Density, and ACPR.
A Spearman's correlation was run to assess the relationship between age and parasite density (at days 1 and 2). Tere was a negative correlation between age and parasite density at days 1 and 2, but this was not statistically signifcant (rho � 0.1415, p � 0.2164 and rho � 0.1415, p � 0.2164, respectively) ( Figure 2). Parasite density at day 1 correlated negatively with ACPR, but the efect was not statistically signifcant (rho � −0.312, p � 0.0054). However, there was a strong negative correlation between parasite density at day 2 and ACPR (rho � −0.4591, p < 0.001).
Signifcant correlation between age and MOI for msp1 or MOI for msp2 was not established in this study (rho = 0.074, p � 0.521 and rho = −0.1289, p � 0.261, respectively) as shown in Table 3. In addition, there was no positive correlation between parasite density at day 1 and MOI of msp1 (rho = 0.133, p � 0.244) ( Table 3 and Figure 2). Tere was no strong positive correlation between parasite density at day 2 and MOI for msp1 (rho = 0.219, p � 0.054) ( Table 3 and Figure 2).
For msp2, there was no strong positive correlation between parasite density at day 1 and MOI, for which there was a statistical signifcance (rho = 0.043, p � 0.712) ( Table 3). Tere was also no strong positive correlation between parasite density at day 2 and the MOI for msp2, which was not statistically signifcant (rho = 0.006, p � 0.957) ( Table 3). Te association between MOI and ACPR was not observed for both msp1 and 2 (chi 2 = 4.0361, p � 0.776 and chi 2 = 4.9152, p � 0.296, respectively).

Discussion
Te genetic diversity of P. falciparum is essential for the parasite to adapt to environmental changes, escaping host immunity through antigenic variation and develop resistance [29,30]. Te levels of parasite allelic diversity, outcrossing and gene fow have been reported to be highest in African populations when compared to South American and Southeast Asian populations [20,31]. Understanding the genetic diversity and population structure of P. falciparum is essential for monitoring and evaluating malaria control strategies and interventions. Te present study focused on the allelic diversity of P. falciparum and its infuence on treatment outcomes utilizing isolates from patients with uncomplicated malaria treated with artemether-lumefantrine.
Te predominant alleles were K1 for the msp1 gene and 3D7 for msp2 the gene. Tese fndings are similar to those in other parts of the world [28,32,33]. Tis observation was not in conformity with previous studies done in Nigeria, Myanmar and Pakistan with regard to msp1, where by the predominant allele was MAD20 [7,34,35]. Te lack of conformity was also observed for the msp2 gene in Nigeria and Sudan, with the predominant allele being FC27 [7,11]. Host immune responses, changing environments and drug pressure may account for the inconsistencies in genetic diversity observed [36,37]. Our fndings suggest that K1 and 3D7 parasite strains are the common genotypes circulating in the study region. A similar observation was recorded in another area of the country (for the msp2 gene) [18], Nigeria and Senegal [38,39].
Plasmodium falciparum infected malaria patients had high frequencies of polyclonal infection for both msp1 and msp2. Te observed high polyclonal infection is not surprising in a malaria endemic and even meso-endemic area where it is common that patients will be infected by more than one distinct parasite genotype [15]. Multiclonal infections could be explained ecologically as a result of Journal of Tropical Medicine 3 cotransmission of diferent parasite variants (coinfection) or superinfection [40,41]. Te high frequency of multiclonal isolates for both msp genes is alarming since multiclonal infections are predicted to be more virulent than single clone infections [15,40] and are likely to be favoured by natural selection; thus, these infections are likely to be dominant in the population [42]. Te P. falciparum mixed clone infections in humans also lead to cross-fertilization and recombination between parasite genomes in the mosquito vector [43]. Tese can lead to the selection of more virulent and competent parasites, endangering the efectiveness of the currently used ACTs.
Te mean MOI for both msp1 and msp2 was high (between 2 and 2.5) at the study site, similar to fndings from Sudan and Uganda [33,44]. Tese values suggest the existence of a high malaria transmission rate. However, these MOI results were comparably lower than those documented more than ten years ago in other areas of the country, namely, Kilombero, Muheza and Ifakara [45,46]. Te difference in the observed MOI among patients with uncomplicated malaria over the past decade could be attributed to the diferences in transmission intensity between the study sites as a result of the scaling up of malaria control and prevention strategies over the years. Te diferences in vector   populations and human host immunity between the study sites and their changes with time could also account for the observed discrepancy. Te correlation between MOI with age has been reported with conficting results; some studies have reported an inverse association between age and MOI, showing lower MOI as age increases [3,7,47]. Tis can be attributed to the acquisition of immunity with age, resulting in some strains being cleared out. Other fndings document an increased MOI with age, which could be a result of the protection of children under fve due to the use of insecticide treated bed nets and other prevention approaches against mosquito bites that could make older children immune naïve [48]. We report an inverse correlation between MOI and age (for msp2) although the fndings lack statistical signifcance. Other studies report a lack of correlation between MOI and age similar to the general fndings from our study [28,49].
Te current study reports a lack of association between P. falciparum allelic families or MOI and clinical outcomes, particularly ACPR similar to studies done in Sudan, Uganda and Gabon [33,50,51]. Our fndings are in contrast with those reported in Uganda where children infected with multiple strains were more likely to experience treatment failure than those infected with a single strain [22]. Tis confict in results could be attributed to the diferences in transmission and vector populations between the study areas. Further evidences such as meta-analyses are required to reach the conclusion on the role of msp polymorphisms on treatment outcomes among malaria patients.
Our fndings suggest a lack of correlation between MOI and parasite density (at days 1 and 2). Te fndings are in match with other previous studies in Africa [52,53] but in contrast with fndings from other previous studies, which report an inverse correlation between parasite density and MOI [28]. An inverse correlation between haemoglobin   Journal of Tropical Medicine values and MOI has been observed in our study. Tis fnding is in contrast to a study done in Congo [53]. Te reason for the inverse correlation between haemoglobin and MOI needs to be established. Te high rate of antimalarial medication before clinical consultation could be the reason behind the observed lack of association between MOI and parasite density [54]. Findings from the present study serve as baseline data for future malaria epidemiological studies on malaria transmission at the study site and an evaluation of the infuence of parasite genotypes on treatment outcomes among P. falciparum malaria patients.

Conclusions
Most malaria patients harbour polyclonal infections harbouring multiple genotypes and the high MOI displays the high genetic diversity of P. falciparum infection in the country. Inverse correlations between day 2 parasitaemia and ACPR and between haemoglobin values and MOI were reported in our study. No correlation was observed between parasite density or age and MOI. MOI did not infuence ACPR among malaria patients. Te observed high number of multiclonal isolates suggests a complex population structure of the parasite and may accelerate the spread of resistance over time.

Abbreviations
ACPR: Adequate clinical and parasitological response ALU: Artemether-lumefantrine COI: Complexity of infection DBS: Dried blood spots ETF: Early treatment failure ITNs: Insecticide treated nets LCF: Late clinical failure LD: Linkage disequilibrium LPF: Late parasitological failure MRDT: Malaria rapid diagnostic test MOI: Multiplicity of Infection msp: Merozoite surface protein WHO: World Health Organization.

Data Availability
Te data used to support the fndings are included within the article.

Ethical Approval
Ethical and study approval was granted by the Joint Catholic University of Health and Allied Sciences (CUHAS)/Bugando Medical Centre (BMC) Institutional Review Board.

Consent
All parents/guardians signed a written informed consent during enrollment in the efcacy study after being informed that the study also covered assessment of parasite genotypes on their treatment outcomes.

Conflicts of Interest
Te authors declare that they have no conficts of interest.

Authors' Contributions
KJM participated in conception, design of the study, feld work, genotyping of msp1 and msp2, data analysis and manuscript drafting. EL carried genotyping of merozoite surface proteins. AK participated in conception, data analysis and manuscript reviewing. EK and GS participated in supervision of the research group, revising and approving the manuscript for submission.