Community-based surveys for Plasmodium falciparum pfhrp2 and pfhrp3 gene deletions in selected regions of mainland Tanzania

Upon successful establishment of blood-stage infection by Plasmodium parasites, various parasite proteins are produced and released into the host blood. Some of these proteins (also referred to as antigens) are targets for malaria rapid diagnostic tests (RDTs). Three antigen targets currently in use include Plasmodium lactate dehydrogenase (pLDH), Plasmodium aldolase (aldolase), and the Plasmodium falciparum-specific histidine rich protein 2 (HRP2) [1‐3]. The use of antigen-based RDTs in many malaria-endemic countries worldwide have profoundly improved malaria case management and surveillance efforts and remains an essential diagnostic tool, especially in Africa [4‐7].

HRP2-based tests are species-specific since the antigen is only produced by P. falciparum, though tests detecting pLDH and aldolase have the potential to detect all human malarias [2, 4, 8]. HRP2 is the most widely used antigen in RDTs either alone or in combination with other antigens, due to its abundance, specificity for P. falciparum infection, and high sensitivity and thermal stability [9]. However, antibodies raised against HRP2 can cross-react with P. falciparum HRP3 antigen due to similarities in amino acid sequences and repeating epitopes [10‐12]. The genes encoding for these two antigens are located on different chromosomes of the P. falciparum genome, with pfhrp2 on chromosome 8 while pfhrp3 gene is on chromosome 13 [13, 14]. A large number of parasites with genetic deletions of pfhrp2 and/or pfhrp3 genes in natural populations of P. falciparum were first reported in Peru and subsequently in other countries including in Africa with potential negative impacts on the performance of currently used RDTs [15‐22].

Sensitivity of HRP2-based RDTs can be affected by transportation and storage conditions outside of manufacturer specifications, operator errors, low density infections, or a mutation or deletion of the pfhrp2 and/or pfhrp3 genes in the infecting parasite strain [15‐23]. In addition, the diversity in parasite population and number of epitopes on HRP2 recognized by the diagnostic test antibodies may modify the sensitivity of the test when dealing with different P. falciparum populations [24‐28].

In Tanzania, RDTs were introduced between 2009 and 2012 and are now widely used in both private and public health facilities throughout the country. Before the introduction of RDTs in Tanzania, a study conducted between 2005 and 2010 (the African Quinine Artesunate Malaria Treatment Trial) found no evidence of pfhrp2/3 gene deletions [29]. However, pfhrp2 gene deletions have been reported in the neighbouring East African countries, including Kenya [21] and Rwanda [27]. In addition, two studies conducted in Tanzania which analysed samples collected in 2010, and between 2016 and 2018 showed evidence of sporadic occurrence of pfhrp2/3 gene deletions in some areas [30, 31]. Although both the initial evidence (sample confirmed as microscopy positive for P. falciparum but negative PfHRP2-detecting RDTs) and confirmatory evidence (molecular approaches) [32] were used to screen for pfhrp2/3 gene deletions, the sample size and the geographic regions covered were limited. In this study, field diagnostic results and a multiplex antigen detection assay were used to investigate potential pfhrp2/3 gene deletions in samples collected in 2017 from four regions of Tanzania with persistently high malaria transmission after five years of introduction of RDTs.

A total of 2520 out of 6207 registered HHs (40.6%) were sampled, with 7313 mainly asymptomatic individuals covered in the cross-sectional survey, which was conducted in 16 villages (in 8 districts) from four regions of Tanzania (Geita, Kigoma, Mtwara and Ruvuma) between July and November 2017 (Table 1). Apart from the 7313 individuals enrolled, an additional of 230 samples were taken from individuals who came to seek health services, but were not from the 120 sampled HHs. Therefore, 7543 blood samples were available for this study, and 3.0% (230/7543) of these were from individuals with incomplete data. The remainder (97.0%, 7313/7543) had complete data with parasitological, clinical and demographic information. The mean age of participants was 22.3 years (SD = 21.0) and 43.4% were male (Table 1).

For all enrolled participants, 38.4% (2897/7543) were positive for any one of the three parasite detection assays: microscopy, RDT, or bead-based multiplex assay. The results showed that 20.6% (1506/7313) of the participants were positive for P. falciparum infection by microscopy, 33.3% (2437/7313) were positive by HRP2 RDT, and 32.0% (2417/7543) were positive for anyone of the Plasmodium antigens tested by the bead-based multiplex assay. For those who were positive by microscopy, the geometric mean parasite density was 575 asexual parasites/µL of blood.

A flow diagram for sample selection for further molecular testing to detect the presence (and potential prevalence) of pfhrp2 and pfhrp3 deletions in this study population is shown in Fig. 2. In selecting specimens warranting molecular assays for pfhrp2 and pfhrp3 genotyping, two types of information were considered: discordance between field microscopy and RDT results for an individual, and the relationship between the pan-Plasmodium antigens and the HRP2 antigen for an individual’s blood sample. From the field tests (microscopy and RDT), 95 persons were found to be microscopy positive but RDT negative (1.3% of the 7313 persons who had data for both tests). Of these 95 persons, 94 had a DBS available for multiplex antigen detection, and 54.3% (51/94) of these were found to be HRP2 antigen positive by bead-based multiplex assay with a typical relationship to the other pan-Plasmodium antigens. Due to the cross-reactivity nature of HRP2 and HRP3, the multiplex assay is non-discriminatory, therefore a positive signal would indicate presence of either HRP2 or HRP3 antigen or both antigens (as illustrated in Fig. 3). Additionally, 28.7% (27/94) of these samples were negative for P. falciparum DNA and could not undergo genotyping. The remaining 16 samples (17.0% of the 94 selected by microscopy/RDT discordance) were all P. falciparum DNA positive, but had no antigens detected (n = 7) or an atypical relationship between the pan-Plasmodium markers and HRP2 (n = 9). Based on these test results, these 16 were selected as warranting pfhrp2 and pfhrp3 genotyping.

Since all DBS were screened by the multiplex antigen assay, samples could be selected for pfhrp2 and pfhrp3 genotyping based solely on bead-based multiplex assay results. Of all 7543 DBS screened by the multiplex antigen assay, malaria antigen could not be detected in 5126 (68.0%) samples. Of the 2417 DBS positive for any antigens, 2398 (99.2%) of these were found to have a typical relationship of the pan-Plasmodium markers with HRP2 (Fig. 3). Of the remaining 19 DBS, which were all positive for the pan-Plasmodium antigens; 6 had a complete absence of HRP2 antigen, and 13 had an atypical relationship between the assay signal for the pan markers and HRP2. All of these 19 were positive for P. falciparum DNA and could thus be utilized for genotyping by PCRs. Together with the 16 samples chosen initially based on field test results, these 35 samples (16 + 19) were gathered as the final set with suspicion of aberrant HRP2 and/or HRP3 antigen production and warranted genotyping to confirm the absence/presence of pfhrp2 and pfhrp3 genes.

Table 2 outlines the pfhrp2 and pfhrp3 genotyping results for these 35 samples, as well as other information regarding the characteristics of the individuals, field test results, and other factors. Most of the samples (31/35, 88.6%) were found to successfully amplify pfhrp2 and pfhrp3 genes for the two exons of each gene. However, one or more of the pfhrp2 and pfhrp3 exon targets could not be amplified in 4 (11.4%) DNA samples. To correctly report the presence of a deletion (i.e. lack of PCR amplification), two other single-copy genes (pfmsp1 and pfmsp2) were chosen and amplified to verify true non-amplification events [32]. For these 4 DNA samples, all were unable to consistently amplify both pfmsp1 and pfmsp2 single-copy genes and were excluded in the analysis. For this reason, non-amplification of pfhrp2 and pfhrp3 targets due to true deletion events could not be verified, and thus, no deletions in these genes could be confirmed in the remaining 31 samples.

The samples used in this study were collected during a community-based survey which was conducted in four regions with persistently high malaria burden over the past 10yrs and used to assess the presence and prevalence of pfhrp2 and pfhrp3 gene deletions in Tanzania. The findings from field RDT and microscopy tests as well as the laboratory multiplex antigen test indicate that the vast majority of P. falciparum infections in Tanzania produced high levels of HRP2 (and HRP3) antigens, which would be recognized by HRP2-based RDTs. Although RDTs did not detect some infections from persons confirmed to be P. falciparum positive by PCR, the evidence presented here suggests that these false negative RDT results were not due to pfhrp2 and pfhrp3 gene deletion.

During the survey, extensive efforts were taken to protect the quality of the RDTs. The experienced study team ensured that RDTs were stored in appropriate conditions as per the manufacturers’ instructions; transported in good conditions (ambient temperature with minimal humidity) and the tests were also performed by experienced technicians. To minimize operator errors, the testing process and reading of RDT results were done in the presence of other members of the team who ensured that any doubtful RDTs results were correctly read and interpreted. Studies conducted elsewhere reported that the performance of HRP2-based RDTs depends on the level of parasitaemia [1, 8, 26], with the lower limit of detection generally around 200 parasites/µL [1]. This community survey included mainly asymptomatic persons, and some had low-level parasitaemia by microscopy (bellow the detection limits of RDTs); this could possibly explain some of the discordances between the field RDT results and the laboratory antigen test. In total, 1998 (27.3% of all) participants were concordant between those two tests, whereas 372 (6.0%) individuals were positive for RDTs alone and 325 (4.4%) were positive only by the bead-based multiplex assay. As RDTs are designed for reliable detection of parasite densities more typical of clinical relevance (200p/µL or greater), their use in asymptomatic population may miss low-density infections with low levels of HRP2 antigens [48]. Even in this community setting, it was encouraging to see that RDTs were able to detect the majority of persons with malaria antigenemia as determined by the bead-based multiplex assay. Of the 2323 total samples found to be antigen positive, 1998 (86.0%) came from persons who were RDT positive. Considering the generally low parasite densities for any P. falciparum-infected persons in this survey, concordance was also good for all three tests; microscopy, RDTs, and the laboratory antigen test. Of the 1158 total microscopy positives, 1063 (91.8%) were also RDT positive, and 1050 (90.7%) were also positive by the bead-based multiplex assay. Concordance among multiple malaria indicators provides greater confidence for the true levels of malaria in a populace.

Of the 35 samples selected for further molecular investigation (from asymptomatic individuals), which were positive for P. falciparum by PCR, the majority (27/35) had relatively low parasite densities ranging from 0 to 1000 parasite/µL. However, eight samples had higher parasite densities (> 1000 asexual parasites/µl), and two of these had very high parasite density (88,600 and 104,000 parasite/µL, respectively). Two of these eight persons gave negative RDT results which could potentially be explained by prozone effect where an excess of antigen leads to false-negative results [49]. All eight of these higher density infections were found to have detectable HRP2 antigen, though at much lower blood concentrations than typical given those levels of P. falciparum parasite densities.

It was surprising to find that 16 out of the 35 samples which were both microscopy and PCR positive with intact pfhrp2 and pfhrp3 gene, had no or atypical antigen levels. Of these, 7 did not have antigen completely and 9 had atypical relationship between the pan-Plasmodium markers and HRP2. The absence of detectable HRP2 antigens in these samples can possibly be due to deterioration/degradation of the antigens, which were analysed after DBS storage for more than 7 months. However, more detailed assessments of such samples might be needed to determine the reasons for such discordant results.

The multiplex antigen screening allowed for reconfirmation of HRP2 (and possibly HRP3) antigen profiles for DBS samples. Of all 7543 samples screened by the multiplex antigen test, few (28, 0.4%) samples had a complete absence of HRP2 or an aberrant relationship between the assay signal for the pan markers and HRP2. This could be due to very low parasite densities or infections with non-P. falciparum species, but this observation could not be explained by deletions of the pfhrp2 and pfhrp3 genes.

This report adds to the literature in the same way as studies in Honduras [47] and French Guiana [50], where no deletion of pfhrp2 gene were reported. However, these findings are in contrast with the results from two recent studies which were conducted in Mbeya, Mtwara, Mwanza and Bagamoyo in Tanzania using samples collected in 2010 and between 2016 and 2018, respectively. The studies found evidence of sporadic occurrence of pfhrp2 and pfhrp3 gene deletion in some areas, with 1.7% of isolates tested reported to have a deletion of either of the genes [30], and 0.7% carried pfhrp2 deletions while another 0.7% carried a deletion of pfhrp3 [31]. Though recommended molecular tests [32] were used in the previous studies, the overall sample sizes were very small with only 176 and 149 Tanzanian samples tested in the two studies, respectively [30, 31]. Malaria endemicity and sampling method could potentially explain the differences between the present study which sampled mainly asymptomatic persons at the community) [33] and the previous studies which sampled symptomatic patients at health facilities [30, 31]. Despite very low estimates of the prevalence of single or double deletions of the genes in the previous studies, there is a potential that those parasite strains are still present at very low-levels in these regions. Thus, future studies will be required to cover these and other areas with varying epidemiological profile and using the World Health Organization (WHO) recommended protocol to confirm the absence or presence of the deletions.

A limitation of the current study is that, the samples used were from a community survey which enrolled mainly asymptomatic individuals and sampling was not done using the protocol recommended by the WHO for pfhrp2/3 gene deletion investigation. The WHO recommends that individuals seeking care for febrile illness at health facilities should be selected as a target population for the investigation of pfhrp2 and/or phrp3 gene deletions. This is because of the increased chances of detecting parasitaemia in symptomatic patients rather than asymptomatic persons. Due to this limitation, these findings are not conclusive rather they provide baseline information of pfhrp2/pfhrp3 gene deletion, and thus needs further studies. Also with the use of multiplex antigen assay, a single gene deletion of pfhrp2 or pfhrp3 alone could be masked by a parasite producing at least one of these antigens. Deletions (or loss-of-function mutations) in both genes would lead to a complete absence of HRP2/HRP3 antigens in a P. falciparum infection. However, if one of these antigens was being produced by the parasites, the algorithm defined here may have not been able to identify such blood samples as suspicious if the assay signal for HRP2 remained high. Additionally, the molecular tests performed here were only for complete gene deletions, and any loss-of-function by point mutations leading to antigen non-expression would not be captured since the pfhrp2 or pfhrp3 specific DNA amplification would still occur using the current protocol. Another limitation from the study is that it does not provide estimates for the entire country, since only 4 out of the 26 regions of mainland Tanzania were included in this study. In addition, these regions have high malaria transmission where parasite mutations including pfhrp2/3 gene deletion could be attenuated by high recombination rates involving different strains, which occur in such areas. Further studies will need to be conducted on P. falciparum isolates collected from symptomatic patients and other geographical regions of Tanzania (especially in low transmission areas) in order to increase the chances of detecting pfhrp2 and pfhrp3 gene deletion in different parts of the country.

Though a low number of false negative RDT results were found in Tanzania, these could not be explained by pfhrp2 or pfhrp3 gene deletions. However, single gene deletions could have been missed by the multiplex antigen assay suggesting parasites with deletion of one of the genes might be circulating in the population. Overall, the study results suggest that HRP2-based RDTs for detection of P. falciparum infection and confirmatory diagnosis of malaria in the surveyed area in Tanzania can be used as a reliable tool for malaria case management. There is a need for continued surveillance to monitor the status of pfhrp2 and/or pfhrp3 deletions in the future.