Background

According to the World Health Organization, in 2021 there were 247 million cases and 619,000 deaths from malaria in 85 endemic countries [1]. In Brazil, 145,188 malaria cases were reported in 2020, showing a reduction of 7.8% compared to the previous year [2]. This reduction reflects the efforts which have been made to eliminate malaria in Brazil through the implementation of the “National Malaria Elimination Plan”, which aims to reduce the number of autochthonous cases to less than 68,000 by 2025, the number of deaths to zero by 2030, and to eliminate the disease from Brazilian territory by 2035. However, there are many challenges to achieving these goals, including the emergence of parasites resistant to anti-malarial drugs, the need for better vector control strategies, human migration, the need for effective surveillance and tools to identify foci of infection in low transmission areas, and the high prevalence of asymptomatic and submicroscopic infections, which can only be detected by molecular techniques [3].

Most human malaria infections are caused by five Plasmodium species: Plasmodium falciparum, Plasmodium vivax, Plasmodium malariae, Plasmodium ovale and Plasmodium knowlesi. Plasmodium knowlesi is a primate malaria parasite that primarily affects long-tailed and pig-tailed macaques, which has emerged as an important human pathogen in recent years, responsible for over 70% of human cases in southeastern Asia [4,5,6,7]. Plasmodium species that cause infection in non-human primates (NHP) are of great interest because they can be naturally-transmitted to humans, thus representing a challenge to malaria elimination.

Besides P. knowlesi, at least other three species, which infect primarily NHP, are involved in the zoonotic transmission of malaria to humans, including Plasmodium cynomolgi in Asia [8] and Plasmodium brasilianum and Plasmodium simium in the Americas [9,10,11,12]. Plasmodium simium, a parasite of a small number of species of Neotropical monkeys, is restricted to the Atlantic Forest from Southeast and South regions of Brazil and it was recently shown to cause zoonotic infections in humans [12]. An outbreak attributed to P. vivax in the Atlantic Forest areas of the state of Rio de Janeiro have been shown to be, in fact, caused by P. simium [12]. Plasmodium simium and P. vivax are genetically, morphologically, and immunologically similar, and the brown howler monkey (Alouatta guariba clamitans) has been suggested to be the main reservoir host of P. simium [9, 13, 14]. However, the reservoir status of howler monkeys is uncertain, since the only evidence currently available from the field is that P. simium infection occurs in this species, but their potential for natural transmission (i.e. infectivity to mosquitoes) in this setting is not yet known.

Despite the importance of parasites with potential for zoonotic transmission, little is known about their epidemiological importance to infect and cause disease in humans. The potential for NHP to act as reservoirs of Plasmodium infection for humans presents a challenge for malaria elimination, as they contribute to the persistence of the parasite and act as source of re-introduction into human populations, in areas where infection has otherwise controlled [15,16,17]. In this context, the occurrence and density of gametocytes, which are the infective stages of Plasmodium for mosquitoes, in reservoir hosts are important factors to estimate their potential for malaria transmission [18,19,20]. Furthermore, this information could be used for modelling the dynamics of zoonotic malaria transmission [16, 21].

Among the proteins expressed in large quantities after gametocyte activation in the midgut of the mosquito are the P25 orthologues Pfs25 and Pvs25, of P. falciparum and P. vivax, respectively [22]. Although the P25 protein is expressed on the surface of malaria parasite stages occurring within the mosquito, the transcription of its gene begins within, and - of the vertebrate host erythrocytic stages - is specific to female gametocytes. The Pvs25 gene is highly conserved among P. vivax isolates [23,24,25], making it a useful molecular marker for detection of sexual stage malaria parasites [16]. Another possible stage-specific target for monitoring the sexual stages of Plasmodium is the P48/45 protein, which is expressed in both male and female sexual stages starting at stage II gametocytes and continuing until fertilization is complete and forms a complex with P230 [26,27,28]. Despite information scarcity in the literature, it is known that P48/45 is involved in male gamete fertility, but not female fertility [29], and is evolutionarily conserved among Plasmodium species, again exhibiting low levels of genetic diversity [30]. A transcriptional study of gametocyte genes from P. vivax showed that they cluster in two groups of co-regulated genes, one includes Pvs25 and the other Pvs48/45, suggesting that the regulation of male and female genes is independent of each other [31].

The aim of the current study was to assess the potential for malaria transmission by naturally-infected howler monkeys through identification of gametocyte transcripts and estimation of gametocyte density through reverse-transcription quantitative PCR. Consequently, it was possible to identify through molecular methods howler monkeys from the Brazilian Atlantic Forest infected with gametocytes of P. simium, which were potentially infective to human beings. This finding can contribute to understanding the occurrence and intensity of malaria transmission from NHP reservoir hosts to mosquito vectors, and so help to define public policies for the control, prevention and eventual elimination of malaria.

Methods

Ethical approval

Capture, handling and blood sampling of free-living primate in the municipality of Joinville in the state of Santa Catarina, Brazil, was approved by the Ethical Committee on the Use of Animals of the Fundação Universidade Regional de Blumenau (FURB) under the protocol nº 012/15. The Brazilian government authorized this study, access to, and transport of, biological samples through the Sistema de Autorização e Informação em biodiversidade (SISBIO) no. 43375–4/2015 (for CPRJ samples) and nº 43375–6 (for Joinville samples).

Non-human primate samples

Whole blood samples were obtained from NHP from fragments of the Brazilian Atlantic Forest located in the states of Santa Catarina and Rio de Janeiro. The sampled animals were free-living NHP from Joinville (n = 32) and captive primates from the Centro de Primatologia do Estado do Rio de Janeiro (CPRJ) (n = 3), comprising two species of the Neotropical primates, the brown howler monkey, Alouatta guariba clamitans (the Atelidae family) (n = 33) and the black-headed uakari, Cacajao melanocephalus (the Pitheciidae family) (n = 2). The samples used here have been both previously diagnosed and published by our group using conventional PCR of 18S rRNA locus [32] and PCR–RFLP of cytochrome c oxidase I locus [33, 34]. The samples chosen for inclusion in this study came from: 11 P. simium-infected NHP, five P. brasilianum-infected NHP, and 13 NHP with mixed-infections (i. e. both P. simium and P. brasilianum), as well as six NHP diagnosed as non-infected. Aliquots of whole blood samples were stored in RNAprotect (Qiagen) at a 1:5 ratio and stored at −20 °C until RNA extraction.

RNA extraction and cDNA synthesis

RNA extraction was performed using different blood in RNAprotect volumes (100 to 1000 μL) using the commercial RNeasy Mini Kit (Qiagen), according to manufacturer’s protocol, which resulted in 30 μL of RNA. DNA removal was done immediately after RNA extraction using the Turbo DNA-free™ Kit (Invitrogen, Life Technologies). In order to synthesize the complementary DNA (cDNA), reverse transcription was performed using the enzyme SuperScript® IV Reverse Transcriptase (SSIV—Invitrogen, Life Technologies) and random primers (Invitrogen, Life Technologies), according to the manufacturer’s instructions. For a final volume of 20 μL, a reverse transcription reaction mix was made containing 1 μL of 50 μM random hexamers, 1 μL of 10 mM dNTPmix, up to 11 μL of template RNA (up to 500 ng mRNA), and up to 13 μL of nuclease-free water. The reaction was incubated at 65 °C for 5 min, and then on ice for at least 1 min. Next, 4 μL of 5 × SSIV Buffer, 1 μL of 100 mM DTT, 1 μL of RNaseOUT™ Recombinant RNase Inhibitor and 1 μL of SSIV were added. The reactions were performed on a Veriti 96-well Thermal Cycler (Thermo Fisher Scientific) at 23 °C for 10 min, 55 °C for 10 min, and then 80 °C for 10 min.

Amplification of 18S rRNA, Pss25 and Pss48/45

For confirmation of Plasmodium infections and detection of P. simium gametocytes, three different quantitative PCR (qPCR) protocols were performed using the cDNA obtained as described above. For the 18S rRNA transcript, the primers used were those described by Wampfler et al. [16]. For the Pss25 (P. simium sexual antigen orthologue to Pvs25) transcript, the primers used were designed for Pvs25 by the same authors, because of the high identity between P. simium and P. vivax (Additional file 1). For the Pss48/45 (P. simium sexual antigen orthologue to Pvs48/45) transcript, new primers were designed using the OligoAnalyzer software based on the sequence available in GenBank (P. vivax transmission-blocking target antigen precursor, putative, Accession Number XM_001614196.1). The best of primer sequences identified were 5'-CTCTACCGGAACCATGTTGAAG-3' (forward) and 5'-GACGTACTTGACCTCTCCTTTG-3' (reverse), which generate a fragment of 109 base pairs. For 18S rRNA transcript amplification, the qPCR reaction was performed using a 10 μL final volume containing 900 nM of each primer, 5 μL GoTaq® qPCR Master Mix, and 1 μL cDNA. The Pss25 and Pss48/45 reactions were performed, separately, using 10 μL final volume containing 200 nM of each primer (forward and reverse), 5 μL GoTaq® qPCR Master Mix, and 1 μL cDNA. The qPCR assays were performed on an automatic thermocycler ViiA7 Real-Time PCR System (Thermo Fisher Scientific) with an initial denaturation at 95 °C for 2 min, followed by 40 cycles at 95 °C for 15 s and 60 °C for 1 min. A final cycling for dissociation curve analysis of 95 °C for 15 s, 60 °C for 1 min and 95 °C for 15 s was used. A sample previously diagnosed as P. vivax by 18S RNA PCR described by Snounou et al. [32], and positive by RT-qPCR for 18S rRNA, Pvs25 and Pvs48/45 was used as a positive control in all qPCR assays, as well as a negative control (without cDNA). The results of the qPCR were analysed using the QuantStudio Real Time PCR Software v1.3.7.

RNA quantification and estimation of P. simium gametocyte density

Plasmids containing the 18S rRNA and Pvs25 fragments of interest were previously prepared by Salazar [35] and were used to perform absolute quantification using a standard curve of P. simium gametocytes. Plasmid DNA concentration was obtained by fluorimetric quantification using a Qubit 4 (Invitrogen). The plasmid copy number (PCN) was calculated to determine the dilutions to be used for the standard curve, which was constructed based on seven ten-fold serial dilutions (ranging from 1 × 106 to 1 × 101). The number of gametocytes based on Pss25 transcript copy number were estimated using a previously published conversion factor (4.17 Pvs25 transcripts/µL is equal to one gametocyte/µL) calculated by Koepfli et al., which was based on a random-effect model from log10-transformed quantities of gametocyte trend-lines [36]. For the quantification analysis, samples with a quantification cycle (Cq) ≥ 34 were not considered, since a large variation was observed between the replicates above this value (Cq SD > 0.3), resulting in unreliable quantification.

Statistical analyses

The statistical analyses were performed using GraphPad Prism 8.0.2 (GraphPad Software, San Diego, CA, USA). Linear regression was used to compare the Cq values obtained for 18S rRNA, Pss25, and Pss48/45, while non-linear correlation was used to compare the copy number between the 18S rRNA and Pss25 transcripts, through the non-parametric Spearman’s rank correlation coefficient. The sample J14 was excluded from the linear correlation analysis. The significance level of 5% was considered for all analysis.

Results

Detection of Plasmodium infection and P. simium gametocytes

Thirty-five samples of NHP previously evaluated for simian malaria were used for detection of Plasmodium infection by RT-qPCR of 18S rRNA, and gametocyte-specific identification through detection of the Pss25 and Pss48/45 transcripts (Table 1). The results for 18S rRNA showed a high percentage of positive samples (74.3%, 26 samples), mostly in agreement with our previous molecular diagnosis (82.9%, 29 of the samples).

Table 1 Detection of asexual (18S rRNA) and sexual (Pss25 and Pss48/45) transcripts of Plasmodium in non-human primates from the Atlantic Forest
Full size table

Considering only the P. simium samples that were positive for the 18S rRNA transcript (n = 21), 13 samples were also positive for Pss25, indicating that 61.9% of the 18S rRNA-positive samples contained P. simium gametocytes, of which seven samples also amplified Pss48/45 transcripts (i. e. 53.8% of the Pss25 positive samples) (Table 1).

Overall, nine samples (25.7%) were negative for all three assayed loci. Another nine samples showed discordant results between our previously molecular diagnosis and the amplification reported here of 18S rRNA transcripts by qPCR (Table 2, highlighted in orange). Surprisingly, one sample previously diagnosed as positive for P. brasilianum gave amplification using the Pvs25 primers (J21, highlighted in orange in Table 2).

Table 2 Transcript detection of Plasmodium 18S rRNA, Pss25 and Pss48/45 in non-human primate samples from the Atlantic Forest
Full size table

Among all the samples positive for at least one of the three assayed transcripts, the observed Cq values were lowest for 18S rRNA, with an average of 24.6 (range 13.8 to 33.5), while those observed for Pss25 were higher with a Cq average of 32.1 (range 24.9 to 38.4) (Table 3). The Pss45/48 target showed a Cq average of 33.5 (range 28.9 to 36.1).

Table 3 Cq values and transcript quantification of the 18S rRNA and Plasmodium simium sexual antigen 25 (Pss25) targets in non-human primate samples
Full size table

A strong positive linear correlation was identified between the Cq values for 18S rRNA and Pss25 (R2 = 0.7205, P = 0.0001) (Fig. 1A), and between those for Pss25 and Pss48/45 (R2 = 0.9032, P = 0.0010) (Fig. 1B). However, a significant linear correlation was not observed between the Cq values for 18S rRNA and Pss48/45 (R2 = 0.5386, P = 0.0604) (Fig. 1C).

Fig. 1
figure 1

Correlation analysis of Cq values and copy number of 18S rRNA and gametocyte-specific transcripts. The linear regression of Cq values between 18S rRNA and Pss25 (A), Pss25 and Pss48/45 transcripts (B), and between 18S rRNA and Pss48/45 (C). Dashed lines indicate 95% confidence interval. A non-linear correlation between estimated copy number of 18S rRNA and Pss25 transcripts using non-linear Spearman’s rank correlation coefficient (D)

Full size image

Quantification of P. simium gametocytes

Absolute quantification was performed using standard curves to estimate the copy number of both 18S rRNA and Pss25. Quantification was not performed for Pss48/45, because amplification of this locus by qPCR gave larger variation, which directly interfered in the accuracy of quantification. The twenty-six RNA samples from the 18S rRNA-positive NHP gave estimated transcript copy numbers ranging from 0.01 to 25,170.71 copies/µL, with an average of 1665.88 copies/µL (Table 3). For Pss25 quantification, the amount of transcript ranged from 0.05 to 24.17 copies/µL, with an average of 3.07 copies/µL (Table 3). However, since only samples with Cq < 34 were considered in the analysis, gametocyte quantification was assessed for only ten samples (with the remaining three Pvs25-positive samples excluded). The non-linear Spearman's rank correlation coefficient revealed a positive correlation between the copy numbers of Pvs25 and 18S rRNA (rs = 0.8645, P = 0.0052, log-transformed values) (Fig. 1D). The estimated number of gametocytes per microlitre (calculated from Pss25 transcript copy number) was very low (less than 1 gametocyte/µL), except for one animal (J14) which showed a high density of gametocytes (5.8 gametocytes/µL) (Table 3).

Discussion

In Brazil, the highest prevalence of malaria cases is in the Amazon region, with over 99% of the national notifications. However, autochthonous infections have also been described in the extra-Amazonian region, mainly in Atlantic Forest areas. The state of Rio de Janeiro recorded an average of four autochthonous malaria cases per year from 2006 to 2014 [37]. However, in 2015 and 2016, these rates increased to 33 and 16, respectively. Mitochondrial genome analyses of the parasites infecting non-human primates and humans from this outbreak revealed that they were P. simium, validating that malaria has a zoonotic transmission in this region [12, 33]. Brown howler monkeys (Alouatta g. clamitans) have been suggested to be the main reservoir of malaria in Atlantic Forest areas. In the current study, more than 60% of Plasmodium-positive samples from this non-human primate species were positive for molecular markers consistent with the presence of gametocytes in their blood, demonstrating the potential of this vertebrate host species to act as a reservoir for malaria transmission. The failure to detect gametocytes in all P. simium-infected howler monkeys may be because the low densities of the gametocytes in the peripheral blood or because of their sequestration in the bone narrow, as showed for P. vivax gametocytes [38, 39]. Interestingly, one howler monkey, previously diagnosed as positive for P. brasilianum, had positive amplification using the primers to Pvs25 locus. This amplification was unexpected because of the many polymorphisms within the primers and probe binding sites in the Pvs25 orthologue of P. brasilianum (Additional file 1). This animal might have had a mixed infection, common in the region it was from [34], with low levels of P. simium. In addition, related to the other NHP species evaluated here, one black-headed uakari (Cacajao melanochephalus) was positive for P. simium infection, but negative for both gametocyte transcripts. Since this species is an endemic of the Amazon, these results suggest that this monkey species might be a dead-end host for P. simium, without epidemiological importance for malaria transmission in the Atlantic Forest. However, because this animal was housed in captivity in the Atlantic Forest, it was not possible to exclude that the gametocytes were in low densities. This requires further investigation, since only two specimens of the black-headed uakari were included here, and large sample sizes are needed.

A deeper understanding of gametocyte carriers is essential to better understand their potential for malaria transmission. There are only a few studies of identifying gametocytes from P. simium and P. brasilianum, all based on the detection of this stage through blood smears under an optical microscope [12, 40, 41]. Molecular identification was not performed for these two Plasmodium species, since gametocyte genes were not previously studied, and their sequences were only recently elucidated with the sequencing of the complete genomes of these two parasites [42,43,44]. Considering the high similarity among P. simium and P. vivax (Additional file 1), the studies about gametocytes detection from the later were used here. Some studies have shown that P. vivax parasitaemia and gametocytaemia are tightly linked [36, 45]. Moreover, infectivity for mosquitoes is known to be positively associated with gametocyte density, with variations between mosquito species and study areas [45,46,47,48,49,50]. In Ethiopia, Golassa et al. [51] showed that 86.4% of asymptomatic P. vivax cases had asexual parasites and 13.6% had both asexual and gametocytes, another study estimated the contribution of asymptomatic for mosquito infections in 79 to 92% [48]. Kosasih et al. [52] showed that the prevalence of gametocytes is higher in individuals with microscopic than sub-microscopic P. vivax infections (92% versus 26%). Imwong et al. [53] used an ultrasensitive PCR (uPCR) method to identify many individuals infected with P. vivax in malaria endemic areas and demonstrated that parasitaemia persists in humans at levels that optimize the probability of generating densities of transmissible gametocytes without causing illness. The authors reported that as gametocytes are eliminated more slowly than asexual stages from the peripheral blood, a significant proportion of the parasites detected by uPCR in asymptomatic individuals are likely to be gametocytes. Additionally, a recent study by Almeida et al. [54] using human samples with very low parasitaemia showed that there is a possibility of mosquito infection. The authors demonstrated, through an artificial feeding assay, that blood from asymptomatic individuals can act as a source of P. vivax transmission to the vector Anopheles (Nyssorhynchus) darlingi, the main vector of P. vivax in the Brazilian Amazon. Although lower infectivity rates were observed for blood from asymptomatic individuals (2.5%) compared to symptomatic patients (43.4%), many asymptomatic carriers maintained parasitaemia for several weeks, indicating their potential role as an infectious reservoir. Symptomatic P. vivax infected individuals are able to infect mosquitoes at variable rates from 20–90% [48, 49, 55,56,57] and asymptomatic individuals can also be infective, but at much lower rates [48, 49, 57]. Although lower gametocyte densities in asymptomatic individuals are obviously much less infectious, they can contribute significantly to transmission by having a higher frequency (> = 80%) in the population [58,59,60,61]. Therefore, in Amazonia sub-microscopic and asymptomatic P. vivax infections constitute the main infectious reservoir of this parasite [62].

Molecular methods have been developed to detect transcripts of gametocyte-specific genes, with a detection limit of 0.02–10 gametocytes per microlitre of blood [63, 64]. According to Bharti et al. [56], the number of Pvs25 transcripts correlates positively with the number of circulating mature gametocytes and can be used as an indirect estimate of gametocyte density in the sample. Koepfli et al. [36] has gone further, calculating a conversion factor for Pvs25 transcripts into the number of gametocytes. Applying this conversion factor, which is defined for P. vivax, a high frequency of low gametocyte densities—less than one gametocyte/μL—was observed here. Therefore, as suggested for P. vivax human infections, the low densities of gametocytes could be compensated by the high frequency of infection. Nonetheless, one howler monkey showed higher levels of gametocytes, which may act as a potential “super-spreader” of P. simium infection. Historical studies have suggested around 10 gametocytes/μL is an infective density [65, 66]. Recently, two studies confirmed that gametocytes as few as 1 gametocyte/μL were able to infect Anopheles dirus and Anopheles stephensi [45, 67]. In the Atlantic Forest, Anopheles from Kerteszia group has been incriminated as the main vector of malaria [9, 68]. High densities of these mosquitoes species, and their highly voracious blood-feeding habits [69], together with their requirement to have more than one blood meal in order to complete their gonothrophic cycle [68], may all increase the chances of their transmitting malaria. However, functional assays, such as skin feeding assays (SFA) or direct membrane feeding assays (DMFA) using different densities of gametocytes remain to be performed in order to identify the infective density of gametocytes required for P. simium transmission to mosquitoes. Previously, this density could not be identified, since the experimental infections performed using laboratory P. simium infection of monkeys only quantified infective gametocytes by microscopy and did not perform mosquito infections using a range of different gametocyte densities [70].

Interruption of malaria transmission is considered a priority task in the process of malaria elimination [71, 72]. Therefore, it is extremely important to understand the epidemiology of gametocytes and the contribution of asymptomatic and sub-microscopic carriers acting as reservoirs, especially in low transmission settings. In this context, high frequencies of infected howler monkeys, can carry gametocytes, even with very low densities, potentially contributing to the infection of mosquitoes and, consequently, human beings. The data presented here could help in the mathematical modelling of the dynamics of zoonotic malaria transmission, which may consider the individual variation in the levels of gametocytes among reservoirs, with a high frequency of howler monkeys with low levels of gametocytes and a few NHP with higher levels. Moreover, these models must take into account the presence of other non-human primate species which maybe a dead-end hosts. The modellers have also to consider that the distribution of gametocyte between different individual hosts within a single reservoir host population/species is expected exponential. However, more studies need to be done to assess mosquito infection rates and to help to understand to what extent NHP can act as reservoir hosts and contribute to the maintenance of the Plasmodium life cycle in the Atlantic Forest.

Conclusion

Gametocytes were detected in P. simium infected brown howler monkeys (Alouatta g. clamitans). This is strong evidence that howler monkeys are acting as Plasmodium reservoirs in the Atlantic Forest. Transmission could be maintained by the high frequency of low-level gametocyte carriers and the low frequency of high-level carriers. This finding will contribute towards the modelling of zoonotic malaria transmission and definition of public policies for malaria control, prevention and elimination.