Parasitological correlates of Plasmodium ovale curtisi and Plasmodium ovale wallikeri infection

Malaria, due to Plasmodium ovale, can be challenging to diagnose due to clinically mild disease and low parasite burden. Two genetically distinct sub-species of P. ovale exist: Plasmodium ovale curtisi (classic) and Plasmodium ovale wallikeri (variant). It is presently unknown if the sub-species causing infection affects performance of malaria diagnostic tests. The aim of this work was to understand how the genetically distinct sub-species, P. o. curtisi and P. o. wallikeri, affect malaria diagnostic tests. Plasmodium ovale-positive whole blood specimens were sub-speciated by PCR and sequencing of 18S rRNA and dhfr-ts. Parasitaemia, morphology, pan-aldolase positivity, 18S copy number, and dhfr-ts sequences were compared between sub-species. From 2006 to 2015, 49 P. ovale isolates were identified, of which 22 were P. o. curtisi and 27 P. o. wallikeri; 80% were identified in the last five years, and 88% were acquired in West Africa. Sub-species did not differ by parasitaemia, 18S copy number, or pan-aldolase positivity. Lack of Schüffner’s stippling was over-represented among P. o. wallikeri isolates (p = 0.02). Several nucleotide polymorphisms between the sub-species were observed, but they do not occur at sites believed to relate to antifolate binding. Plasmodium ovale is increasing among travellers to West Africa, although sub-species do not differ significantly by parasitologic features such as parasitaemia. Absence of Schüffner’s stippling may be a feature specific to P. o. wallikeri and is a novel finding.


Background
Plasmodium ovale is one species of malaria parasites that is pathogenic to humans [1]. Prevalent in West and East Africa and the Asia-Pacific [1], P. ovale infections can be challenging to diagnose in part due to the typical presentation of clinically mild disease and low parasite burden [2]. Consequently, conventional diagnostic tools, such as microscopy of blood films and rapid diagnostic tests (RDT), have limited performance in detecting P. ovale infection [3][4][5], often resulting in the use of molecular techniques to confirm the diagnosis [5].
falciparum and Plasmodium vivax. However, investigations into this region have only recently begun for P. ovale isolates [2,8,11]. Previous reports have implicated specific sites in the peptide sequence that may interfere with antifolate binding at the active site, which include D53, F57, S58, T61, S113, S116, I169, and T190 [8,11]. R499/503 and C519/523 (P. ovale curtisi/P. ovale wallikeri) are also believed to be significant for thymidylate synthase activity, specifically [8]. As atovaquone-proguanil, the proguanil component of which is an antifolate, is one of the most commonly and empirically used antimalarials for travel-acquired malaria, it is important to understand if mutations in regions of dhfr-ts known to confer resistance in other species of malaria are present in P. ovale isolates, and if these mutations are specific to one sub-species over another.
Further understanding of parasitologic correlates of P. ovale sub-species may have important implications for diagnosis and management. The potential differences in parasite burden, RDT performance, morphology, and molecular diagnosis between P. ovale curtisi and P. ovale wallikeri were investigated. The DHFR-TS peptide sequence was also compared between isolates to investigate potential mutations in regions known to confer antifolate resistance in other species of malaria.

Specimens
Plasmodium ovale-positive, whole blood specimens stored in the malaria biobank at Public Health Ontario Laboratories (PHOL) from October 2006 to July 2015 were identified and retrieved. For the staining procedure, 2 mL of standard Giemsa stain was diluted in 48 mL of phosphate buffer (pH 7.1) prior to use. The phosphate buffer consists of 65 mL of Na 2 HPO 4 , 35 mL of NaH 2 PO 4 , and 900 mL of distilled water, resulting in 4.35 and 2.33 mM concentrations, respectively. For thin films, the slide was immersed in methanol for 10 s, allowed to dry, and then immersed in the working Giemsa solution for 10 min. The slide was then immersed in phosphate buffer solution for 10 s and air dried. Thick films were immersed in the working Giemsa solution for 10 min, immersed in the phosphate buffer for 30 s, and air dried. With each working solution prepared, a control P. falciparum slide was prepared and examined. Parasitaemia, morphological features, pan-aldolase antigen-positivity as determined by the BinaxNOW Malaria test (Alere, Ottawa, ON, Canada), year of import, and country of acquisition that were recorded in the biobank database following initial diagnostic processing were collected and analysed. The study was approved by the Research Ethics Board of Public Health Ontario.

DNA extraction
DNA of all specimens was extracted using the DNA Mini Kit Blood or Body Fluid Spin Protocol (Qiagen, Germantown, MD, USA). Prior to use, DNA was stored at −20 °C.

Qualitative and quantitative real-time PCR
Plasmodium falciparum/P. vivax species-specific duplex [6] and Plasmodium malariae/P. ovale species-specific duplex real-time PCR (qPCR) assays were conducted to confirm microscopy species identification as previously described [12]. All qPCR assays were run under the following conditions using the ABI 7900HT qPCR system: 50 °C for 2 min, 95 °C for 10 min, 95 °C for 15 s, and 60 °C for 1 min (45 cycles); 12.5 µl of TaqMan universal PCR master mix (Life Technologies, Burlington, ON, Canada) and 5 µl of DNA primers and probes with concentrations as previously reported [6,12,13] were used, for a final volume of 25 µl per reaction. All qPCR amplification curves were analysed using a manual threshold cycle of 0.02 and an automatic baseline. A cycle threshold (Ct) of <38 was considered to be a positive result.
18S rRNA gene copy numbers were quantified by running the P. malariae/P. ovale species-specific duplex qPCR assay and including serial dilutions of a P. ovale clone (ATCC, Manassas, VA, USA) alongside specimen DNA in triplicates. A linear regression was then constructed based on the logarithm of the gene copy number and Ct values for each concentration of the clone. This equation could then be used to calculate the gene copy number for each banked specimen.

Sequencing of 18S rRNA and dihydrofolate reductase-thymidylate synthase
Endpoint PCR of target regions, visualization of the amplicons on agarose gel, and Sanger sequencing were conducted as previously described [12]. Specifically, endpoint PCR of a 396-bp product from the 18S rRNA region was conducted with high-fidelity polymerase AccuPrime Pfx Supermix (Life Technologies, Burlington, ON, Canada) and 200 nM (each) of the primers Plasmo 18S forward (5′-ATTCAGATGTCAGAGGTG AAATTCT-3′) and Plasmo 18S reverse (5′-TCAATCC TACTCTTGTCTTAAACTA-3′). Using an ABI Veriti fast thermal cycler, the cycling conditions were 95 °C for 5 min, followed by 95 °C for 15 s, 58 °C for 30 s, and 68 °C for 30 s for 45 cycles, and then 68 °C for 5 min. Amplicons were visualized on 1% agarose gels with ethidium bromide prior to Sanger sequencing. For sequencing, the same forward and reverse primers were used along with a BigDye Terminator v3.1 cycle sequencing kit (Life Technologies, Burlington, ON, Canada), which was run according to the manufacturer's recommended conditions. The products were purified by a BigDye XTerminator (Life Technologies, Burlington, ON, Canada) and analysed with an ABI 3130xl genetic analyzer. Forward and reverse sequences of each sample were aligned using Vector NTI software (Life Technologies, Burlington, ON, Canada), and the sequences were verified using a BLAST search to confirm the sub-species identification [14].
Primers for sequencing the dhfr-ts gene were newly designed for this study. All primers were designed using Primer3 [15,16] and predominantly conserved regions of the complete sequence of the P. ovale dhfrts gene (GenBank: EU266602). Degenerate nucleotides were included based on the dhfr-ts gene for P. o. curtisi (GenBank: KP050414) and P. o. wallikeri (Gen-Bank: KP050415). Endpoint PCR of the dihydrofolate reductase-thymidylate synthase region was conducted on all specimens with high-fidelity polymerase Phusion (New England BioLabs, Ipswich, MA, USA) and 200 nM (each) of the primers PocPowDHFR-exF (5′-YTCWACC TTCAGGGGTATCG-3′) and PocPowDHFR-exR (5′-AG TTTWAGCGTGGGRAAAGG-3′), generating an approximately 1700-bp product that covers the 99th amino acid to 1814-1826th amino acid of the sequence, where the end target varies based on the sub-species. The cycling conditions were 98 °C for 30 min, followed by 98 °C for 10 s, 64 °C for 30 s, and 72 °C for 1 min for 40 cycles, and then 72 °C for 10 min using an ABI Veriti fast thermal cycler. Visualization and Sanger sequencing of the PCR products were performed identically compared to the 18S rRNA methodology with the exception of the primers used. For sequencing, PocPowDHFR-exF, PocPowDHFR-exR, and a third internal primer, PocPowDHFR-inR (5′-TTTCCATTKGTTTCCCCTC T-3′) were used. Sequences for each specimen were aligned and verified using a BLAST search to further confirm the sub-species identification. Alignments of P. ovale sequences were performed using Mega6 software [17]. Genetic and peptide sequences between the enrolled specimens and complete P. ovale sequences that were previously reported (GenBank: EU266601, EU266604-EU266618, GQ250090, GQ250091, KP050409, and KP05413-KP050415) [2,7,18] were compared.

Statistical analysis
Reported parasitaemia, morphological features (including Schüffner's stippling), pan-aldolase antigen-positivity, 18S gene copy number, year of import, and region of acquisition were compared between the two sub-species. Parasitaemia was treated as categorical data by recording if the percentage of parasites observed were <0.1 or ≥0.1%, as the former category is a convention used by PHOL laboratory technicians. Categorical variables were compared using Yates' corrected Chi squared analysis or Fisher's exact test, and continuous variables using Mann-Whitney U tests using GraphPad Prism 6 (Graph-Pad Software, CA, USA). Level of significance was set at p < 0.05.

Results
Of 49 isolates of P. ovale identified from October 2006 to July 2015, ten (20.4%) were identified through 2010, while 39 (79.6%) were identified in or after the year 2011, with 24 (49%) identified in 2013 or 2014, the last two full years in the enrolment period. Country of exposure data were available for 17/49 (34.7%) specimens, with Nigeria being the most well represented source country (n = 9, 18.4%). Of the 17 isolates with an attached travel history, 15 (88.2%) were acquired in West Africa, with represented source countries of Nigeria, Ghana (n = 2), Ivory Coast (n = 1), Liberia (n = 1), Sierra Leone (n = 1), Congo (n = 1), and Cameroon (n = 1). Two isolates were imported from East or Saharan Africa, one each from Tanzania and the Sudan.
When comparing morphological features (Table 2), it was noted that all eight P. ovale parasites without Schüffner's stippling were P. ovale wallikeri (p = 0.02) (Fig. 2). Parasitaemia (p = 0.20), 18S rRNA gene copy number (p = 0.50), and pan-aldolase antigen positivity (p = 0.40), were not found to be differentially associated with Schüffner's stippling. Data on time between specimen collection and receipt were missing for two P.

Dihydrofolate reductase-thymidylate synthase (dhfr-ts)
Partial sequences of dfhr-ts could be obtained in 16 P. ovale wallikeri and seven P. ovale curtisi. In 26/49 isolates, the dhfr-ts region selected could not be amplified by the selected primer set. Several differences in the peptide sequences between P. ovale curtisi and P. ovale wallikeri were observed (Table 3). Non-synonymous mutations were observed within the P. ovale wallikeri isolates as previously reported [2]. Specifically, deletions of two or four amino acids (i.e., TA252-253 or TATA252-255) were identified in six and one P. ovale wallikeri isolates, respectively (Table 3). Furthermore, due to an ambiguity in the nucleic acid sequence, potential Y359F and Y366F mutations were observed. Finally, for one specimen, a H285Q mutation was recorded for one P. ovale wallikeri isolate. While Y359F and Y366F mutations are not expected to change any amino acid properties, a H285Q indicates a change from a basic histidine to a polar, uncharged glutamine. Among the sequences analysed in this study, polymorphisms specific to either sub-species were not observed (Table 4). One P. ovale wallikeri specimen did have a I169L mutation, although this does not reflect any changes in amino acid properties.

Discussion
Plasmodium ovale is increasingly observed among travellers, and due to its ability to relapse months to years after primary infection; unlike the more common and severe P. falciparum, its ability to inflict acute morbidity continues long after the immediate post-travel period. The data reported herein underscore the anecdotal impression among laboratory personnel that P. ovale is increasing NA not available    [12]. Although it appeared in this study that many more P. ovale wallikeri isolates were imported in the past five years compared to P. ovale curtisi, this difference was not significant, although with higher numbers over a longer enrolment period such a difference may emerge. Due to widespread attrition of trained microscopists in laboratories of non-malaria endemic regions, such as North America, diagnosis of malaria is becoming increasingly reliant on RDT and molecular methods. However, given the extremely poor sensitivity of standard RDT for P. ovale infection [19], as well as typically low parasitaemia, under-and delayed diagnosis are increasingly likely. In addition, diagnostic morphologic features may not be demonstrable in the setting of extremely low parasitaemia, prompting empiric use of anti-malarials for a patient whose malaria screen is initially reported as "Plasmodium spp. " pending confirmatory molecular analysis. Standard choice of empiric oral therapy for uncomplicated malaria in North America is usually either atovaquone-proguanil (AP), or more rarely, artemether-lumefantrine. As resistance to the proguanil component of AP is conferred in P. falciparum by step-wise mutations of the Dhfr gene, understanding whether or not similar mutations can accumulate in P. ovale, and whether such mutations are over-represented among a particular sub-species of P. ovale, is important for informing empiric choice of antimalarials in a patient with unspeciated malaria.
Among peptide sequences of DHFR-TS in isolates of P. ovale curtisi and P. ovale wallikeri studied herein, mutations known to confer antifolate resistance in other species of malaria, notably P. falciparum were not demonstrated Furthermore, non-synonymous mutations within each sub-species, P. ovale wallikeri in particular, were also observed, confirming previous reports [2]. While there are variations in the peptide sequences between sub-species, it is unknown how these polymorphisms influence enzyme activity. Amino acids in the DHFR-TS that are believed to be involved in antifolate binding, and which are believed to be synonymous with amino acids found in other Plasmodium species, were found to be consistent throughout all examined P. ovale specimens, regardless of sub-species. These findings suggest that the lack of clinical treatment failures observed in patients with P. ovale correlate well to gene sequences free of known markers of resistance, and are reassuring from the perspective of empiric choice of anti-malarials. It would be very unlikely, and unfortunate, for a traveller from West Africa with unspeciated malaria to be started empirically on chloroquine, which is actually the drug of choice for P. ovale, rather than AP.
Many reports have documented the existence of the two sub-species of P. ovale; it was previously unknown if parasitaemia, 18S rRNA gene copy number and the detection of the pan-aldolase antigen in BinaxNow RDT could differ between them. The results reported herein indicate that P. ovale curtisi and P. ovale wallikeri do not differ significantly by parasitaemia, pan-aldolase antigenpositivity or 18S rRNA gene copy number. A general low pan-aldolase antigen-positivity rate among all specimens was observed, which is consistent with previous literature [19,20]. A lack of discernible Schüffner's stippling may be a feature specific to P. ovale wallikeri, and may   ND  ND  ND  ND  S  S  I  T  R  C   28  ND  F  S  T  S  S  I  T  R  C   29  ND  ND  ND  ND  S  S  I  T  R  C   30  D  F  S  T  S  S  I  T  R  C   33  ND  F  S  T  S  S  I  T  R  C   34  ND  F  S  T  S  S  I  T  R  C   35  D  F  S  T  S  S  I  T  R  C   37  D  F  S  T  S  S  I  T  R  C   39  ND  F  S  T  S  S  I  T  R  C   41  D  F  S  T  S  S  I  T  R  C   42  ND  F  S  T  S  S  I  T  R  C   43  ND  ND  ND  ND  S  S  I  T  R  C   44  ND  ND  ND  ND  S  S  I  T  further challenge microscopists who rely on this as one of the features to distinguish P. ovale from the similarly low parasitaemic P. malariae, particularly when not all differentiating morphologies are present on the slide. As previously noted, P. ovale wallikeri cases appeared to increase more so than P. ovale curtisi in the past five years, and if this trend were to continue, the lack of Schüffner's stippling may further hinder interpretation of microscopy and RDT results. The lack of Schüffner's stippling among P. ovale wallikeri isolates joins other findings that the subspecies have distinguishing clinical characteristics, such as latency and degree of thrombocytopaenia, in addition to their genotype [9,10]. That said, molecular techniques are still the most accurate and specific method for determining sub-species of P. ovale.
Limitations of the study include a relatively small sample size which limited the ability to comment on the possible emergence of variant over classic P. ovale, an inability to investigate DHFR-TS enzyme activity, and possible variation in Schüffner's stippling visualization due to factors such as the age of the Giemsa stain or operator experience. As well, the attempt to account for time for specimen processing may be limited in terms of precision, as the exact time when blood was collected from patients was unavailable, meaning that this variable could only be analysed on the order of days rather than hours, the latter of which would have been ideal. Full clinical linkage of specimens was lacking and therefore data on latency of P. ovale infection or full travel details were unavailable. The limited availability of gene sequences in GenBank also restricted the ability to investigate the full dhfr-ts region and other potential virulence factors that could theoretically differ between the sub-species.
Future studies, including whole genome sequencing of P. ovale, will facilitate identification of genes of other known Plasmodium virulence factors that have not been previously investigated in P. ovale. In particular, nucleotide and peptide heterogeneity between the two sub-species can be further explored across a larger sample size to extrapolate potential consequences in protein structure. Clinical outcomes can also be investigated prospectively to determine whether specific polymorphisms have any effect on delayed parasite clearance and treatment failure. Future prospective investigation of the Schüffner's stippling finding in P. ovale wallikeri versus P. ovale curtisi by diligent exclusion of competing factors, such as time to specimen processing, would further add to the literature around morphological identification, considering that microscopy remains the gold standard diagnostic test in malaria, and is often the only diagnostic test available in under-resourced settings.

Conclusions
This study demonstrates that the two P. ovale sub-species, P. ovale curtisi and P. ovale wallikeri, are increasing among travellers to West Africa, and do not differ significantly by parasitaemia, RDT positivity, or 18S rRNA gene copy number, and may therefore be treated similarly from a clinical management perspective. A lack of discernible Schüffner's stippling on microscopic examination may be a feature specific to P. ovale wallikeri, and one which may significantly challenge the diagnosis of this typically low parasitaemic species, especially as loss of technical expertise in diagnostic laboratories continues. Although nucleotide polymorphisms in dhfr-ts were identified, it is unknown if they could lead to antifolate resistance, as they did not occur in regions known to confer antifolate resistance in other species such as P. falciparum. Further interrogation of the relevance of such genetic differences is warranted.