Introduction

Malaria is triggered by an apicomplexan parasite, Plasmodium spp. 215 millions of cases and 429 000 deaths occurred in the world in 20151. Six Plasmodium species infect humans: Plasmodium falciparum (Pf), Plasmodium vivax (Pv), Plasmodium malariae (Pm), Plasmodium knowlesi (Pk) and Plasmodium ovale spp (P. ovale spp) including Plasmodium ovale curtisi (Poc) and Plasmodium ovale wallikeri (Pow). Although malaria transmission was absent, France was the most exposed non endemic country for malaria2. Imported malaria cases diagnosed in France are reported to the National French Malaria Reference Center (FNMRC). In 2015, whereas Pf infection represented most of malaria cases (89%), P. ovale spp were involved in 6.8% of cases (4690 cases total)3, all contracted in Africa. Moreover, seven cases of Plasmodium ovale spp relapse (defined by malaria disease more than one year after the return from an endemic area) occurred in these patients (46% of the 15 observed relapses)3.

Malaria symptoms are nonspecific and the observation of thick and thin stained blood smears remained the reference method of diagnosis4,5. This technique is rapid and unexpensive, but required an experienced microscopist to ensure the quality of the result. However, species discrimination could be challenging, as Poc and Pow which cannot be reliably differentiated using this methodology.

Alternative ways to identify Plasmodium infection, such as Rapid Diagnosis Test (RDT) and PCR methods, have recently become available. Molecular biology diagnosis based methods are in constant expansion, in particular real-time methods. In the last decades, the malaria researcher community has developed several primers sets for species identification, to increase sensitivity in Plasmodium diagnosis.

First, nested-PCR methods with agarose gel revelation were developed. Probe based real-time PCR assays such as Taqman have been developed later on, and increased sensitivity and specificity. However, the sample processing and the technique itself are time consuming and cannot be applied in routine diagnosis procedures.

Up to 2010, scientists and biologist referred to Poc and Pow as P. ovale spp6. It is now admitted that Poc and Pow are two distinct species and not subspecies7. Poc and Pow are minor forms of malaria, however severe disease forms and deaths have been reported8. Differences in clinical implication, specific geographic repartition and response to antimalarial therapy might also exist between Poc and Pow, but the respective implication of each has not been completely studied yet. In addition, Poc and Pow physiopathology is unknown, especially mechanisms leading to the dormant form hypnozoite formation and relapses9,10. In fact, some differences in the latency period of the hypnozoites were highlighted11, higher thrombocytopenia and nausea with Pow was reported by another study12. Yet, microscopic examination of blood smear does not allow differentiation of the two species, even if a recent study described a lack of discernible Schüffner’s stippling in Pow samples13. Moreover, RDT efficiency was low and no differences in RDT positivity were found between Poc and Pow 13. Therefore, differentiation of Poc and Pow remains challenging, and the development of convenient tools to differentiate Poc and Pow is crucial to investigate the burden on malaria infections of each species, as well as their clinical implication.

Poc and Pow can be differentiated based on differences in their DNA sequences. Nested-PCR technique has been developed14 as well as qPCR protocols15,16,17,18. However, nested-PCR techniques are time-consuming requiring a two-step amplification followed by a PCR product migration on agarose gel. The entire process takes almost a day, and needs a dedicated environment for post-PCR products’ manipulation. Taqman probe based techniques are highly accurate and specific, but are expensive.

Here, we developed a new qPCR technique with High Resolution melting (HRM) (qPCR-HRM) revelation with the aim to reliably differentiate Pow and Poc, using samples from patients with imported malaria received at the FNMRC.

Results

Of the 356 samples were included in the study period, 25 were negative for Plasmodium infection and 331 were positive according to qPCR-Taqman results (Supplementary Table S1): 219 were positive for Pf, 69 Po (subsequently identified as 33 Pow, and 36 Poc with nested PCR), 20 for Pm, 17 for Pv. 6 samples present mixed infection: 4P. ovale spp + Pf (subsequently identified for two P. ovale spp as Poc), 1P. ovale spp (subsequently identified as Poc) + Pm and 1P. ovale spp + Pm. In microscopy, among 343 observed by microscopy, we detected 46 negative, 6 Plasmodium spp, 182 Pf, 20 Pm, 70 P. ovale spp, 17Pv, 1 Pf + Pm, 1 Pf + P. ovale spp.

For specificity analysis, we also included 6 other parasitological positive samples from blood parasites (Toxoplasma gondii, Leishmania infantum, Loa loa, Onchocerca volvulus, Wuchereria bancrofti and Babesia divergens).

qPCR-HRM development

In silico prevision

In silico alignment was performed from predictive PCR product using the primers chosen for qPCR-HRM analysis. Plasmo-1F and Plasmo-2R primers amplify a 160-based pair (bp) and 157 bp sequence for Poc and Pow respectively. As a sequence variation in predictive PCR product for Poc and Pow was observed, HRM reaction could highlight a difference between both species. Using the specific uMELT online tool for PCR product melting temperature (Tm) analysis, predicting melting temperature from the HRM phase based on an algorithm analyzing DNA bounds strength, Tm could be predicted for the amplified sequences (77.4 °C for Pow and 76.8 °C for Poc). Predicted Tm as well as theorical derivated melting curves for other Plasmodium species are displayed in Supplementary Figure S2.

Distinction of P. ovale wallikeri and Plasmodium ovale curtisi

Among the 69P. ovale spp samples confirmed by qPCR-Taqman, Tm values for each species were reproducible (see Table 1 for average Tm and interquartile interval for Poc and Pow). 32 Pow and 34 Poc were identified by qPCR-HRM and these results were compared with nested-PCR followed by an electrophoresis in agarose gel and species were determined with 100% of concordance (Supplementary Table S1). However, by qPCR-HRM, 3 samples were misidentified as Plasmodium malariae: two were Poc and one was Pow by nested-PCR.

Table 1 Median melting temperature and interquartile range for each species.
Full size table

Specific geographic repartition of Poc and Pow in the included samples is presented in Fig. 1 and in Supplementary Table S1. Pow seems to be overrepresented in Cameroon whereas Poc seems to be overrepresented in Central African Republic. Poc and Pow are co-endemic in Ivory Coast, Cameroon, Central African Republic, Mali, Congo.

Figure 1
figure 1

Map representing the origins of P. ovale spp samples included in the study. Orange represent Poc and purple Pow. Sign size is proportional to the number of samples from a given area. The smaller sign represents one sample, and the bigger one five samples. This map is modified from this URL: http://www.mapsopensource.com/africa-countries-outline-map-black-and-white.html whice provided open source map with no permission required (https://creativecommons.org/licenses/by/3.0/).

Full size image

Mixed artificial samples with both Poc and Pow

Both Poc and Pow were distinguished in artificial samples, except when Pow was present in a strong majority (ratio 80% Pow/20% Poc and 90% Pow/10% Poc). Derivative melt curves of those samples are presented in Supplementary Figure S3. Same conclusions were made for Pf, Pv. Pm-Po mixed infections which could not be detected in mixed artificial samples.

qPCR-HRM characteristics

Repeatability and reproducibility

Repeatability was evaluated for Poc and Pow using three different parasite’s densities. Repeatability was between 0.38 standard deviation (SD) and 0.49SD of Ct for Poc and, between 0.55 to 1.26 SD of Ct for Pow. Reproducibility was not evaluated as the assay was not quantitative. However, comparing the Tm for 10 samples on 10 distinct reactions, we observed a constant Tm in the different assay. Regarding Poc, coefficient variation was between 0.18 to 0.25% (70.90+/−0.132 to 70.90 +/− 0.162) for Tm1 and between 0.17 to 0.54% (73.98 +/− 0.13 °C to 74.31 +/− 0.405 °C) for Tm2. Regarding Pow, coefficient variation for Tm1 was between 0.20 to 0.47% (71.446 +/− 0.144 °C to 41.434 +/− 0.314 °C) and between 0.19 to 0.24% (73.708 +/− 0.144 °C to 73.834 +/− 0.174 °C) for Tm2.

Sensitivity and specificity

LOD was the lowest parasite concentration that had repeated positive signal detection with qPCR-HRM19 and was evaluated at 1 parasite/µL for both Poc and Pow.

No cross-reaction with Toxoplasma gondii, Leishmania infantum, Loa loa, Onchocerca volvulus, Wuchereria bancrofti, and Babesia divergens was detected with the qPCR-HRM technique. We observed significant amplification for Toxoplasma gondii and Wuchereria bancrofti but none of the Tm were specific to any Plasmodium (Supplementary Table S4).

As in silico prevision for the ability of the qPCR-HRM to distinguish P. ovale spp from other Plasmodium species, we applied the qPCR-HRM on 356 Plasmodium positive samples to confirm experimentally in silico data. Concordance between the qPCR-HRM and the qPCR-Taqman FTD Malaria Differentiation was observed for 336 samples (93.6%). Among the 23 (6.4%) discordant samples, there were with qPCR-HRM 11 false negative (sensitivity = 97%), 5 undetected species mixed (only one of both species detected), and 7 false positive results with crossing reactions between Pf and Pv (2 cases), P. ovale spp and Pm (3 cases), Pm and Pv (2 cases). No false positive were reported (specificity of 100%). We could distinguish Plasmodium species because Pf and P. ovale spp displayed a profile pattern with two specific Tm (Table 1) whereas melting curves obtained with samples positive for Pm, Pv and Pk displayed a profile with one specific Tm. Poc and Pow melting curves derivative plot showed two distinct profiles with Tms specific of each (70.529 °C and 73.750 °C for Poc, and 71.165 °C and 73.387 °C for Pow) (Fig. 2). In our batch of sample, we described three P. ovale spp with one remaining Tm (2 Poc and 1 Pow) with no relation with the Ct (mean Ct equal to 25.755 +/− 2.188).

Figure 2
figure 2

Plasmodium ovale wallikeri and Plasmodium ovale curtisi in derivative (a) and aligned melt curves (b) obtained after HRM phase in Viia7. X axis represents the temperature (°C). Y axis represent the fluorescence (derivative melt curve, a) or the aligned flurorescence (aligned melt curves, b). Plasmodium ovale curtisi profile is displayed in blue, and Plasmodium ovale wallikeri is displayed in orange.

Full size image

All isolated species, except Pm, are in good concordance between qPCR-HRM and qPCR-Taqman (Table 2).

Table 2 Comparison between qPCR-Taqman, nested-PCR and qPCR-HRM results, by species, alone or in association.
Full size table

Discussion

The existence of the two species Poc and Pow has been known since 2010, but has currently no impact on clinical care of the patient. It is not possible to differentiate them by optical microscopy. However, some clinical and biological differences have been noticed, such as higher proportion of thrombocytopenia and nausea with Pow 12. It also appears that Poc is more frequently asymptomatic20. For these reasons, a cost-effective qPCR-HRM method was developed for the differentiation of Poc and Pow species.

qPCR-HRM allowed us to discriminate Poc and Pow in confirmed P. ovale spp positive samples. To our knowledge, this is the first qPCR-HRM that allows Poc and Pow discrimination. LOD is similar to other qPCR-HRM already published21. Repeatability and reproducibility are really satisfying. From selected primer already published22, 18S RNA amplicons were aligned with the chosen primer to ensure the sequences divergence, and therefore the Tm variation, before setting the qPCR-HRM assay. The sequence presents noticeable variation between the Pow and Poc (98% consistency), and each parasites contains four to eight copies guaranteeing a good sensitivity to the PCR23 but this also complicates quantitative approaches. In silico prediction shows us different predictive Tm for Poc and Pow that convinced us to test Plasmo1-F and Plasmo2-R. In that in silico prediction, distinction between P. ovale spp and other Plasmodium species was also possible.

From 69 samples of P. ovale spp: 66 were identified by qPCR-HRM, and among those, concordance with nested PCR for species determination was perfect (100% concordance). This qPCR-HRM is fully efficient in Poc and Pow discrimination, after P. ovale spp confirmed identification: identification of other species than Poc or Pow don’t have to be considered because we demonstrated false identification. qPCR methods have already been published to distinguish species of P. ovale spp using specific probes in qPCR-Taqman18 but we present here an innovate way of PCR revelation. In our study we highlighted that Poc and Pow are co-endemic in some African countries such as Congo as previously described15. We demonstrated that Poc and Pow are also co-endemic in Ivory Coast, Cameroon, Central African Republic and Mali.

To assess the in silico data for the predicted Poc and Pow discrimination with other Plasmodium species infecting humans, qPCR-HRM was performed on other malaria positive samples. We couldn’t find the right specie in every sample. 18S RNA gene copies are located in 4 chromosomes among Plasmodium genome (chromosome 1, 5, 7 and 13). And for Pf, copies present 2 different sequences of 18S RNA gene. It explains the existence of two Tms as described by Kassaza et al.24 and shown in Supplementary Figure S2. We assume that the double Tm profile for P. ovale spp correspond also to genomic sequence variation (Fig. 2). Pm and Pv exhibit one specific Tm, with a difference in Tm of at least 0,5 °C between the two species. We tested two samples of Pk to assess the specificity of Po detection because P. ovale spp and Pk are co-endemic in Asia. Distinction of Pk and P. ovale spp was possible in all cases.

We reported 23 discordances in 356 samples (6.4%) in our data set between qPCR-HRM and qPCR-Taqman (FTD Malaria Diagnosis). There were eleven false negative results (nine for Pf and two for Pm) and discordances in species determination. We didn’t detect species combination in clinical samples whereas it was possible in artificial sample calibrated in laboratory, related to the difficulty to reveal the minor species. But, according to FNMRC data, P. ovale spp is rarely present in Plasmodium coinfection (3% in 2016)3 and co-infection malaria infections appear rare in import malaria (less than 1% of malaria infections in 2014 in United States25).

The rationale behind the development of our qPCR-HRM method was cost-efficiency for separating minor DNA sequence variation. It also provides an affordable molecular tool for laboratories, since only standard primers and protocols are required, and a same mix and protocol allows several analyses. qPCR-HRM allows examination of mutations in 400 bases pairs fragment whereas probe-based qPCR assays examine only twenty bases pairs fragment, are more expensive and present technical pitfalls which must be considered in large scale studies (e.g. probe conservation). Compared to nested-PCR, qPCR-HRM is faster, easier to interpret and limit experimental errors.

This qPCR-HRM allowed us to distinguish Poc and Pow in 100% cases. However, the technique does present some limits, especially for Pm with a risk of misdiagnosis. Here, some technical improvements could be evaluated, such as Mutant Allele Amplification Bias (MAAB)26 or cold-PCR27 to discriminate the minor species in mixed infection, using the qPCR-HRM method we have described here. Those technical innovations should be conducted in the context of Plasmodium genomes and may be used to detect minor species in a mixed sample. Additional solutions may be considered, such as an increase DNA concentration to lower sensitivity threshold, or increase time of PCR (with a higher risk of false positive).

To conclude, we developed a qPCR-HRM for Poc and Pow discrimination, that can be used in epidemiological studies on the geographical distribution of each species or in studies aimed at detecting differences in clinical, prognostic and sensitivity to anti-malaria drugs. In general, this will facilitate a better understanding of the biological differences between Poc and Pow and their incidence in pathology.

Materials and Methods

Sample collection

Between January 2015 and August 2015, fresh blood samples collected on EDTA received in the FNMRC for expertise from patients were included. We also tested 2 Pk positive samples.

No specific consent was required because of, in coordination with Sante Publique France organization for the care and surveillance of malaria, the parasitological data were collected in the FNMRC database and analyzed in accordance with the common public health mission of all French National Reference Centers. The study of the biological samples obtained in the medical care context was considered as a non-interventional research (article L1221-1.1 of the French public health code) only requiring the non-opposition of the patient during sampling (article L1211-2 of the French public health code). All data collected were anonymized before analysis.

DNA extraction

DNA was extracted from a sample of 200 µL of whole blood sample using Magnapure® (Roche diagnosis, Bale, Switzerland) and was eluted in 100 µL of buffer following manufacturer’s instructions. DNA was stored at −20 °C until species analyses.

Species identification

On each included sample, microscopy diagnosis and qPCR-Taqman for Pf, Pv, Pm, P. ovale spp determination. On positive samples for P. ovale spp with qPCR-Taqman, Poc and Pow distinction was determined by nested-PCR. We then performed qPCR-HRM on each sample.

Microscopic investigation

Thick blood smears were considered positive if one or more malaria parasites were visualized and negative if no parasites were detected after examining 1000 white blood cells. Thin blood smears were read by two distinct operators. Parasite species was determined using thin blood smears. Parasitaemia was expressed in parasite density (parasites/µL) following this formula:

$${\rm{parasite}}\,{\rm{density}}={\rm{45000}}\ast {\rm{parasitaemia}}\,(\mathrm{parasites}/\mu {\rm{L}}).$$
(1)

qPCR-Taqman method

Species identification was confirmed by qPCR-Taqman (FTD Malaria, Fast-track Diagnostics, Launch diagnostics®) following manufacturer’s instruction. All samples were run in qPCR-Taqman.

Nested-PCR for P. knowlesi determination

Pk infection was confirmed by PCR as described by Singh et al.28.

Nested-PCR for determination of Poc and Pow

All P. ovale spp positive samples in qPCR-Taqman were run in nested-PCR. We used nested-PCR previously described for Poc and Pow determination29. The first PCR employed primer pair rPLU1 (5’-TCA-AAG-ATT-AAG-CCA-TGC-AAG-TGA-3′) and rPLU5 (5′-CCT-GTT-GTT-GCC-TTA-AAC-TTC-3′) primers for a Plasmodium genus reaction30. The second PCR reaction was performed with primer pair rOVA1 (5′-ATC-TCT-TTT-GCT-ATT-TTT-TAG-TAT-TGG-AGA-3′)/rOVA2 (5′-GGA-AAA-GGA-CAC-ATT-AAT-TGT-ATC-CTA-GTG-3′) for Poc identification31 and primer pair rOVA1v (5′-ATC-TCC-TTT-ACT-TTT-TGT-ACT-GGA-GA-3′)/rOVA2v (5′-GGA-AAA-GGA-CAC-TAT-AAT-GTA-TCC-TAA-TA-3′) for Pow identification14,29. PCR products were visualized on 2% agarose gels stained with Gelred® (10000X, Biotium, inc).

qPCR-HRM development

Primers used in the qPCR-HRM were Plasmo1-F (5′-GTT-AAG-GGA-GTG-AAG-ACG-ATC-AGA-3′) and Plasmo2-R (5′-AAC-CCA-AAG-ACT-TTG-ATT-TCT-CAT-AA-3′), targeting the 18S RNA gene which contains both highly conserved and variable regions in all Plasmodium species infecting humans, previously described by Rougemont et al.22,32 (Fig. 3).

Figure 3
figure 3

Alignement of 18S RNA amplified sequence from Plasmodium ovale curtisi and Plasmodium ovale wallikeri using MultiAlin. Primer position are displayed on 18S RNA gene sequence (Plasmo1-F and Plasmo2-R). The insertion of two Guanine in position 103 and 104 allows species differentiation using HRM detection.

Full size image

To set up a one-step PCR method for P. ovale spp species determination, differences in the PCR products for a given set of PCR primers were investigated to know if they could present enough differences to be noticeable with a HRM based PCR product detection. To this end, we first aligned the PCR products from the in silico PCR, using the multialin online tool (http://multalin.toulouse.inra.fr/multalin/)33. Then, we performed a melting temperature simulation, using uMELT online tool (available at https://www.dna.utah.edu/umelt/umb.php), to assess if the differences observed in the PCR products would be sufficient or not to distinguish Poc and Pow using HRM detection34 and to distinguish P. ovale spp from other Plasmodium species.

PCR reaction

The PCR mixture included 0.3 µM of each primer, 12.5 µL of MeltDoctor® HRM master mix (Life Technologies, Carlsbad, USA), 5 µL of DNA extract and water for a final volume of 20 µL. The thermal profile used for qPCR was as follows: 10 min initial denaturation at 95 °C; 40 cycles of 10 s at 95 °C denaturation and 1 min at 60 °C; 10 s at 95 °C followed by 1 min at 60 °C before HRM phase. HRM phase consisted on a 0.025 °C/s temperature increase. Melt curve plot was generated and analysed with Viia 7 software (Applied Biosystems) to determine Tm.

Mixed artificial samplesh

Artificial samples were performed with both Poc and Pow to determine the ability of qPCR-HRM to differentiate Poc and Pow in the same isolate. We prepared 9 artificial samples with various ratio of Poc and Pow (9/1; 8/2; 7/3; 6/4; 5/5; for both). Initial parasite density of Poc and Pow was 3000 parasites/µL and 6000 parasites/µl. Similar artificial samples (Poc and Pow with other Plasmodium species in the same isolate) were prepared for Pf, Pv and Pm.

qPCR-HRM performance

Repeatability and reproducibility

Repeatability was determined by using a unique sample thirty times in the same run for three different parasite density for both Poc and Pow at 3200 parasites/µL, 640 parasites/µL and 64 parasites/µL. We calculated standard deviation of Ct for repeatability as recommended in the MIQE guidelines19. We evaluated reproducibility as concordance of Tm between different run for the same isolate at three different parasite density for Poc and Pow (6400 parasites/µL, 640 parasites/µL and 64 parasites/µL).

Sensitivity and specificity

A serial sample dilution was performed to obtain eleven points ranging from 4000 to 1 parasites/µL from one Poc and one Pow positive sample, to determine the limit of detection (LOD). We calculated sensitivity of the method following this formula:

$${\rm{Sensitivity}}=(1-\frac{False\,Negative}{Total\,samples})\ast 100.$$
(2)

Ct amplification and melting curve were used to define LOD and samples were considered as negative if no specific Tm was observed.

Specificity was evaluated in two different way. First, we tested Toxoplasma gondii, Leishmania infantum, Loa loa, Onchocerca volvulus, Wuchereria bancrofti, and Babesia divergens patient positive blood samples from Bichat’s hospital to prove the existence of Poc and Pow specific Tm. Second, we evaluated the ability of the qPCR-HRM of distinguish P. ovale spp from other Plasmodium species in clinical samples.

Ethic approval

Participants’ consent was not required since samples were already collected for other medical purposes, and according to French legislations on non-interventional research.