Multicopy targets for Plasmodium vivax and Plasmodium falciparum detection by colorimetric LAMP

First, we explored conserved regions of PvCOX1, Pvr47, Pfr364, and PfEMP1 by aligning sequences available from PlasmoDB (plasmoDB.org) and the NIH genetic sequence database GenBank (https://​www.​ncbi.​nlm.​nih.​gov/​genbank/​). Pvr47 and Pfr364 sequences were obtained from Demas et al. [37], composed of 14 Pvr47 copies (P. vivax Sal-I strain) and 41 copies of Pfr364 (subfamilies 1 and 2 of P. falciparum 3D7 strain). For the COX1 and EMP1 genes, we used the complete sequences reported in GenBank.

The LAMP primers were designed using the free software Primer Explorer V5 (http://​primerexplorer.​jp/​lampv5e/​index.​html) and were located in a conserved region of each target. Only the primers for the COX1 target were designed by checking manually the main point of primer design described in EIKEN GENOME SITE (http://​loopamp.​eiken.​co.​jp/​e/​lamp/​primer.​html). Primer specificity was tested using the NCBI Basic Local Alignment Tool (BLAST) tool https://​www.​ncbi.​nlm.​nih.​gov/​tools/​primer-blast/​) and potential secondary structures were analysed with OlygoAnalizer 3.1 (https://​www.​idtdna.​com/​calc/​analyzer). Finally, four pairs of primers for each target were selected for LAMP assays (Table 1).

The colorimetric LAMP assays were optimized to provide a clear visual detection based on the drop in pH caused by DNA target synthesis. When using the neutral red dye, the pH drop produces a change in the colour solution from yellow to fuchsia. For this, the reaction mixtures were prepared at a final volume of 25 μL, containing 0.5 × Bst DNA polymerase reaction buffer (10 mM Tris–HCl, 5 mM (NH₄)₂SO₄, 25 mM KCl, 1 mM MgSO₄, 0.05% Tween 20, pH 8.8 New England Biolabs Inc., MA, USA), 0.2 μM each external primer F3 and B3, 2 μM each inner primer FIP and BIP, 12U Bstv2.0 DNA polymerase (New England Biolabs Inc., MA, USA), 1.4 mM of each dNTPs, 0.003% of neutral red dye and 5 μL of a test sample. The reaction mixture for the Pvr47-LAMP was adjusted to 6 mM MgSO₄, the PvCOX1-LAMP was adjusted to 6 mM MgSO₄ and 0.4 M betaine, the Pfr364-LAMP was adjusted to 8 mM MgSO₄ and the PfEMP1-LAMP was adjusted to 8 mM MgSO₄ and 0.4 M betaine.

A positive reaction was assessed by visual inspection and considered as such only if the reaction turned from yellow to fuchsia after 65 min of incubation at 60 °C. Incubation steps were performed on a Bio-Rad T100 TM Thermal Cycler. (Bio-Rad Laboratories, CA, USA).

To analyse the LOD of each colorimetric LAMPs in target copies, plasmids for each target were assembled through PCR amplification using the external primer F3/B3. The amplicons were cloned using a TOPO TA cloning kit (Life Technologies). Recombinant plasmid DNA was purified using a Qiagen plasmid mini kit, and the concentration of the extracted plasmid was determined using a Qubit dsDNA BR Assay Kit (ThermoFisher Scientific).

The LOD was assessed in two formats: (1) LOD in terms of number of copies per µL, for which serial dilutions of each recombinant plasmid with the specific target segment were used; and (2) LOD in terms of number of parasites per µL, for which serial dilutions of a cultured sample of P. falciparum 3D7 and pooled samples of whole blood from patients infected with P. vivax (quantified previously by microscopy and qPCR) were used. Each dilution series was tested 15 times by each colorimetric LAMP. The LOD was calculated using Probit regression whit the MedCalc Statistical Software version 19.2.6 (MedCalc Software bv, Ostend, Belgium; https://​www.​medcalc.​org; 2020).

The reference samples from the P. falciparum 3D7 culture and a pool of blood samples infected with P. vivax (microscopy and PCR-confirmed) were supplied by the malaria laboratory at the Universidad Peruana Cayetano Heredia (UPCH). DNA controls were supplied by the Centers for Disease Control and Prevention (CDC). The quality control distribution 4260 of WHO and UK National External Quality Assessment Service (UK NEQAS) Parasitology were used for validation of genus and species. The 4260 material contained five samples: specimen 4307 (Plasmodium ovale 200,000 parasites/mL), specimen 4308 (Plasmodium vivax 20,000 parasites/mL), specimen 4309 (P. falciparum 20,000 parasites/mL), specimen 4310 (No Plasmodium nucleic acid detected), and specimen 4311 (Plasmodium knowlesi 200,000 parasites/mL).

A total of 312 blood samples were selected from the sample bank of the Malaria Laboratory at UPCH. The samples were collected from 10 communities (Quistococha, Santo Tomas, Rumococha, Gamitanacocha, Libertad, 1 de Enero, Salvador, Lago Yuracyacu, Puerto Alegre, and Urcomiraño) from Loreto (Peru) during an active sample collection in 2018. All samples were collected after an informed consent form was signed according to the ethical clearance from the Ethics Review Board of the UPCH, Lima, Peru (SIDISI code 101645).

These samples were categorized into three groups. The first group included 28 samples with infections detected by both microscopy and real-time PCR with 21 P. vivax infections and 7 P. falciparum infections. All samples showed asexual parasitaemia with more than 100 parasites/µL as assessed by microscopy, with the exception of three P. vivax infections that had < 100 parasites/µL. The second group included 101 submicroscopic infections (83 P. vivax and 18 P. falciparum infections) detected only by real-time PCR as described by Mangold et al. [40]. The third group included 183 negative samples as determined by both microscopy and real-time PCR; this group of samples was included to assess interference of human DNA with the LAMP primers. Figure 1 shows the flowchart of the experimental procedure and details of the diagnosis of clinical samples are available in Additional file 1.

All the selected samples have a result of the thick drop and thin film for malaria diagnosis. Parasite densities were quantified by counting asexual parasites and gametocytes separately with the number of thick drop leukocytes based on an estimated mean count of 8000 leukocytes per microliter of blood. The counting criteria were applied according to the Peruvian Minister of Health [5] and were carried out with the following formula:

The DNA was extracted using the EZNA® Blood DNA (Omega Bio-tek, Norcross, GA, USA), following the manufacture’s protocol. In brief, 40 µL blood was transferred to a sterile microcentrifuge tube adding 250 µL 2X TEN buffer pH 8.0 (0.2 M NaCl, 0.02 M Tris–HCl pH 8.0, 0.002 M EDTA pH 8.0), 180 µL distilled water, 50 µL of 10% SDS; and 25 µL OB protease solution. The mix was incubated at 65 °C for 60 min and centrifuged at 13,000g for 5 min; the supernatant was transfer to a new tube, adding 250 µL of BL buffer and 260 µL of 100% ethanol. Then, 500 µL of this homogenate was transferred to a mini HiBind® DNA column inserted into a 2 mL collecting tube and centrifuged at 10,000g for 1 min. The filtrate was discarded, and the mini column was inserted into a new 2 mL collection tube, 500 µL of HBC buffer was added, then centrifuged at 10,000g for 1 min and the filtrate was discarded. Then, 700 µL of wash buffer was added to the column and centrifuged at 10,000g for 1 min; this process was repeated once more. The empty HiBind mini DNA column was centrifuged at maximum speed for 2 min to dry the column matrix. The DNA from the column was eluted in 60 µL elution buffer preheated to 65 °C. Eluted DNA was stored at − 20 °C. Hereafter, the DNA extracted using this method is referred to as MSC DNA.

The 18S rRNA gene was chosen for species-specific detection using the qPCR performed by Mangold et al. [40]. The reaction mixtures were prepared at a final volume of 25 µL with a final concentration of 1X master mix of PerfeCTa SYBR® Green Fastmix, and 0.3 μM each primer PL1473F18 5′-TAACGAACGAGATCTTAA-3′ and PL1679R18 5′-GTTCCTCTAAGAAGCTTT-3′ [40].

The PCR conditions consisted of an initial denaturation step at 95 °C for 2 min, followed by 45 cycles of 20 s at 95 °C, 20 s at 50 °C, and 20 s at 68 °C, with fluorescence acquisition at the end of each extension step. Amplification was immediately followed by a melting curve program consisting of 2 min at 68 °C, and a stepwise temperature increase of 0.5 °C/s until 90 °C, with fluorescence acquisition at each temperature transition. All of the samples with a quantification cycle (Cq) of less than 36.2 were considered positive according to the LOD determined using a serial dilution of the 1st WHO International Standard for Plasmodium falciparum DNA Nucleic Acid Amplification Techniques (NIBSC code 04/176) available at the malaria laboratory, and the same LOD for P. vivax was assumed, considering that it is the same amplification target. The melting curve analysis was used to determine the species-specific: a melting peak of 73.5 °C ± 0.5 was considered to define P. falciparum infection and a peak of 77 °C ± 0.5 was considered to determine P. vivax infection.

Correlation between the diagnotics test were evaluated using the Cohen kappa test. Analyses of sensitivity and specificity were performed using the free software package reportROC in RStudio.

Using a serial dilution of plasmid for each target, Pfr364-LAMP showed an LOD of 27.3 copies of plasmid/µL and PfEMP1-LAMP showed an LOD of 28.9 copies of plasmid/µL. For P. vivax detection, PvCOX1-LAMP showed an LOD of 23.9 copies of plasmid/µL and the Pvr47-LAMP showed an LOD of 15.6 copies of plasmid/µL (Table 2). In addition, sensitivity analyses were performed using serial dilutions from a 3D7 cultured blood sample for P. falciparum to define LOD based on parasites/µL. The results showed an LOD of 3.7 parasites/µL for Pfr364-LAMP and of 3.3 parasites/µL for PfEMP1-LAMP. For P. vivax, using a pooled samples of blood from infected patients (quantified as parasite/µL by PCR), the PvCOX1-LAMP showed an LOD of 2.4 parasites/µL and the Pvr47-LAMP showed an LOD of 3 parasites/µL (Table 2).

Overall, colorimetric LAMPs with PfEMP1 and Pfr364 targets showed high specificity for detecting P. falciparum 3D7 and the corresponding reference samples. These results are shown in Fig. 2. The colorimetric LAMPs designed for P. vivax with the COX1 and Pvr47 targets showed a positive reaction with P. vivax strain Sal-I, but it also showed a positive reaction with P. knowlesi, likely due to the high genomic similarity.

The PvCOX1-LAMP displayed the best sensitivity for Plasmodium vivax detection. Using MSC DNA, the PvCOX1-LAMP detected 100% (21/21) of microscopic samples and 86.7% (72/83) of submicroscopic samples, whereas the Pvr47-LAMP detected only 71.4% (15/21) and 36.1% (30/83) of these samples, respectively (Table 3). Likewise, using BS DNA, the PvCOX1-LAMP detected 81% (17/21) of microscopic samples and 36.1% (30/83) of submicroscopic samples. The Pvr47-LAMP only detected 52.4% (11/21) and 22.9% (19/83) of these samples, respectively. In the microscopic infection group, only four BS samples produce negative results in PvCOX1-LAMP; two of these samples had parasite densities of 324 and 514 parasites/µL, while the other two showed the lowest parasite density detected by microscopy with 32 parasites/µL.

Overall, 11 of 83 submicroscopic P. vivax infections (13.3%) were negative in PvCOX1-LAMP using sample from both preparation methods (see Additional file 1). Of these samples, one quantified by PCR with 6 parasites/µL was above the 95% confidence interval of the LOD (1.8–4.8 parasites/µL), another six samples had parasitaemia values within the interval and four had parasitaemia below this range. Finally, 30 submicroscopic infections were also detected by PvCOX1-LAMP using both types of extracted DNA, whereas 42 were detected only using MSC DNA (Fig. 3a).

Five additional positive samples to P. vivax by PvCOX1-LAMP were found in the third group of negative samples, by using MSC DNA (Fig. 3a). Upon further examination, these samples were considered to have had lower parasitaemia levels to P. vivax, with the Taqman PCR, described by Rougemont et al. [41]. In addition, in Pv47-LAMP, nine samples were positive for P. vivax using MSC DNA, five of them were also positive using BS DNA, and one additional sample was positive using BS DNA (Fig. 3b). However, negative results were always obtained for these 10 samples using different real-time PCR methods.

The sensitivity and specificity of the Pfr364 and the PfEMP1 targets were similar using all three groups of samples and the PCR as a reference method. Both LAMPs showed a moderate agreement, with Cohen kappa indexes between 0.74–0.82 when using MSC DNA and 0.64–0.57 when using BS DNA (Table 4).

Both, Pfr364-LAMP and the PfEMP1-LAMP detected all infections positive that were identified by microscopy (n = 7) using DNA from both sample preparation methods. In the submicroscopic group, two of nine P. falciparum infections with parasitemia above the 95% confidence interval of Pfr364-LAMP LOD (2.8–7.5 parasites/µL) showed a negative result when tested by Pfr364-LAMP using MSC DNA. One submicroscopic infection with parasitemia above the 95% confidence interval of PfEMP1-LAMP LOD (2.4–6.7 parasites/µL) showed a negative result when tested by PfEMP-LAMP using MSC DNA. However, both assays exhibited difficulty detecting submicroscopic infections using BS DNA. Moreover, the Pfr364-LAMP detected 6 out of 18 PCR positive samples, whereas PfEMP1-LAMP detected only four samples (Fig. 4). One P. vivax infection detected by PCR with low parasitaemia of 40 parasites/µL was additionally detected as P. falciparum infection by the two LAMPs with both types of extracted DNA. This sample was also positive for P. vivax in PvCOX1-LAMP using MSC DNA, and later it was confirmed by Taqman PCR (as described by Rougemont et al. [41]) to be a mixed sample.

In PfEMP1-LAMP using MSC DNA, two additional P. falciparum were detected. One of these was a P. vivax submicroscopic infection, and the other one was a negative sample.

Considering the P. vivax and P. falciparum diagnosis scenario with two individual LAMP assays and the PCR as a reference method, PvCOX1-LAMP and Pfr364-LAMP together showed a sensitivity of 84.5% compared to the 87.6% sensitivity of both PvCOX1-LAMP and PfEMP1-LAMP together using MSC DNA; the values using BS DNA were 47.3% and 45.7%, respectively (Table 5).

PvCOX1-LAMP and Pfr364-LAMP combined system displayed slightly more specificity than PvCOX1-LAMP and PfEMP1-LAMP combination using MSC DNA, at 97.3% and 96.7% respectively (Table 5). This slight difference indicates that PfEMP1-LAMP may decrease the specificity of the system when the two individual LAMP assays are used for both P. vivax and P. falciparum detection.