Streamlined, PCR-based testing for pfhrp2- and pfhrp3-negative Plasmodium falciparum

Background Rapid diagnostic tests (RDTs) that detect histidine-rich protein 2 (PfHRP2) are used throughout Africa for the diagnosis of Plasmodium falciparum malaria. However, recent reports indicate that parasites lacking the pfhrp2 and/or histidine-rich protein 3 (pfhrp3) genes, which produce antigens detected by these RDTs, are common in select regions of South America, Asia, and Africa. Proving the absence of a gene is challenging, and multiple PCR assays targeting these genes have been described. A detailed characterization and comparison of published assays is needed to facilitate robust and streamlined testing approaches. Results Among six pfhrp2 and pfhrp3 PCR assays tested, the lower limit of detection ranged from 0.01 pg/µL to 0.1 ng/µL of P. falciparum 3D7 strain DNA, or approximately 0.4–4000 parasite genomes/µL. By lowering the elongation temperature to 60 °C, a tenfold improvement in the limit of detection and/or darker bands for all exon 1 targets and for the first-round reaction of a single exon 2 target was achieved. Additionally, assays targeting exon 1 of either gene yielded spurious amplification of the paralogous gene. Using these data, an optimized testing algorithm for the detection of pfhrp2- and pfhrp3-negative P. falciparum is proposed. Conclusions Surveillance of pfhrp2- and pfhrp3-negative P. falciparum requires careful laboratory workflows. PCR-based testing methods coupled with microscopy and/or antigen testing serve as useful tools to support policy development. Standardized approaches to the detection of pfhrp2- and pfhrp3-negative P. falciparum should inform efforts to define the impact of these parasites. Electronic supplementary material The online version of this article (10.1186/s12936-018-2287-4) contains supplementary material, which is available to authorized users.

The methods required to identify and confirm pfhrp2/3 gene deletions are challenging, due to the difficulty of proving the absence of a gene. While PCR assays that target pfhrp2/3 are expected to yield negative results when applied to parasites lacking the gene(s), PCR failure can occur for other reasons. Testing of parasites with intact pfhrp2/3 genes may yield false-negative results due to DNA concentrations below the assay's limit of detection, poor quality DNA, variable reagent performance, or other factors. Cheng et al. published useful guidelines to standardize the reporting of pfhrp2/3 gene deletions [21]. However, the specific methods employed for the detection and confirmation of deletions continue to vary between laboratories, and recent evidence suggests that atypical elongation temperatures may improve amplification of AT-rich regions of both genes [14,22]. This manuscript seeks to address these issues by comparing the performance of published PCR assays for pfhrp2 and pfhrp3, exploring the impact of reduced elongation temperatures on assay sensitivity, assessing assay specificity, and describing a streamlined testing algorithm.

Methods
The performance characteristics of six published PCR assays, including four designed to amplify pfhrp2 and two designed for pfhrp3, were compared [3,5,9,23]. After determining the optimal annealing temperatures for each assay, we assessed their lower limits of detection (LOD) using DNA extracted from cultured P. falciparum 3D7 strain parasites. DNA was quantified using the Qubit 2.0 instrument with dsDNA high sensitivity reagents (Ther-moFisher Scientific, Waltham, MA) and serially diluted in nuclease-free water to achieve concentrations ranging from 10 −1 to 10 −7 ng/µL (seven tenfold dilutions). For each dilution, the assay was performed in triplicate, using different elongation temperatures of 60, 65, and 72 °C. PCR assays were performed on Mastercycler thermocyclers (model AG 22331; Eppendorf, Hamburg, Germany) using 25 µL reaction volumes containing 12.5 µL HotStarTaq Master Mix (Qiagen, Venlo Netherlands), 200-400 nM primers synthesized by Eurofins Genomics (Louisville, KY) with salt-free purification, nuclease-free water, and 3 µL of DNA template (Table 1). For nested reactions, first-round product was diluted 100-fold in nuclease-free water prior to second-round amplification. PCR products were visualized using electrophoresis with 1% agarose gels in TBE buffer (Tris/Borate/EDTA). Finally, LOD testing was repeated using optimized reaction conditions and serial dilutions of 3D7 DNA from a separate stock.
The specificity of the best performing assays, including those with targets spanning exon 1 and 2 (exon 1/2) and exon 2 alone, was then evaluated. Assays were performed using control DNA from P. falciparum Dd2 (MRA-150G) and HB3 (MRA-155G) strain parasites, which lack pfhrp2 and pfhrp3, respectively. Control DNA was obtained from the Malaria Research and Reference Reagent Resource Center ([MR4], BEI Resources, Manassas, Virginia) and diluted to a concentration of 0.1 ng/ µL after initial quantification using Qubit as above. For assays that yielded an unexpected result using optimized reaction conditions (i.e. bands from a pfhrp2 assay performed using pfhrp2-deleted Dd2 DNA or bands from a pfhrp3 assays performed using pfhrp3-deleted HB3 DNA), amplicons were sequenced using Sanger sequencing at Eton Bioscience (Research Triangle Park, NC), and assays were repeated at the other elongation temperatures (60, 65, and/or 72 °C). For PCR products with multiple bands appreciated by gel electrophoresis, individual bands were excised, and DNA was extracted using the QIAquick Gel Extraction Kit (Qiagen, Hilden, Germany) before sequencing. Gel extraction was performed according to the manufacturer's instructions, with the exception of the final DNA elution step, in which we performed two separate elutions using 30 µL aliquots of Buffer EB through the column, followed by a final elution step using the combined 60 µL initial eluate to maximize DNA yield. Raw sequence reads were processed using Sequencher 5.4 (Gene Codes Corporation, Ann Arbor, MI), trimming bidirectional sequences based on confidence values and visual inspection of the chromatograms. We used EMBOSS Water for pairwise nucleotide alignments to the 3D7 (v3.0) pfhrp2 and pfhrp3 reference sequences. Sequence identification was based on sequence homology and alignment score, using default settings (DNAfull matrix, gap open penalty 10, gap extend penalty 0.5) [24].

Assay performance
Reduced elongation temperatures improved the sensitivity of five of the six assays (Additional file 1: Figures S1,  S2). An elongation temperature of 60 °C reduced the LOD by tenfold and/or produced darker bands for all exon 1/2 targets and for the first-round reaction of a single exon 2 target (assay 4). Using optimized elongation temperatures and under our lab conditions, the LOD of published PCR assays for pfhrp2 and pfhrp3 varied, with lower limits ranging from approximately 0.01 pg/ µL to 0.1 ng/µL of 3D7 DNA, or approximately 0.4-4000 parasite genomes/µL (Table 1). Additionally, while adding a second round of amplification in the best performing nested PCR for pfhrp2 exon 1/2 (assay 1) resulted in darker bands, the assay's LOD was unchanged.

Amplification of paralogous genes by exon 1/2 assays
Unexpectedly, paralogous amplification by assays targeting exon 1/2 but not exon 2 of both genes was observed. Bands were visualized from assays targeting exon 1/2 of pfhrp2 (assay 1) using Dd2 (pfhrp2-deleted) control DNA (Additional file 1: Figure S3) and those targeting exon 1/2 of pfhrp3 (assay 5) using HB3 (pfhrp3-deleted) control DNA (Additional file 1: Figure S4), respectively. Assay 1 produced unexpected bands at all three extension  [32,33] temperatures tested, while assay 5 produced bands when tested using 60 and 65 °C extension temperatures but not at 72 °C. All tested exon 2 assays for both genes (assay 3, 4, and 6) produced negative results using Dd2 or HB3 DNA, as expected. Pfhrp2 exon 1/2 assay 2 was not tested due to its poor performance during initial LOD testing, presumably a result of a single base insertion near the 3′ end of the reverse primer compared to the 3D7 reference sequence. The sequence homology of pfhrp2 and pfhrp3 exon 1/2 primer binding sites (Fig. 1) suggested that amplification of paralogous genes had occurredi.e., the amplicons generated by the pfhrp2 exon 1/2 assay using Dd2 strain control DNA represented pfhrp3 amplification and that those generated by the pfhrp3 exon 1/2 assay using HB3 strain control DNA represented pfhrp2 amplification.
Sequencing results confirmed amplification of pfhrp3 by the pfhrp2 exon 1/2 assay, and vice versa. With Dd2 strain (pfhrp2-deleted) template, the pfhrp2 exon 1/2 assay (assay 1) unexpectedly produced a single band with a fragment length of approximately 300 bp. The amplicon's sequence aligned to the pfhrp3 gene with 92% sequence homology (Additional file 1: Figure S5). With HB3 strain (pfhrp3-deleted) template, the pfhrp3 exon 1/2 assay (assay 5) unexpectedly produced two clear bands with fragment lengths of approximately 300 and 800 bp and a faint band at approximately 400 bp. Sequences generated using DNA extracted from each band aligned to the pfhrp2 gene, with 98, 99, and 97% sequence homology for the 300, 400, and 800 bp fragments, respectively (Additional file 1: Figure S6). However, when applied to 3D7 strain (pfhrp2/3-positive) control template, both exon 1/2 assays produced the expected result: a single band with a sequence that aligned to pfhrp2 with 96% homology or pfhrp3 with 99% sequence homology (for assays 1 and 5, respectively). Fig. 1 Pfhrp2 and pfhrp3 gene sequence homology. Alignment of reference sequences from the consensus 3D7 (v3.0) genome, 5′→3′, with expected binding sites for pfhrp2 assays (white boxes) and pfhrp3 assays (gray boxes). The reverse primer sequence for Assay 2 includes an a single-base insertion (cytosine) at the location indicated by an asterisk (*). Identical bases are indicated by a period (.), missing bases by a dash (-), substitutions by the discordant base, and non-coding regions by lower case font

Streamlined testing algorithm
These findings were used to develop a streamlined, PCRbased testing pipeline for pfhrp2/3-negative P. falciparum (Fig. 2) that incorporates optimized elongation temperatures, LOD testing results, and assay specificity.

Discussion
By lowering the LOD and employing assays that distinguish pfhrp2 from pfhrp3, this testing algorithm provides an improved approach to PCR-based detection of pfhrp2/3-negative P. falciparum. Importantly, PCR-based approaches for identification of pfhrp2-and pfhrp3negative parasites must be coupled with verification of P. falciparum parasitaemia and confirmation that parasite DNA is present at concentrations above the LODs of the pfhrp2 and pfhrp3 assays. These goals were achieved by employing assays targeting two P. falciparum-specific, single-copy genes, lactate dehydrogenase (pfldh) and P. falciparum beta tubulin (PfBtubulin), as the initial and final steps of the testing pipeline.
Lowering the elongation temperature improved the LOD of all published assays with exon 1/2 targets on either gene. This finding likely represents improved amplicon extension across the AT-rich intron between the exons as suggested by previous reports [22]. Unexpected amplification of paralogous gene targets by the pfhrp2 and pfhrp3 exon 1/2 assays was observed, presumably due to sequence homology at the primer binding sites. In regions where co-existing pfhrp2 and pfhrp3 deletions are common, the impact of non-specific amplification is expected to be reduced [25][26][27]. Additionally, the absence of paralogous amplification of 3D7 control DNA suggests that the availability of abundant, completely homologous primer binding sites early in PCR cycling reduces the likelihood of exponential amplification after mispriming. To reduce the risk of unintentional amplification of paralogous Fig. 2 Proposed testing pipeline. Single-step assays are favoured to reduce the risk of contamination, and assays should be performed in duplicate. In addition to positive P. falciparum (e.g. 3D7 strain) DNA and no template controls, either *pfhrp2-negative (e.g. Dd2 strain) or **pfhrp3-negative (e.g. HB3 strain) P. falciparum DNA controls should be used genes, this testing algorithm uses assays targeting exon 2 of both genes. This approach also permits analysis of the repetitive sequences that encode epitopes recognized by anti-PfHRP2 antibodies [28].
A broad range of LOD results was observed for published pfhrp2 and pfhrp3 assays, spanning over 4 orders of magnitude under the laboratory conditions employed during this study. These differences were addressed in the resulting testing pipeline (Fig. 2) by defining an initial threshold DNA concentration tenfold higher than the LOD of the downstream pfhrp2 and pfhrp3 assays. In addition, a stringent, final singlecopy-gene PCR that meets the same LOD requirement was included, providing confirmation that sample degradation has not occurred during the testing process. The typical workflow employed in this laboratory includes assays 3 and 6 for pfhrp2 and pfhrp3 testing, respectively, performed in duplicate. For discordant results (i.e. one of two replicates positive), samples are called positive if there is a clear band of appropriate fragment length. If not, the assay is repeated, and the final call is based on the third result. Because the first-round of the nested assays achieved LODs below the initial and final confirmatory, falciparum-specific assays, their use as single-step assays is favoured in this laboratory to reduce the risk of contamination and improve work flow.
In settings where real-time PCR is not feasible, the proposed initial lactate dehydrogenase (pfldh) quantitative PCR assay and the final confirmatory P. falciparum beta tubulin (PfBtubulin) assays could be replaced with traditional PCR assays with LODs above the LOD of the pfhrp2 and pfhrp3 assays. Because assay performance can vary from laboratory to laboratory and with different reagents or equipment, it is essential to confirm the LOD of each assay using the reagents and laboratory infrastructure at hand.
In addition to PCR-based testing, current guidelines recommend independent confirmation of P. falciparum parasitaemia using microscopy or a non-PfHRP2based RDT, such as an RDT that detects P. falciparum lactate dehydrogenase (pf-pLDH), before making deletion calls [19,21]. Quantification of circulating PfHRP2 antigen is also a valuable tool that can be particularly useful for assessing PfHRP2-RDT-negative but pfhrp2/3-PCR-positive isolates with impaired protein expression [29]. Additionally, novel assays under development such as those targeting regionally specific deletion breakpoints or employing droplet digital PCR, have potential to improve throughput [30].

Conclusions
Surveillance of pfhrp2-and pfhrp3-negative P. falciparum requires careful laboratory workflows. PCR-based testing methods, coupled with microscopy and/or antigen testing, serve as useful tools to support policy development. Standardized approaches to the detection of pfhrp2-and pfhrp3-negative P. falciparum should inform efforts to define the impact of these parasites [20,21].