Development and validation of a duplex real-time PCR for the rapid detection and quantitation of HTLV-1

HTLV-1 is the first RNA retrovirus found to be mainly transmitted by mother-to-child, sexual contact and injecting drug users [1, 2]. It also can be transmitted by the transfusion of infected blood components, liver, kidney and lung transplantation [3]. The two most prevalent diseases derived from HTLV-1 are adult T-cell leukemia (ATL) and HTLV-1-associated myelopathy/tropical spastic paraparesis (HAM/TSP), but HTLV-2 is infrequently linked to a disease [4, 5]. At present, no vaccine is yet available and poor prognosis has been reported for ATL and HAM/TSP. In this sense, screening blood and organ donors has been shown to be a successful tactic for halting the spread of HTLV-1.

According to the report of a statewide multicenter large-scale survey on the seroprevalence of HTLV infection among Chinese blood donors, the majority of mainland China is classified as an HTLV non-endemic area [6]. The National Health Commission of the People's Republic of China has recommended that beginning in 2022, ELISA testing for the presence of the HTLV-1/2 antibody be done on 100% of blood donors in areas with a relatively high prevalence of HTLV (higher than 0.028‰) and on 10–30% of blood donors in areas with a moderate prevalence of HTLV (higher than 0.01‰). And then the blood centers would confirm any initially reactive samples. This program will lessen the burden, lower the cost of screening, and minimize the hazards of HTLV-1 transmission through blood transfusion. For some other countries, especially those with a high need for testing or a shortage of laboratory supplies, pooling of samples for seroepidemiological surveillance of HTLV-1/2 is another strategy to reduce the cost and technician time. However, there will inevitably occur numbers of false positives in the initial screening for the non-endemic regions [7]. Western blot (WB) and line immunoassay (LIA) with high specificity are two of the most frequently used confirmatory assays for the confirmation and differentiation of HTLV-1 and HTLV-2 infections. Whereas, the serologic confirmatory tests are too expensive to be employed in every blood center. In addition, both of them were able to produce a considerable high proportion of inconclusive (indeterminate and untypeable) results, especially when samples obtained from patients with HTLV-2 infection and those from patients coinfected with HIV, and/or HBV, and/or HCV [8]. Molecular assays, such as real-time quantitative PCR (qPCR), with high sensitivity and low cost should be performed to detect the nucleic acids of HTLV-1/2 provirus for confirmation. Since HTLV-1 infects only through cell-to-cell transmission, most previous studies used peripheral blood mononuclear cells (PBMC) for HTLV-1 quantification, and few reports used whole blood [9‐12]. Unfortunately, there are presently no commercial kits or sensitive method available to determinate the proviral load (PVL) and assess the real status of HTLV-1 infections in China. More importantly, it has been revealed that more than 4 copies/100 PBMCs is an independent risk factor for progression of ATL, and other investigations have confirmed the significance of quantifying PVL for the assessment of the emergence of HTLV-1-associated diseases [13, 14]. To address the shortcomings of the present confirmative test and offer accurate PVL for better monitoring the affected patients, it is important to create a nucleic acid test with high sensitivity, specificity, and reproducibility.

PVL quantification was initially performed using semiquantitative PCR and southern blots, however neither method was able to determine the precise copy number. In recent years, qPCR has been performed for HTLV-1 or/and HTLV-2 viral quantification. However, the published qPCR techniques had a few shortcomings. (1) the preparation of quantitative standards, which typically involved sequential reference DNA diluents derived from HTLV-1-infected cell lines, was a labor-intensive technique that made it challenging to replicate the findings [9, 10]. (2) the standard curves were typically created using a combination of two distinct plasmids. However, it was challenging to precisely dilute two plasmids, which led to inconsistent findings between laboratories [11]. (3) Prior validation for the majority of the established technologies was insufficient, which was required before the technique could be used in ordinary laboratory settings [11]. (4) the performance of the established methods has rarely been assessed in blood donors, especially in non-endemic areas. (5) the majority of the primer–probe set used for internal reference in earlier approaches came from commercial kits, or the reference gene needed to count cells had to be found independently, which significantly reduced the cost-effectiveness of HTLV-1 PVL quantification [9, 11].

Due to this, we decided to develop a single-tube qPCR assay to quantify HTLV-1 PVLs in PBMCs, which was normalized by means of the quantitation of a cellular gene RNase P (RPPH1). Multiple pairs of primers and probes were designed and screened based on the HTLV-1 conserved region (pol region and tax region) and the internal reference RPPH1 gene. A plasmid carrying both the HTLV-1 of interest and a fragment of the RPPH1 gene was used to establish the standard curves. The sensitivity, specificity, reproducibility and accuracy of this assay were analyzed and preliminary results on its use for specimens with LIA-indeterminate and positive results from blood donors were reported.

The qPCR is one of the most important molecular confirmatory tools for HTLV-1/2 diagnosis. Serological confirmatory assays for HTLV-1/2 (WB and LIA) are expensive and tend to provide indeterminate and untyped results. In China, no molecular assays for HTLV-1/2 diagnosis are available. The purpose of the present study was to develop and validate a duplex qPCR HTLV-1 assay to simultaneously detect the presence of HTLV-1 and the RPPH1 reference gene, so as to solve the problems related to time constraints, cost, and data absent about PVLs in HTLV-1 infected patients in China.

The LTR, pol, gag, tax and rex genes are the regions used most frequently for HTLV-1 DNA quantitation. But deletions or mutations in the rex region have been reported in HAM/TSP cases and defection of 5’LTR frequently occurred in ATL cases [15‐17]. Therefore, we screened specific primers and probes from a range of primer–probe sets targeting the pol and tax gene, which are relatively conservative in HTLV-1. In our study, a set of primer–probe with the best specificity for pol gene was finally used for our later work. It should be emphasized that we also screened specific primers and probes for internal reference RPPH1 gene, rather than using commercial kits described in previous studies [12, 18], which helped reduce detection costs and avoided having to independently detect the reference gene to count the number of cells. After optimizing the concentrations and annealing temperature of the selected primers and probes, we validated that neither of the two sets of primer-probes affects each other's fluorescence signal in a single tube.

Previous research identified HTLV-1 PVLs using successive dilutions of reference extracted DNA from cell lines with known HTLV-1 virus copy numbers [9]. This protocol was time-consuming, inconvenient and required certain cell-culturing facilities, making it challenging to use in a clinical setting. Synthetic oligonucleotides were utilized by Bandeira et al. to create a standard curve for measuring HTLV-1 PVL, and they underlined the fact that obtaining and maintaining such reference material was less expensive than collecting DNA from lineages of HTLV-1-infected cells [19]. However, plasmids are easier to build the standard curves for reproducible results than synthetic oligonucleotides and are more stable over time. In addition, two identical plasmids separately contain the HTLV-1 and internal control gene were often utilized to measure the PVLs. Such protocol necessitates accurate dilution of the standard curve of each reference plasmid, otherwise the results will be difficult to interpret due to the minimal discrepancies between the dilution format. In our study, a pcDNA3.1(-)-H1-pol-tax-RPPH1 that contained one copy of the RPPH1 gene and HTLV-1 target region was constructed and used to establish the standard curves, which was simple to prepare, highly reproducible and stable over time.

In fact, the duplex qPCR established in our study demonstrated a broad linear dynamic range, presented acceptable efficiency and was capable of detecting low-level proviral load (1 copy/reaction). To evaluate the detection limit of the method for the real sample, Huh102 cells were serially diluted with Jurkat cells and then were detected by our duplex qPCR for 20 times. Our method was sensitive enough to detect 2.5 copies per 10,000 cells. Although earlier research using mqPCRs revealed greater sensitivity for detecting and quantifying HTLV-1 than the current study, those earlier studies detected the reference gene independently to estimate the number of cells. For instance, Moens et al. developed a triplex qPCR for simultaneous detections of 4 copies of HTLV-1 per 10⁵ cells and 2.5 copies of HTLV-2 per 10⁵ cells, using 5 specific target primer-probes sets and RPPH1 as reference gene [12]. However, the reference gene to quantify cells number was detected separately by using the commercial kit and five different dyes are restricted to equipment and reagents.Lee et al. developed one duplex qPCR assay using primers that targeting the tax region shared by HTLV-1 and HTLV-2, and simultaneously detect the HLA-DQ alpha to normalize the amount of cellular DNA in another single qPCR tube; all of them based on SYBR Green dye. Although they displayed excellent sensitivity (1 copy per reaction), the duplex qPCR could not able to distinguish HTLV-1 from HTLV-2 and two tubes were required for each sample [20]. Another duplex qPCR developed by Estes and Sevall also used a singleplex assay to detect human β-globin for PVL quantification [21].As for the specificity, our duplex assay disclosed 100% specificity when applied in samples of patients infected with TP or HBV or HCV or HIV or HTLV-2 infected cell lines. Regarding the reproducibility, both intra- and inter-assays showed that the method developed in this study was surprisingly reproducible for the high, medium, and low plasmid concentrations, with CV of Ct values for HTLV-1 lower than 1.88 and 3.06%, respectively, and for RPPH1 lower than 1.95 and 2.90%, respectively. The accuracy was evaluated using independent extractions of standard Hut102 dilutions with accurate HTLV-1 proviral loads (quantitated by ddPCR). The analysis of six independent extractions within the same qPCR run showed samll variations for samples with high PVL. It is crucial to note that although though the PVL of the diluted Huh102 was as low as 0.026 copies/100 cells, the results detected by our method (0.0275 ± 0.0092 copies/100 cells) were very close to the theoretical values, reflecting a high accuracy.

In order to verify the feasibility of this method, LIA-positive and indeterminate samples with blood cells available were performed with WB and our duplex qPCR simultaneously. As shown in the results, 90.0% of LIA-positive, 33% of LIA-untypeable and 2% of LIA-indeterminate samples were confirmed as qPCR-positive, indicating that our qPCR method could successfully detect HTLV-1 in the majority of seropositive samples and was able to distinguish partial untypeable and indeterminate samples. The median PVL value of the qPCR-positive samples in our study was 1.621 copies/100 cells, which was either higher or lower or similar to previous reports [9, 22, 23]. In our study, HTLV-1-infected individuals with a broad range of PVLs, the values ranged from 0.001 to 11.397 copies per 100 PBMCs (Additional file 1: Table S1). This distribution was similar to the that observed in asymptomatic carriers (range: 0.018–12.434 copies/100 100 PBMCs) as well as in blood donors (about range: 0.001–10 copies/100 100 PBMCs approximated by their figures) [18, 22]. According to another study by Allison et al. [9], the PVL ranged from 0.0148 to 2.760 copies per 100 PBMCs in asymptomatic carriers. The conflicting results may be caused by the various geographic backgrounds and incomparable methodologies used in the various investigations. The median proviral load of LIA-positive samples measured by our method was approximately 100-fold higher than that of LIA-indeterminate and untypeable samples (1.780 copies/100 cells versus 0.186 copies/100 cells), which is in comparation to previous authors’ findings[18]. However, the number of samples with qPCR-positive but LIA-untypeable or indeterminate results in our study was too small, therefore, more samples need to be analyzed in the future to confirm this finding. It is interesting to note that one of the WB-indeterminate samples had a PVL of 11.4 copies per 100 cells, which was significantly higher than previously reported. However, Andrea et al. noted that one blood donor from Argentina had WB-indeterminate results while having a rather high PVL value (4.5copies/100 cells) [24]. This possibly due to the mutation in the HTLV-1 provirus, which reduces the expression of antigen and antibody. Such mutated sites resulting an antigenic alteration similar to what has been reported in cases of HBV infection [25].

In addition, 87% of LIA-positive, 17% of LIA-untypeable and none of the LIA-indeterminate samples were confirmed WB-positive in this investigation, demonstrating the duplex qPCR was more sensitive than WB. Actually, in our study, HTLV-1 provirus was found in 96.7% (27/28) of WB-positive blood donors, exceeding the detection rates of the qPCRs created by Moens et al. and Waters et al. for the seropositive blood donors (91% and 90%), and being close to the rate of a multiplex qPCR created by Gonçalves et al. (97.4%) [9, 10, 26]. However, the sensitivity was decreased to 74.6–77.1% when it employed in samples from HIV/HTLV-coinfected patients. The negative results may largely due to the use of highly active antiretroviral therapy (HAART), low proviral load, and punctual mutations in the genomic regions employed in PCR assays or defective provirus circulating in such patients [26, 27]. Unfortunately, our assay was not applied in such populations, but in the future, sensitivities of this assay in different populations should be analyzed. Furthermore, 33.3% (2/6) of WB-indeterminate samples were confirmed as qPCR-positive. This ratio is higher than those that have been previously recorded in other areas of the word, such as Brazil (9.2%), Iran (12.5%) and Argentina (14.7%) [28‐30], but is lower than that in the United States (44%) [31], indicating the ratio may differ geographically. However, few data are available on the proportion of qPCR-positive samples among LIA-untypeable or indeterminate samples. LIA has a greater sensitivity and accuracy in comparison to WB for confirmation of HTLV-1/2 infection [7, 32]. But it usually yields more inconclusive results than WB, which may partial due to the very low proviral load of the carriers. Our findings also showed that the majority of the qPCR-negative samples had much lower Murex-ELISA and Roche-ECLIA cut-off values than the qPCR-positive samples, indicating that the inconclusive results may also due to cross-reactivity against other retroviruses or microbial agents, or to the seroconversion period. The real status of the blood donors need to be followed-up to ensure the blood safety.

In conclusion, the simplicity, applicable, reproducible and low-cost characteristics of the present assay for PVL quantification make it potentially an effective tool to instead of WB as the first-line test in confirming the ELISA reactive samples. It also offers a way for better compare of the PVLs obtained from different laboratories, and the results obtained support its use for monitoring infection and stratifying the risk for HTLV-1-associated disease.