Digital droplet PCR (ddPCR) for the precise quantification of human T-lymphotropic virus 1 proviral loads in peripheral blood and cerebrospinal fluid of HAM/TSP patients and identification of viral mutations

Abstract

An elevated human T cell lymphotropic virus 1 (HTLV)-1 proviral load (PVL) is the main risk factor for developing HTLV-1-associated myelopathy/tropical spastic paraparesis (HAM/TSP) in HTLV-1 infected subjects, and a high cerebrospinal fluid (CSF) to peripheral blood mononuclear cell (PBMC) PVL ratio may be diagnostic of the condition. However, the standard method for quantification of HTLV-1 PVL-real-time PCR-has multiple limitations, including increased inter-assay variability in compartments with low cell numbers, such as CSF. Therefore, in this study, we evaluated a novel technique for HTVL-1 PVL quantification, digital droplet PCR (ddPCR). In ddPCR, PCR samples are partitioned into thousands of nanoliter-sized droplets, amplified on a thermocycler, and queried for fluorescent signal. Due to the high number of independent events (droplets), Poisson algorithms are used to determine absolute copy numbers independently of a standard curve, which enables highly precise quantitation. This assay has low intra-assay variability allowing for reliable PVL measurement in PBMC and CSF compartments of both asymptomatic carriers (AC) and HAM/TSP patients. It is also useful for HTLV-1-related clinical applications, such as longitudinal monitoring of PVL and identification of viral mutations within the region targeted by the primers and probe.

Fig. 1

Representative two-dimensional ddPCR plots of individuals with increasing HTLV-PVL. a PBMC DNA from a normal blood donor. b PBMC DNA from an HTLV-1-infected individual with a PVL of 3.25 %. c PBMC DNA from an HTLV-1-infected individual with a PVL of 8.46 %. d PBMC DNA from an HTLV-1-infected individual with a PVL of 19.50 %

Fig. 2

Statistically significant differences in HTLV-1 PBMC PVL between HAM/TSP patients, asymptomatic carriers, and healthy donors. Mann-Whitney tests were performed to compare differences between groups in sets of pairs. HAM/TSP PBMC PVL were statistically different from asymptomatic carrier PBMC PVL, p = 0.0093 (**). HAM/TSP PBMC PVL were also statistically different from healthy donor PBMC PVL, p < 0.0001 (****). Asymptomatic carrier PBMC PVL versus healthy donor PBMC PVL were also statistically different, p = 0.002 (***)

Fig. 3

Strong correlation between ddPCR and qPCR for HTLV-1 PBMC PVL quantitation, though ddPCR has lower inter-assay variability. a Replicate, non-normalized PBMC PVL values obtained for five HAM/TSP patients using qPCR and ddPCR were similar, though the values obtained by ddPCR were consistently higher. b A strong correlation exists between the normalized PVL values for ddPCR and qPCR (Pearson correlation coefficient, R ² = 0.97). c An ROC analysis of the normalized PVL values for each method indicates that the assays perform similarly for HTLV-1 PVL estimation. d Comparison of each method’s normalized values demonstrates that ddPCR has a lower coefficient of variation compared to qPCR

Fig. 4

Low intra-assay variability by ddPCR over a wide range of HTLV-1 PBMC PVL. The intra-assay variability of ddPCR for low, medium, and high PVL groups was evaluated by running ten replicates of PBMC DNA from eight patients (three AC and five HAM/TSP). For the low PVL group (<5 %), the mean PVL was 1.85 % with a mean CV of 13.7 %. For the medium PVL group (5–10 %), the mean PVL was 9.77 % with a mean CV of 7.44 %. For the high PVL group (>10 %), the mean PVL was 32.245 with a mean CV of 3.4 %

Fig. 5

Elevated CSF to PBMC PVL in HAM/TSP patients demonstrated by ddPCR. HTLV-1 PVL is elevated in the CSF of seven HAM/TSP patients relative to matched PBMC, by approximately 2.8-fold. For each patient, there is a significant difference between PBMC and CSF HTLV-1 PVL (paired t test, p = 0.0036)

Fig. 6

HAM/TSP patients’ PBMC PVL are relatively stable over a period of approximately 4 years. a HAM/TSP patient 15 had a mean PVL of 18.13 %. b HAM/TSP patient 13 had a mean PVL of 12.3 %. c HAM/TSP patient 14 had a mean PVL of 9.16 %. d HAM/TSP patient 3 had a mean PVL of 10.25 %

Fig. 7

DdPCR fluorescence profile indicates mutations in viral region targeted by primers/probe. a A typical one-dimensional droplet profile for HAM/TSP patients has a positive HTLV-1 tax population at a fluorescence amplitude of 9,000 (red arrow). b The HTLV-1 tax-positive droplets for HAM/TSP patient 16 were observed around an amplitude of 3,500 (red arrow). c Sequence alignment revealed that HAM/TSP patient 16 has eight point mutations (indicated in red) in the region of HTLV-1 tax targeted by the ddPCR primers/probe, one in the probe-binding region. The primer-binding regions are italicized, and the probe-binding region is underlined. d A typical one-dimensional droplet profile for HAM/TSP patients has a positive HTLV-1 gag population at a fluorescence amplitude of 11,000 (red arrow). e The HTLV-1 gag positive droplets for HAM/TSP patient 19 were observed around an amplitude of 4,000 (red arrow). f Sequence alignment revealed that HAM/TSP patient 19 has one point mutation (indicated in red) in the region targeted by the gag probe (underlined)