Development of a high-performance multi-probe droplet digital PCR assay for high-sensitivity detection of human papillomavirus circulating tumor DNA from plasma

Abstract

Objectives: To develop a high-performance droplet digital PCR (ddPCR) assay capable of enhancing the detection of human papillomavirus (HPV) circulating tumor DNA (ctDNA) in plasma from patients with HPV-associated oropharyngeal squamous cell carcinoma (HPV+ OPSCC).

Materials and methods: Plasma samples from subjects with HPV+ OPSCC were collected. We developed a high-performance ddPCR assay designed to simultaneously target nine regions of the HPV16 genome.

Results: The new assay termed 'ctDNA HPV16 Assessment using Multiple Probes' (CHAMP- 16) yielded significantly higher HPV16 counts compared to our previously validated 'Single-Probe' (SP) assay and a commercially available NavDx® assay. Analytical validation demonstrated that the CHAMP-16 assay had a limit of detection (LoD) of 4.1 copies per reaction, corresponding to < 1 genome equivalent (GE) of HPV16. When tested on plasma ctDNA from 21 patients with early-stage HPV+ OPSCC and known HPV16 ctDNA using the SP assay, all patients were positive for HPV16 ctDNA in both assays and the CHAMP-16 assay displayed 6.6-fold higher HPV16 signal on average. Finally, in a longitudinal analysis of samples from a patient with recurrent disease, the CHAMP-16 assay detected HPV16 ctDNA signal ∼ 20 months prior to the conventional SP assay.

Conclusion: Increased HPV16 signal detection using the CHAMP-16 assay suggests the potential for detection of recurrences significantly earlier than with conventional ddPCR assays in patients with HPV16+ OPSCC. Critically, this multi-probe approach maintains the cost-benefit advantage of ddPCR over next generation sequencing (NGS) approaches, supporting the cost-effectiveness of this assay for both large population screening and routine post-treatment surveillance.

Keywords: Anal cancer; Biomarker; Cell-free DNA; Cervical cancer; Circulating tumor DNA; HPV; Head and neck cancer; Liquid biopsy; Oropharyngeal cancer; ctDNA.

Figure 1.. Design and Experimental Testing of Multiple Primer/Probe Sets Targeting the HPV16 genome.

(A) Flowchart outlining the steps for design and development of the HPV16 multi-probe assay. Briefly, computational analysis was used to scan the entire HPV16 genome for primer and probe binding regions complying with our primer/probe design criteria to select combinations predicted to specifically bind the HPV16 genome. A schematic based on The PapillomaVirus Episteme (PaVE) coordinates depicts the HPV16 genomic regions targeted by primer/probe combinations at various stages of selection. Of the 49 primer/probe sets selected from computational screens (forward primer = green and reverse primer = red), 18 non-overlapping sets were selected with examples circled and 1 primer/probe set #4 overlapping with another primer set at the E6/E7 junction was selected based on higher GC content, making a total of 19 primer/probe sets for functional screens. (Note: ‘SP assay’ = previously validated, single-probe assay with a primer/probe set targeting a 77 bp region of HPV16; [29]). (B) ddPCR droplet profile displaying the FAM signal (blue) for the 19 selected primer/probe candidates, tested in duplicates at final primer/probe concentration of 900 nM/250 nM using 400 GEs of HindIII digested HPV16 cell line UM-SCC-104 genomic DNA (gDNA) or 200,000 GEs of HindIII digested non-HPV human genomic DNA (hgen DNA) as template. The threshold (pink line) was kept constant for all primer/probe combinations except for #17 and #18 where the signal intensity was not remarkably high compared to the background (gray). HPV16 copies detected (average of duplicates) with each primer/probe combination compared to the SP assay (orange) are tabulated at the bottom. In the case of primer/probe combination #14, the background could not be distinguished from the FAM signal. No signal was observed with the negative control non-HPV hgen DNA as template, for any of the primer/probe combinations (droplet plot on the right).

Figure 2.. Primer/Probe Pool Selection for Multi-Probe Assay Development.

(A) Pooling schema for a multi-probe assay development using primer/probe sets targeting different regions of HPV16 and selected based on a high signal intensity compared to background during screening. The schematic shows a combination of 3, 5, 10, or 14 primer/probe sets, termed as 3-Pool, 5-Pool, 10-Pool, and 14-Pool, respectively. (B) ddPCR droplet profile (left) displaying the FAM signal (blue) in a single-plex assay for 5 primer/probe sets individually (#4, #10, #13, #15 and #16), as well as for 3-Pool, 5-Pool, 10-Pool, and 14-Pool, with a concentration of 450 nM of each primer and 125 nM of each probe. In the case of 3-Pool and 5-Pool, separation of the signal from the background (gray) allowed for the calculation of HPV16 copies but in the case of 10-pool and 14-pool the signal overlapped with the background and HPV16 copies could not be calculated. The HPV16 copies obtained for individual primer/probe combinations, 3-Pool and 5-Pool with sheared UM-SCC-104 gDNA (500 GEs) or water control as template are tabulated at right. Data shown is representative of different primer/probe concentrations that were tried to test if their signal intensity to background ratios can be improved for the pools (Fig. S2). (C) VIC labeled probes #2, #4, #13, #15, and #16 were tested in various combinations of 4 primer/probe sets to screen for a 4-Pool using sheared UM-SCC-104 gDNA template (100 GEs) at final primer/probe concentration of 450 nM/125 nM each. The droplet profile shows signal intensity (green) and the background (gray). HPV16 copies are listed and 4-Poolv20 (highlighted in light green) was chosen for further experimental evaluation.

Figure 3.. Screening and Selection of a Dual-Colored Multi-Probe Pool.

2-D ddPCR plots showing droplet profiles of (A) single-probe FAM assay #6 and (B-I) different 5-Pool combinations containing FAM labeled probes (blue) screened with 4-Poolv20 (#2+#4+#13+#15) containing VIC labeled probes (green) at final primer/probe concentration of 450 nM/125 nM each, using sheared UM-SCC-104 gDNA as template (50 GEs). HPV16 copies along with fold enhancement of the signal measured with the 5-Pool FAM combinations over the single-probe (#6) assay are tabulated at the bottom. The best separation of the background (gray) from signal (blue) was observed in 5-Poolv39FAM (#5+#6+#7+#16+#19) (panel G; highlighted in light red) when comparing the FAM labeled single-probe (#6) assay to the different 5-Pool FAM combinations. The threshold of 6200 for 4-Poolv20VIC was kept constant to ensure that the FAM pool was not changing the background intensity of the 4-Poolv20VIC in the case of dual-channel measurement. In cases where FAM pools altered the background (e.g., 5-Poolv35 and 5-Poolv41), HPV16 copies could not be determined and are listed as ‘nd’.

Figure 4.. Analytical Validation of the Dual-Colored Multi-probe CHAMP-16 Assay.

(A and B) Linearity plots showing ddPCR assay results from a 2-fold dilution series (average of 3 replicates) to determine the Limit of Detection (LoD) and reportable range of HPV16 ctDNA using (A) a template comprising 9 synthetic HPV16 dsDNA duplexes corresponding to the genomic regions targeted by the 9 primer/probe combinations pooled in the CHAMP-16 assay and (B) sheared UM-SCC-104 gDNA as template. The synthetic HPV16 target DNA pool and sheared UM-SCC-104 gDNA were tested in a background of sheared hgen matrix (`44,000 diploid GEs per well) and (A) the expected copies (x-axis) or (B) the GEs tested (x-axis) were plotted against the measured copies (y-axis). The LoD was analyzed and calculated to be 4.1 copies per 20ul reaction (see supplemental methods). No non-specific signal was observed when only sheared hgen DNA (~44,000 diploid GEs per 20 μl reaction) was used as a non-HPV template (n = 61). (C and D) % CV was plotted for measured copies of HPV16 at different dilutions (y-axis) versus (C) the expected copies (converted to log₁₀) of synthetic pool of 9 DNA targets tested (x-axis) or (D) the number of GEs (converted to log₁₀) of sheared UM-SCC-104 gDNA tested (x-axis). An arbitrary 20% threshold is indicated by a dashed line.

Figure 5.. Case Study Comparing the CHAMP-16 Assay to the Conventional SP Assay for Early Detection of Recurrence.

Plasma samples collected from a biopsy-confirmed HPV+ OPSCC patient undergoing chemoradiotherapy (CRT) were analyzed for HPV16 ctDNA using the SP assay (orange triangles) and CHAMP-16 assay (green circles). Number of HPV16 copies detected (y-axis) were plotted as a cumulative of duplicate plasma cfDNA samples tested from blood drawn over several months (x-axis). PET (+) indicates a positive PET scan, whereas PET (-ve) indicates a negative PET scan with respect to detecting presence of cancer. Closed symbols represent HPV16 molecules detected above LoD (black dotted line), and open symbols represent copies that were below the LoD. Notably, the CHAMP-16 assay was able to detect HPV16 ctDNA signal ~20 months prior to clinical recurrence and signal detection by the SP assay.