HPV Population Profiling in Healthy Men by Next-Generation Deep Sequencing Coupled with HPV-QUEST

Abstract

Multiple-type human papillomaviruses (HPV) infection presents a greater risk for persistence in asymptomatic individuals and may accelerate cancer development. To extend the scope of HPV types defined by probe-based assays, multiplexing deep sequencing of HPV L1, coupled with an HPV-QUEST genotyping server and a bioinformatic pipeline, was established and applied to survey the diversity of HPV genotypes among a subset of healthy men from the HPV in Men (HIM) Multinational Study. Twenty-one HPV genotypes (12 high-risk and 9 low-risk) were detected in the genital area from 18 asymptomatic individuals. A single HPV type, either HPV16, HPV6b or HPV83, was detected in 7 individuals, while coinfection by 2 to 5 high-risk and/or low-risk genotypes was identified in the other 11 participants. In two individuals studied for over one year, HPV16 persisted, while fluctuations of coinfecting genotypes occurred. HPV L1 regions were generally identical between query and reference sequences, although nonsynonymous and synonymous nucleotide polymorphisms of HPV16, 18, 31, 35h, 59, 70, 73, cand85, 6b, 62, 81, 83, cand89 or JEB2 L1 genotypes, mostly unidentified by linear array, were evident. Deep sequencing coupled with HPV-QUEST provides efficient and unambiguous classification of HPV genotypes in multiple-type HPV infection in host ecosystems.

Keywords: HPV; HPV-QUEST; Linear Array; Papillomavirus Episteme; asymptomatic men; deep sequencing; multiple-type HPV infection.

Figure 1

Primer sets used to generate human papillomaviruses (HPV) L1 amplicon libraries. A cocktail of PGMY11/09 PGMY consensus primers was used in 1st-round amplification to produce a product of 464 nucleotides (nt). GP5+ and GP6+ consensus primers conjugated at 5’ end with adaptors A or B respectively, were used in 2nd-round amplification to produce a product of 155 nt (106 nt w/o adaptors, multiplex identifier (MID) and primers). Primer positions are based on HPV8 genome (M12737.1) [44,45]. Distinct MIDs were added between adaptor A and GP5+ for sample identification. Sequences were read from adaptor A.

Figure 2

Bioinformatic pipeline of sequence analysis. HPV L1 was amplified by PCR using PGMY or nested PCR using PGMY/GP+ from genomic DNA from each individual and subjected to Linear Array (LA) and deep sequencing/HPV-QUEST, respectively, to compare sensitivity and coverage in HPV genotype capturing. Raw deep sequencing reads generated from pooled HPV L1 libraries were separated and assigned to each individual according to MID sequence codes. Quality control removed low quality reads. Quality sequences of an HPV16 clone were aligned to HPV16 reference sequence (GI│333031) to study errors due to PCR and deep sequencing. To study HPV L1 sequences at variant level, quality sequences were clustered at 3% pairwise distance [51]. Consensus sequence derived from each cluster represented an HPV variant, and relative proportion of each HPV genotype within a sample served as a reference rather than a quantification of viral population structure. Consensus sequences were queried in HPV-QUEST to obtain HPV genus, species, genotype and oncogenicity. Papillomavirus Episteme (PaVE) Specific Blastn was used to confirm HPV-QUEST genotyping results, and Nationa Center for Biotechnology Information (NCBI) Blastn was applied to evaluate discrepancies of genotyping by HPV-QUEST and PaVE. NGS/HPV-QUEST was compared with LA with respect of sensitivity and coverage in identifying HPV genotypes and variants.

Figure 3

HPV genotyping by deep sequencing/HPV-QUEST. Each HPV genotype captured by deep sequencing within an individual is depicted as a color-filled block: Pink represents HPV type defined by ≥95% of sequences; orange, sequences ≥50% but <95%; yellow, sequences ≥1% but <50%; green, sequences ≥0.1% but <1%; and blue, <0.1% sequences. Red-outlined squares represent HPV types detected by LA. All HPV types, except for HPV32 and JEB2 outlined by blue, are covered by LA probe set. Symbols: *, sample untyped by LA; ‡, HPVcand85; #, HPVcand89; ¶, unknown. Reference sequences for genotypes identified in the study samples include: HPV16 (GI|333031), HPV18 (GI|60975), HPV31 (GI|333048), HPV35h (GI|396997), HPV51 (GI|333087), HPV52 (GI|397038), HPV56 (GI|39053), HPV59 (GI|557236), HPV66 (GI|1020290), HPV70 (GI|1173493), HPV73 (GI|1491692), HPVcand85 (GI|4574720), HPV6b (GI|60955), HPV11 (GI|333026), HPV32 (GI|396981), HPV62 (GI|577400), HPV81 (GI|40804509), HPV83 (GI|5059324), HPV87 (GI|14475578), HPVcand89 (GI|22095322), or JEB2 (GI|45925861) [53,61,62,63].

Figure 4

HPV L1 variants within genotypes. HPV L1 nucleotide sequence(s) of genotype variant(s) were aligned to corresponding reference genotype sequences with grey color highlighting variants identical to reference sequences. Synonymous and nonsynonymous nucleotide differences were highlighted by green and red, respectively, and boxed to indicate the same nucleotide substitution in more than one variant of an HPV genotype in one or multiple individuals. Nucleotide positions in HPV L1 sequences were based on L1 region of reference sequences (nucleotides 1123 to 1212 for HPV16; 1195 to 1290, HPV18; 1051 to 1137, HPV31; 1006 to 1098, HPV35h; 1052 to 1141, HPV59; 1006 to 1095, HPV70; 1006 to 1095, HPV73; 1012 to 1098 for HPVcand85; 1000 to 1089 for HPV6b; 67 to 156 for HPV62; 1018 to 1107 for HPV81; 1018 to 1107 for HPV 83; 1066 to 1153 for HPVcand89, and 15 to 95 for HPVJEB2 (#: partial L1 reference sequence)). * sample with variant(s) failed to be typed by LA.

Figure 5

Longitudinal changes of HPV genotypes in two individuals. Each HPV genotype captured by deep sequencing within an individual is depicted as a color-filled block: Pink represents HPV type defined by ≥95% of sequences; orange, sequences ≥50% but <95%; yellow, sequences ≥1% but <50%; green, sequences ≥0.1% but <1%; and blue, <0.1% sequences. Red-outlined squares represent HPV types detected by LA. H: high-risk; L: low-risk.