Higher Expression of HPV16 Derived E7_LI Transcript Observed in Men With HIV and Recurrent Anal Cancer

Abstract

Squamous cell carcinoma of the anus (SCCA) or anal cancer (AC) is an understudied cancer with a high occurrence rate in people with HIV (PWH), especially men having sex with men (MSM). Furthermore, AC recurs in approximately one-fourth of patients who undergo standard care with chemoradiation therapy (CRT). Using bulk RNA sequencing data of AC obtained from 12 patients with non-recurrent (NR, N = 9) or recurrent (R, N = 3) cancer, we previously showed upregulated expression of key immune genes in the NR compared to the R group. Although the main causative agent of AC is high-risk human papillomavirus (HPV), association of host and viral RNA transcript expression contributing to AC recurrence has not been extensively studied. The objective of the current study was to determine whether enrichment of specific HPV genotypes and/or HPV gene expression patterns differentiate the two groups and if any specific viral (HPV) and host (human) immune mediators correlate with each other. Using bulk RNA sequencing data and VIRTUS 2, we detected viral RNA reads mapping to seven high-risk and six low-risk HPV types, of which the high-risk HPV16 observed in 83% (10/12) AC tumors (7/9 NR and 3/3 R). Rate of all HPV genomes trended toward a decrease in NR AC isolates and correlation between HPV types was more commonly observed in low-risk ones. Analysis of HPV 16 gene expression profile showed a significantly lower positivity rate for a polycistronic transcript encoding for E7^L1 in the NR group (1/9, NR vs. 3/3, R, p < 0.05). An unbiased correlation analysis of HPV-human transcript expression showed a direct correlation between HPV transcripts and human genes involved in cell growth. The data also identified human transcripts showing an inverse correlation with HPV gene expression. These included genes involved in negative regulation of growth, proliferation, and immune response. Taken together, these data indicate that concurrent analyses of viral and host factors in the same tumor can identify potential new therapeutic targets to ameliorate cancer recurrence post-treatment.

Keywords: HIV; HPV; RNA‐seq; anal cancer; cancer recurrence; transcriptome profile; treatment response; virome; virtus2.

Figure 1

HPV16 is the predominant HPV type observed in anal cancer isolates. Unmapped nonhuman reads were mapped to all HPV types included in concatenated Virtus2 reference genomes through a modified version of the Virtus2 paired‐end workflow using a wrapper to define sample groups. Coverage and viral to human rate (v/h mapped read %) were determined for all HPV types and those which were detected are shown as indicated. The full name and source of each reference sequence are provided. Significance p value results of Mann−Whitney's U‐test performed internally through Virtus2 between indicated groups is also shown on left side of graph. HPV types identified are also divided into high‐risk and low‐risk.

Figure 2

Rate of all HPV genomes trends toward increase in recurrent anal cancer isolates and correlation is more commonly observed in low‐risk HPV types. (A) Box‐plot graph representing the rate of HPV16 in both groups, p‐value represents the Mann−Whitney t‐test comparison between groups. (B) Correlation matrix representing the correlation between the presence of each of the two compared HPV types (high and low risk). Numbers indicated within each box represent the r value for correlation. Higher numbers and deeper blue indicate a more positive correlation. HPV types are separated into high or low risk.

Figure 3

Relative abundance of HPV16 polycistronic transcripts in relation to anal cancer treatment outcome. Reads are aligned to polycistronic transcripts as opposed to a full ORF sequence. (A) Results in rate (v/h mapped read %) and coverage represented via bubble plot. Transcripts which significantly differ between groups are outlined in red (with red arrow) while those showing a trend are indicated with an orange circle and arrow (p < 0.1). Bar graphs indicating positive frequency in each group of the indicated transcript 20 (B) or transcript 1 (C) are shown. HPV16 sequence position numbers and transcript splice site as compared to HPV16 ORF sequence indicated for each transcript below the bar graph.

Figure 4

HPV 16 rate correlates negatively with HLA‐A and positively with DEAD‐Box protein (DDX24) gene expression. (A) Heatmap of HPV16/18 genome rate alongside expression levels of indicated HLA's or DEAD‐Box Protein family genes (indicated in RPKM, or rate of expression per kilobase exon per million reads). Mann−Whitney's U test (within Virtus2) or EdgeR differentially expressed gene p values are indicated between recurrent and nonrecurrent sample groups as indicated. Correlation analysis of HPV16 whole genome rate with indicated (A) HLA class I and II molecules or (B) DEAD‐Box proteins. R, R2, and p values for Pearson's correlation are indicated (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 respectively).

Figure 5

An unbiased HPV‐human gene correlation analysis indicates gene expression of HPV16 and its transcripts positively correlate with markers of cell growth in anal cancer. RPKM of human gene mRNA was correlated in an unbiased bioinformatic pipeline with the viral rate (viral/human reads) of all detected whole HPV type genomes as well as individual HPV16 transcripts. Then human genes which were directly or inversely correlated with HPV genome or transcript rates were enriched for gene ontology (GO) terms separately. (A) Pyramid plot of human genes which were inversely and directly correlated with HPV16 whole genome rate. # of genes enriching to GO term represented as bars, green for direct and red for indirect. (B) Correlation plot for highest overall directly correlated human gene with HPV16 rate (top row) as well as top gene correlated with HPV16 rate from top two GO terms (left, in green box). Lowest gene (top row), or highest inversely correlated gene with HPV16 rate as well as top inversely correlated gene with top two inversely correlated GO terms (right, in red box). Clustered radar chart of GO enrichment for all HPV transcripts (C) directly or (D) indirectly correlated human genes by transcript. # Genes in each GO term maps to radius of lines from center. Top (E) most directly or (F) most indirectly correlated gene for the highest gene # GO term of every transcript is represented as a faceted line plot with only line of best fit represented. Correlation and p‐value of each are also shown beneath each plot. “T#” refers to Transcript # of HPV16 individual transcripts identified in Figure 3.