Detection and significance of HPV16 E5-specific productive transcripts in oropharyngeal carcinoma

Abstract

Background

HPV16 is implicated in around 30% of oropharyngeal cancers (OPCs). HPV-associated OPCs generally show a better prognosis, but 20% deviate from this trend, indicating a need for better molecular profiling. HPV16-E5 is an oncoprotein encoded by an mRNA that undergoes extensive splicing, with only one specific transcript being translatable. The prognostic significance of this E5-productive transcript in HPV-related OPCs is not well-studied.

Methods

We retrospectively analysed 74 HPV16-positive OPC cases diagnosed between 2011 and 2021. FFPE tissue samples were used for p16, EGFR, and HLA-I analysis by immunohistochemistry while E5, productive E5, E6 and E7 transcripts were detected by qPCR. Survival analysis was conducted using Kaplan-Meier curves and log-rank tests.

Results

Productive HPV16-E5 transcripts were detected in 11.3% of OPCs. There was no significant association between HPV16-E5 transcripts and EGFR or HLA-I expression. However, the presence of the productive E5 transcript was associated with worse progression-free survival (PFS) (p = 0.0024). EGFR or HLA-I expression was not significantly associated with PFS (p = 0.17 and p = 0.93, respectively).

Conclusions

This study demonstrates for the first time the presence of E5-productive transcripts in HPV16-positive OPCs and its association with poorer PFS, highlighting its potential as a prognostic marker. Further research with larger cohorts is needed to validate these findings.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12885-025-14506-0.

Background

Head and neck cancers (HNCs) represent the fifth most common cancer worldwide. High-risk Human Papillomaviruses (HR-HPV), particularly HPV16, are implicated in approximately 30% of oropharyngeal cancers (OPCs). Though there are notable geographical differences, the prevalence of OPCs is increasing worldwide [1, 2]. While HPV-associated OPCs generally show a better prognosis and response to therapy compared to HPV-negative OPCs, about 20% of HPV-positive OPCs do not exhibit these favourable characteristics, suggesting a need to improve molecular profiling and patient stratification for tailored therapies [3, 4].

HPV16-E5 is the least characterized of the three HR-HPV oncoproteins, namely HPV E5, E6 and E7. HPV16-E5 expression has been detected in almost 80–90 % of cervical low-grade (LSILs) and high-grade squamous intraepithelial lesions (HSILs), and 60% of cervical carcinomas [5]. Several functions of E5 have been reported, including anti-apoptotic activity, promotion of epidermal growth factor receptor (EGFR) recycling, and immune evasion [6]. The role of HPV16-E5 in promoting immune evasion has been mainly attributed to its ability to negatively affect HLA class I (HLA-I) expression, thus affecting antigen processing and presentation [7]. Interestingly, integration of HR-HPV DNA into the host genome leads to loss of the E5 open reading frame (ORF), suggesting that this oncoprotein might play a role in the early stages of malignant transformation.

HPV16-E5 mRNA undergoes extensive splicing and importantly only one of the several mRNA transcripts with coding potential for E5 protein can effectively be translated into the E5 protein, namely the E5-specific productive transcript. Only a few studies have evaluated the expression of E5 in HPV-related OPCs; [8, 9] however, none have analysed the prognostic significance of the E5-specific productive transcript.

Currently, anti-EGFR therapies are being evaluated in de-escalation clinical trials, yet the prognostic significance of E5 expression in relation to EGFR expression is largely unknown [9]. Thus, a better understanding of the prognostic and predictive value of EGFR expression in relation to HPV oncogenes may help to guide safer de-escalation and improve management of patients with HPV-associated OPC. Similarly, downregulation of HLA-I has been observed in many cancers, including HNCs, although inconsistently. HLA-I has also been investigated as a possible prognostic marker in HNCs, though results have been conflicting [10, 11].

This study aimed to assess, for the first time, the presence of the specific productive transcript of HPV16-E5 in HPV16-associated OPCs and investigate its association with EGFR and HLA-I expression. In addition, we sought to explore the correlation of HPV16-E5 mRNA, EGFR and HLA-I expression with clinical outcomes.

Methods

Study population

Cases of histologically-confirmed OPCs diagnosed between January 2011 and November 2021 at the IRCCS Regina Elena National Cancer Institute were retrospectively evaluated. The present study included OPCs fulfilling the following inclusion criteria: (i) primary and untreated OPC (ICD-10 codes: base of the tongue; C01; soft palate, C05.1; tonsil, C02.4, C09.0–9; uvula, C05.2; other oropharynx parts, C10.0–9; Waldeyer’s ring, C14.2); (ii) simultaneous positivity for HPV16-DNA and p16ink4a (p16) immunostaining; (iii) sufficient residual tissue for study analyses. Formalin-fixed paraffin-embedded (FFPE) tissue blocks were retrieved from the Pathology Department archives. Demographic, clinical and histopathologic features were obtained from medical records. Staging used AJCC 7th edition for diagnoses between 2011 and 2016 and 8th edition for diagnoses from 2017 on. The 8th edition introduced novel pathologic and clinical staging systems for HPV-associated OPCs, after several studies had demonstrated that the 7th edition was inadequate for predicting outcomes of these tumours. Patients with HPV-associated OPCs at advanced stage, for instance, do not experience the same poor outcome as those with HPV-negative tumours. To reflect better the understanding of tumour biology and clinical behaviour of HPV-associated vs. HPV-negative OPCs, the 8th edition introduced a different staging system for OPCs caused by HPV [12]. All enrolled patients signed an informed consent to the protocol approved by the IRCCS Regina Elena National Cancer Institute, Istituti Fisioterapici Ospitalieri, Institutional Review Board (RS 1443/20).

HPV-DNA testing

HPV-DNA detection and genotyping on FFPE tissue samples were performed using INNO-LiPA genotyping Extra and Extra II assays (Fujirebio) according to the manufacturer's instructions. SPF10 primers were used to amplify a 65 bp fragment in L1 ORF, providing qualitative detection of mucosal HPVs, including HR-HPV types.

FFPE tissue block sectioning

Sections of FFPE tissue blocks were obtained as follows: (i) 1 to 3 × 5-µm sections for RNA extraction; (ii) 2 × 3-µm sections for EGFR and HLA-I immunohistochemical staining; (iii) 1 × 3-µm for hematoxylin and eosin staining and diagnosis confirmation. To avoid cross-contamination, the microtome surface was cleaned with DNAZap (Invitrogen, Thermo Fisher Scientific) and a new blade was used for each case.

Immunohistochemical staining

p16, EGFR and HLA-I expression were evaluated by immunohistochemistry. p16 expression was assessed by CINtec^® Histology kit (clone E6H4, Roche Diagnostics) and classified as positive when a strong nuclear expression was observed in > 75% of tumor cells [13]. EGFR and HLA-I expression were detected using CONFIRM anti-EGFR (clone 3C6, Roche Diagnostics) and anti-HLA Class 1 ABC antibody (clone EMR8-5, Abcam), respectively. EGFR staining was positive if > 10% of tumor cells showed membrane staining with weak/moderate or strong intensity [9]. HLA-I expression was defined as: diffuse loss (< 10% positive cells), subclonal loss (10% to < 80% positive cells), or intact (≥ 80% positive cells). Two investigators independently evaluated the staining for each biomarker.

Extraction and purification of RNA from FFPE tissues

Total RNA was extracted from FFPE sections using the RNeasy FFPE kit (Qiagen) following the manufacturer’s instructions.

DNA synthesis and real-time quantitation

RNA was reverse-transcribed into cDNA using the Quantitect Reverse Transcription Kit (Qiagen), yielding a 20µL cDNA solution per sample. The g-DNA wipeout buffer eliminated contaminating genomic DNA. cDNA was amplified by Real-time PCR (qPCR) with iQ Supermix (Bio-Rad) containing SYBR Green and β-actin primers to test RNA quality. Specific E5 primers (forward: 5’-GCGACGTGAGAGCAACG-3’, reverse: 5’-AGGGGTTTCCGGTGTCTGG-3’) were designed to target the splicing site SD226–SA3358, previously reported to generate an mRNA with coding potential for E5 protein, using Beacon Design software (Bio-Rad). Additional primer pairs used to detect total E5, E6, and E7 transcripts were adopted from previous studies [14–16]. Complete primer sequences, expected amplicon sizes, and PCR conditions for all transcripts analyzed in this study are provided in Supplementary Material S1. qPCR analysis was performed with iCycler IQ software (Bio-Rad), using the ΔΔCt method normalized to β-actin. RNA from W12 cell line was used as a positive control.

Statistical analysis

Descriptive statistics summarized the study population characteristics. OPCs with and without HPV16-E5 mRNA (total and specific transcripts) were compared in terms of EGFR and HLA-I expression by chi-square test. Kaplan-Meier curves were used for the survival analysis (progression-free survival, PFS), considering disease progression or death as the events of interest. Patients alive and recurrence-free at last follow-up were right-censored. Survival curves were compared using Log-rank test. Statistical significance was set at p ≤ 0.05. Analyses were conducted using MedCalc Statistical Software version 22.013 (MedCalc Software Ltd, Ostend, Belgium; https://www.medcalc.org).

Results

Study population

Over the study period, 374 primary untreated OPCs were available and all of them had been tested for HPV-DNA. Figure in Supplementary Material S2 shows how the final study group, which included 74 HPV16-associated OPCs, was reached. These 74 patients had a median age of 61 years (IQR: 54–69). Most were men (54/74, 72.9%), with tonsil being the most common cancer subsite (41/74, 55.4%). The majority of the patients had T1-T2 disease (51/74, 68.9%), and nodal involvement (64/74, 86.5%). Patients diagnosed in 2011–2016 were mainly stage IV (28/44, 63.7%), whereas those diagnosed in 2017–2021, after implementation of the 8th edition of the AJCC Cancer Staging Manual, were most commonly stage I (13/30, 43.3%). (Table 1). Treatment predominantly involved multimodal therapy, followed standard care for each period. Median follow-up was 51 months (IQR: 43–57 months).

Table 1.

Characteristics of the 74 patients with HPV16-associated OPC overall and according to the period of diagnosis

Characteristic	HPV16-driven OPC patients n, %
	2011–2016 N = 44	2017–2021 N = 30	Total N = 74
Sex
Male	33, 75.0	21, 70.0	54, 72.9
Female	11, 25.0	9, 30.0	20, 27.1
Cancer subsite
Tonsils	24, 54.5	17, 56.7	41, 55.4
Base of tongue	19, 43.2	12, 40.0	31, 41.9
Epiglottic vallecula	1, 2.3	1, 3.3	2, 2.7
T category
1–2	33, 75.0	18, 60.0	51, 68.9
3–4	11, 25.0	12, 40.0	23, 31.1
N category
0	4, 9.1	6, 20.0	10, 13.5
N+	40, 90.9	24, 80.0	64, 86.5
TNM stagea
I	0, 0.0	13, 43.3	-
II	3, 6.8	9, 30.0	-
III	13, 29.5	8, 26.7	-
IV	28, 63.7	0, 0.0	-
Treatment
S	4, 9.1	4, 13.4	8, 10.8
S-RT	2, 4.5	3, 10.0	5, 6.7
S-CRT	12, 27.3	0, 0.0	12, 16.2
RT	6, 13.6	9, 30.0	15, 20.3
CRT	20, 45.5	12, 40.0	32, 43.2
CT	0, 0.0	1, 3.3	1, 1.4
Other (palliative care)	0, 0.0	1, 3.3	1, 1.4

^aStaging was based on AJCC 7th edition for diagnoses between 2011–2016 and on 8th edition for diagnoses from 2017 on. S Surgery, S-RT Surgery + radiotherapy, S-CRT Surgery + chemoradiation, RT Radiotherapy, CRT Chemoradiation, CT Chemiotherapy

Detection of HPV16-E5, -E6, and -E7 mRNAs

Three of the 74 cases (4.1%) lacked β-actin mRNA amplification and were excluded from the analysis. The remaining 71 OPCs all exhibited E6/E7 polycistronic transcripts, which also encompassed the E5 ORF. Specifically, HPV16-E6 was detected in all samples; HPV16-E7 in 65/71 samples (91.6%); and HPV16-E5 in 53/71 samples (74.6%). Eight cases (11.3%) tested positive for the E5-specific productive transcript (Fig. 1 and Figure S1).

Fig. 1.

Proportions of HPV16-associated OPC samples positive for the presence of HPV16 E5, E6, E7, and E5 productive mRNA.This figure highlights the differential expression of these viral transcripts in HPV16-positive OPCs, with the E5-specific productive mRNA being the least commonly detected

EGFR and HLA-I expression

Expression of EGFR and HLA-I was evaluated by immunostaining, Representative examples of OPCs displaying negative and positive staining for these two biomarkers are shown in Fig. 2.

Fig. 2.

Immunostaining for EGFR and HLA-I in HPV16-associated OPCs. HPV16-associated tonsillar squamous cell carcinoma: A Hematoxylin and Eosin staining (H&E); B Negative EGFR staining; C Intact HLA-I expression. HPV16-associated base of the tongue squamous cell carcinoma: D H&E staining; E Positive EGFR staining; F HLA-I diffuse loss. 3′,3′ diaminobenzidine (DAB) was used as a chromogenic substrate. 200× magnification

EGFR expression was observed in 35/74 OPCs (47.3%), with similar rates in tumours positive and negative for HPV16-E5 total (43.4% vs. 50.0%, p = 0.63) and productive mRNA (37.5% vs. 46.0%, p = 0.65), respectively (Table 2).

Table 2.

Results of the immunohistochemical staining for EGFR and HLA-I in 71 HPV16-associated OPCs stratified according to the expression of HPV16E5-mRNA and HPV16E5-productive mRNA

	EGFR n (%)		HLA-I n (%)			Totaln (%)
	negative	positive	intact	subclonal loss	diffuse loss
HPV16E5-mRNA
negative	9 (50.0)	9 (50.0)	1 (5.6)	12 (66.7)	5 (27.8)	18 (100)
positive	30 (56.6)	23 (43.4)	6 (11.3)	23 (43.4)	24 (45.3)	53 (100)
HPV16E5-productive mRNA
negative	34 (54.0)	29 (46.0)	5 (7.9)	33 (52.4)	25 (39.7)	63 (100)
positive	5 (62.5)	3 (37.5)	2 (25.0)	2 (25.0)	4 (50.0)	8 (100)
Total, n (%)	39 (54.9)	32 (45.1)	7 (9.9)	35 (49.3)	29 (40.8)	71 (100)

HLA-I expression loss was observed in 67/74 cases (90.5%), with 31/74 (41.9%) showing diffuse loss and 36/74 (48.6%) subclonal loss. The majority of the OPCs positive for HPV16-E5 total (47/53, 88.7%) and productive transcript (6/8, 75.0%) displayed HLA-I expression loss. (Table 2). No significant differences in HLA-I expression (intact, subclonal loss, diffuse loss) were observed based on the presence of HPV16-E5 total (p = 0.23) or productive mRNA (p = 0.19).

Survival analysis of patients with HPV16-driven OPC

Kaplan-Meier survival analysis was conducted to evaluate the association of each biomarker with PFS for the 71 patients with both a valid RNA analysis and immunohistochemistry results. Patients positive for HPV16-E5 total transcripts showed a better PFS compared to those negative for these transcripts (hazard ratio [HR] for recurrence or death 0.45, 95% CI: 0.16 to 1.26; p = 0.13) (Fig. 3a), whereas patients positive for the HPV16-E5 productive transcript showed a significantly worse PFS than those negative for this transcript (HR 13.80, 95% CI: 2.53 to 75.03; p = 0.0024) (Fig. 3b). At the last available follow-up, one of the patients harbouring the productive transcript had died from the OPC. EGFR-positive patients showed a worse PFS compared to EGFR-negative patients, but this difference was not significant (HR 1.88, 95% CI: 0.77 to 4.62; p = 0.17) (Fig. 3c). No significant difference between patients with subclonal/diffuse loss of HLA-I and those with intact HLA-I expression was observed (HR 0.93, 95% CI: 0.21 to 4.24; p = 0.93) (Fig. 3d).

Fig. 3.

Unadjusted Kaplan Meier curves illustrating progression-free survival (PFS) outcomes for the 71 HPV16-associated OPC patients with both a valid mRNA analysis and immunohistochemistry results. Patients are stratified according to the expression of a HPV16-E5 mRNA, b HPV16-E5 productive mRNA, c EGFR, and d HLA-I; p values for the log-rank test used to compare the survival curves are also reported

Discussion

In the present study, we investigated the presence of two types of HPV16-E5 transcript in HPV16-driven OPCs. We also evaluated the expression of EGFR and HLA-I. Indeed, two of the main functions of HPV16-E5 are the recycling of activated EGFR on the cell surface and the prevention of MHC surface expression [6]. Expression levels of HPV16-E5 mRNA in OPCs show considerable variability across studies. Um et al. reported that HPV16 E5-encoded transcript levels in HPV16-positive OPCs are more variably expressed than E6-E7 gene transcript levels and that E5 transcript levels are correlated with EGFR expression [8]. They also noted that E5 and EGFR each independently predict relapse-free survival (RFS), though in opposing directions. Taberna et al. measured HPV16-E5 mRNA with no comparisons to E6-E7 transcripts, and similarly found HPV16-E5 expression in the majority of HPV16-positive OPCs [9]. More recently, Miyauchi et al. reported that HPV16-E5 and HLA expression levels may impact patient outcomes or responses to therapy [17]. Although HPV16-E5 expression level alone did not significantly affect disease-free survival (DFS) or overall survival (OS), when high versus low expression of HLA was included, OPCs with high E5 and low HLA-I expression had worse DFS, with a trend toward a decrease in OS. However, these studies all overlooked the polycistronic nature of HPV16-E5 mRNA and its extensive splicing processes which Schwartz et al. highlighted as critical to cellular responses during virus replication and transformation [18]. Remarkably, one of the several mRNA transcripts with coding potential for E5 protein can effectively be translated into the E5 protein. Nilsson et al. demonstrated that HPV16-E5 mRNAs are translated via the scanning mechanism, with upstream AUGs from E1 and E7 inhibiting E5 ORF translation. In particular, an early mRNA spliced from SD226 to SA3358 produced high levels of E5 protein [19]. Using specific primers encompassing this splicing site, we have previously detected HPV16-E5 specific transcript in cervical LSILs and HSILs [20]. Here, we utilised different primers encompassing the same splicing sites and demonstrated for the first time that this early E5-specific productive transcript is present in HPV16-positive OPCs. The low number of OPCs positive for this transcript may indicate that only a small subset of tumours retains the E5 protein, which is normally expressed in the early stages of cervical carcinogenesis [6]. In addition, in an on-going study of oral rinses from cancer-free individuals we did not detect this specific transcript in any of the HPV16 positive samples [21]. Results on the effect of E5 mRNA on EGFR up-regulation and HLA-I down-regulation are conflicting, perhaps due to the detection of all E5 transcripts, many of which cannot be translated into E5 protein, and to the differences in assessing EGFR up-regulation or HLA-I inhibition. Indeed, there is little consensus on methodologies to evaluate and quantitate EGFR or HLA-I on the cell surface. Our investigation into how E5 total transcripts and EGFR expression affect survival revealed a non-significant effect on PFS in opposite directions, with EGFR positivity resulting in a worse PFS, and E5 mRNA positivity in a better PFS. In contrast, patients harbouring E5 productive transcript showed a significantly worse PFS compared to those negative for this mRNA, although the CI for the Hazard Ratio was particularly wide due to the low number of the patients. It is also worth noting that in the group with the E5 productive transcript, three patients were lost to follow-up. Interestingly, no significant difference in EGFR expression was observed between HPV16-E5 mRNA positive and negative tumours, either considering the total or the specific productive transcripts, suggesting that EGFR expression in HPV16-driven OPCs can be modulated by mechanisms other than E5 expression. Similarly, no significant association was found between HLA-I expression loss and either E5 total or specific productive transcripts. Indeed, around 90% of the cases displayed HLA-I loss, indicating this is a common feature of HPV-associated OPCs.

Our study has some limitations. Survival analysis was only univariate. Kaplan–Meier curves were not adjusted for confounders such as age, sex, stage of the tumour, etc.; thus, the observed effects on the survival could be due to other characteristics with prognostic significance. The low number of OPCs positive for E5 specific productive transcript hampered stratification of patients by other clinically meaningful variables, namely age, stage, and therapy. E5 protein expression was not assessed since, to the best of our knowledge, antibodies for the immunohistochemical evaluation of E5 protein in tissue samples are not available. The possible integration of HPV16 genome in our set of OPCs was not investigated. Nonetheless, data on the physical status of the viral genome might provide valuable insights with regard to the presence of E5 productive transcripts.

Conclusions

The E5-specific productive transcript is associated with a statistically significant worse PFS in HPV-associated OPCs, but the very limited number of patients displaying this transcript does not allow us to draw firm conclusions. Further investigations on larger patient cohorts, which will likely require a multi-institutional collaborative study, are pivotal to confirm our findings regarding the influence of the E5 productive transcript on the survival of patients with HPV16-associated OPC.

Authors’ contributions

MGD, FP, AV designed the study, analyzed and interpreted the patient data and were major contributors in writing the manuscript. MB, FR, EM, EG, CB, RV performed the histological examinations and HPV PCR tests, FC, BP, RP enrolled OPC patients, FL, FC, RA, ILM, GDE, enrolled health subjects. AS, ACDF were contributor in writing the manuscript. All authors read and approved the final manuscript.

Funding

This work was partially funded by LILT 2020-21 Program and MAECI-Call for Joint Project Proposals Italy-Brazil # BR22GR03.

Data availability

The raw data are available from the corresponding author upon reasonable request and are uploaded into the GARR Box repository.

Declarations

Ethics approval and consent to participate

All enrolled patients signed an informed consent to the protocol approved by the IRCCS Regina Elena National Cancer Institute, Istituti Fisioterapici Ospitalieri, Institutional Review Board (RS 1443/20). The study was conducted in accordance with the Declaration of Helsinki.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.