Evaluation of type-specific antibodies to high risk-human papillomavirus (HPV) proteins in patients with oropharyngeal cancer

Oral Oncology 70 (2017) 43–50

Contents lists available at ScienceDirect

Oral Oncology journal homepage: www.elsevier.com/locate/oraloncology

Evaluation of type-specific antibodies to high risk-human papillomavirus (HPV) proteins in patients with oropharyngeal cancer q Martina A. Broglie a,⇑, Wolfram Jochum b, Angelika Michel c, Tim Waterboer c, Diana Foerbs b, René Schoenegg b, Sandro J. Stoeckli a, Michael Pawlita c, Dana Holzinger c a

Department of Otorhinolaryngology, Head and Neck Surgery, Kantonsspital St. Gallen, 9007 St. Gallen, Switzerland Institute of Pathology, Kantonsspital St. Gallen, 9007 St. Gallen, Switzerland Division of Molecular Diagnostics of Oncogenic Infections, Research Program Infections, Inflammation and Cancer, German Cancer Research Center (DKFZ), 69120 Heidelberg, Germany b c

a r t i c l e

i n f o

Article history: Received 19 January 2017 Received in revised form 20 April 2017 Accepted 17 May 2017 Available online 25 May 2017 Keywords: Oropharyngeal cancer Human papillomavirus Antibody response

a b s t r a c t Objectives: High risk human papillomavirus (HR-HPV) infection leads to a subgroup of oropharyngeal cancer (OPSCC) characterized by improved treatment response. However an universally accepted definition of an HR-HPV-attributable cancer is lacking. Methods: Detailed, type-specific HPV antibody responses were analyzed by multiplex serology in HRHPV-attributable OPSCC patients, defined by p16INK4A overexpression and HR-HPV DNA detection by PCR amplification and sequencing. Results: Fifty patients were prospectively enrolled. 26/50 (52%) tumor samples were positive for both p16INK4A expression and HR-HPV DNA (22 HPV16, 4 HPV33). Seropositivity was present in 26/26 HPVattributable OPSCC and one p16INK4A-positive/HPV DNA-negative case. The sensitivity and specificity to diagnose an HR-HPV-attributable tumor was 100% and 96%, respectively for anti-E6 reactivity, 82% and 100%, respectively for anti-E2 reactivity, and clearly lower for anti-E7, anti-E1, anti-E4 and anti-L1reactivity. 3 yr-overall (OS) and disease specific survival (DSS) was higher in patients with HR-HPVattributable tumors (OS 88% vs 64%, p = 0.02; DSS 90% vs 80%, p = 0.07) and seropositive patients (OS 88% vs 62%, p = 0.01; DSS 92% vs 78%, p = 0.05) than HR-HPV-negative or seronegative patients. Conclusions: Detection of HR-HPV type-specific antibodies highly correlated with HPV-attributable OPSCC and was associated with better survival. HR-HPV antibodies are promising diagnostic, prognostic and potentially screening markers in HR-HPV-attributable OPSCC. Ó 2017 Elsevier Ltd. All rights reserved.

Introduction Active infection with high risk human papillomavirus (HR-HPV) is an established risk factor for the development of oropharyngeal squamous cell carcinoma (OPSCC) [1,2–5]. Tumors with biologically active HR-HPV-infection (HR-HPV-attributable tumors) are known to have distinct biology and improved treatment response leading to better patient outcome [3,6–10]. The definition of an HR-HPV-attributable cancer is still a matter of debate. It is based on the detection of HR-HPV DNA by polymerase chain reaction (PCR) or in situ hybridization (ISH), overexpression of the cellq Study results were presented at the AHNS Meeting, July 26–30, 2014, New York, USA and at the AHNS Meeting, July 16–20, 2016, Seattle, USA. ⇑ Corresponding author at: Department of Otorhinolaryngology, Head and Neck Surgery, Kantonsspital St. Gallen, Rorschacherstrasse 95, 9007 St. Gallen, Switzerland. E-mail address: [email protected] (M.A. Broglie).

http://dx.doi.org/10.1016/j.oraloncology.2017.05.010 1368-8375/Ó 2017 Elsevier Ltd. All rights reserved.

cycle regulator p16INK4A as a surrogate marker [11], or a combination of these methods. Most of the PCR based assays are extremely sensitive, but are not specific for biologically relevant oncogenic infections [12]. For p16INK4A overexpression as a single marker, insufficient sensitivity [3,13,14] and specificity [3,14,15] have been demonstrated. Therefore, an algorithm using a combination of p16INK4A expression and HR-HPV DNA positivity has been proposed [15]. This algorithm achieved an accuracy of 98% using HR-HPV RNA detection as the gold standard [11]. Serum antibodies directed against HR-HPV proteins have been detected in patients with HR-HPV-attributable OPSCC [16–21]. In a large case–control study the agreement between HPV16 antibody detection and HR-HPV DNA-positive, p16INK4A-overexpressing OPSCC was 67% [17]. Moreover, seropositivity was associated with better survival [21,22]. Recent studies demonstrated strong antibody responses against the HPV16 E6 protein, but also against other early proteins and the

44

M.A. Broglie et al. / Oral Oncology 70 (2017) 43–50

late protein L1 in patients with HR-HPV-attributable OPSCC [21]. Antibodies were detected more than 10 years prior to OPSCC diagnosis [21]. Despite numerous therapeutic advances, the long term survival of patients with OPSCC, has remained modest, especially since these tumors are often diagnosed in advanced stages [23]. Therefore, screening strategies could possibly improve long-term outcomes. Whereas oral premalignant lesions such as leukoplakia are well described [23], HR-HPV-related premalignant oropharyngeal lesions have not been detected so far. A first attempt to evaluate an oropharyngeal Pap-test equivalent for HPV-positive cancers was not feasible [24], likely due to limitations in sampling the relevant tonsillar crypt epithelium. The question arises whether seroreactivity to HR-HPV proteins is a reliable biomarker in HR-HPV-attributable OPSCC and could possibly serve as a diagnostic and predictive as well as a possible marker for cancer screening. The aims of our study were (1) to analyse type-specific HPV antibody responses to early and late proteins in newly diagnosed HR-HPV attributable OPSCC, and (2) to assess a potential role of HPV antibodies as prognostic marker. Patients and methods Patients Fifty patients undergoing endoscopic evaluation at the Kantonsspital St. Gallen, Switzerland of a histologically confirmed and untreated OPSCC were enrolled. The study protocol had been approved by the local ethics committee and written informed consent was obtained from all patients prior to study entry.

p16INK4A immunostaining Two-mm thick FFPE tissue sections were deparaffinized followed by antigen retrieval (EDTA pH 9.0, 95 °C, 30 min). p16INK4A staining (clone E6H4, dilution 1:10, 30 min; Roche Diagnostics, Rotkreuz, Switzerland) was performed on a Leica BOND MAX instrument (Leica) using the Bond Polymer Refine detection kit (Leica). Positivity for p16INK4A overexpression was defined as 70% tumor cells with strong nuclear and cytoplasmic staining [31]. Statistical analysis HR-HPV-attributable tumors were defined as positive for both p16INK4A overexpression and presence of HR-HPV DNA [15]. Cohen’s Kappa was used to analyse the concordance between HR-HPV seropositivity and other parameters. Overall survival (OS) and disease specific survival (DSS) were estimated by Kaplan-Meier analysis and reported for the entire patient cohort and the subgroups. For comparisons between groups, the logrank test was used. Cox proportional hazard models were used to estimate the importance of HPV seropositivity, molecular HPV-status of the tumor tissue and different risk factors for outcome. Hazard Ratios (HR) and corresponding 95% confidence intervals (95% CI) were calculated using SAS 9.4. Further statistical analyses were evaluated using the R package, Sigma Plot 11.0 as well as SPSS 22 for windows. A two-tailed P value  0.05 was considered statistically significant. Results HR-HPV positivity of tumor tissues

Detection of antibodies against early and late proteins of HPV high risk types Blood samples were taken prior to treatment initiation. Serum was prepared, stored frozen and transferred to the German Cancer Research Center (DKFZ), Heidelberg, Germany for antibody determination by multiplex serology as reported previously [25,26]. Briefly, antigens were bacterially expressed. Affinity-purified viral proteins were fused to N-terminal Glutathione S-transferase. Serum samples were analyzed at 1:100 dilution for antibodies to the major capsid protein (L1), the early oncoproteins (E6, E7), and other early proteins (E1, E2, E4) of the following mucosal HR-HPV types: 16 and 18 (L1, E1, E2, E4, E6, and E7), and 31, 33, 35, 45, 52, and 58 (L1, E6, and E7); and of the non-carcinogenic mucosal HPV types 6 and 11 (L1, E6, and E7). Median fluorescence intensity (MFI) values were dichotomized as antibody positive or negative using predefined cutoff values [21]. Strong (MFI > 1000) positivity against a type specific HPV E6 protein or the simultaneous positivity against 3 or more different type-specific early proteins were used for the serological definition of HPV-attributable OPSCC. The definition of the cutoff values of >1000 MFI was based on earlier studies to allow for direct comparisons [21,27,28]. The accuracy of a simultaneous positivity against 3 or more different type-specific early proteins for the serological definition of HPVattributable OPSCC was recently demonstrated by Holzinger et al. [29]. HPV detection in the tumor tissue Tumor cells were micro-dissected from three to five 4 mm thick serial sections. Genomic DNA was prepared. HPV DNA testing was performed using the L1C1/L1C2 consensus primer set as previously described [30].

Fifty patients were enrolled. Twenty-nine OPSCC samples (58%) were HR-HPV DNA positive (HPV16 n = 25, HPV33 n = 4). In thirty samples (60%), carcinoma cells displayed strong p16INK4A protein expression. Twenty-six OPSCC (52%) displayed both HR-HPV DNA positivity and p16INK4A overexpression. These HR-HPV DNA/ p16INK4A double-positive tumors were categorized as HR-HPVattributable OPSCC. Three HPV16 DNA positive tumors were p16INK4A protein negative. Four p16INK4A-positive OPSCC were negative for HR-HPV DNA. Seventeen samples (34%) were negative for both markers. These tumors together with the p16INK4A or HR-HPV DNA single positive samples tumors were combined into the HRHPV-negative group (24/50, 48%). The concordance rate between HR-HPV DNA detection and p16INK4A expression was 86%. Patient characteristics The median age of patients was 62 years (range 30–81 years). Thirty-eight patients (75%) were male. HR-HPV-attributable tumors were predominantly located in the tonsils (73% vs 46%), equally distributed in the base of the tongue (27% vs 25%) and non-existent in other oropharyngeal subsites (0% vs 31%) compared to HR-HPV-negative patients. Furthermore, lower Tcategories (85% vs 54% T1-T3), higher prevalence of nodal involvement (85% vs 54% N1-3), lower frequencies for smoking history (31% vs 92%) or increased alcohol consumption (19% vs 58%) were revealed in patients with HR-HPV-attributable tumors compared to the HR-HPV-negative subgroup (Table 1). Serological findings HR-HPV-attributable OPSCC were defined by the presence of serum antibodies against a single HR-HPV E6 protein or 3 or more

M.A. Broglie et al. / Oral Oncology 70 (2017) 43–50 Table 1 Patients and tumor characteristics according to molecular HPV-status of the tumor. HR-HPV

Age, median (range) Gender Male Female History of tobacco smoking (>10 py) Yes No

*

Negative (n = 24)

Attributable (n = 26)

61.5 (46–80)

62.5 (30–81)

17 (71%) 7 (29%)

21 (81%) 5 (19%)

pvalue

0.4 0.3

0.0001* 22 (92%) 2 (8%)

8 (31%) 18 (69%) 0.007*

History of alcohol consumption (>3U) Yes No

14 (58%) 10 (42%)

5 (19%) 21 (81%)

Tumorsite Tonsils Base of tongue Other subsite

11 (46%) 6 (25%) 7 (31%)

19 (73%) 7 (27%) –

T-Category T1-T3 T4

13 (54%) 11 (46%)

22 (85%) 4 (15%)

N-Category N0 N1-N3

11 (46%) 13 (54%)

4 (15%) 22 (85%)

UICC-Stage I/II III/IV

6 (25%) 18 (75%)

4 (15%) 22 (85%)

Performed treatment Surgery only Surgery plus adjuvant RT/CRT Primary RT/CRT

7 (29%) 8 (33%) 9 (38%)

3 (11%) 14 (54%) 9 (35%)

3-year overall survival rate (%)

64

88

0.02*

3-year disease-specific survival rate (%)

80

90

0.07

0.02*

0.02*

0.02*

0.4

0.23

Statistically significant.

different HR-HPV early proteins. In 27/50 (54%) sera, antibodies with the predefined pattern typical for HR-HPV-attributable OPSCC were found. In more detail, strong HR-HPV type specific E6 seroreactivity (MFI > 1000) was present in all 22 HPV16-attributable OPSCC, and in all four HPV33-attributable OPSCC, but only in 1/4 of p16INK4A-positive/HR-HPV DNA-negative (seropositive to HPV16 and HPV33 E6), in 1/17 p16INK4A/HPV DNA double negative (seropositive to HPV31 E6) and in 0/3 HPV16 DNA-positive/ p16INK4A-negative OPSCC. Antibodies directed to at least three HR-HPV proteins were present in 26/26 HR-HPV-attributable OPSCC and in the one p16INK4Apositive/HR-HPV DNA-negative case (Fig. 1). The sensitivity and specificity of strong E6 seroreactivity (MFI > 1000) to diagnose an HR-HPV-attributable tumor were 100% (95% CI 85–100) and 96% (95% CI 80–99), respectively, followed by anti-E2 reactivity (82% and 100%, respectively), while the sensitivity and specificity of anti-E7 reactivity was clearly lower (70% and 88%, respectively). Moreover, seropositivity to the late protein L1 showed low sensitivity (50%, Table 2). The accuracy of the predefined cutoff values was evaluated for each antigen by calculating the sensitivity and specificity of different possible cutoff values and by creating receiver operating characteristic curves (Tables and ROC curves depicted in the supplement). Accuracy of the predefined cutoff values could be confirmed for all different antigens. The superiority of HPV16 E6, E1 and E2 seroreactivity over seropositivity to other HR-HPV pro-

45

teins as reliable diagnostic markers were confirmed by a value of the area under the curve of >0.9. Concordance between HR-HPV seropositivity (strong antibody response to E6 [MFI > 1000] or seropositive to at least 3 early proteins) and p16INK4A overexpression in the tumor as a single marker was higher (Cohens Kappa 0.88, 95% CI 0.73 – 1.00) than concordance between HR-HPV seropositivity and HR-HPV DNA positivity as a single marker (Cohens Kappa 0.80, 95% CI 0.63 – 0.95). Highest concordance was observed between seropositivity and HR-HPV DNA/ p16INK4A double positivity (Cohens Kappa 0.96, 95% CI 0.87 – 1.00). With regard to type-specific antibody responses the agreement between HPV16 DNA positivity in tumor tissue and HPV type 16 serum antibodies was high (Cohens Kappa 0.88, 95% CI 0.72 – 1.00). Similar results were achieved for HPV33 positive cases (Cohens Kappa 0.88, 95% CI 0.48 – 1.00). HPV33 E1, E2 and E4 proteins were not available for antibody analysis, but no potential cross-reactions with the closely related HPV16 E1, E2 and E4 proteins was observed. Among the 22 HPV16 E6 seropositive sera from HPV16attributable tumor patients, cross-reactivity was frequent with E6 from HPV33 (59%) and HPV58 (55%) but rare (<20%) with E6 from any of the other 7 HPV types analyzed. Among the 17 HPV16 E7 seropositive sera, cross-reactivity with HPV33 (65%) and HPV58 (41%) was even more frequent, but also with HPV31 (47%), HPV35 (47%) and HPV52 (33%). In contrast, among the HPV33-attributable tumor patients, 4 HPV33 E6 and 2 HPV33 E7 seropositives, none cross-reacted with the homologous HPV16 proteins but some E6 cross-reactivity was present with HPV52 and HPV58 and for E7 again with HPV58 but also HPV6, HPV35 and HPV45. Survival analysis The median follow-up time of the patient cohort was 42 months (range 6 – 64 months) with a minimal uneventful follow up time of two years. The 3-year OS and 3-year DSS rates were 77% and 86%, respectively. Patients with HR-HPV-attributable tumors had better 3-year OS and a trend to better 3-year DSS (OS 88% vs 64%, p = 0.02; DSS 90% vs 80%, p = 0.07, Fig. 2A+B). Using HR-HPV serology findings as criterion, seropositive patients had improved survival compared to seronegative patients (OS 88% vs. 62%, p = 0.01; DSS 92% vs. 78%, p = 0.05; Fig. 2C+D). P16INK4A overexpression and HR-HPV DNA positivity as single markers did not have an impact on OS and DSS (Fig. 2E+F). When analyzing the association between OPSCC risk factors and survival, univariate but not multivariate Cox proportional hazard regression revealed lower hazard ratios (HR) in OS and DSS for HR-HPV-attributable tumor and HR-HPV seropositivity. Other factors including p16INK4A overexpression as a single marker, T- or Ncategory, as well as smoking or alcohol consumption did not play a role for survival (Table 3). Discussion HR-HPV-attributable tumors are known to have a distinct tumor biology and improved treatment responses leading to better patient outcomes with up to 72% reduction in risk of death [6– 9,32]. Recent guidelines have suggested to include HR-HPV testing in the diagnostic work-up of OPSCC [33]. Various prospective studies are underway investigating the impact of HPV tumor positivity on treatment decisions. However, the adequate definition of an HRHPV-attributable cancer is still a matter of debate and universally accepted tests are lacking. PCR-based detection methods of HR-

46

M.A. Broglie et al. / Oral Oncology 70 (2017) 43–50

Fig. 1. Cluster analysis of molecular and serological HPV markers in OPSCC patients (heat-map). Patients were clustered by seropositivity (red) to the early and late antigens of HPV16 and HPV33 (E6, E7, E1, E2, E4, L1). Patients were further grouped by their serum positivity (sero+), HPV-attributable tumors (HPV+), HPV DNA positivity (DNA+) and p16INK4A overexpression (p16+). For E6, E7 and L1 antibodies, seroreactivity to type-concordant proteins is shown. HPV33 E1, E2 and E4 antibodies were not tested (displayed in grey). Seronegativity is shown in black. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Table 2 Sensitivity and specificity of serum antibodies to HPV proteins to diagnose a HPV-attributable tumor. HPV protein

E6 E7 E1* E2* E4* L1 *

Molecular HPV status of the tumor

Seropositive Seronegative Seropositive Seronegative Seropositive Seronegative Seropositive Seronegative Seropositive Seronegative Seropositive Seronegative

Positive

Negative

26 0 18 8 17 5 18 4 16 6 13 13

1 23 3 21 1 23 0 24 5 19 1 23

Sensitivity% (95% CI)

Specificity% (95% CI)

100 (85–100) 70 (52–87) 77 (57–90) 82 (62–93) 73 (52–87) 50 (31–69)

96 (80–99) 88 (69–96) 96 (80–99) 100 (86–100) 79 (60–91) 96 (80–99)

Only available for HPV16 attributable tumors.

HPV DNA are extremely sensitive and may lead to false-positive results [12]. Studies assessing sensitivity [3,13,14] and specificity [3,14,15] of p16INK4A overexpression as a single marker are scarce. According to Jordan et al. [34] p16INK4A immunohistochemistry (IHC) p16 IHC or HPV16 in situ hybridization (ISH) alone result in misclassification of approximately one fifth of tumors. The validation of an algorithm using a combination of p16INK4A IHC and HR-HPV DNA detection by PCR [15] demonstrated an accuracy of 98% compared to HR-HPV RNA detection serving as gold standard [11]. Similar results have been demonstrated by Schache et al. [13]. Since serum antibodies directed against HR-HPV proteins have been detected in patients with HR-HPV-attributable antibodies were found to be specific markers for HPV associate OPSCC

[17,18,35], the aim of our study was to prospectively analyse the diagnostic application of HPV type-specific antibody responses directed against early and late HPV oncoproteins. In this study the algorithm by Smeets et al. [15] for the molecular definition of an HR-HPV-attributable tumor was used since in contrast to HRHPV-RNA detection it is already part of clinical routine. To the best of our knowledge, this is the first study to demonstrate a high concordance between HR-HPV type specific seropositivity using multiplex serology [25,26] and p16INK4A-positive/HRHPV DNA-positive tumors. The prevalence of HPV16-attributable tumors in our cohort was 44% and HPV33-attributable tumors added further 8%. None of the patients with molecularly defined HR-HPV-attributable tumors were seronegative. The sensitivity

M.A. Broglie et al. / Oral Oncology 70 (2017) 43–50

47

Fig. 2. Kaplan-Meier curves of overall survival (OS) and disease specific survival (DSS)). A, OS stratified by HPV positivity; B, DSS stratified by HPV positivity; C, OS stratified by HPV antibodies; D, DSS stratified by HPV antibodies; E, OS stratified by p16INK4A overexpression; F, DSS stratified by p16INK4A overexpression.

and specificity of seropositivity to diagnose an HR-HPVattributable tumor was with 100% and 96%, respectively very high. Moreover a Cohens Kappa value of 0.96 demonstrated a high agreement between HR-HPV positivity in tumor tissue and HRHPV seropositivity in blood samples. The sensitivity, specificity and correlation of type-specific antibody responses in HPV16 and HPV33-attributable tumors was comparable to the results found in the entire cohort. Tachezy et al. [36] showed a moderate to good correlation between the presence of HR-HPV DNA in tumor tissue and the presence of antibodies to HPV16 E6 (Kappa value = 0.75 (0.61– 0.88)) and antibodies to HPV16 E7 (Kappa value = 0.56 (0.41– 0.71)) using an enzyme linked immunosorbent assay (ELISA) based

on GST-fusion proteins as used in multiplex serology here. They also detected HPV33 in 6% of tissue samples. Unfortunately, HPV33 type specific antibody responses were not outlined in their study. Smith et al. [37] only found 67% of HPV16 DNA tumorpositive cases to be HPV16 E6 and/or E7 seropositive using the same ELISA. Dahlstrom et al. [38] used in vitro translated HPV proteins as antigens also bound on beads and analyzed seropositivity also for additional viral proteins E1, E2, E4, E5, L1 and L2 in patients with HPV16 DNA positive tumors identified by PCR. In their study only 76/96 (79%) were HPV16 E6 seropositive and 90% (86/96) were positive for any early protein, while in the HPV DNAnegative subgroup 8/18 were E6 seropositive and 13/18 positive for any early protein, leading to a sensitivity and specificity of

48

M.A. Broglie et al. / Oral Oncology 70 (2017) 43–50

Table 3 Hazard Ratios (HR) for OS and DSS of different risk factors. Overall survival

Age Tumor size Lymph node involvement Alcohol consumption Smoking history p16 expression HR-HPV DNA HR-HPV status (p16high/DNA + ) HR-HPV serology Multivariate models* p16 expression HR-HPV DNA HR-HPV status (p16 high/DNA+) HR-HPV serology

Disease specific survival

HR

95%CI

p-value

HR

95%CI

p-value

T4 vs T1-T3 N + vs N0  3U vs < 3U  10py vs < 10py high vs low pos vs neg pos vs neg pos vs neg

1.03 2.53 1.72 1.52 2.90 0.42 0.37 0.27 0.25

0.95–1.06 0.92–7.00 0.48–6.11 0.55–4.20 0.82–10.30 0.15–1.18 0.13–1.04 0.09–0.85 0.08–0.79

0.9 0.07 0.4 0.4 0.1 0.1 0.06 0.03 0.02

0.98 2.93 3.07 2.86 4.93 0.38 0.34 0.25 0.23

0.92–1.05 0.73–11.71 0.38–25.10 0.68–12.00 0.60–40.36 0.09–1.60 0.08–1.41 0.05–1.24 0.05–1.15

0.6 0.1 0.3 0.2 0.1 0.2 0.1 0.09 0.07

high vs low pos vs neg pos vs neg pos vs neg

0.76 0.34 0.31 0.29

0.09–5.98 0.07–1.83 0.04–2.17 0.04–2.10

0.6 0.1 0.09 0.07

0.65 0.39 0.29 0.27

0.15–2.78 0.11–1.30 0.07–1.22 0.06–1.11

0.8 0.2 0.2 0.2

Bold values are statistically significant. * HR for models with p16, DNA, HPV, HPV serology were always adjusted by age, T and N stage, alcohol and tobacco consumption.

90% and 66% for E6 antibody responses and 87% and 66% for any E protein antibody response, respectively. In contrast to this study [38] demonstrating that some patients were negative for E6 but positive for E1 and E2 antibodies all of our patients with a HR-HPV-attributable tumor were E6 seropositive. In all but one of our only p16INK4A or HPV16 DNA positive cases antibodies to HPV16 E6 were found leading to a high sensitivity and specificity of 100% and 96% respectively. We speculate that the one HPV16 and HPV33 E6 positive, p16INK4A-positive/HPV DNA-negative case probably is an example for a false negative DNA testing since follow up sera demonstrated similar results and there was no evidence for an HR-HPV attributable tumor elsewhere such as in the genital tract. The remarkable difference between the sensitivity and specificity reported by Dahlstrom et al. [38] and our own results may be explained first by methodological reasons due to a rapid loss of E6 immunoreactivity after purification and succeeding false negative serological findings and second by an overestimated prevalence of HR-HPVattributable tumors based on PCR amplification of HR-HPV DNA as a single marker. In our study, antibodies directed to HR-HPV E6 proteins showed high levels of sensitivity and specificity (100% and 96%, respectively) to identify HR-HPV-attributable OPSCC. In contrast, antibodies to HR-HPV E2, E7, E1, E4 and L1 revealed lower sensitivity and specificity. These findings are similar to the results reported by Dahlstrom et al. [38] and an earlier study of the same group [39]. It illustrates that E7 alone is not a reliable biomarker. It further suggests that antibodies to regulatory (E proteins), but not late proteins are useful to diagnose HR-HPV-attributable tumors. The accuracy of the predefined cutoff values could be confirmed for all different antigens. Antibody titers were high in all HR-HPV-attributable tumor patients irrespective of the HR-HPV type. Furthermore, all but one patient were seropositive to other early proteins than E6. This is in line with the report by Zumbach et al. [40], Anderson et al. [41] and a recent publication by Lang Kuhs et al. [27] with a large cohort of healthy controls who described low HR-HPV E6/E7 antibody titers and absence of E6/E7 double positives cases among the few seropositive control patients. Similar findings have been demonstrated in cervical carcinoma patients [40,42,43]. In a recent study by Zhang et al. [44] an overall low HPV16 E6 seroprevalence of 3.2% in oral HPV16 DNA positive and 0.7% in oral HPV16 DNA negative participants was found compared to 0.6–1.2% in previous studies [21,27,41] suggesting a potential utility of HR-HPV E6 antibodies as screening marker.

The second aim of our study was to assess the impact of HRHPV antibody responses on survival and tumor control. It is a limitation of our monocentric study that we could include only 50 patients leading to statistical restrictions in subgroup analyses and insignificant results in multivariable analysis. Nevertheless, we were able to demonstrate that in univariate analysis a positive HR-HPV status of the tumor as well as the detection of HR-HPV antibodies positively influenced survival whereas p16INK4A overexpression and HR-HPV DNA as single markers did not have an impact on any survival estimate. In a study by Liang et al. [45], HPV16 E6 and E7 antibodies were superior to all other biomarkers in predicting survival in OPSCC patients and p16INK4A overexpression or HPV16 DNA, as individual biomarkers, were not predictive. E6 and/or E7 seropositive patients showed a significant reduction in overall mortality regardless of DNA status. According to Dahlstrom et al. [38] seropositivity to HPV16 E antibodies was a strong prognosticator for survival. They were able to separately analyse the impact of antibody responses to different oncoproteins and found E1, E2, and E6 antibody positivity to be strongly associated with improved survival. Long term survival of patients with OPSCC, has remained modest, especially due to advanced stages at the time of diagnosis [23]. Therefore, screening strategies would be mandatory, but in contrast to oral premalignant lesions such as leukoplakia [23], HRHPV-related premalignant oropharyngeal lesions have not been detected so far. The evaluation of an oropharyngeal Pap-test equivalent for HPV-positive cancers was not successful possibly due to sampling limitations [24]. Since HR-HPV antibodies especially to E6 highly correlated with the diagnosis of an HR-HPVattributable tumor and were detected more than 10 years prior to OPSCC diagnosis [21], seroreactivity to HR-HPV proteins could be a possible marker for cancer screening. In summary, detection of HR-HPV antibodies especially to E6 highly correlated with the diagnosis of an HR-HPV-attributable tumor as defined by HR-HPV DNA detection by PCR and p16INK4A overexpression, whereas the sensitivity and specificity of antibodies to other early proteins (E1, E2, E4 and E7) alone as well as to the late protein L1 was insufficient to serve as a diagnostic biomarker but could help to increase the specificity of the serological diagnosis. In our cohort a considerable proportion of HR-HPV-attributable tumors were driven by HPV33 instead of HPV16 for which reason the serological analysis of different HR-HPV types is mandatory. Moreover, our results underline the fact that HPV DNA detection by PCR or p16INK4A overexpression as single markers have limited accuracy for the definition of an HR-HPV-attributable tumor.

M.A. Broglie et al. / Oral Oncology 70 (2017) 43–50

The definition of an universally accepted and validated test would be mandatory as soon as treatment decisions are based on HRHPV-status of the tumor. Furthermore, we found that seropositivity could be a strong predictor for survival. Our study therefore suggests that antibodies to HR-HPV E6 and additional early proteins could serve as reliable markers for diagnosis and prognosis. Whether HR-HPV antibodies could be used as possible biomarkers for cancer screening in HR-HPV-attributable OPSCC has to be further elucidated. Conflict of interest statement There is no conflict of interest to declare. Funding support This study was supported by the CTU Kommission, Kantonsspital St. Gallen and the ‘‘Drei Könige” foundation of the Rotary Club Zürcher Unterland, Switzerland. DH and MP were supported by the EU through the HPV-AHEAD consortium (FP7-HEALTH-2011-282562). Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.oraloncology. 2017.05.010. References [1] Snijders PJ, Cromme FV, van den Brule AJ, Schrijnemakers HF, Snow GB, Meijer CJ, et al. Prevalence and expression of human papillomavirus in tonsillar carcinomas, indicating a possible viral etiology. Int J Cancer 1992;51 (6):845–50. [2] Andl T, Kahn T, Pfuhl A, Nicola T, Erber R, Conradt C, et al. Etiological involvement of oncogenic human papillomavirus in tonsillar squamous cell carcinomas lacking retinoblastoma cell cycle control. Can Res 1998;58 (1):5–13. [3] Holzinger D, Schmitt M, Dyckhoff G, Benner A, Pawlita M, Bosch FX. Viral RNA patterns and high viral load reliably define oropharynx carcinomas with active HPV16 involvement. Can Res 2012;72(19):4993–5003. [4] Castellsague X, Alemany L, Quer M, Halec G, Quiros B, Tous S, et al. HPV involvement in head and neck cancers: comprehensive assessment of biomarkers in 3680 patients. J Natl Cancer Inst 2016;108(6). [5] Gillison ML, Koch WM, Capone RB, Spafford M, Westra WH, Wu L, et al. Evidence for a causal association between human papillomavirus and a subset of head and neck cancers. J Natl Cancer Inst 2000;92(9):709–20. [6] Vermorken JB, Psyrri A, Mesia R, Peyrade F, Beier F, de Blas B, et al. Impact of tumor HPV status on outcome in patients with recurrent and/or metastatic squamous cell carcinoma of the head and neck receiving chemotherapy with or without cetuximab: retrospective analysis of the phase III EXTREME trial. Ann Oncol 2014;25(4):801–7. [7] Ang KK, Harris J, Wheeler R, Weber R, Rosenthal DI, Nguyen-Tan PF, et al. Human papillomavirus and survival of patients with oropharyngeal cancer. N Engl J Med 2010;363(1):24–35. [8] Lassen P, Eriksen JG, Hamilton-Dutoit S, Tramm T, Alsner J, Overgaard J. Effect of HPV-associated p16INK4A expression on response to radiotherapy and survival in squamous cell carcinoma of the head and neck. J Clin Oncol 2009;27 (12):1992–8. [9] Fakhry C, Westra WH, Li S, Cmelak A, Ridge JA, Pinto H, et al. Improved survival of patients with human papillomavirus-positive head and neck squamous cell carcinoma in a prospective clinical trial. J Natl Cancer Inst 2008;100(4):261–9. [10] Lindel K, Beer KT, Laissue J, Greiner RH, Aebersold DM. Human papillomavirus positive squamous cell carcinoma of the oropharynx: a radiosensitive subgroup of head and neck carcinoma. Cancer 2001;92(4):805–13. [11] Rietbergen MM, Leemans CR, Bloemena E, Heideman DA, Braakhuis BJ, Hesselink AT, et al. Increasing prevalence rates of HPV attributable oropharyngeal squamous cell carcinomas in the Netherlands as assessed by a validated test algorithm. Int J Cancer 2013;132(7):1565–71. [12] Snijders PJ, van den Brule AJ, Meijer CJ. The clinical relevance of human papillomavirus testing: relationship between analytical and clinical sensitivity. J Pathol 2003;201(1):1–6. [13] Schache AG, Liloglou T, Risk JM, Filia A, Jones TM, Sheard J, et al. Evaluation of human papilloma virus diagnostic testing in oropharyngeal squamous cell carcinoma: sensitivity, specificity, and prognostic discrimination. Clin Cancer Res 2011;17(19):6262–71.

49

[14] Hoffmann M, Tribius S, Quabius ES, Henry H, Pfannenschmidt S, Burkhardt C, et al. HPV DNA, E6⁄I-mRNA expression and p16INK4A immunohistochemistry in head and neck cancer - how valid is p16INK4A as surrogate marker? Cancer Lett 2012;323(1):88–96. [15] Smeets SJ, Hesselink AT, Speel EJ, Haesevoets A, Snijders PJ, Pawlita M, et al. A novel algorithm for reliable detection of human papillomavirus in paraffin embedded head and neck cancer specimen. Int J Cancer 2007;121 (11):2465–72. [16] Herrero R, Castellsague X, Pawlita M, Lissowska J, Kee F, Balaram P, et al. Human papillomavirus and oral cancer: the International Agency for Research on Cancer multicenter study. J Natl Cancer Inst 2003;95(23):1772–83. [17] Anantharaman D, Gheit T, Waterboer T, Abedi-Ardekani B, Carreira C, McKayChopin S, et al. Human papillomavirus infections and upper aero-digestive tract cancers: the ARCAGE study. J Natl Cancer Inst 2013;105(8):536–45. [18] Smith EM, Ritchie JM, Pawlita M, Rubenstein LM, Haugen TH, Turek LP, et al. Human papillomavirus seropositivity and risks of head and neck cancer. Int J Cancer 2007;120(4):825–32. [19] Zumbach K, Hoffmann M, Kahn T, Bosch F, Gottschlich S, Gorogh T, et al. Antibodies against oncoproteins E6 and E7 of human papillomavirus types 16 and 18 in patients with head-and-neck squamous-cell carcinoma. Int J Cancer 2000;85(6):815–8. [20] D’Souza G, Kreimer AR, Viscidi R, Pawlita M, Fakhry C, Koch WM, et al. Casecontrol study of human papillomavirus and oropharyngeal cancer. N Engl J Med 2007;356(19):1944–56. [21] Kreimer AR, Johansson M, Waterboer T, Kaaks R, Chang-Claude J, Drogen D, et al. Evaluation of human papillomavirus antibodies and risk of subsequent head and neck cancer. J Clin Oncol 2013;31(21):2708–15. [22] Smith EM, Rubenstein LM, Ritchie JM, Lee JH, Haugen TH, Hamsikova E, et al. Does pretreatment seropositivity to human papillomavirus have prognostic significance for head and neck cancers? Cancer Epidemiol Biomark Prev 2008;17(8):2087–96. [23] Lingen MW. Screening for oral premalignancy and cancer: what platform and which biomarkers? Cancer Prev Res (Phila) 2010;3(9):1056–9. [24] Fakhry C, Rosenthal BT, Clark DP, Gillison ML. Associations between oral HPV16 infection and cytopathology: evaluation of an oropharyngeal ‘‘pap-test equivalent” in high-risk populations. Cancer Prev Res (Phila) 2011;4 (9):1378–84. [25] Waterboer T, Sehr P, Michael KM, Franceschi S, Nieland JD, Joos TO, et al. Multiplex human papillomavirus serology based on in situ-purified glutathione s-transferase fusion proteins. Clin Chem 2005;51(10):1845–53. [26] Waterboer T, Sehr P, Pawlita M. Suppression of non-specific binding in serological Luminex assays. J Immunol Methods 2006;309(1–2):200–4. [27] Lang Kuhs KA, Anantharaman D, Waterboer T, Johansson M, Brennan P, Michel A, et al. Human papillomavirus 16 e6 antibodies in individuals without diagnosed cancer: a pooled analysis. Cancer Epidemiol Biomark Prev 2015;24 (4):683–9. [28] Zhang Y, Waterboer T, Haddad RI, Miles BA, Wentz A, Gross ND, et al. Human papillomavirus (HPV) 16 antibodies at diagnosis of HPV-related oropharyngeal cancer and antibody trajectories after treatment. Oral Oncol 2017;67:77–82. [29] Holzinger D, Wichmann G, Baboci L, Michel A, Hofler D, Wiesenfarth M, et al. Sensitivity and specificity of antibodies against HPV16 E6 and other early proteins for the detection of HPV16-driven oropharyngeal squamous cell carcinoma. Int J Cancer 2017. [30] Broglie MA, Jochum W, Forbs D, Schonegg R, Stoeckli SJ. Brush cytology for the detection of high-risk HPV infection in oropharyngeal squamous cell carcinoma. Cancer Cytopathol 2015. [31] Singhi AD, Westra WH. Comparison of human papillomavirus in situ hybridization and p16 immunohistochemistry in the detection of human papillomavirus-associated head and neck cancer based on a prospective clinical experience. Cancer 2010;116(9):2166–73. [32] Posner MR, Lorch JH, Goloubeva O, Tan M, Schumaker LM, Sarlis NJ, et al. Survival and human papillomavirus in oropharynx cancer in TAX 324: a subset analysis from an international phase III trial. Ann Oncol 2011;22(5):1071–7. [33] Pfister DG, Spencer S, Brizel DM, Burtness B, Busse PM, Caudell JJ, et al. Head and neck cancers, version 1.2015. J Natl Compr Canc Netw 2015;13(7):847–55. quiz 56. [34] Jordan RC, Lingen MW, Perez-Ordonez B, He X, Pickard R, Koluder M, et al. Validation of methods for oropharyngeal cancer HPV status determination in US cooperative group trials. Am J Surg Pathol 2012;36(7):945–54. [35] Furniss CS, McClean MD, Smith JF, Bryan J, Applebaum KM, Nelson HH, et al. Human papillomavirus 6 seropositivity is associated with risk of head and neck squamous cell carcinoma, independent of tobacco and alcohol use. Ann Oncol 2009;20(3):534–41. [36] Tachezy R, Klozar J, Rubenstein L, Smith E, Salakova M, Smahelova J, et al. Demographic and risk factors in patients with head and neck tumors. J Med Virol 2009;81(5):878–87. [37] Smith EM, Pawlita M, Rubenstein LM, Haugen TH, Hamsikova E, Turek LP. Risk factors and survival by HPV-16 E6 and E7 antibody status in human papillomavirus positive head and neck cancer. Int J Cancer 2010;127(1):111–7. [38] Dahlstrom KR, Anderson KS, Cheng JN, Chowell D, Li G, Posner M, et al. HPV serum antibodies as predictors of survival and disease progression in patients with HPV-positive squamous cell carcinoma of the oropharynx. Clin Cancer Res 2015;21(12):2861–9. [39] Anderson KS, Wong J, D’Souza G, Riemer AB, Lorch J, Haddad R, et al. Serum antibodies to the HPV16 proteome as biomarkers for head and neck cancer. Br J Cancer 2011;104(12):1896–905.

50

M.A. Broglie et al. / Oral Oncology 70 (2017) 43–50

[40] Zumbach K, Kisseljov F, Sacharova O, Shaichaev G, Semjonova L, Pavlova L, et al. Antibodies against oncoproteins E6 and E7 of human papillomavirus types 16 and 18 in cervical-carcinoma patients from Russia. Int J Cancer 2000;85(3):313–8. [41] Anderson KS, Dahlstrom KR, Cheng JN, Alam R, Li G, Wei Q, et al. HPV16 antibodies as risk factors for oropharyngeal cancer and their association with tumor HPV and smoking status. Oral Oncol 2015;51(7):662–7. [42] Meschede W, Zumbach K, Braspenning J, Scheffner M, Benitez-Bribiesca L, Luande J, et al. Antibodies against early proteins of human papillomaviruses as diagnostic markers for invasive cervical cancer. J Clin Microbiol 1998;36 (2):475–80.

[43] Combes JD, Pawlita M, Waterboer T, Hammouda D, Rajkumar T, Vanhems P, et al. Antibodies against high-risk human papillomavirus proteins as markers for invasive cervical cancer. Int J Cancer 2014;135(10):2453–61. [44] Zhang Y, Waterboer T, Pawlita M, Sugar E, Minkoff H, Cranston RD, et al. Human Papillomavirus (HPV) 16 E6 seropositivity is elevated in subjects with oral HPV16 infection. Cancer Epidemiol 2016;43:30–4. [45] Liang C, Marsit CJ, McClean MD, Nelson HH, Christensen BC, Haddad RI, et al. Biomarkers of HPV in head and neck squamous cell carcinoma. Can Res 2012;72(19):5004–13.