Introduction

Laryngeal squamous cell carcinoma (LSCC) is one of the most common type of head and neck cancer and, although tobacco smoking and alcohol abuse are the principal risk factors, some viral infections are implicated [1, 2]. Human papillomavirus (HPV) is a group of oncoviruses that have been widely associated with the development and clinical prognosis of LSCC tumors [3]. According to their oncogenic potential, HPV are categorized into low-risk (LR-HPV) or high-risk (HR-HPV) genotypes [4]. LR-HPV are associated with warts or skin papillomas, whereas HR-HPV are directly associated with the development of invasive carcinomas [5, 6]. Prevalence studies have reported that HPV-16 and -18 are the most common HR-HPV genotypes found in LSCC although, other HR-HPV, such as 31, 33, 35, 39, 45, 51, 52, 56, 58, and 59 are also important [7, 8]. Epstein–Barr virus (EBV) is strongly associated with onset and progression of several types of epithelial cell tumors such as nasopharyngeal carcinoma or gastric adenocarcinoma [9, 10]. The role of EBV in LSCC is controversial since difficulties have been reported in detecting its presence in this kind of cancer or it has shown a low prevalence [11]. Both HPV and EBV produce oncoproteins that have the potential to induce processes of carcinogenesis and tumor progression [12]. E7 is an oncoprotein coded by HPV that promotes retinoblastoma protein (pRb) degradation by the ubiquitin–proteasome pathway [13]. Degradation of pRb consequently induces the overexpression of the tumor suppressor protein p16INK4a, which is essential in cell cycle progression [14]. Latent membrane protein 1 (LMP1) is one of the major EBV oncoproteins that inactivates transcription factors essential for the expression of p16INK4a [15]. The alteration of p16 INK4a expression leads to hyperphosphorylation of pRb that promotes uncontrolled cellular proliferation [16]. High levels of p16INK4a expression have been suggested as a surrogate marker of HPV and EBV infection in pharyngeal, nasopharyngeal, and oropharynx tumors, whereas in laryngeal tumors are still controversial; therefore, pRb expression may be considered as an additional biomarker, particularly in tumors with infection by HPV and EBV [12, 17,18,19]. This work is aimed to compare the expression of pRb and p16INK4a as possible biomarkers in tumor tissues with and without infection by EBV or different genotypes of HPV from patients with LSCC attending a third-level care referral hospital.

Materials and methods

Samples

Formalin-fixed and paraffin-embedded (FFPE) tissue specimens from patients with LSCC were included in a retrospective study. Samples were taken from a previous study evaluating the HPV and EBV prevalence in patients with LSCC [20]. All patients had undergone surgical resection of the laryngeal tumor between 2012 and 2015 at the Unidad Medica de Alta Especialidad (UMAE) No. 25, a tertiary referral hospital of the Instituto Mexicano del Seguro Social (IMSS) located in northeastern Mexico. The information obtained from clinical records included age, sex, history of alcohol and tobacco consumption, and laryngeal subsite location of the tumor (subglottic, glottic, or supraglottic). Histological and pathological grading of tumors was made using the Broder’s classification [21]. The National Committee of Scientific Research of the IMSS granted ethical approval to carry out the study (R-2014-785-055). Individual informed consent was not obtained as specimens were retrospectively collected.

HPV genotyping and EBV detection

Tumor sections from each FFPE sample were carefully punched off and conducted to a DNA isolation using the NucleoSpin DNA FFPE (Macherey–Nagel, Düren, Germany). The presence and genotype of HPV were evaluated using the INNO-LiPA HPV Genotyping Kit Extra II Amp (Innogenetics, Gent, Belgium), a line probe assay designed for the identification of 32 genotypes of the HPV, including thirteen HR-HPV (HPV-16, HPV-18, HPV-31, HPV-33, HPV-35, HPV-39, HPV-45, HPV-51, HPV-52, HPV-56, HPV-58, HPV-59, and HPV-68), six possible HR-HPV (HPV-26, HPV-53, HPV-66, HPV-70, HPV-73, and HPV-82), nine LR-HPV (HPV-6, HPV-11, HPV-40, HPV-42, HPV-43, HPV-44, HPV-54, HPV-61, and HPV-81), and other four not yet classified HPV genotypes (HPV-62, HPV-67, HPV-83, and HPV-89) [22]. To detect DNA of EBV, a 69 bp fragment flanking the 679–748 position of the viral LMP2 glycoprotein was amplified using the primers 5′-AGC TGT AAC TGT GGT TTC CAT GAC-3′ and 5′-GCC CCC TGG CGA AGA G-3′ with the probe 6FAM5′-CTG CTG CTA CTG GCT TTC GTC CTC TGG-3′TAMRA. All the reactions for EBV detection were performed after the HPV genotyping assays using the TaqMan Universal PCR MasterMix (Applied Biosystems, Foster City CA, USA) in the Light Cycler (Roche). Synthetic DNA fragment (gBlocks Gene Fragment, IDT) containing the 69 bp sequence of the amplicon of EBV was used as amplification control and to evaluate the efficiency and detection limits of the assay.

p16INK4a and pRb expression

Immunohistochemical (IHC) staining method for p16INK4a and pRb was performed on 5-µm-thick FFPE tissue cross-sections using the mouse and rabbit specific HRP/DAB (ABC) detection IHC kit (Abcam, Cambridge, UK) with the monoclonal anti-CDKN2A/p16INK4a antibody (clone EPR1473, ab108349, dilution 1:100; Abcam, Cambridge, UK) and the monoclonal anti-Rb antibody (clone EPR17512, ab181616, dilution 1:250; Abcam, Cambridge, UK). Hematoxylin–eosin (HE) stains were also prepared for histological evaluation of tumors. Two independent pathologists who were blinded to the clinical information and to the HPV or EBV status of the samples evaluated all the IHC and HE stains. Evaluations were made based on staining intensity as other previous reports have done. Specimens were considered positive for p16INK4a expression if they showed ≥ 75% of stained cells, whereas for pRb, ≥ 25% of stained cells were interpreted as a positive expression [23,24,25,26].

Statistical analysis

We used the Mann–Whitney U test, Pearson's X2 test, and Fisher's exact test. A p value < 0.05 was considered statistically significant. All analyzes were performed using the IBM SPSS Statistics program (Version 20, SPSS. Inc., Chicago, USA).

Results

One hundred and three FFPE specimens from the previously published study with sufficient tumor tissue for IHC were eligible for the present work. The age range of the included patients was 40 to 89 years, with a median of 64, and most were 60 years or older (73.8%). Most samples were from men (96.1%), and tobacco smoking was common (98.0%). In contrast, alcohol intake frequency was low (6.8%). The glottis subsite was the most frequent location of tumors (49.5%). We had 57 (55.3%) tumor samples positive for HPV DNA, and 42 (40.8%) of these had at least one HR-HPV genotype. EBV DNA was present in 28 (27.2%) samples. Of these, 16 (15.5%) were in co-infection with HPV, 11 (10.7%) corresponded to co-infections with at least one HR-HPV, and 12 (11.7%) cases had only EBV infection (Table 1). Clinical characteristics of patients with LSCC tumor samples positive for at least one HR-HPV are summarized in Table 2. HPV-52 was the most frequently HR-HPV genotype, with 34 cases corresponding to 59.6% of all the HPV-positive samples. Immunohistochemistry were assessed considering a positive expression of p16INK4a if ≥ 75% of tumor cells were stained and for pRb if ≥ 25% of stained cells were observed (Fig. 1).

Table 1 Correlation of p16INK4a and pRb expression with clinicopathological characteristics of 103 patients with laryngeal squamous cell carcinoma
Full size table
Table 2 Clinical data and status of p16INK4a and pRb expression in 42 laryngeal squamous cell carcinoma patients positive for at least one high-risk human papillomavirus
Full size table
Fig. 1
figure 1

Representative examples of immunohistochemical (IHC) staining of tumors from patients with laryngeal squamous cell carcinoma. a Sample of patient No. 175 without infection of HPV or EBV; b sample of patient No. 046 that was positive for the 11 and 16 HR-HPV and negative for EBV; and c sample of patient No. 021 that was negative HPV and positive for EBV (× 40 microscopic magnification)

Full size image

Of the 103 tumor samples, p16INK4a expression was positive in 55 (53.4%). According to the surgical source of the sample, biopsies (64.5%) had a higher frequency of expression of this protein than laryngectomies (p = 0.009 by the Pearson's chi-squared test). Samples cataloged as Grade III by Broder's classification showed a higher frequency of p16INK4a expression (80.9%) compared with those of grade I (50.0%) and II (42.1%) (p < 0.05 by the Mann–Whitney U test). Only one of the seven alcohol consumers showed p16INK4a expression (14.3%) (p = 0.048 by the Fisher's Exact Test). No significant difference was observed when comparing p16INK4a expression with the presence of HPV or EBV (p > 0.05 by the Fisher's Exact Test). Regarding pRb expression, 78 (75.7%) samples were positive. Specimen’s obtained from biopsies (84.2%) showed a higher frequency of pRb expression than laryngectomy samples (p = 0.025 by the Pearson's chi-squared test). The frequency of pRb expression was higher in HPV negative samples (87.0%) (p = 0.021 by the Fisher's Exact Test). A higher frequency of pRb expression was also observed in the group of HR-HPV negative samples (85.2%) (p = 0.010 by the Fisher's Exact Test). No significant differences (p > 0.05) were observed in the expression of p16INK4a or pRb by age and sex of patients, laryngeal tumor subsite, and tobacco consumption habits of the patients (Table 1).

Discussion

To assess the expression of p16 INK4a and pRb in LSCC specimens with and without infection by EBV or different genotypes of HPV, 103 FFPE samples retrospectively collected from patients were evaluated by IHC analysis. The epidemiological characteristics of the studied sample are consistent with those previously described for this disease [11, 19, 27,28,29,30,31,32,33]. p16INK4a status has been well characterized in HPV-positive head and neck tumors such oropharyngeal carcinoma and has been suggested as a surrogate marker for infection of HPV and progression, but this has not been definitively established for LSCC [34, 35]. In this study, 55 (53.4%) of 103 patients were positive for p16INK4a. Tong et al. found that 115 (54%) of 211 LSCC specimens were positive for p16INK4a [36]. Elhadj et al. show that the expression positive of p16INK4a was found in 36 (51.43%) of 70 LSCC cases [37]. Expression of p16INK4a might be affected by genetic or epigenetic mechanisms [38]. Some studies suggest hypermethylation may represent an early event in carcinogenesis of different neoplasms, such as endometrial cancer [39]. In LSCC, expression level of p16INK4a is significantly reduced and hypermethylation has been shown to be a common mechanism causing this downregulation [40]. We found that 80.9% of our specimens classified as grade III presented a higher p16INK4a expression. In contrast to our results, Tong and et al. found that p16INK4a expression was observed more commonly in well-differentiated samples [36]. In our study, only 7 (6.8%) patients were alcohol consumers, and one of them expressed p16INK4a. Although this low number of cases precludes reliable statistical comparison, most alcohol consumers were significantly negative for p16INK4a expression (p = 0.048). Elhadj et al. found no association between alcohol consumption and p16INK4a expression when evaluating 30 alcohol-consuming patients [37]. Otherwise, we found that 32 (56.1%) of the fifty-five positive samples for p16 INK4a were also positive for HPV infection. Stephen et al. reported the expression of p16INK4a in 21 (26.0%) of 80 patients, furthermore, 12 (57.0%) of these were positive for HPV infection [34]. Hernandez et al. observed p16INK4a expression in 8 (7.9%) of 101 cases, and only 2 (25.0%) of these were also HPV DNA positive [41]. Sanchez et al. found that 48 (39.0%) of 123 samples were positive for p16INK4a expression, and 14 (29.2%) of these were positive for HPV infection [42]. Dogantemur et al. observed p16INK4a expression in 18 (20.0%) of 90 cases and only 6 (33.3%) of these were also positive for DNA of HPV [43]. The fact that no association between the expression of p16INK4a and HPV infection was observed could support the suggestion that this protein may not be reliable as a surrogate marker for HPV infection [44].

Modifications of the p16INK4a and pRb expression have been delineated in several types of human cancers [45, 46]. We found that pRb expression was present in 78 (75.7%) of 103 specimens. Morshed et al. found that 90 (69%) of 130 patients with LSCC had expression of pRb [47]. Krecicki et al. reported that 51(88%) of 58 samples were positive for pRb expression [48]. Some studies suggest that high pRb expression may be due to different factors. Soares et al. suggested that it can be due to an increase in the proportions of proliferating cells, which is supported by the fact that the hyperphosphorylated pRb inactive form increases during the G2/M phases [49]. IHC evaluation using antibodies against phosphorylated and non-phosphorylated pRb forms could solve this. As pRb is a negative regulator of p16INK4a, its inactivation results in overexpression of p16INK4a. Thus, a positive test for HPV combined with p16INK4a expression has been described as evidence of biologically relevant infection [50, 51]. The frequency of pRb expression was significantly higher in the HPV-negative specimens; we found 40 (87.0%) with positive pRb expression in 46 HPV-negative tumors. However, the clinical and prognostic significance remains poorly described in LSCC. Soares et al. found a significant difference between pRb expression and HPV-positive (9/11, 81.8%) and HPV-negative (15/22, 68.2%) oral squamous cell carcinoma (OSCC) samples [49]. Shaikh et al. demonstrated that pRb expression was predominant in 51 (77.3%) of 66 HPV-negative head and neck cancers cases, but they did not find significant differences [51]. We also observed a significantly higher frequency of pRb expression in HR-HPV negative specimens. The pRb expression was present in 51 (85.2%) of 61 HR-HPV negative samples. By contrast, Nemes et al. reported no statistical difference in 37 (84%) of 44 HR-HPV negative OSCC samples with pRb expression [52]. Various studies have shown that HPV-positive cancers generally exhibit decreased expression of pRb [53]. In addition, the oncoprotein E7 of HPV is known to participate in the degradation of pRb through a ubiquitin–proteasome and other pathways [54]. Thus, the absence of HPV infection could be related to the pRb expression in our samples.

Conclusion

We detected p16INK4a in half of LSCC cases, and in half of these, HPV was also detected. However, the proportion of samples with p16INK4a expression was similar between those with HPV and EBV. Although the proportion of LSCC samples with p16INK4a expression was higher than reported in several previous studies, our findings suggest little value of this marker as a reliable surrogate for identifying HPV in laryngeal cancer. On the other hand, as in previous studies, most of our LSCC samples had pRb expression, this being more frequent in tumors without the presence of HPV, even in those positive for high-risk HPV. Therefore, pRb expression seems to indicate HPV negativity, which could justify not performing other molecular tests for HPV detection in this type of cases, as previously suggested. However, further studies with a larger number of cases, including controls without laryngeal cancer, involving other tumor suppression pathways or other molecular markers, are necessary to determine the real role of p16INK4a and pRb in LSCC.