Purpose: Patients with human papillomavirus (HPV)-containing oropharyngeal squamous cell carcinomas (OSCC) have a better prognosis than patients with HPV-negative OSCC. This may be attributed to different genetic pathways promoting cancer. Experimental Design: We used comparative genomic hybridization to identify critical genetic changes in 60 selected OSCC, 28 of which were associated with HPV-16 as determined by HPV-specific PCR and fluorescence in situ hybridization analysis and positive p16^INK4A immunostaining. The results were correlated with HPV status and clinical data from patients. Results: Two thirds of OSCC harbored gain at 3q26.3-qter irrespective of HPV status. In HPV-negative tumors this alteration was associated with advanced tumor stage (P = 0.013). In comparison with HPV-related OSCC, the HPV-negative tumors harbored: (a) a higher number of chromosomal alterations and amplifications (P = 0.03 and 0.039, respectively); (b) significantly more losses at 3p, 5q, 9p, 15q, and 18q, and gains/amplifications at 11q13 (P = 0.002, 0.03; <0.001, 0.02, 0.004, and 0.001, respectively); and (c) less often 16q losses and Xp gains (P = 0.02 and 0.03). Survival analysis revealed a significantly better disease-free survival for HPV-related OSCC (P = 0.02), whereas chromosome amplification was an unfavorable prognostic indicator for disease-free and overall survival (P = 0.01 and 0.05, respectively). Interestingly, 16q loss, predominantly identified in HPV-related OSCC, was a strong indicator of favorable outcome (overall survival, P = 0.008; disease-free survival, P = 0.01) and none of these patients had a tumor recurrence. Conclusions: Genetic signatures of HPV-related and HPV-unrelated OSCC are different and most likely underlie differences in tumor development and progression. In addition, distinct chromosomal alterations have prognostic significance.

B lymphocytes are important players in immune responses to cancer. However, their composition and function in head and neck squamous cell carcinoma (HNSCC) has not been well described. Here, we analyzed B cell subsets in HNSCC (n = 38), non-cancerous mucosa (n = 14) and peripheral blood from HNSCC patients (n = 38) and healthy controls (n = 20) by flow cytometry. Intratumoral B cells contained high percentages of activated (CD86⁺), antigen-presenting (CD86⁺/CD21⁻) and memory B cells (IgD⁻/CD27⁺). T follicular helper cells (CD4⁺/CXCR5⁺/CD45RA⁻/CCR7⁻) as key components of tertiary lymphoid structures and plasma cells made up high percentages of the lymphocyte infiltrate. Percentages of regulatory B cell varied depending on the regulatory phenotype. Analysis of humoral immune responses against 23 tumor-associated antigens (TAA) showed reactivity against at least one antigen in 56% of HNSCC patients. Reactivity was less frequent in human papillomavirus associated (HPV⁺) patients and healthy controls compared to HPV negative (HPV⁻) HNSCC. Likewise, patients with early stage HNSCC or MHC-I loss on tumor cells had low TAA responses. Patients with TAA responses showed CD4⁺ dominated T cell infiltration compared to mainly CD8⁺ T cells in tumors without detected TAA response. To summarize, our data demonstrates different immune infiltration patterns in relation to serological TAA response detection and the presence of B cell subpopulations in HNSCC that can engage in tumor promoting and antitumor activity. In view of increasing use of immunotherapeutic approaches, it will be important to include B cells into comprehensive phenotypic and functional analyses of tumor-associated lymphocytes.

Characterization of tumor-associated T-lymphocyte subsets and immune checkpoint molecules in head and neck squamous cell carcinoma Characterization of tumor-associated T-lymphocyte subsets and immune checkpoint molecules in head and neck squamous cell carcinoma

HPV‐related HNSCC generally have a better prognosis than HPV‐negative HNSCC. However, a subgroup of HPV‐positive tumors with poor prognosis has been recognized, particularly related to smoking, EGFR overexpression and chromosomal instability. Viral integration into the host genome might contribute to carcinogenesis, as is shown for cervical carcinomas. Therefore, all HPV16‐positive HNSCC cell lines currently available have been carefully analyzed for viral and host genome parameters. The viral integration status, viral load, viral gene expression and the presence of aneusomies was evaluated in the cell lines UD‐SCC‐2, UM‐SCC‐047, UM‐SCC‐104, UPCI:SCC090, UPCI:SCC152, UPCI:SCC154 and 93VU147T. HPV integration was examined using FISH, APOT‐PCR and DIPS‐PCR. Viral load and the expression of the viral genes E2, E6 and E7 were determined via quantitative PCR. All cell lines showed integration‐specific staining patterns and signals indicating transcriptional activity using FISH. APOT‐ and DIPS‐PCR identified integration‐derived fusion products in six cell lines and only episomal products for UM‐SCC‐104. Despite the observed differences in viral load and the number of viral integration sites, this did not relate to the identified viral oncogene expression. Furthermore, cell lines exhibited EGFR expression and aneusomy (except UPCI:SCC154). In conclusion, all HPV16‐positive HNSCC cell lines showed integrated and/or episomal viral DNA that is transcriptionally active, although viral oncogene expression was independent of viral copy number and the number of viral integration sites. Because these cell lines also contain EGFR expression and aneusomy, which are parameters of poor prognosis, they should be considered suitable model systems for the development of new antiviral therapies.

Infection with high-risk human papillomavirus (HPV) type 16 is an independent risk factor for the development of oropharyngeal squamous cell carcinomas (OSCC). However, it is unclear whether viral integration is an essential hallmark in the carcinogenic process of OSCC and whether HPV integration correlates with the level of viral gene transcription and influences the expression of disrupted host genes. We analyzed 75 patients with OSCC. HPV16-positivity was proven by p16^INK4A immunohistochemistry, PCR and FISH. Viral integration was examined using DIPS- as well as APOT-PCR. Viral E2, E6 and E7 gene expression levels were quantified by quantitative reverse transcriptase (RT-q)PCR. Expression levels of 7 human genes disrupted by the virus were extracted from mRNA expression profiling data of 32 OSCCs. Viral copy numbers were assessed by qPCR in 73 tumors. We identified 37 HPV16-human fusion products indicating viral integration in 29 (39%) OSCC. In the remaining tumors (61%) only episome-derived PCR products were detected. When comparing OSCC with or without an integration-derived fusion product, we did not find significant differences in the mean RNA expression of viral genes E2, E6 and E7 or the viral copy numbers per cell, nor did the RNA expression of the HPV-disrupted genes differ from either group of OSCC. In conclusion, our data do not support the hypothesis that integration affects the levels of viral and/or HPV-disrupted human gene transcripts. Thus constitutive, rather than a high level, of expression of oncogene transcripts appears to be required in HPV-related OSCC.

Human papillomavirus (HPV) is a risk factor for the development of benign and malignant mucosal head and neck lesions. P16^INK4A is often used as a surrogate marker for HPV‐infection, although there is still controversy with respect its reliability. Our aim was to determine if p16^INK4A overexpression can accurately predict both high‐risk and low‐risk‐HPV‐presence in (pre)malignant and benign head and neck lesions. P16^INK4A immunohistochemistry was performed on paraffin‐embedded tissue sections of 162 oropharyngeal squamous cell carcinomas (OPSCC), 14 tonsillar and 23 laryngeal dysplasias, and 20 tonsillar and 27 laryngeal papillomas. PCR, enzyme‐immunoassay and FISH analysis were used to assess HPV‐presence and type. Of the 162 OPSCC and 14 tonsillar dysplasias, 51 (31%) and 10 (71%) were HPV16‐positive, respectively. All tonsillar papillomas were HPV‐negative and four laryngeal dysplasias and 26 laryngeal papillomas were positive for HPV6 or −11. P16^INK4A immunohistochemistry revealed a strong nuclear and cytoplasmic staining in 50 out of 51 HPV16‐positive and 5 out of 111 HPV‐negative OPSCC (p < 0.0001) and in all HPV16‐positive tonsillar dysplasias, whereas highly variable staining patterns were detected in the papillomas and laryngeal dysplasias, irrespective of the HPV‐status. In addition, the latter lesions generally showed a higher nuclear than cytoplasmic p16^INK4A immunostaining intensity. In conclusion, our data show that strong nuclear and cytoplasmic p16^INK4A overexpression is a reliable surrogate indicator for HPV16 in OPSCC and (adjacent) dysplasias. For HPV6 or −11‐positive and HPV‐negative benign and premalignant lesions of the tonsil and larynx, however, p16^INK4A immunostaining is highly variable and cannot be recommended to predict HPV‐presence.