Abstract
Human papillomavirus (HPV)-positive tonsillar- and base of tongue cancer is increasing epidemically and has much better outcome than corresponding HPV-negative cancer and most other head and neck cancers with around 80% 3-year disease free survival with conventional radiotherapy and surgery. Consequently, most HPV-positive cancer patients may not require the intensified chemoradiotherapy given to many head and neck cancer patients and would, with tapered treatment, avoid several severe side-effects. Moreover, intensified therapy has not improved survival and treatment alternatives are needed. To identify patients eligible for tapered or targeted therapy, additional biomarkers are required. Several studies have, therefore, focused on finding predictive markers, some of which are also potentially targetable. To conclude, better-tailored therapy, either as tapered or targeted, is important for increasing numbers of patients with HPV-positive tonsillar- and base of tongue cancer. This review deals with some of these issues and presents some promising markers.
- Tonsillar cancer
- base of tongue cancer
- oropharyngeal cancer
- head and neck cancer
- HPV
- MHC class I
- CD8+ TIL
- CD44
- FGFR3
- VEGF-A
- review
In 2007, human papillomavirus (HPV) was recognized as a risk factor for oropharyngeal squamous cell carcinoma (OPSCC) by the International Agency for Research against Cancer (IARC), besides the traditional risk factors smoking and alcohol (1-4). Concurrently, an increased incidence of OPSCC, more specifically tonsillar and base of tongue cancer (TSCC and BOTSCC, the two main HPV-positive OPSCC sites) was described in many developed countries, and found to be due to a sharp rise of HPV-positive cases (5-18). These HPV-positive tumors had a much better clinical outcome than their HPV-negative counterparts and head and neck squamous cell carcinoma (HNSCC) in general when treated with radiotherapy and surgery (1-3, 15-17, 19, 20). Treatment of HNSCC including OPSCC has, however, gradually become more intensified with induction chemotherapy or concurrent chemo-radiotherapy and sometimes the administration of Erbitux, leading to more severe side effects (16, 20, 21). This is likely unnecessary for most HPV-positive TSCC and BOTSCC patients since they generally respond well to less therapy and furthermore, intensified therapy does not improve survival of HPV-positive TSCC and BOTSCC (1-3, 15-17, 20, 22, 23). It is, therefore, important to identify patients with HPV-positive TSCC and BOTSCC eligible for less treatment or targeted therapy and efforts focus now on finding predictive and targetable markers for these patients (24-39).
Here, a short background of the field and an update on recent attempts to find predictive and targetable markers with different methodologies for individualized therapy is presented.
Human Papillomavirus in Tonsillar and Base of Tongue Squamous Cell Carcinoma and an Epidemic Increase of these Tumors
In 2000, an association between HPV and OPSCC and TSCC was reported, and with radiotherapy and surgery, these HPV-positive cases had much better outcome than corresponding HPV-negative cases (80% vs. 40-50% 5-year survival) (1, 2). In 2004, a similar relationship of HPV to BOTSCC, but not to mobile tongue cancer was disclosed (40). Based on differing characteristics, HPV-positive and HPV-negative cancers were suggested to be separate entities, with the latter mainly caused by smoking and alcohol (1-4). Most HPV-positive tumors expressed normal p53 and overexpressed p16Ink4a in contrast to HPV-negative cases (1, 41). HPV-positive tumors were often aneuploid, less differentiated, and had, like cervical cancer, chromosome 3q amplification. Yet, independent of stage, differentiation, or ploidy, patients had better outcome than those with HPV-negative tumors (1-4, 42, 43). HPV-positive cancer patients were also often never-smokers, and never-smokers with HPV-positive cancer had better outcome than smokers (22, 44).
For HPV detection, HPV DNA and RNA analysis was done by e.g. PCR-based methods, in situ hybridization, and Hybrid Capture 2 assays, and p16Ink4a overexpression by immunohistochemistry (IHC) was used as a surrogate marker for HPV (45-52). HPV E6 and E7 mRNA expression, indicative of active HPV is however the golden standard, but in formalin fixed paraffin embedded (FFPE) tissue, presence of HPV DNA with p16Ink4a overexpression in combination, but not each one alone, is almost as sensitive (52).
OPSCC HPV prevalence varied depending on the quality, the methodology, the time period and geographical region of the analyzed samples, but much of the variation was eventually found due to an epidemic of HPV-positive TSCC and BOTSCC in some countries (5-18).
In 2006, in Stockholm, Sweden, a rise in the incidence of TSCC was reported from 1970 to 2002 and in parallel an increase in the percentage (23% to 68%) of HPV-positive TSCC was found. Similar trends were subsequently described in the US (6, 8-10). Furthermore, in 2009 and 2010, reports from Sweden on TSCC and BOTSCC respectively and in 2011, from the U.S. on OPSCC showed increases of HPV-positive cases and decreases of HPV-negative cases, the latter due to reduced smoking in these countries (11, 12, 17). Reports from some other Western countries showed similar trends (5, 7). The increase was eventually attributed to changes in sexual habits, and a significant correlation between HPV-positive OPSCC, early sex debut and numbers of oral or vaginal partners was reported (53).
Treatment of Tonsillar and Base of Tongue Cancer
Most TSCC and BOTSCC patients present symptoms and seek care when their tumors are fairly large similar to HNSCC in general, and they often receive the gradually intensified therapy given to other HNSCC (16, 20). Treatment consists of radiotherapy at high doses for 6-7 weeks and patients with advanced stage (III-IV) also have induction or concomitant chemotherapy and sometimes EGFR blockers depending on the tumor burden and whether they are medically fit (16, 20, 21). This treatment can give acute side-effects such as pain, difficulties in swallowing and eating caused by radiation as well as nausea, mucositis and sometimes local and/or systemic infections and fatigue. Patients with remaining lymph node metastases after radiotherapy have to complete their treatment with modified neck node surgery. This can result in additional side effects, with increased fibrosis and stiffness of the neck, impaired shoulder mobility and sometimes worsening of the swallowing difficulties caused by radiotherapy. Surviving patients may suffer from late side effects e.g. xerostomia, taste alterations, difficulties with swallowing, trismus and hearing impairment that can worsen with time. Some patients suffer from radio-osteonecrosis and need reconstruction surgery (54-56). Upon recurrence, patients can sometimes obtain salvage surgery or re-irradiation with or without chemotherapy. For palliative treatment, chemotherapy or biological therapies can be given (20, 21).
Notably, in parallel, to the intensification of HNSCC treatment, paradoxically a growing number of HPV-positive TSCC and BOTSCC patients are included in this group, of which most do not need intensified therapy, which does not even improve their survival and gives more side effects (20, 23). Clinical attempts to reduce radio- or chemotherapy are discussed and have been initiated (57, 58). However, to further enhance tapered or targeted therapy, improve survival and quality of life, better means to identify HPV-positive TSCC and BOTSCC patients responding well to treatment or eligible for targeted therapy are needed.
Studies on Other Markers and Re-staging of HPV-positive Tonsillar and Base of Tongue Cancer and Response to Treatment
Several approaches have been made to identify cancer patients that will respond to present therapy or need additional e.g. targeted therapy. For OPSCC, TSCC and BOTSCC, initial reports focused on differences between HPV-positive and HPV-negative tumors, while later studies investigated markers with prognostic value in other tumor types and later evaluated the use of combining these markers in mathematical models (1-3, 23-26, 28-39, 42-44, 59). More recently, next-generation sequencing (NGS), the evaluation of miRNA expression, proteomic and transcriptome profiling and the microbiome, have been explored in different tumors including HNSCC and sometimes OPSCC, TSCC and BOTSCC (37-39, 60-83). Finally, based on accumulated data, previous staging has been re-classified for HPV-positive OPSCC (84, 85). Below, some of these attempts are described in more detail.
Studies Disclosing Differences between HPV-positive and HPV-negative Oropharyngeal Cancer, Tonsillar and Base of Tongue Cancer
To disclose differences between HPV-positive and HPV-negative OPSCC and TSCC and follow survival, initial studies investigated e.g. HPV DNA/RNA status, p16Ink4a expression, p53 expression, age, gender, smoking, aneuploidy and comparative genomic hybridization (1-3, 22, 41-44). Non-smoking patients with tumors that were HPV DNA/RNA positive and/or overexpressed p16Ink4a had the best survival, while ploidy, or having specific chromosomal changes were secondary to HPV status (1-3, 22, 41-44). Physical state (episomal/integrated) and load of the viral genomes were also examined, and a high viral load was associated to better survival in some, but not all studies, while physical state had no influence on survival (22, 86, 87).
The location of HPV in OPSCC also mattered. The prognostic significance of HPV was limited to TSCC and BOTSCC, the two subsites originating from Waldeyers lymphatic ring, while HPV was rare and its prognostic influence unclear in other OPSCC sites (88, 89).
Finally, using the AJCC/UICC TNM staging systems has not been useful for HPV-positive OPSCC prognostication, since already very low T stages often exhibited lymph node metastasis, resulting in the tumors being defined in higher stages. Therefore, a new staging system was proposed, evaluating patients with p16Ink4a positive and p16Ink4a negative OPSCC separately (84). This staging system was later validated (85) and is now included in the eighth edition of the AJCC/UICC TNM staging system manual.
Studies Focusing on Immunological and Stem Cell Markers in Tonsillar and Base of Tongue Cancer
Similar to other tumor types, immunological and other markers have often been examined by IHC in HPV-positive OPSCC, TSCC and BOTSCC (24-36). HPV-positive tumors generally had higher numbers of CD8+ lymphocytes infiltrating the tumor or in surrounding stroma than corresponding HPV-negative tumors, and having high numbers of CD8+ cells was correlated to a better survival for both HPV-positive and HPV-negative tumors (32-35). The prognostic value of CD4+ as well as of FoxP3+ lymphocytes and Cox 2 was also analyzed, but none alone correlated to survival, while having a low CD4+/CD8+ ratio or a high CD8+/FoxP3+ ratio did correlate to better survival irrespective of HPV status (32, 33).
Recently, the potential of PD-L1 expression as a biomarker was analyzed in parallel to that of CD8+ T-cells and CD68+ macrophages in HPV-positive and HPV-negative OPSCC (35). In HPV-positive tumors, high CD8+ T-cell counts indicated a good response to therapy post-radiation, while in HPV-negative tumors increased numbers of CD68+ macrophages expressing PD-L1 may reflect a more favorable immune milieu (35).
Unexpectedly, in some but not all reports absent/low MHC class I expression was of positive prognostic value for HPV-positive TSCC and BOTSCC, whereas the reverse, as would be expected was found for those with HPV-negative cancer (26, 27). TAP1 and 2, and LMP 2, 7 and 10 expression was also examined, since these factors influence peptide antigen processing and MHC class I expression (29, 30). TAP2, LMP2 and LMP7 expression frequently decreased in HPV-positive tumors, and decreased LMP7 expression and absent/low compared to medium/high nuclear LMP10 staining correlated to better survival (29, 30). MHC class II expression did not influence survival (27).
Down-regulation of MHC class I, and components of the antigen processing machinery in HPV-positive tumors could be a consequence of the strong immune defense against these tumors, necessitating a stronger need for immune escape mechanisms. The decrease in MHC class I expression was assumed to be due to active HPV infection and more specifically the downregulation of MHC class I by HPV E5 and E7 (23, 90). MHC class I expression was suggested to increase upon irradiation and make the tumors more sensitive to the immune system of the host (91). Indeed, experimentally, irradiation increased MHC class I expression in vitro in some, but not all, HPV-positive OPSCC cell lines and in one HPV-positive cell line a decrease in HPV16 E5 mRNA expression was also noted (91). Increased MHC class I expression upon irradiation could in part explain why HPV-positive tumors with absent/low MHC class I expression still respond readily to therapy and are recognized by the immune system (91). Eliciting a vigorous immune response in this context, would be in line with data obtained in mice, where HPV-positive tumors were curable after cisplatin or radiation therapy only in immunocompetent and not in immunoincompetent mice (92). In that system, it was proposed that being able to mount an immune response was required to eradicate these tumors (92).
The role of HPV mRNA expression was also analyzed in correlation to survival, due to a possible correlation to MHC class I expression (90). The presence of HPV16 E2, E5 or E7 mRNA did not correlate to MHC class I expression. However, the absence of HPV16 E2 mRNA correlated with worse prognosis similarly to cervical cancer (23, 90).
Other molecular markers studied by IHC were CD44 and CD98, and having low CD44 intensity expression or CD98 expression was beneficial for survival for patients with HPV-positive and HPV-negative cancer (31) For tumor suppressor genes LRIG 1-3, only high expression of LRIG1 was correlated to better survival in HPV-positive, but not HPV-negative TSCC and BOTSCC (36).
Taken together the immunological and IHC analysis show that some markers could be specific for only HPV-positive OPSCC, TSCC and BOTSCC, or HPV-negative OPSCC, while others could be of more general use.
Sequencing of Head and Neck Cancer, Oropharyngeal, Tonsillar and Base of Tongue Cancer and the Unraveling of Additional Differences Between HPV-positive and HPV-negative Cancers and Targetable Markers
The application of next-generation sequencing (NGS) technology for mutational analysis of tumors for clinical and research purposes has increased exponentially the past 10 years, thanks to reasonable costs and the extremely small amount of DNA needed (even from low-quality FFPE-derived DNA).
The NGS approach to HNSCC has revealed some distinct differences between HPV-positive and HPV-negative tumors and is providing valuable additional information on HPV affected genes and on potential therapeutic targets for HNSCC patients (37-39, 60-64). More specifically, HPV-positive OPSCC, mostly TSCC and BOTSCC, show increased mutation rates in PIK3CA, NOTCH1 and the fibroblast growth factor 3 (FGFR3), while HPV-negative tumors are dominated by mutations in TP53 and CDKN2A/B and to a lower degree in PIK3CA (37-39, 60-64).
Mutated TP53 was found to be of prognostic value for survival in HPV-negative OPSCC in one, but not in another study (38, 39). This discrepancy might be due to differences in sequencing methods, classification of TP53 mutation as well as number and type of patients included in the analysis (38, 39). PRIMA-1MET is the only drug currently under clinical trials targeting mutant p53 (93). Mutation in PIK3CA gene has also been associated with poorer prognosis in human cancer (94-96). However, no conclusive data are reported for HNSCC (66, 97). Nevertheless, a relevant number of inhibitors targeting PIK3CA or the PI3K pathway are available and on trial (98).
Another gene of interest, FGFR3, is often mutated in HPV-positive HNSCC (7-11%) compared to HPV-negative HNSCC (2%) and represents an interesting therapeutic target since specific inhibitors are available (39, 60, 64). Furthermore, the FGFR3 p.S249C variant has been primarily found in HPV-positive OPSCC and cervical cancer and has been associated with poorer prognosis (39, 60, 98-100). Targeted therapy with FGFR inhibitors has shown promising therapeutic effect in pre-clinical and clinical studies (101-105) and this applies also to the FGFR3-TACC3 translocation found in HPV-positive HNSCC (105).
To summarize, gene sequencing is disclosing both new and already known mutations in HPV-positive and HPV-negative tumors, where targeted therapy could be available.
MicroRNA Expression in Oropharyngeal Cancer in Relation to HPV-status and Clinical Outcome
MicroRNA (miR) expression in OPSCC has been mainly analyzed in normal tissue, HPV-positive vs. HPV-negative tumor status and to a lesser extent to outcome (67-73). The results varied, with many miRs proposed to correlate with tumor HPV status or outcome, with few miRs found common between studies. Mir-9, 155 and 163b were in several studies found overexpressed in HPV-positive vs. HPV-negative OPSCC, while miR-31 and 193b were downregulated (68-71, 73). Also, a 25 miR signature discriminating HPV-positive vs. HPV-negative OPSCC was suggested, but few of the indicated miRs were common with other studies (67). These discrepancies can depend on the assessment of large miR numbers in small sample sizes and lack of validation cohorts.
Few studies have analyzed miR expression in OPSCC in relation both to HPV and clinical outcome and these miRs differ between the studies (71-73). Overexpression of miR-142-3p, 146a, 26b was positively correlated to overall survival, while miR- 31, 24, 193b was negatively correlated in a study by Gao et al. (73). In another study on clinical outcome in relation to different chemo-radiotherapy combinations, miR-146a, 155, 200b were correlated to treatment outcome, with the first two independent of tumor HPV status (72).
To summarize, there are still limited numbers of studies relating miR expression to clinical outcome of OPSCC and in relation to tumor HPV status. Given the large variation in miRs found, more studies on large cohorts, including validation sets, are needed before miRs can be used in the clinical practice for prediction of clinical outcome.
Transcriptome Analysis in OPSCC
Some studies have described differences between HPV-positive and HPV-negative OPSCC with regard to RNA expression and differences have mainly been reported in genes involved in immune response, apoptosis, proliferation/cell cycle and differentiation (74-77, 106). Notably, some studies have, based on transcriptional and HPV data, suggested molecular sub-grouping of HNSCC. Two distinct subtypes were e.g. described among HPV-positive tumors and these had different clinical outcomes, suggesting heterogeneity among the HPV-positive tumors (77). Moreover, genes related to immune defense, e.g. CD8a, have also been linked to HPV status and better survival by transcriptional analysis (78).
Protein Profiling in Oropharyngeal Cancer
While several studies have analyzed differences in mRNA expression between HPV-positive and HPV-negative OPSCC, few proteomic analyses have been performed. Sewell et al. used reverse-phase protein array profiling on a restricted set of 137 total and phosphorylated proteins (66). Differences were found in e.g. PI3K/AKT/mTOR and receptor kinase pathways. Protein expression was also evaluated in relation to PIK3CA mutations in HPV-positive OPSCC and found to be associated with activation of mTOR. Another study presented proteomic profiling by mass spectrometry and described differences between HPV-positive and HPV-negative OPSCC in a large number of pathways and a specific finding was the increased expression of ASS1 in HPV-positive OPSCC (65).
Microbiome Studies in Head and Neck Cancer
Recently the human microbiome has been established to connect functionally to diseases such as cancers, enteric infectious diseases, aging, obesity, diabetes and neurological disorders (79, 80). Although not as well characterized as in the gut, a growing body of evidence hints that the microbiota in the oral tracts also plays a role in modulating our health and e.g. changes in abundance of Firmicutes (Streptococcus) have been suggested to be associated with oral pre-cancers (81). Recently, a study on HNSCC patients showed a significant loss in diversity of microbiota species compared with healthy controls (82). In addition, bacteria diversity also significantly decreased after surgery (83). Furthermore, it has been noted that mucosal toxicity of different treatments for HNSCC are related to patient overall survival, also including HPV status in the analysis (107). If microbiome or related metabolic changes after therapy are related to the mucosal toxicity shift is of interest to evaluate further. So far, research on microbiome changes associated to HPV infection in HNSCC is limited, but HPV infection has been associated with microbiota changes in HPV related cervical intraepithelial neoplasia (CIN) and cervical cancer (108). There is hint that the abundance of Veillonella, Prevotela and Streptococcus fluctuated when HPV status changed (83). However, all data in HNSCC are still preliminary and the sample sizes are small. More investments are urgently needed before one can apply bacteria such as probiotic Lactobacillus into HNSCC prevention and treatments. Together these results suggest that HNSCC associated shifts in the composition of the microbial community may provide us another possible explanation for HNSCC cancer progression.
Conclusion
HPV-positive TSCC and BOTSCC have much better outcome than corresponding HPV-negative cases and have been increasing epidemically in incidence the past decades. Individualizing treatment for this group of patients is important and many attempts in finding markers useful for tapering or targeting treatment are being made of which some are successful. Hopefully in the future, many of these patients will benefit from some of these efforts.
Acknowledgements
This work was supported in part by the Swedish Cancer Foundation, the Stockholm Cancer Society, the Cancer and Allergy Foundation, Svenska Läkarsällskapet (SLS), Magnus Bergvalls stiftelse, stiftelsen Tornspiran, the Stockholm City Council, the Sigurd and Elsa Goljes Foundation and the Karolinska Institutet, Sweden.
Footnotes
This article is freely accessible online.
Conflicts of Interest
No conflicts of interests declared.
- Received August 10, 2017.
- Revision received September 1, 2017.
- Accepted September 5, 2017.
- Copyright© 2017, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved