Research team Tina Dalianis
Studies on head and neck cancer
Background and previous results
There are >150 HPV types of which some types or species are potentially oncogenic. Best known are HPV16 and 18, which cause cervical cancer and included in today's HPV vaccines. However, besides smoking and alcohol, HPV16 has a causal link to oropharyngeal cancer, where tonsillar and base of tongue cancer (TSCC and BOTSCC) dominate. We and others have shown that HPV-positive TSCC and BOTSCC have a much better survival than corresponding HPV-negative cancers and head/neck cancers in general (80% vs. 40% 5-year disease-specific survival). In addition, we were the first to show that the incidences of HPV-positive TSCC and BOTSCC have increased epidemically the past decades. Due to its poor prognosis, therapy for head and neck squamous cell carcinoma (HNSCC) generally includes chemotherapy and intensified radiotherapy (RT), with many more severe side effects. Intensified therapy has not improved survival for those with a poor prognosis and is at the same time likely unnecessary for 80% of patients with HPV-positive TSCC and BOTSCC who do well with conventional RT and account for >35% of all HNSCC, so new therapeutic options are needed. However, to better individualize therapy it is important to find markers that along with HPV can better predict a favorable response to conventional treatment, or that can be useful for targeted therapy.
Preliminary results and research plan
In a TSCC and BOTSCC patient cohort from 2000-2011 at the Karolinska University Hospital, we have identified such markers (e.g. HLA class I, CD44, LRIG, CD8 tumour-infiltrating lymphocyte counts, expression of HPV16 E2 mRNA), which individually or together identify 20-30% of patients with 90-95% probability for complete response. Using two mathematical models we have combined HPV-positive status with several prognostic biomarkers and can identify 40-56% of the patients that respond to therapy. Now we want to find additional markers to be able to combine old and newer markers in our mathematical models and understand their mechanisms of action to identify virtually all patients with a very good outcome. Our research will enable us to identify patients that respond to therapy safely and to start prospective randomized trials with less therapy. For those that do not respond to therapy, we hope to find other alternatives such as specific targeted therapy.
One step in this direction and to find additional biomarkers was to examine both HPV-positive and HPV-negative tumors for mutations in oncogenes and suppressor genes and to perform protein profiling. A second step just initiated with Dr Juan Du and prof. Lars Engstrand is to test our tumours for specific microbiome profiles, should they be useful as biomarkers.
Examining hotspot mutations by Next Generation Sequencing of 50 cancer related genes in 300 HPV+ TSCC/BOTSCC and 50 HPV- TSCC/BOTSCC and found that the genes PIK3CA and FGFR3, both targetable proteins, were among the most mutated ones in HPV+ TSCC/BOTSCC, while p53 was most frequently mutated in HPV-TSCC/BOTSCC (Fig.1). Notably, we also found that mutated FGFR3 in HPV positive tumours correlated to worse prognosis.
This indicates that FGFR3 could be very useful as an additional predictive marker to combine with the markers already mentioned earlier in the mathematical models described above to distinguish patients with very favourable or poor prognosis. Furthermore, 36% of those with FGFR3 mutations had PIK3CA mutations, thus possibly targetable with two drugs.
When performing protein profiling done by a sensitive proximity elongation assay (PEA) we found that some proteins related to angiogenesis, e.g. VEGF-A, CYR61 and ESM-1 and IGF1R, the latter conferring resistance to EGFR inhibitors, are up-regulated, while hK-8 is low, in HPV+ TSCC/BOTSCC that do not respond to treatment, indicating that these proteins can be used to predict effect of therapy.
More recently, we performed whole exome sequencing (WES) on tumors of patients with and without relapse and found some genes that were mutated in both cohorts and some genes CDC27, BCLAF1, and AQP7 that were mutated only in the group with recurrent disease (e.g. a specific CDC27 variant). Hopefully some of the data here will contribute to finding new options for targeted therapy.
To acquire more data on targeted therapy, we have employed in vitro models tissue culture models for targeted therapy. HPV positive and HPV negative TSCC and BOTSCC cell lines with or without e.g. FGFR3 and PIK3CA mutations have been obtained and have been tested for their sensitivity to different targeted therapies by viability, proliferation, apoptosis and cytotoxicity assays. Initial experiments that have been performed were very promising. We have for example found that our cell lines show dose dependent responses to PI3K and FGFR inhibitors and moreover when these two inhibitors are combined, synergistic effects are disclosed. We are now adding additional inhibitors to our repertoire.
With this approach we hope to identify whether drugs used today for targeted therapy in other types of tumors, or against newly disclosed mutated genes by WES also can be used for TSCC and BOTSCC. Useful drugs will also be further analyzed in experimental animal PDX systems.
Studies of childhood cancer
Recently, we have also screened a number of childhood cancers such as a number of different childhood brain tumors and neuroblastomas for the presence of FGFR3 and PIK3CA mutations. These types of mutations are rare, but have been found in at least one neuroblastoma (NB). We are therefore also testing some NB cell lines with the same assays that have been described above for the tonsillar and base of tongue cancer (TSCC and BOTSCC) cell lines. We have found some positive effects and these are investigated further by team Ourania Kostopoulou in our group.
By delivering novel tools to better tailor therapy for growing numbers of patients with TSCC and BOTSCC, allowing some patients to receive tapered therapy, and others to receive specific targeted therapy we will significantly increase survival as well as quality of life, issues of both ethical and socioeconomic importance.
For childhood cancer, if targeted therapy can be used synergistically this may be useful for those with tumors that are resistant to conventional chemotherapy and help increase survival as well as avoid some side effects associated with conventional chemotherapy.
Tina Dalianis, MD.PhD, Professor em. Group leader
Torbjörn Ramqvist, Assistant Professor, associated
Stefan Holzhauser, PhD, Post doc
Ourania Kostopoulou, PhD, post doc
Linnea Haeggblom, PhD, post doc, associated
Juan Du, PhD, Post doc associated
Michael Mints MD, PhD, Post doc, associated
Andrea Vlastos, MD, PhD, ENT specialist, associated