Skip to main content

The role of Sox2 in epithelial homeostasis and cancer

Periodic Report Summary 1 - SOX (The role of Sox2 in epithelial homeostasis and cancer)

During the outgoing phase in the Fuchs lab, I was able to comprehensively profile the SOX2 regulatory network in the context of tumorigenesis. First, in collaboration with the lab of Jonathan Weissman, we were able to establish a protocol for in vivo epidermis-specific ribosome profiling. Using this technique, I characterized the transcriptional and translational landscape of SOX2. We found that SOX2 expression resulted in a substantial reduction of protein synthesis. Remarkably, however, certain key oncogenic mRNAs were spared from this suppression suggesting that SOX2 reprograms the translational apparatus to favor the translation of certain oncogenic mRNAs. In an unbiased in vivo shRNA screen, I could reveal the underlying mechanisms involved in these changes. By dissecting the regulatory nodes within the translational apparatus, I documented the role of different eukaryotic initiation factors in a novel switch induced by SOX2. Finally, by conducting CRISPR/CAS9-mediated genetic analyses of different eukaryotic initiation factors, I could demonstrate a critical role of the alternative initiation factor eIF2A in tumor initiation. eIF2A knockout abrogated almost completely squamous cell carcinoma formation. Finally, increased eIF2A levels correlate with shorter survival and disease-free survival in human head and neck SCC patients, suggesting that eIF2A is a bona fide target for therapeutic strategies.

This work was published in Nature as an Article.

Sendoel A, Dunn J, Gonzales E, Naik S, Gomez S, Hurwitz B, Levorse J, Dill B,
Schramek D, Molina H, Weissman J, Fuchs E. Translation from unconventional 5’ start sites drives tumor initiation. Nature (2017). doi:10.1038/nature21036

The work was featured in:
− Pederson MT, Jensen KB. Unconventional translation in cancer. Nature (2017). doi:10.1038/nature21115
− Research, A. A. F. C. Translation from Alternate Start Sites Enriches Oncogenic Proteins. Cancer Discov (2017). doi:10.1158/2159-8290.CD- RW2017-014

Below is the summary regarding specific progress towards objectives in the proposal:

Aim 1: Profile the Sox2 network
-established an in vivo ribosome profiling protocol for the epidermis
-set up and characterized the system to induce SOX2 in the mouse epidermis using a Rosa26CAG-loxp-stop-loxp-Sox2-IRES-Egfp mouse model
-profiled the epidermis upon SOX2 expression by RNA-seq, ribosome profiling and proteomics
-set up a bioinformatics pipeline to analyze ribosome profiling data in our lab
-characterized the transcriptional and translational landscape upon SOX2 expression

Aim 2: Identify the physiological relevance of Sox2 target gene changes
-based on the results from Aim 1, I compiled a shRNA library of about 800 shRNAs targeting different translation factors
-injected this library into E9.5 amniotic cavity in vivo of wild-type and SOX2 embryos.
-to ensure a total of > 400 x coverage (cells per shRNA), > 30 embryos per genotype were injected
-P0 skins were then collected, genomic DNA was isolated, library was amplified and sequenced on a HiSeq 2500 platform
-bioinformatics pipeline was set up to analyze relative shRNA representation in wild-type and SOX2 epidermis

Aim 3: Functionally characterize novel Sox2 targets
-Novel hits from the screen were tested then in an allograft assay
-I used a newly established allograft model where oncogenic HRASG12V together with loss of TGFbeta receptor 2 results in fast and robust squamous cell carcinoma formation (SCC)
-using the CRISPR/CAS9 system in SCC cells, I deleted different eukaryotic initiation factors that were uncovered in the shRNA screen (Aim 2)
-established clonal cell lines with complete loss of function of different eukaryotic initiation factors in SCC and followed up on eIF2A, the top hit in our screen
-tested the clonal cell lines in allograft and limited dilution assay, additionally, rescue experiments were performed to show specificity
-bioinformatically analyzed TCGA human cancer data set, stratification of patients dependent on eIF2A mRNA levels

In conclusion, we found that translational reprogramming in SOX2 cells was critical for tumor formation. Intriguingly, by analyzing The Cancer Genome Atlas (TCGA) data sets, we found that increased mRNA levels of the alternative initiation factor eIF2A correlated significantly with shorter patient survival and disease free survival. These findings form a foundation into whether translational control factors can be exploited for future therapeutic interventions and we are currently screening for small molecule inhibitors which could be used to treat human squamous cell carcinomas.