Tobias Sjöblom's projects on mutations causing colorectal cancer

Integrated data and sample collection in clinical cancer care

Per-Henrik Edqvist, Chatarina Larsson, Lucy Mathot, Tony Hansson, Emelie Bladin

Translational cancer research is dependent on high quality patient materials. We address this by coordinating a longitudinal collection of patient data, tissues, and imaging before, during, and after cancer therapy at Uppsala Academic Hospital and Umeå University Hospital (www.u-can.uu.se) with support from the Swedish Government.

At the end of 2016, more than 12,000 patients with colorectal, brain, prostate, gynecological, neuroendocrine, breast, lung, lymphoma or haematological cancers had been included in U-CAN (Tobias Sjöblom, Program Director; Tony Hansson/Per-Henrik Edqvist, Administrative Director; Chatarina Larsson, Project Administrator; Emelie Bladin, technician; and U-CAN clinical partners).

Following the Excellence reviews in the evaluation of Governmental Strategic Research Areas (SFO/SRAs) in 2015 U-CAN has continued to support high patient inclusion rates during 2016 but has also made efforts to gradually re-focus to increase the information density and quality for each patient and support research based on the collected materials.

The sample collection has received increasing interest from researchers and companies and applications for withdrawal of materials from the U-CAN collection tripled in numbers during 2016 compared to 2015. The first publication encompassing patients from U-CAN was published in 2016 (Ljungström et al, Blood 2016).

The spin-out company ExScale Biospecimen Solutions AB, founded in 2012 based on technology developed by us (Mathot et al, 2011; Mathot et al, 2013) had first sales of its CE/IVD reagent system for automated serial extraction of DNA and RNA from FFPE samples in clinical diagnostics in 2016.
 

Mutational studies of candidate cancer genes

Tom Adlerteg, Ivaylo Stoimenov, Lucy Mathot, Viktor Ljungström, Veronica Rendo, Snehangshu Kundu

Somatic mutations are the basis for modern cancer diagnostics and therapeutics development. Somatic mutation status is derived from sequencing of DNA from cancer patient samples and depending on tumour cell content of the samples this can be a challenging task.

We have developed software tools for rapid and accurate mutational analysis of deep sequencing data from solid tumours with significant content of normal cells. These tools have superior indel calling capabilities, a major challenge in mutational analysis, as compared to state of the art (Stoimenov, Adlerteg et al, manuscript). For this application, novel statistical mathematics has been developed and patented (Swaminathan et al, Pattern Recognition 2016; WO2016043659). Current work in the project focuses on improved alignment algorithms. Preparations for commercialization is currently undertaken together with UU Innovation (Björn Ingemarsson)

Using these tools, we have completed deep mutational analyses of 676 genes in cancer pathways in 107 colorectal cancers (Mathot, Ljungström et al, Cancer Res 2017). While the expected frequencies and types of mutations were observed in known CRC genes such as APC, KRAS, and TP53, we noted an enrichment of mutations in the Ephrin receptor tyrosine kinase gene family in tumours giving rise to metastasis.

Ephrin receptors have previously been associated with metastatic disease development, however, no mutational evidence was previously available to explain the downregulation of Eph proteins associated with metastasis of CRCs. Functional data indicate that the mutations identified by us confer a phenotype on colorectal cancer cells. These findings could be used to identify patients that require close monitoring to detect recurrence and to stratify CRC patients that would benefit most from adjuvant treatments.

Follow-up work includes identification of additional Eph receptor mutations as well as novel scalable assays for assessment of Eph mutation phenotypes and drug responses. Current mutational detection efforts also include whole genome sequencing of CRC cases in U-CAN where longitudinal blood samples are available (Viktor Ljungström).
 

Functional studies of novel candidate cancer genes

Snehangshu Kundu, Chatarina Larsson, Tatjana Pandzic, Ivaylo Stoimenov, Veronica Rendo

Gene mutation prevalence is not sufficient to prove cancer gene status - functional and phenotypic studies comparing mutant and wild-type alleles in relevant model systems are required for ultimate proof. Such analyses may be accomplished through genome editing in human cancer cells.

We have developed scalable experimental and computational tools for designing rAAV gene targeting constructs to all genes in the human genome (Stoimenov, Akhtar Ali et al, NAR, 2015). We previously identified novel mutations in 12 genes by sequencing of breast cancer patient samples (Jiao et al. 2012) and using rAAV technology we knocked out one of these, the putative breast cancer gene DIP2C, obtaining evidence for activation of cancer promoting traits by DIP2C gene inactivation (Larsson et al. in peer review).

We have targeted the transcriptional modulator ZBED6 in colorectal cancer cells and demonstrated effects on cell growth rate and regulation of genes in CRC pathways (Akhtar Ali et al, PNAS 2015). We have also generated knock-ins of colorectal cancer genes (PRDM2, MLL3, and KRAS) that are currently being characterized by us and used by collaborators in drug discovery efforts (Pandzic et al, in peer review; Larsson et al, Kundu et al, manuscripts).

To better understand which genes belong to the Ras pathway in human CRC, we have adapted technology for forward genetics by transposon mutagenesis in human cells to map the RAS pathway in human colorectal cancers by a phenotypic screen, assigning 163 recurrently targeted genes to the Ras pathway. Out of the 15 genes selected for further validation, 3 genes showed phenotypes associated with Ras pathway activation in CRC following knock-down. Two of the three genes controlled the level of pERK in CRC cells, providing independent evidence of them being components of the Ras pathway (Kundu et al, in review).

A similar forward genetic screen based on transposon mutagenesis was used to identify genes involved with resistance to Fludarabine in chronic lymphocytic leukemia in collaboration with Mats Hellström and Rickard Rosenquist (Pandzic et al. Clin Cancer Res 2016).
 

Exploiting loss of heterozygosity for a novel anti-cancer therapy

Veronica Rendo, Ivaylo Stoimenov

The success of anti-cancer therapy is based on finding conditions resulting in selective killing of cancer cells, while the normal tissues of the patient are spared. We propose a concept which is based on exploitation of the genetic variation (SNPs) naturally occurring in the human population and the cancer specific phenomenon loss of heterozygosity (LOH) to identify tumors that are sensitized to certain drugs relative to the normal tissues.

Using the 1000 Genomes database, we identified human enzymes having variant amino acids in their active sites as result of SNPs and ranked the 20 putative targets according to the prevalence of SNPs and LOH in common human cancers. For the top candidate NAT2, a known drug metabolic enzyme, we estimate that >3 % of patients with CRC could benefit from a tailored drug therapy, which translates to >35,000 cases worldwide per year.

We therefore constructed and validated CRC cell model systems for this candidate in two independent genetic backgrounds. Subsequent drug discovery efforts uncovered a compound with 3-fold increased cytotoxicity in cells lacking NAT2 in vitro and in vivo (Ivaylo Stoimenov, Veronica Rendo et al., in peer review).

A similar project where inactivating polymorphisms (STOPs and indels) have been systematically investigated led to the identification of a promising candidate target enzyme for which cell models are being developed (Veronica Rendo, Ivaylo Stoimenov).