Recent Progress 2015

The CEP group places emphasis on the discovery, development and validation of circulating biomarkers to facilitate drug development and to aid cancer patient treatment decision making. This year we also placed considerable emphasis on the development of lung cancer patient Circulating tumour cell Derived eXplant models (CDX) which faithfully recapitulate patient responses to standard of care chemotherapy. CDX are being used to explore the biology of both Small Cell and Non Small Cell Lung Cancer, to identify new drug targets and to test novel treatments and biomarkers in collaboration with Pharma partners and the CRUK Centre for Drug development. Promising novel drugs and drug combinations will be translated to clinical trials within our newly established Manchester and UCL CRUK Lung Cancer Centre of Excellence. This year the nucleic acids biomarkers team has developed our capability for single circulating tumour cell molecular analysis facilitating exploration of tumour heterogeneity. We have also developed sensitive ctDNA assays for deployment in the upcoming Christie NHS Foundation Trust TARGET protocol, working with our clinical colleagues to optimise selection of the most appropriate phase I trial for their patients.  

Preclinical Pharmacology Team – Research Highlights
(i) CDX models
CTCs are highly prevalent in SCLC patients. We asked whether CTCs, enriched from Small Cell Lung Cancer (SCLC) patients and implanted into immune-compromised mice, could give rise to tumours and whether these tumours reflected the response of the donor patient to therapy. To date we have implanted CTCs from 35 extensive stage SCLC patients prior to treatment, with follow-up relapse samples from 10 patients. These gave rise to 13 CTC derived explant (CDX) models. CellSearch enumeration of CTCs in a parallel blood sample revealed a range of EpCAM+/CK+ CTCs in patients who gave rise to CDX was 160-7687 CTCs/7.5 ml blood; 8/9 samples with >400 CTCs/7.5 ml generated CDX. When CDX were treated with cisplatin and etoposide, the standard of care for SCLC patients, they responded in the same way as their donor patient (see Hodgkinson et al., 2014, Figure 1) and represent improved models for therapy testing in a disease where tumour biopsy is very challenging. We have initiated academic collaborations with the Dana Farber Cancer Institute, Memorial Sloan Kettering Cancer Center and the Karolinska Institute to test novel agents in our CDX panel and to evaluate tumour epigenetics. We are also testing novel therapeutic strategies in collaboration with a number of pharmaceutical companies, including those targeting DNA damage repair pathways. We initiated a CDX programme for Non Small Cell Lung Cancer (NSCLC) where the number of CellSearch EpCAM+/CK+ CTCs is considerably lower. The first NSCLC CDX we developed was derived from a patient with no CellSearch detectable EpCAM+/CK+ CTCs. However, when their blood sample was filtered and stained with epithelial and mesenchymal markers, >150 CTCs/ml blood were detected with ~80% expressing the mesenchymal marker indicative of an epithelial to mesenchymal transition (EMT).

 

Figure 1. SCLC CDX models recapitulate patient response to treatment with cisplatin and etoposide.  Mice bearing CDX3, 2 or 4 were treated with cisplatin and etoposide or vehicle control and the length of time taken to reach 4 times initial tumour volume (4xITV) monitored. Patient 3 survived 9.7 months and only 2 CDX3 tumours reached 4xITV after treatment with cisplatin and etoposide. Patient 2 survived for 3.5 months and CDX2 exhibited an intermediate response to cisplatin and etoposide while patient 4 progressed through therapy and survived 0.9 months and CDX4 did not respond to therapy.

(ii) Lowering the threshold for apoptotic cell death in SCLC

We have a long standing interest in the regulation of apoptosis in SCLC and the action of the BH-3 mimetic class of drugs. PIM kinases promote cell survival by disrupting the interactions between the pro- and anti-apoptotic Bcl-2 family members as well as stabilising c-Myc and causing increased c-Myc driven transformation. Given this co-operation between PIM kinase signalling and Myc driven oncogenesis, together with the survival promoting phosphorylation by PIM kinases of the pro-apoptotic Bcl-2 family member Bad, we investigated the effect of the PIM kinase inhibitor AZD1208 in combination with the BH3 mimetic AZD4320 (which targets the BH3 domain of anti-apoptotic Bcl-2 family members) in collaboration with colleagues at AstraZeneca. Furthermore, both MYC and Bcl-2 are frequently amplified in SCLC. We found that the PIM kinase inhibitor AZD1208 caused cytostasis across a panel of 9/11 SCLC cell lines, down-regulated global protein translation and although it reduced the pro-survival phosphorylation of the pro-apoptotic protein BAD, it did not (as a single agent), induce cell death. In five SCLC cell lines, the combination of AZD1208 and AZD4320 caused significant sensitisation to cell death and significantly impaired the growth of H2171 SCLC xenograft tumours in vivo. These data suggest that PIM kinase inhibition combined with BH-3 mimetic mediated lowering of the apoptotic threshold could hold promise for the treatment of SCLC.

(iii) Vasculogenic Mimicry in SCLC
Vasculogenic mimicry (VM) describes the ability of aggressive tumour cells to ‘mimic’ properties of endothelial cells and enable de novo generation of tumour-derived vascular networks that provide micro-circulation independently of non-cancer cells/stroma. VM is evaluated in clinical specimens by immunohistochemical analysis of Periodic Acid Shiff (PAS) positive, CD31 negative vessels. In common with angiogenesis, VM can be driven by a hypoxic tumour microenvironment providing an alternative to angiogenesis for the delivery to tumour of oxygen and nutrients and a potential escape mechanism for tumours treated with anti-angiogenic therapies.VM was first associated with poor prognosis in aggressive melanomas. Using a SCLC tissue microarray from 41 limited stage (LS) patients, we have shown for the first time that VM occurs in SCLC and that a high VM score is associated with worse patient overall survival (Figure 2). We also observed VM in 11 SCLC CDX models established from patients with extensive stage (ES) disease. Furthermore, circulating tumour cells (CTCs) from 37/38 SCLC patients contained a subpopulation of CTCs expressing Vascular Endothelial-Cadherin (VE-cadherin), a VM associated protein. Using shRNA VE-cadherin knockdown, we demonstrated its functional role information of VM-like networks by the SCLC cell line NCI-H446 in matrigel. VM warrants further investigation in SCLC where it may contribute to the prevalence of CTCs in this highly metastatic disease, and where components of the VM pathway may be targets for SCLC therapeutics.

Figure 2. VM networks are present in SCLC.Representative images of a human SCLC TMA stained with modified PAS/CD31 (central panel) and higher magnification panels showing PAS+/CD31 VM vessels (left) and PAS+/CD31+ endothelial vessels (right).

Update from the Nucleic Acids Biomarkers Team (NAB)
Tumour heterogeneity and subsequent emergence of drug resistance is increasingly recognised as a major barrier for improving cancer patient treatment. The NAB team is developing and applying molecular profiling methods suitable for monitoring tumour heterogeneity and evolution in a patient blood sample. Use of circulating biomarkers to determine the molecular status of the patient’s tumour also reduces reliance on tumour biopsies. In some cancers as little as 1 ml of patients’ blood samples may contain 0-1000s CTCs amongst 108 normal white blood cells (WBCs) along with ng quantities of circulating DNA (cfDNA) of which some is tumour-derived. The approaches being evaluated within the NAB team now include:

  • Processing of a single blood sample to deliver CTC mRNA, CTC gDNA and cfDNA up to 4 days post blood sample draw
  • Marker dependent and independent enrichment and isolation of single CTCs to enable molecular analysis
  • Routine whole genome amplification of single patient derived CTCs and subsequent next generation sequencing (NGS) based copy number analysis (CNA)
  • Application of targeted and whole exome sequencing (WES) NGS to identify potential clinically addressable mutations in patient CTCs
  • Routine plasma cfDNA isolation and quantification
  • Tailored NGS library generation to maximise sensitivity and reproducibility of molecular analysis of cfDNA
  • Targeted pull-down and PCR based NGS panels for mutational analysis of +600 cancer associated genes from patient cfDNA
  • Both WES and targeted NGS of cfDNA including detection of potential mechanisms of resistance to targeted therapies
  • Single cell mRNA profiling by Fluidigm, Nanostring and RNA-Seq
  • A bioinformatics pipeline for NGS analysis in collaboration with Crispin Miller

NAB place emphasis on improving blood collection and ensuring minimal loss or change in cellular and extracellular nucleic acids following storage of whole blood for four days at room temperature, ensuring compatibility with multisite clinical trials. To this end we have developed protocols for improving the detection of the circulating tumour DNA (ctDNA) fraction of cfDNA by avoiding the inadvertent lysis of WBCs and consequent dilution of low abundance ctDNA. We confirmed that our protocol avoids WBC lysis and delivers ctDNA amenable to both targeted sequencing and genome-wide NGS analysis (Figure 3A). We increased the sensitivity of NGS capture and developed a targeted 600-gene panel of cancer associated genes. This panel is ready for application in 2015 to identify clinically important mutations in cfDNA isolated from cancer patients on The Christie NHS Foundation Trust Early Clinical Trial Unit TARGET protocol.

Figure 3A. Heat map showing CNA of SCLC patients. CNA profiles were generated from isolated CTCs, EDTA cfDNA, CellSave cfDNA, two CDX tumours, germline gDNA and isolated WBC from a SCLC patient. Matching patterns of gain (red) and loss (blue) were seen across all tumour material. Chromosomal locations are shown below.

In collaboration with Professor Richard Marias, we have performed targeted and WES analysis of cfDNA from melanoma patients and identified potential mechanisms of resistance to targeted therapies. In collaboration with Dr Fiona Blackhall, we have accumulated and analysed a unique collection of CTCs from SCLC patients collected prior to and post chemotherapy treatment. NGS analysis of these samples has identified potential molecular signatures of resistance that are currently undergoing further analysis (Figure 3B).  Similar NGS analysis is underway comparing SCLC CTC samples from patients who responded well to initial treatment to those CTC samples from SCLC patients who failed to respond.

Figure 3B. Venn diagram showing distribution of somatic mutations identified in baseline and relapse samples from a SCLC patient.

Using our protocol for the accurate and reproducible transcriptional profiling at the single cell level (Rothwell et al., in press), we are RNA profiling CTCs. These studies include analysis of RNA signatures which may help guide the choice of treatment for prostate cancer within our Movember Centre for Excellence with partners in Belfast.  In addition, we are also examining whether CTC mRNA profiling can be used to identify transcriptional changes in CTC that arise as a consequence of treatment. To this end we are developing both focussed approaches, utilising the Fluidigm and NanoString platforms to analyse pre-defined transcriptional signatures and whole genome RNA-Seq approaches to identify global transcriptional changes. We are currently refining approaches to enable this analysis on both enriched populations and at the single cell level.

Underpinning the NGS studies in NAB is strong bioinformatics collaboration support from both CEP-based bioinformaticians and Dr Crispin Miller’s group at CRUK MI. Regular interactions between the groups have led to the development of robust pipelines for the analysis of NGS-based CNA that enables the accurate identification of tumour cells and detailed mutational profiling of samples. We became the oncology hub of the Manchester MRC Single Cell Centre of Excellence following this award in late 2014. 

A Snapshot of our Clinical Trials Biomarker Portfolio
The CEP Clinical Trials Portfolio encompasses 60 active clinical trials and experimental medicine projects, with another 13 at the planning stage. To assist effective project management and to provide user-friendly rapid access to key project and trial information, a customised ‘Dashboard’ has been developed. This allows the user to navigate quickly to one-page trial summaries using a range of queries, for example disease area, investigational drug, biomarker type, etc. The Dashboard ‘front page’ is shown in Figure 4. As our portfolio continually expands we are revising our processes and document management systems, to more easily achieve and monitor our compliance to Good Clinical Practice for Laboratories (GCPL). As part of this process, we are implementing Q-Pulse, a proven compliance solution providing a suite of integrated modules to manage business functions effectively and efficiently, including:

  • Document Control
  • Audits and Findings
  • Issues and Corrective / Preventative Action (CAPA)
  • Staff Competency, Training and Development
  • Asset Management

 

Figure 4. The CEP ‘dashboard’ front-page. This web-based interface provides the user with easy access to information contained with CEP portfolio of clinical trials with biomarkers based on a range of searches, such as disease area or biomarker type.

During the last year our specific user requirements have been comprehensively ‘mapped’ and the training and experience gained during QPulse configuration will ensure we have an effective solution in place for ongoing maintenance and future developments including our ambitious plans to build and develop, over the next two-four years, a CRUK Manchester Centre for Biomarker Sciences.

There has been significant activity this year assessing the value of enumerating and characterising CTCs as prognostic, predictive and pharmcodynamic biomarkers. We have also worked with a number of CTC technology providers in the past year, evaluating new approaches to CTC enrichment and isolation. Notably, in collaboration with the CRUK Manchester Institute Molecular Oncology Group, melanoma patient samples have been processed, enumerated and stored for detailed molecular analysis. Here we have worked to facilitate the development of the Clearbridge ClearCell FX device, a novel marker-independent spiral chip which has been used successfully to detect and enrich melanoma CTCs.