Cancer arises as a result of the acquisition of a series of abnormalities and mutations, typically involving oncogenes and tumor suppressor genes, which ultimately confer a growth advantage upon the cells in which they have occurred. A wide spectrum of types of genetic alteration can contribute to the promotion of cell growth, ranging from single amino acid mutations to deletions and chromosomal translocations, with many of these mutations representing potential therapeutic targets.

For example, mutation of K-Ras leads to aberrant growth signaling, loss of PTEN tumor suppressor leads to enhanced survival, and deletion or mutation of p53 leads to a defect in response to DNA damage.  Vital clues to how these and other mutated oncogenes and tumor suppressor genes contribute to cancer comes from studying the roles played within the cell by the normal counterparts of these genes.

Having the right preclinical tools to effectively interrogate novel targeted agents or treatment regimens before moving into the clinic is therefore fundamental for all drug discovery programs. Moreover, in recent years in vivo cell-based screening has proven to be a valuable tool for PK/PD and efficacy characterization in the vast majority of drug discovery programs. 

Scientists at Horizon Discovery have applied their understanding of the key genetic drivers of cancer to develop a panel of cancer-relevant human isogenic cell line pairs, using the company’s proprietary rAAV-based gene editing platform. Because these cell line pairs share the same genetic background, differing only by the mutation of interest, it is possible to carry out unambiguous studies of the biological impact of defined genetic differences between mutant and normal, without the typical confounding complications seen when performing such studies on unrelated cancer cell lines.

Taking this capability to interrogate the effects of specific genetic mutations further, a collaborative partnership with Crown Bioscience has been formed. Through this partnership, many X-MAN™ isogenic lines are being transitioned into animals and incorporated into Crown’s in vivo compound screening program, designed to aid pharma and biotech in the preclinical evaluation of novel anti-cancer agents.


click to enlarge
Figure 1: Growth of DLD-1 wild-type (parental) and KRAS G13D knock-out cells in vitro and in vivo. (A) Cells were grown in 2D culture in standard 96-well plates for 72h and assayed using alamar Blue®.  

Over 500 X-MAN™ isogenic cancer models with mutations in a wide variety of genes including K-RAS, PIK3CA, PTEN, IDH1 and p53 have been generated to date by Horizon, and many have been established as xenografts in immuno-deficient mice. Isogenic pairs are first characterized in vivo for their growth kinetics at the subcutaneous site to ensure the lines generate robust tumor growth, with the matched tumors differing only in the genetic composition of target gene.

This is particularly powerful as traditional xenograft models, while they may have a specific mutation of interest, don’t have the archetype wild-type (WT) as the comparative cell line control. Researchers are limited to using cell lines that harbor other background mutations or drivers as controls, and this can complicate interpretation of the outcome.

K-Ras is the most frequently mutated oncogene in solid tumors, with up to 40 percent of colorectal cancer patients harboring a K-Ras mutation. However, the identification of the critical effectors of K-Ras-mediated tumorigenesis and the development of clinically effective therapeutic K-Ras-directed strategies has remained challenging, with K-Ras mutant cancers being among the most refractory to current treatments such as Cetuximab and Panitumumab. There is, therefore, an acute clinical need to identify pharmacologically tractable components of K-Ras driven tumorigenesis.

In general, colon cancer cell lines available for research are broadly classified as K-Ras-dependent or K-Ras-independent, and many differential gene signatures and profiles have been established to distinguish these two groups using panels of such cell lines. However, such profiles are often based on in vitro data which may not be representative of differences seen at the clinical level. For example, the colorectal cell line DLD-1 contains an endogenous heterozygous K-Ras G13D mutation. Using GENESIS gene engineering, the mutant K-Ras allele was knocked-out. This, along with its parental, then offers the ideal K-Ras-dependent/independent model to interrogate the biological effectors in this oncogenic pathway with little difference between the cell lines except for activated K-Ras pathway signalling.

In vitro, the isogenic pair of cells grows with similar rates, showing little dependency on the K-RAS G13D allele (Figure 1). However in vivo this is not the case. In a proof of concept study, Crown Bioscience has shown that knocking-out the K-Ras G13D allele in the DLD-1 colorectal cancer cell line leads to low tumorigenicity in nude mice compared to the wild-type cell line. This clearly demonstrates the addiction to the oncogenic K-Ras mutation in this cell line and the advantage of modelling these lines in vivo.

The same principle can be applied to tumor suppressor genes. The PTEN gene is one of the most commonly lost/mutated tumor suppressors in human cancer occurring in in prostate, endometrium, thyroid, colon, lymphatic and breast tumors with high frequency. A host of alterations in the other genes in the PTEN pathway have also been shown to activate the PTEN/PI3K network in tumors.

click to enlarge
Figure 2: Growth of HCT116 wild-type (parental) and PTEN null (-/-) derived tumors in vivo.  

Investigating the loss of PTEN phenotype is of particular importance, as activation of the PI3K pathway, which can be caused by PTEN loss, has been linked to resistance to targeted therapies including Herceptin® and Iressa™, and in colon cancer predicts a poor response to Erbitux®. Horizon has generated a homozygous knock-out of PTEN by deleting exon 5 which encodes the active site of the protein in the colorectal cancer cell line HCT116. In vivo, evaluation of the isogenic pair shows slightly increased tumor growth of the PTEN(-/-) relative to the wild-type cell line (Figure 2). This pair may provide an ideal model to screen for the next generation of compounds that may overcome resistance mechanisms induced by loss of PTEN.


Whilst the subcutaneous xenograft model is easy to monitor and quantify, it is nevertheless heterotopic (i.e. out of the native). Although some tumor microenvironment pressures such as pH and hypoxia can be mimicked, invasion, tumor architecture and vascular characteristics of the tumor are not optimally recapitulated. Orthotopic tumor models present a degree of complexity superior to classical subcutaneous models, due to their exhibiting higher vascular volume and permeability. The implantation of tumor cells in their original tissue therefore allows the development of models which are more comparable to clinically presented tumors, allowing better prediction of the therapeutic effect of targeted agents.

Orthotopic models can also lead to the production of metastases which are not seen in subcutaneous xenografts. Metastasis continues to be the most significant problem in cancer, with a poor prognosis for patients who present with metastatic disease or those who develop metastases after successful management of the primary tumor. The invasive and metastatic stages of cancer progression are a complex process involving resistance and epithelial-mesenchymal transition (EMT). A broader understanding of the biology of metastases is necessary to better design targeted treatments, with the biological complexity that characterizes metastasis requiring relevant animal models. The phenotypic changes involved in metastasis are often recapitulated in orthotopic in vivo models, facilitating the characterisation of emerging targets in drug discovery.

A high level of skill is required for the initiation of surgically-induced orthotopic and metastatic tumor models. The incidence of establishment as well as propensity to metastasise may vary between tumors. Orthotopic implantation of colorectal lines, for example, requires injection of a cell inoculum or transplantation of tissue fragments into the caecum, resulting in local invasion and metastasis to the liver. Similarly, intra-thoracic injection of cells into the deep alveolar bed of lung tissue allows establishment of tumors representative of lung adenocarcinomas.

Patient-relevant orthotopic and metastatic models, such as those offered by Crown Bioscience, therefore offer a valuable tool in the study of metastasis. Using Horizon’s precisely engineered isogenic cell line pairs at the orthotopic site provides the most relevant model and environment to study key mutations and test targeted agents.



click to enlarge
Figure 3: An orthotopic model of NSCLC can be imaged in real-time using the IVIS® Spectrum CT imaging system.  

Therapeutic read-out from orthotopic and metastatic sites can be achieved through the optical imaging of engineered fluorescent or luminescent cell lines (Figure 3). In some models, such as NSC lung cancer and in the caecum in colon cancer, resection of the primary tumor can be performed to allow disease progression and metastases to be tracked real-time. Tracking the key stages in the progression of cancer could allow the optimal timing of drug administration to be determined based on the biological signals measured. In some models, such as colon cancer, resection of the primary tumor may also be possible allowing secondary sites to outgrow which could be monitored in real-time.

Introduction of such detection technologies into isogenic cell line pairs can provide a mechanism by which to fully utilise these models in orthotopic and metastatic studies. Horizon is in the process of incorporating luciferase into a number of its isogenic cell line pairs, allowing any variation in growth to be monitored in real-time, and progression to be tracked at orthotopic and metastatic sites. Additionally, and in collaboration with Promega, they have generated several cell lines that incorporate Promega’s NanoLuc™, a very bright luciferase, as a reporter of endogenous promoter activity or protein dynamics of key oncology-relevant genes. These reporter cell lines have the potential to enable the monitoring of a diverse range of biological processes and pathways involved in the tumor and its microenvironment - from proliferation to DNA damage and hypoxia.

A diverse array of genetic mechanisms drives cancer, resistance, metastasis and the eventual failure of treatment. The explosion of knowledge and expansion of our understanding of these genetic drivers of human cancer represents a significant shift towards a personalized medicine approach using targeted therapies. In order to translate this knowledge into the clinic there is a pressing need to model, in the preclinical setting, these fundamental genetic mechanisms as a key step towards developing therapeutic strategies, biomarker identification and stratification of patient recruitment. The unique and clinically-relevant disease models developed by Horizon Discovery and Crown Bioscience offer powerful tools for the identification of new targets, alongside development of first in class novel oncology targeted agents for early stage drug discovery programs.