Gene therapy for colorectal cancer by an oncolytic adenovirus that targets loss of the insulin-like growth factor 2 imprinting system

Malignant neoplasms are responsible for nearly 7.5 million deaths representing 13% of all mortality and 5% of the global burden of disease in terms of disability-adjusted life years lost[1]. Colorectal cancer is the third most common cancer in men (10% of the total) and the second in women (9.4% of the total) worldwide[2]. Conventional chemotherapeutic agents, although often effective, are highly toxic because of their lack of selectivity for cancer cells. As a result, efforts have focused on developing interventions that target tumor-specific genes using techniques to construct tumor-selective replicating viruses.

Genomic imprinting is an epigenetic modification of a gene based on its parental origin, which results in monoallelic expression[3]. For most genes, both paternal and maternal alleles are expressed, but for some genes, only one allele is expressed, which is known as maintenance of imprinting (MOI). In contrast, loss of imprinting (LOI) is the reactivation of the silenced allele of an imprinted gene, leading to perturbation of monoallelic expression. LOI is closely related to the occurrence of malignant tumors[4]. The first endogenous imprinted gene identified was mouse insulin-like growth factor 2 (IGF2) encoding a critical fetal-specific growth factor that is only expressed from the allele inherited from the father[5]. IGF2 and H19 genes are reciprocally imprinted, and regulated by the enhancer, DNA differentially methylated domain (DMD), and promoter. The current model for the mechanism underlying this reciprocal imprinting proposes a critical role for CCCTC binding factor (CTCF) that binds to the unmethylated maternal DMD. Binding of target sequence elements by CTCF can block the interaction between enhancers and promoters, thereby limiting the activity of enhancers for certain functional domains. In recent years, researchers have elucidated the important roles of methylated regions and CTCF complexes that combine with the unmethylated maternal DMD of IGF2 imprinting[6‐9]. Hu et al.[6] transferred nuclei from human tumor cells to show loss of IGF2 imprinting in enucleated mouse and human fibroblasts that had maintained normal IGF2 imprinting. After nuclear transfer, the abnormal tumor epigenotype was corrected by in vitro reprogramming, suggesting that LOI is associated with the loss of activity of non-CTCF trans-imprinting factor(s) that are either inactivated or mutated in tumors. In addition to blocking enhancers, CTCF can act as a chromatin barrier by preventing the spread of heterochromatin structures. In mice, it has been shown that CTCF serves as a strategic protein that implements DNA loops[7, 8] and helps to silence DNA regions by binding and recruiting the polycomb repressive complex 2[9]. Recently, Zhang et al.[10] extended the examination of the role of CTCF in orchestrating long-distance intrachromosomal looping in the human IGF2/H19 imprinting domain, and demonstrated that interruption of intrachromosomal looping by CTCF decoy proteins abrogates genomic imprinting of human IGF2. This observation confirmed that the inactivation or mutation of CTCF complexes is closely associated with the epigenetic IGF2 LOI in tumor cells. In our previous study, we found that loss of IGF2 imprinting is present in both colorectal tissues and cancer cell lines (HCT-8, HT-29, HCT15, and SW1222). No such phenomenon has been observed in normal tissue cells[6]. Furthermore, we successfully confirmed targeted tumor gene therapy based on IGF2 LOI[11].

Adenoviruses rapidly infect a broad range of human cells, including proliferating and quiescent cells, which tend to yield high levels of adenoviral particles[12]. In addition, adenoviral vectors are relatively easy to manipulate, purify and manufacture using recombinant DNA techniques, and transgenes generally do not undergo changes during successive rounds of viral replication. On the other hand, adenoviral vectors have poor targeting toward tumors and a low efficiency of gene transfer[13]. Abnormal biological characters shared by multiple cancers have enabled optimization of genetically engineered viruses that replicate exclusively in tumor cells (termed replication-selective adenoviral vectors), while not affecting normal cells. Adenovirus H101 has been clinically approved in China for the treatment of several malignancies. E1B and parts of the E3 regions have been deleted in this virus that cannot normally replicate and only lyses cells in which p53 is inactive. The cell cycle gatekeeper p53 is commonly observed to be inactive in tumors. Our present study includes four cell lines that have various IGF2 imprinting and p53 mutations (Table1), and infection with H101 serves as the positive control.

In our previous study, a recombinant replication-defective adenovirus carrying the IGF2 imprinting system and the DT-A gene was successfully constructed, in which Ad-DT-A effectively kills tumor cells showing IGF2 LOI[11]. Replication-defective adenoviral vectors cannot continuously renew after initial activation, resulting in only transient expression of the DTA gene in LOI colorectal cancer cells.

One of the critical functions of the human adenovirus early region 1A (E1A) protein is to activate transcription of early viral genes[14]. This region activates viral gene expression by interacting with and recruiting cellular transcription machinery to the regulatory regions of early viral genes. Recent evidence indicates that the Ad5 E1A gene has a tumor suppressor role and exhibits a dual anti-tumor effect[15]. The E1A protein inhibits over-expression of erbB2 and protease genes, blocks NF-Kβ activity and then, as an antitumor effect, increases the expression of nm23 and p53 proteins[16]. Furthermore, E1A protein induces cell cycle progression from G0/G1 to S phase by binding and repressing the function of intracellular proteins such as pRb and p300, leading to cancer cell lysis by massive up-regulation of viral replication[17]. In this study, we constructed a novel replication-selective adenovirus based on the IGF2 LOI system to target colorectal cancers.

In the present study, we successfully constructed a novel replication-selective adenovirus, Ad315-E1A, and a replication-deficient adenovirus, Ad315-EGFP. Our data show a significant effect of Ad315-E1A against tumor growth both in vitro and in vivo. This observation suggests that replication-selective adenoviruses driven by the IGF2 imprinting system may be used as a novel anti-cancer agent with a high therapeutic potential.

Using viruses as a genetic carrier offers an attractive alternative or promising integration with traditional therapeutic regimens for cancer treatment[18]. Among viral-based therapies and oncolytic agents, adenoviruses have emerged as a promising vector that is already being used for the treatment of solid tumors in humans[19]. Adenoviruses are non-enveloped viruses with linear double-stranded DNA, which infect cells by binding to the coxsackie and adenovirus receptor expressed on the surface of target cells[20]. As vectors for oncolytic therapies, these viruses have many advantages over other vectors, including the capability of transducing and replicating in dividing and non-dividing cells, their ease of manipulation, and a naturally lytic replication cycle[21‐24]. All of these features highlight the utility of adenoviruses for in vitro production and in vivo curative effects[25]. Since 1993, more than 300 clinical trials based on adenoviral vectors have been performed[20] with promising outcomes. The first clinical results of trials based on adenoviruses as an oncolytic therapeutic agent have been promising, and show the clinical safety and feasibility of this approach[26]. In our study, we constructed an oncolytic adenoviral vector and demonstrated its effective killing of cancer cells both in vitro and in vivo.

IGF2 encodes the mitogenic anti-apoptotic peptide IGF-II that is overexpressed in several human cancers cells. In colorectal cancer, it has been reported that LOI of IGF2 occurs in 44% of informative colorectal cancers that are linked to microsatellite instability[27]. Additional studies have reported a range from 33% to 87% of sporadic colon cancers that were related to LOI of IGF2[28, 29].

Based on these findings, we theorized that we could build a kind of oncolytic adenovirus specific for replication in LOI tumor cells for colorectal cancer therapy. In this proof-of-concept study, we constructed adenoviral vectors, Ad315-E1A and Ad315-EGFP, which carry the three enhancer elements, DMD and promoter based on IGF2 LOI. Initially the utility of our expression system was tested using EGFP reporter assays. The results showed that green fluorescence was detected only in LOI cells. By increasing the multiplicity of infection and extending the time of infection, we found that the results were unchanged. To further demonstrate the selective efficacy of the LOI system, we used the oncolytic adenovirus H101, which has been clinically approved in China to treat malignancies, as a positive control for specific lysis of tumor cells by targeting the inactivated p53 in tumors. We used RT-PCR and western blotting to detect E1A expression at mRNA and protein levels, respectively, in Ad315-E1A-infected cells, and confirmed E1A expression in LOI colorectal cell lines. Furthermore, in H101-infected cells, E1A mRNA and protein was only detected in the p53 mutant cell line HT-29. All of the above results indicate that the replication-selective system that we used is practicable and effective.

In this study, the therapeutic potential of Ad315-E1A was tested by an MTT assay and flow cytometry, and the results clearly support the rationale of our hypothesis, in which Ad315-E1A effectively suppressed the growth of tumor cells and induced obvious cytotoxicity in LOI tumor cells. Therefore, we compared the therapeutic effect between IGF2 regulatory sequences and H101 for colorectal cancer treatment. The data showed that there was no detectable cytotoxicity in HCT-116 and GES-1 cells infected with Ad315-E1A or H101 (10 PFU/cell), whereas Ad315-E1A infection achieved significant growth inhibition of the IGF2 LOI and p53-active cell line (HCT-8) relative to that by H101 infection. In addition, both oncolytic viruses exhibited superior efficiency to inhibit growth of the HT-29 (p53 mutant and LOI) cell line. These results suggest that the IGF2 LOI system in our viral vector enabled efficient adenovirus replication in host cells using the E1A endogenous promoter.

Using the human colorectal cancer cell line HCT-8 as a model, we showed that this therapeutic strategy significantly increased cancer cell death and apoptosis caused by virus replication in vivo. Despite Ad315-E1A treatment significantly inhibiting tumor growth and prolonging the survival of tumor-bearing mice, the tumors were not eliminated completely. Previous research also found that oncolytic viruses have a limited potential to eradicate tumors when used as a monotherapy[30‐32]. Thus, oncolytic viruses are often used in combination with other modalities, such as chemotherapy[33], radiotherapy[34], or arming oncolytic adenoviruses with therapeutic genes[35], to improve the anti-tumor efficacy. In our next study, we will investigate appropriate methods to enhance the potency of the novel recombinant adenovirus for cancer gene therapy.

In summary, our results show that an oncolytic adenovirus regulated by the IGF2 LOI system confers a significant anti-tumor effect by induction of apoptosis in vitro and in vivo in human colorectal cancer cells. LOI exists in a wide variety of tumors, and although more investigations are required for future use, our preliminary data support the use of oncolytic viruses in the context of the IGF2 LOI system as a novel approach for cancer therapy.