CMV enhancer may not be suitable for tissue-specific enhancement of promoters in cancer gene therapy

Abstract

Gruh et al. [1] reported that when the human cytomegalovirus (CMV) immediate-early enhancer was integrated into lentiviral vectors before tissue or cell-specific promoters human atrial natriuretic factor (ANF), human ventricular myosin light chain (MLC2v), or type II alveolar epithelial cell (AT-2)-specific human surfactant protein C (SP-C), it could confer efficient cell-type-specific gene expression in cardiomyocytes, lung AT-2 cells, and nontarget cells. Based on their work, the authors claimed that lentiviral vectors containing the human CMV enhancer may be promising in gene therapy. With great interest, we were attracted by this article which introduced a much more simple but potent transcriptional enhancement approach to overcome the low efficiency limitation of most tissue/cell-specific promoters used in current gene therapy vectors. We used an adenoviral vector, which infection is less selective, in our work of cancer gene therapy, and the expression of gene of interest must be restricted in cancer cells under the control of tumorspecific promoters to ensure the therapeutic efficacy and safety. We adopted the same sequence of the human CMV enhancer as reported by Gruh et al. [1] to enhance the transcription efficiency of two most widely used tumorspecific promoters, carcinoembryonic antigen (CEA) promoter, and alpha-fetoprotein (AFP) promoter. Luciferase reporter vectors were constructed in pGL4.10[luc2] and cotransfected with reference vector pGL4.74[hRluc/TK] into CEA and AFP positive carcinoma cells and negative control cells respectively, and the dual luciferase assay system was used to evaluate the transcription efficiency. Our results showed that the CMV enhancer not only greatly increased CEA or AFP promoter activity in CEA or AFP positive cells, but also significantly increased promoter activity in CEA or negative control cells, which means that CMV enhancer could not maintain the tumor-specific transcriptional activity of CEA or AFP promoter [2, 3]. Since the ‘advanced’ generation of lentiviral vector that Gruh et al. used should not have any effect on exogenous gene expression specificity, we considered the structural differences between the promoters used in their work and our work. Gruh et al. [1] used a 564 bp sequence that covers the region of −473 to +91 bp of ANF gene (GenBank accession number K02043) for cardiomyocyte-specific expression. Previous studies showed that human and rat ANF genes share high homology in the proximal 5′-upstream sequences [4, 5]. Two highly-conserved regulatory elements, CArG[CC(A/T)6GG]/serum response element (SRE)-like sequences at −405/−406 and −114, respectively, confer basal and inducible cardiac muscle-specific transcriptional activity by interaction with serum response factor-related proteins [6–8]. These two upstream elements of the human ANF gene is sufficient to display atriumspecific transcription in cells cultured in vitro [9], and were included in the reported 564 bp sequence to restrain cardiomyocyte-specific transcription. MLC-2v gene exhibits cardiac ventricular cell specificity through a 250 bp MLC-2v promoter which depends on two adjacent positive regulatory elements (HF-1a and HF-1b/ MEF-2), as well as a negative regulatory element (HF-3) that acts to suppress expression in other muscle subtypes [9]. The 28 bp HF-1 element, which contains both HF-1a and HF-1b sites, is sufficient to confer cardiac cell-specific and inducible expression [10]. AP-1, AP-2, CArG/SRE, and MEF-2 regulatory motifs are found in the −92 to +12 The authors contributed equally: Jing Jia, Shuzhi Zhang