Safety, efficacy and enhanced transgene expression are the primary concerns while using any vector for gene therapy. One of the widely used vectors in clinical trials is adenovirus which provides a safe way to deliver the therapeutic gene. However, adenovirus has poor transduction efficiency in vivo since most tumor cells express low coxsackie and adenovirus receptors. Similarly transgene expression remains low, possibly because of the chromatization of adenoviral genome upon infection in eukaryotic cells, an effect mediated by histone deacetylases (HDACs). Using a recombinant adenovirus (Ad-HSVtk) carrying the herpes simplex thymidine kinase (HSVtk) and GFP genes we demonstrate that HDAC inhibitor valproic acid can bring about an increase in CAR expression on host cells and thereby enhanced Ad-HSVtk infectivity. It also resulted in an increase in transgene (HSVtk and GFP) expression. This, in turn, resulted in increased cell kill of HNSCC cells, following ganciclovir treatment in vitro as well as in vivo in a xenograft nude mouse model.
Among the several cancer gene therapy strategies the prodrug activation strategy using herpes simplex virus thymidine kinase gene (HSVtk)/ganciclovir (GCV) has been widely used.1 Owing to its high affinity for HSVtk, GCV can kill HSVtk expressing cells. Cell kill is further enhanced by a phenomenon known as “bystander effect” which results in apoptosis of neighboring cells2 presumably due to the toxic metabolite GCV-triphosphate passing through gap junctions into neighboring cells3, 4 or phagocytosis of apoptotic vesicles generated from dying cells2, 4–6 thereby interfering with the DNA synthesis.
Human adenovirus Serotype 5-based vectors have been used extensively to deliver therapeutic genes in a large number of clinical trials.7 Features like high efficiency of gene transfer, ability to infect nonmitotic cells, and the established general safety in preclinical and early clinical investigations, give adenoviruses a great advantage over other vectors.8 Since recombinant adenovirus also remains as a nonreplicating extra-chromosomal entity, there are remote chances of it activating a dormant oncogene or interrupting a tumor suppressor gene.
Adenovirus enters the host cell by a receptor-mediated endocytosis mechanism.9 Attachment of the virus depends on its fiber protein binding to a primary tethering receptor known as coxsackie and adenovirus receptor (CAR).10–12 In addition it requires co-receptors which are αvβ3- and αvβ5- integrins which promote internalization of the virus.13, 14 CAR levels vary across the epithelial cell lines and so does the adenovirus infectivity.15 It has been shown that head and neck squamous cell carcinoma (HNSCC) cell lines of the same or similar histological characteristics have significant variations in the transduction efficiency of adenoviruses.16
Acetylation of histone-like adenoviral proteins have been reported.17, 18 It may be speculated that like the eukaryotic genome, regulation of adenoviral genes can be influenced by the compaction or opening of chromatin, an effect mediated by histone acetyl transferases (HATs) and histone deacetylases (HDACs).19 HDACs are enzymes which alter the chromatin structure by deacetylating histones and result in repression of gene expression by inhibition of transcription. Hence compounds that not only augment but also cause sustained gene expression by viral vectors via alterations in the chromatin structure are of interest to researchers.20, 21
Various groups have demonstrated increased infection and transgene expression from viral vectors on treatment with inhibitors of HDACs (HDACi).22–24 HDACi like phenyl butyrate has been reported to modulate glial fibrillary acidic protein and Connexin 43 expression and enhance gap-junction communications in glioblastoma cells.25 There is also a report that phenyl butyrate increases bystander killing of HSVtk transfected glioma cells in vitro via increase in gap junctions.26 Another group has reported that depsipeptide (FK228) an HDACi, enhances adenoviral infection in rat renal cancer allograft model systems.27 Valproic acid (VPA), an anti-epileptic drug and a potent HDACi as well as an anticancer drug, has been shown to increase viral infectivity as well as transgene expression in cell lines as well as in vivo.28 Another HDACi panobinostat has been used in a clinical trial in 10 patients with cutaneous T-cell lymphoma and shown to induce clinical response.29
The present study was carried out to investigate whether VPA can result in increased infection of replication deficient, Serotype 5-based, recombinant adenoviruses carrying HSVtk, in a HNSCC cell line NT8e, cause elevated transgene expression and finally result in enhanced cell kill in vitro and in vivo.
Ad-HSVtk: recombinant adenovirus carrying herpes simplex virus thymidine kinase gene; CAR: coxsackie and adenovirus receptor; GFP: green fluorescence protein; GFU: green fluorescence unit; HDACi: histone deacetylase inhibitor; HNSCC: head and neck squamous cell carcinoma; qRT-PCR: quantitative real time PCR; VPA: valproic acid
Material and Methods
Cell culture and xenografts
Head and neck squamous cell carcinoma cell line NT8e, established in our laboratory30 was used for in vitro as well as in vivo experiments. NT8e cells were maintained in DMEM containing 10% FCS. To obtain xenograft tumors 1 × 107 NT8e tumor cells in 100 μl hanks balanced salt solution were injected subcutaneously on the dorsal right flank of the mouse. The tumor volumes were calculated as described earlier31 using the formula (a × b2)/2, where a, b are 2 longest perpendicular diameters.
HEK293T (human embryonic kidney) cells were used for adenoviral preparation. The cells were maintained in DMEM containing 10% FCS.
Preparation of recombinant Ad-HSV TK
Human adenovirus, Serotype 5 was used in the present study. Recombinant adenovirus carrying the thymidine kinase gene from herpes simplex virus (HSVtk) and green fluorescence protein (GFP) under 2 separate CMV promoters was prepared using AdEasy system (kind gift from Dr. B. Vogelstein) by the strategy given by He et al.32 Briefly HSV-TK gene was excised as a 1.8 kb BamHI fragment from pTK26 (kind gift from Dr. S. Rhode)33 and cloned in pAdTrack-CMV. The resultant plasmid (pAdTrack-HSVtk) was linearized with PmeI and cotransformed with the adenoviral backbone vector pAdEasy-1 into the BJ5183 E coli cells by electroporation and spread over plates containing kanamycin. Colonies thus obtained were analyzed by restriction digestion with EcoRI and positive clones were further analyzed by PacI digestion.
Recombinant TK adenovirus (pAd-HSVtk) was then transformed into E.coli XL-1 Blue cells for large-scale amplification. The PacI-digested pAd-HSVtk was transfected into HEK-293T cells and the HSVtk adenovirus was expanded, purified and tittered as described earlier.32 Crude stock of viruses was prepared by lysing the infected cells by freeze thaw cycles and this was used to infect HEK293 cells in bulk. Cells were collected, lysed in 5 ml of Tris Buffer (10 mM Tris, pH 8) and centrifuged. The supernatant was loaded on CsCl density gradient and centrifuged at 32,000 rpm/16°C/1 hr. Virus band was carefully removed from the tubes and resuspended in 2× glycerol buffer (10 mM Tris pH 8, 100 mM NaCl, 0.1% BSA, 1 mM mgCl2, and 50% glycerol). Viral titers were calculated by infecting serially diluted virus particles in HEK293 cells, and analysis of GFP positive cells by flow cytometry 24-hr post infection by the method described by Gueret et al.34 using the following formula:
*GFU is green fluorescence units used as a unit for measuring infective viral particles.
In vitro studies
The 2 × 103 NT8e cells were plated per well in 96 well plates and treated with 1 mM VPA. One-day post VPA treatment, cells were infected with Ad-HSVtk at 3 × 104 GFUs. At 24-hr post virus infection the prodrug, Ganciclovir (GCV, Roche pharmaceuticals, Switzerland) was added to the wells at 1 μg ml−1 or 2 μg ml−1 final concentration, and cells were further incubated for 3 days. Viability was measured using Sulphorhodamine B (SRB) assay. Cells were fixed with 50% TCA, stained with SRB and the O.D. readings taken at 515 nm.
For terminal dUTP nick end labeling assay NT8e cells were plated on cover slips in petridishes and treated with VPA/Ad-HSVtk/GCV or Ad-HSVtk/GCV as described above. One day post treatment cells were fixed and TUNEL assay performed using DeadEnd™ Fluorometric TUNEL System, (Promega, Medison, CA) following manufacturer's protocol. Images were captured on Laser Confocal Microscope (Zeiss, LSM 510META, Germany).
Confocal microscopy and flow cytometry
105 NT8e cells were plated in 60-mm plates and treated with 1 mM VPA. One day after VPA treatment, cells were infected with Ad-HSVtk at 5 × 105 GFUs. At 24-hr post virus infection, images of the cells were captured on Laser Confocal Microscope (Zeiss, LSM510META, Germany) and analyzed using LSM image analyzer. The fluorescence intensities were represented in the form of peak height and peak shifts. For analysis purposes, multiple fields from each plate were captured, and representative graphs are shown. To quantitate the effect of VPA on Ad-HSVtk infectivity, NT8e cells after treatment were trypsinized, washed once with PBS and subjected to flow cytometry using FACSCalibur (Becton Dickinson, San Jose, CA). GFP was detected at 530/30 nm. The data was analyzed using CellQuest software.
RT-PCR analysis and real-time quantitative RT-PCR (qRT-PCR)
Total RNA was isolated from cells using Trizol reagent (Invitrogen, Carlsbad, CA). cDNA synthesis was performed using SuperScript™ first strand cDNA synthesis kit (Invitrogen, Carlsbad, CA). The primers used for RT-PCR analysis are shown in Table 1.
Table 1. Primer sequences used in RT-PCR
Expression of the genes encoding CAR, Connexin 43 and viral thymidine kinase (HSVtk) was analyzed by qRT-PCR on ABI-Prism 7900 (Applied Biosystems, Foster city, CA). The final reaction mixture contained 20 ng cDNA in 2 μl and 2 μl (from 2.5-μM stock) primer pairs, using Syber green detection system (provided in 2× dynamo mix, Finnzyme) as a reporter, and ROX (Finnzyme) as a passive reference. Hot start Taq polymerase, dNTPs, PCR buffer and MgCl2 were provided in 2× dynamo mix. RNA was first checked for DNA contamination by carrying out PCR without Reverse Transcriptase, and then qRT-PCR was done. Dissociation curves showed very specific peaks for all the genes including internal control. Primer pairs used for qRT-PCR are shown in Table 2.
Table 2. Primer sequences used in qRT-PCR
Fold increase in the Connexin 43, and CAR transcripts in NT8e cells after VPA treatment was evaluated. Connexin 43 and CAR transcript levels in untreated NT8e cells were considered as basal, and fold increase after VPA treatment was calculated as log 2 ratio. Similarly, fold increase in HSVtk levels resulting from Ad-HSVtk infection with or without prior treatment of VPA was evaluated by qRT-PCR as described above. For HSVtk, transcript levels resulting from only Ad-HSVtk infection was considered as basal and fold increase in HSVtk levels resulting from VPA + Ad-HSVtk treatment was calculated as log 2 ratio. For quantification, expression of these transcripts was normalized to house keeping gene RPL35A.
In vivo studies
1 × 107 NT8e cells were injected in NMRI strain of female nude mice and treatment was started when the tumors reached 5–7 mm in size uniformly in all mice. The treatment protocol comprised of VPA (100 μl, 10 mg ml−1, 50 mg kg−1 body weight) given intraperitoneally 1 day prior to viral injection, followed by 2.5 × 108 GFUs Ad-HSVtk given intratumorally. This was repeated twice with a gap of 3 days. GCV (100 μl of 22 mg ml−1, 110 mg kg−1 body weight) was injected 1 day after viral injection and continued everyday for 8 days. Tumor sizes were measured 15 days after GCV treatment began. Tumor bearing mice were divided into following groups with at least 4 mice in control groups: NT8e (untreated control); Ad-HSVtk (virus control); VPA control; GCV control, VPA and GCV control, VPA and Ad-HSVtk control; and 7 mice in treatment groups i.e. Ad-HSVtk and GCV; Ad-HSVtk, GCV and VPA. For statistical analysis, Ad-HSVtk/GCV group was compared to NT8e control, while VPA/Ad-HSVtk/GCV group was compared to both NT8e control as well as Ad-HSVtk/GCV treatment group.
All statistical tests were done using SPSS software. For in vitro and in vivo studies, comparisons between treatment group and controls were done using student's paired t test.
Effect of VPA treatment on infectivity of Ad-HSVtk in NT8e cells
To be able to detect differences in infectivity of Ad-HSVtk after VPA treatment, a lower titer of Ad-HSVtk that could infect about 50% of NT8e cells was used for the experiments. To check the effect of VPA on Ad-HSVtk infectivity, 105 NT8e cells were plated in 60-mm tissue culture plates and treated with or without VPA. After 24 hr, cells were infected with 5 × 105 GFU of Ad-HSVtk and examined under a Laser Confocal microscope with GFP as a read out. The infection in terms of GFP positivity in presence of VPA was higher as compared to virus alone (Fig. 1a). The same cells were further quantified using flow cytometry. While infection of NT8e cells with Ad-HSVtk alone resulted in 44.70% GFP positive cells, prior treatment with 1 mM VPA resulted in 79.91% positivity (Fig. 1b). Moreover the transgene expression, as determined by mean fluorescence intensity was 538.07 U which on treatment with 1 mM VPA increased to 1279.25 U.
Effect of VPA treatment on gene expression
Semi quantitative RT-PCR was carried out first to check expression of genes involved in binding and internalization of adenoviruses on cell membrane including CAR and αv integrin genes, as well as genes involved in gap junctions including Connexin-26 and Connexin-43 on NT8e cells in the presence of VPA. Transgene expression in VPA treated, Ad-HSVtk infected NT8e cells, was also studied. An increase in expression of CAR, Connexin-43 and HSVtk transgene was observed in NT8e cells after VPA treatment (Figs. 2a and b). There was a marginal increase in the expression of Connexin-26 (Fig. 2b). Expression levels of αv integrin, a co-receptor for adenovirus remained unaltered (Fig. 2b). Further quantification of the fold increase in Connexin-43, CAR and HSVtk expression was done by real time quantitative RT-PCR (Fig. 2c). VPA treatment resulted in increase in Connexin-43 (log 2 ratio 2.1) and CAR (log 2 ratio 1.3) transcripts. Ad-HSVtk infection in VPA-treated NT8e cells resulted in enhanced viral thymidine kinase gene expression (log 2 ratio 2.48) compared to Ad-HSVtk infection in absence of VPA treatment (Fig. 2c).
Effect of VPA on tumor cell kill in Ad-HSVtk infected cells in vitro
The effect of VPA treatment on HSVtk/GCV mediated tumor cell kill in vitro was studied. NT8e cells were treated with VPA and infected with Ad-HSVtk followed by GCV treatment as described above. While 62.77% cells were viable in Ad-HSVtk/GCV treated cells, the viability reduced to 24.1% in VPA-treated cells (Fig. 3) indicating a 2.5-fold higher cell kill in the presence of VPA. No significant increase in the cell death was observed when GCV concentration was increased to 2 μg ml−1. Enhanced cell kill was also apparent as indicated by TUNEL assay, with apoptotic cells visible in VPA/Ad-HSVtk/GCV treated cells, compared to only Ad-HSVtk/GCV treated cells (Fig. 4). Absence of apoptotic cells in only VPA-treated cells clearly indicated that 1 mM concentration is not toxic to the cells. Cell cycle profile of VPA-treated cells was similar to untreated cells (Supporting Information Fig. 1s), suggesting that this concentration was nontoxic and did not alter the cell cycle. At higher concentrations, VPA arrested the cells in G1 phase (data not shown).
Effect of VPA on Ad-HSVtk/GCV-mediated tumor cell kill in a xenograft model
Tumor bearing mice were divided into different treatment groups as described in Material and methods. Tumor size was measured and expressed as a ratio of tumor volumes after completion of the treatment and before starting the treatment. In case of Ad-HSVtk/GCV group there was no change in the tumor volume at the end of the treatment indicating that the treatment had suppressed tumor growth compared to control mice (1.1 ± 0.6, p = 0.01) (Fig. 5, Supporting Information Fig. 2s). In VPA/Ad-HSVtk/GCV group treatment resulted in a significant regression of tumors (0.3 ± 0.1, p = 0.01) as compared to untreated control (5.5 ± 1.3). Further we observed a significant decrease in ratio of volumes between Ad-HSVtk/GCV- and VPA/Ad-HSVtk/GCV-treated groups (p = 0.01) (Fig. 5). Mice treated with VPA alone did not show tumor regression. Similarly, Ad-HSVtk alone, GCV alone or VPA in combination of Ad-HSVtk without GCV did not show any tumor regression. The reduction in tumor volumes in the VPA/AdHSVtk/GCV group was significant when compared to both untreated control and Ad-HSVtk/GCV treatment groups (Fig. 5).
Although gene therapy studies in animal models have been quite promising, in human clinical trials they have failed in most cases. This has resulted in reassessment of the various gene therapy strategies. According to Gottesman35 there is a need to address 3 major recurring themes in cancer gene therapy: strategies to kill or slow down growth of cancer cells; development of vectors and delivery systems; translation of preclinical studies into clinical protocols. One of the potential strategies to kill tumor cells is to introduce the prodrug activation gene such as HSVtk into the tumor cells followed by the prodrug GCV.36, 37 This strategy has the added advantage of “Bystander effect” wherein neighboring cells are killed in addition to the cells transduced with AdHSVtk. However, efforts to increase transgene expression, target the transgene to maximum number of tumor cells and increase bystander effect are required. Adenoviral vector system has been used in majority of clinical trials (www.wiley.co.uk/genmed/clinical). High degree of transduction of the virus in vivo and optimal expression of therapeutic gene inside the tumor is the key to the success of gene therapy approaches.
Several clinical studies utilizing HSVtk suicide gene therapy have been reported showing safety of the gene therapy treatment and improvement in quality of life in small number of brain tumor patients.38 However, the only Phase III trial where retroviral vector producing cells carrying HSVtk were injected intratumourally into brain tumors showed complete lack of benefit to patients.39 Possible reasons for treatment failure were due to limited transduction of viral vectors and low expression of thymidine kinase gene. Various approaches have been designed to boost the HSVtk gene expression and its effect on neoplastic cells. Introducing mutation in HSVtk has been shown to increase its activity by increasing the Km for thymidine binding and reducing the competition between thymidine and ganciclovir.40, 41 Use of compounds that increase the Gap junctions has been shown to augment the tumor cell kill by GCV in HSVtk transduced cells.42 Hydroxyurea, an inhibitor of ribonucleotide reductase, has been shown to increase bystander effect in cells in the absence of gap junctions, by decreasing the endogenous dGTP pool and thereby decreasing competition with GCV-triphosphate for DNA incorporation.43
Various HDAC inhibitors have been reported to increase CAR levels44 as well as enhance transgene expression,28, 45 which makes them suitable for use in conjunction with adenoviral vector-based therapies. Valproic acid, an HDACi, which is used to treat mood disorders and epilepsy has been shown to display potent antineoplastic46 and antiangiogenic47, 48 activities. Reports relating to the effect of VPA in tumor regression when used with Ad-HSVtk/GCV-mediated suicide gene therapy in vivo are inadequate. Here we demonstrate that VPA treatment prior to adenovirus infection in a HNSCC cell line NT8e results in enhanced infection, transgene expression, as well as tumor cell kill in vitro and in vivo. Fan et al.28 have demonstrated enhanced gene expression from viral gene transfer vectors in the presence of VPA in vivo at the site of injection as well as in the liver. In the present study Ad-HSVtk was injected intratumourally and no toxicity was observed although expression of the transgene in the liver was not determined.
A dose of 1 mM VPA delivered to NT8e cells was well tolerated and did not induce apoptosis, as determined by flow cytometry and TUNEL analysis. The 1 mM VPA concentration has been reported to be nontoxic and closest to in vivo therapeutic range.28 However, it has been reported that Trichostatin A, an HDACi can relieve repression of transduced viral genes in cell lines.49 Whether VPA can also relieve expression of transduced viral genes remains to be determined. Treatment with 1 mM VPA increased CAR expression in NT8e cells. This was also reflected by enhanced infectivity of recombinant adenoviruses (Ad-HSVtk). Further, VPA also increased the transgene expression (GFP and HSVtk) which could be due to improved infectivity of the virus. An alternate explanation for increase in transgene expression could be that HDAC inhibitors exert their “chromatin opening” effect even on adenoviral genome, thus allowing more transcription factors binding to the regulatory elements and increase in the transgene expression. Histone like proteins in adenoviral genomes are shown to be regulated by acetylation events similar to eukaryotic genome.17
A higher expression of GFP compared to HSVtk was seen which may be due to promoter competition since both genes were cloned under 2 separate CMV promoters. However, densitometry analysis (data not shown) as well as qRT-PCR revealed a clear increase in the expression of HSVtk transcripts. This increase in HSVtk expression was sufficient to result in increased cell death after GCV treatment in vitro and in vivo.
There was also an increase in Connexin 43 transcripts which is known to be associated with GAP junctions between the cells.50, 51 Gap junctions play an important role in mediating bystander effect.50, 52 Treatment with VPA in vitro was able to enhance the Ad-HSVtk-mediated cell kill significantly when compared to Ad-HSVtk/GCV treatment alone. This could be attributed to increase in gap junctions resulting in increased bystander effect and passage of toxic metabolites to neighboring cells.
Thus VPA treatment appears to affect the viral genome as well as the host genome and favor tumor cell kill. VPA not only increases infectivity of adenovirus to tumor cells and increases transgene expression but probably also increases gap junctions on the tumor cells thereby increasing bystander-mediated cell kill. VPA has been safely used in the long-term therapy of patients with epilepsy over decades. In a dose escalating Phase 1 study in cancer patients, VPA at 60 mg kg−1 body weight per day for 5 days was well tolerated.53 All these findings together suggest that combining valproic acid with Ad-HSVtk/GCV could be more effective even at lower virus concentrations, in human cancer gene therapy clinical trials.
V.K. was recipient of a fellowship from Council for Scientific and Industrial Research, India.