Molecular Biotechnology

, Volume 37, Issue 1, pp 38–47

The Rational Design of β Cell Cytoprotective Gene Transfer Strategies: Targeting Deleterious iNOS Expression

Authors

  • Cillian McCabe
    • Regenerative Medicine Institute, National Centre for Biomedical Engineering SciencesNational University of Ireland
    • Department of MedicineNational University of Ireland, Galway
    • Regenerative Medicine Institute, National Centre for Biomedical Engineering SciencesNational University of Ireland
    • Department of MedicineNational University of Ireland, Galway
Original Paper

DOI: 10.1007/s12033-007-0049-6

Cite this article as:
McCabe, C. & O’Brien, T. Mol Biotechnol (2007) 37: 38. doi:10.1007/s12033-007-0049-6

Abstract

Islet transplantation represents a promising therapeutic strategy for the treatment of type 1 diabetes mellitus (T1DM) [Hakim and Papalois (Ann Ital Chir 75:1–7, 2004); Jaeckel et al. (Internist (Berl) 45:1268–1280, 2004); Sutherland et al. (Transplant Proc 36:1697–1699, 2004)]. The insulin-secreting pancreatic β cells of the islet allograft are, however, subject to recurrent immune-mediated damage. Principal among the molecular culprits involved in this destructive process is the proinflammatory cytokine IL-1β. IL-1β-induced β cell destruction may be mediated by the generation of NO and/or ROS, although the relative importance of NO and ROS in this process remains unclear. This study broadly encompassed three arms of investigation: the first of these was geared toward the establishment of a robust in vitro cell system for the study of IL-1β-induced pathophysiology; the second arm aimed to provide a comparative analysis of the gene transfer profiles of the three most commonly used gene transfer vehicles, namely plasmid vectors, adenoviral vectors, and lentiviral vectors, in the aforementioned cell system; the final arm aimed to screen an array of potentially cytoprotective gene transfer strategies incorporating the optimal gene transfer vectors. Briefly, we established an in vitro β cell system that accurately reflected primary β cell cytokine-induced pathophysiology. That is, IL-1β exposure (100 U/ml) induced a time-dependent decrease in rat insulinoma (RIN) cell viability, which coincided with an induction in iNOS expression and nitrite accumulation. Gene transfer studies using plasmid, adenoviral, or lentiviral vectors underscored the superiority of viral vector-based gene transfer strategies for the manipulation of this β cell line. Using these vectors, we provide evidence that NF-κB-based iNOS inhibition confers significant protection against IL-1β-induced damage whereas antioxidant overexpression fails to provide protection. Conferred cytoprotection was associated with a suppression of iNOS expression and nitrite accumulation. From a therapeutic standpoint, gene transfer strategies employing efficient viral vectors to target iNOS activation may harbour therapeutic potential in preserving β cell survival against proinflammatory cytokine exposure.

Keywords

β CellsCytokinesiNOSROSAntioxidantsNFκBGene transferCytoprotectionAdenovirusshRNALentivirusRat insulinoma cellsCytokinesIL-1β

Introduction

Islet transplantation represents a promising therapeutic strategy for the treatment of type 1 diabetes mellitus (T1DM) [13]. The insulin-secreting pancreatic β cells of the islet allograft are, however, subject to immune-mediated damage. Principal among the molecular culprits involved in this destructive process is the proinflammatory cytokine interleukin-1β (IL-1β). While the precise signal transduction pathways employed by IL-1β remain poorly understood, it is believed that the generation of nitric oxide (NO) and/or reactive oxygen species (ROS) ultimately culminates in β cell dysfunction and death [4, 5], although the relative roles of the former and the latter remain unclear. However, the activation of inducible nitric oxide synthase (iNOS) and the accumulation of NO represent hallmarks of IL-1β exposure that highlight the central role of free-radical attack in β cell destruction [6, 7]. In corroboration, previous data from our group has highlighted the cytoprotective utility of lentiviral vector-based RNAi strategies targeting deleterious iNOS expression in β cells exposed to proinflammatory cytokines [8]. Similarly, our group has previously demonstrated the cytoprotective potential of adenoviral vector-mediated overexpression of the NF-κB inhibitor, IκBα, in a β cell line exposed to cytokine-induced toxicity [9]. The fact that β cells exhibit heightened susceptibility to cytokine-induced oxidative damage [10] is considered a consequence of the poor antioxidant defense status of this cell type.

Gene transfer to pancreatic β cells may prove useful in preventing allograft cell destruction and prolonging islet graft survival after transplantation in patients with T1DM. Potentially, a host of therapeutically relevant transgenes may be incorporated into an appropriate gene delivery vehicle and used for cytoprotection. An increasing understanding of the molecular pathogenesis of immune-mediated β cell death has served to highlight molecules that have become suitable candidates for promoting β cell survival against proinflammatory cytokines [11].

To date, an array of non-viral and viral gene transfer vectors have demonstrated acceptable gene transfer profiles in both purified β cells and intact islets [11]. Viral vectors have offered the most efficient means of gene delivery to β cells given their quiescent status. Successful application of gene therapy to β cell modification in the clinical setting is currently limited by the availability of a single, ideal gene delivery vehicle. However, for the purposes of pilot studies aimed at evaluating the cytoprotective potential of candidate molecules, plasmid lipofection, adenoviral vectors, and lentiviral vectors have remained the gene transfer strategies of choice. In this report, we have performed an evaluation of these vector systems with respect to their β cell-specific gene transfer profiles.

As a result of the increasing incidence of T1DM worldwide, recent years have spawned a renewed interest in the delineation of cytokine-induced pathophysiological pathways in the β cell. Many studies have suggested a myriad of potentially cytoprotective genes of interest. Recently, a substantial body of conflicting evidence has been amassed pertaining to the cytoprotective potential of antioxidant overexpression strategies in β cells exposed to cytokines or conditions of oxidative stress scenarios. While a number of reports describe a cytoprotective benefit of antioxidant overexpression in β cells [1221], a number of contrasting reports highlight the shortcomings of antioxidant overexpression strategies in cells exposed to comparable cytotoxic scenarios [12, 1618]. Recent evidence, however, suggests that antioxidant overexpression fails to suppress NF-κB activation and, for this reason, does not alter cytokine-induced damage in whole transgenic islets [22].

The regulation of NF-κB activity remains a determinant of β cell fate. Cytokines induce NF-κB activation in human [23] and rodent islets [24] in vitro. IL-1β-mediated iNOS induction in RINm5F cells [25], and ex vivo human and rodent islets in culture [25], is dependent on NF-κB activation. Modulating the NF-κB-dependent induction of iNOS in β cells exposed to cytokines may represent a more promising cytoprotective strategy when compared to antioxidant overexpression. Recently a mutant non-degradable form of the endogenous NF-κB inhibitory subunit, IκBα has been engineered, rendering it resistant to phosphorylation and degradation [24].

In light of these facts, this investigation also aimed to screen an array of potentially cytoprotective gene transfer strategies focussing either on the augmentation of the β cell’s endogenous antioxidant profile and the inhibition of the deleterious expression of iNOS. Our data clearly indicates that rat insuloma (RIN-r) cells represent an invaluable tool for the study of cytokine-induced β cell pathophysiology and that viral vectors confer significantly enhanced gene transfer efficiencies in RIN cells relative to plasmid lipofection. Importantly, we demonstrate that iNOS inhibition, achieved either through adenoviral-mediated NF-κB blockade or lentiviral vector-based RNAi-mediated iNOS gene silencing, conferred significant protection against IL-1β-induced damage whereas antioxidant overexpression failed to provide protection. Cytoprotection was associated with a suppression of iNOS activation and nitrite accumulation. In corroboration with the hypothesis that iNOS alone represents a prime target for therapeutic intervention strategies is the finding too that administration of the specific iNOS inhibitor, L-NIO, provided complete protection from cytokine exposure. Thus, from a therapeutic standpoint, we suggest that viral vector-based strategies aimed at targeting the activation of iNOS may harbour therapeutic potential in preserving β cell survival in the face of proinflammatory cytokine exposure.

Research Design and Methods

Materials

RIN cells were a gift from the Steno Diabetes Centre, Gentofte, Denmark. Tissue culture plastics were supplied by Nunc and Sarstedt. RPMI 1640 culture media and supplements were supplied by Cambrex. Primary antibodies against human MnSOD and CuZnSOD were supplied by Stressgen. Primary antibodies against human catalase, α-actin, and IκBα were supplied by Sigma. Primary antibodies against human iNOS, nNOS, and GAPDH were supplied by Transduction Laboratories. Secondary anti-rabbit and anti-mouse antibodies conjugated to horseradish peroxidase were supplied by Sigma and GE Healthcare, respectively. Recombinant human IL-1β was obtained from PromoCell. BCA protein assay reagents were obtained from Pierce. Cell viability was assayed using the Cell Proliferation Kit 1 (MTT) (Roche Diagnostics, Mannheim, Germany). All other reagents were obtained from Sigma.

Cell Culture and Cytokine Treatment

RIN cells were cultured at 37°C in RPMI 1640 supplemented with 10% (v/v) FBS, 20 mM HEPES, 2 mM l-glutamine, 24 mM NaHCO3, 100 U/ml penicillin, 100 μg/ml streptomycin under a humidified atmosphere of 5% CO2. Cells were passaged every 3–4 days by trypsinization. Cells were allowed to recover for 24 h after plating before exposure to IL-1β (100 U/ml). HEK 293FT cells and Hela cells were cultured in T-75 cell culture plates in DMEM supplemented with 10% (v/v) FBS (GIBCO), 2 mM l-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin under a humidified atmosphere of 5% CO2.

Vector Preparation

The plasmid vector encoding the green fluorescent protein (GFP) reporter gene (pCMV-GFP) was obtained commercially (Plasmid Factory). Recombinant replication-incompetent E1E3-deleted adenoviral vectors encoding null transcripts, human MnSOD, human CuZnSOD, human catalase, green fluorescent protein (GFP), or IκBα transgenes were generated as previously described [26]. All transgenes were under the control of the cytomegalovirus (CMV) promoter. Recombinant adenoviruses were propagated in transformed HEK293 cells, isolated, purified by double cesium gradient ultracentrifugation, and dialysed against 10 mmol/l Tris, 0.1 mmol/l MgCl2, 1.0 mM HEPES, and 10% glycerol for 4 h at 4°C. Viral titre was determined by plaque assay and expressed as plaque forming units per ml (pfu/ml). Viral stocks were stored at −80°C. The lentiviral vectors expressing iNOS shRNA were generated using the BLOCK-iT Lentiviral RNAi Expression System (Invitrogen). Two vectors were prepared, one expressing shRNA specific for rat iNOS and the other expressing shRNA specific for the unrelated control gene, lamin A/C. The cDNA sequences encoding the shRNAs were designed by Ambion and commercially synthesized by Invitrogen (top strand: 5′-CACCGGATGACCCTAAGAGTCACTTGAGAAAGTGACTCTTAGGGTCATCC-3′ and bottom strand: 5′-AAAAGGATGACCCTAAGAGTCACTTTCTCAAGTGACTCTTAGGGTCATCC-3′). Virus-containing supernatants were harvested 48 h post-transfection, isolated by centrifugation at 3,000 rpm for 5 min at 4°C, and concentrated by centrifugation at 22,500 rpm for 2 h before resuspension in basal RPMI medium. Viral particles were titred according to the manufacturer’s guidelines using blasticidin resistance in Hela cell colonies before aliquoting and storing at −80°C.

Gene Transfer to RIN Cells

The optimal transfection parameters were determined empirically, in 70% + confluent RIN cells, to be a DNA (μg): lipofectamine 2,000 (μl) ratio of 1:3.5 via reverse transfection. The optimal MOI for adenoviral and lentiviral vector-mediate gene transfer was predetermined to be 100 MOI. Under these conditions, lipofection, adenoviral, and lentiviral vector-mediated gene transfer was found to confer maximal gene transfer efficiencies of 46.28 ± 1.61%, 83.38 ± 1.33%, and 82.56 ± 1.67%, respectively. Gene transfer efficiencies were calculated based on the number of GFP-positive cells as quantified by flow cytometry either at 24 h or 96 h after gene transfers via lipofection or transductions, respectively. For all experiments evaluating the cytoprotective potential of the candidate cytoprotective genes, RIN cells were plated in 6-well plates at 3 × 105 cells/well (Western blots) or in 48-well plates at 2.5 × 104 cells/well (Griess and MTT studies) and allowed to attach for 24 h before transfection or transduction. IL-1β incubations commenced 48 h after transfection and 96 h after viral vector transduction. Following exposure to IL-1β cell viability, nitrite accumulation and iNOS expression were analysed. Transgene activity, where appropriate, was measured using the appropriate enzyme activity assay. SOD activity assay was measured using an ‘‘In Gel’’ assay providing a measure of enzymatic activity in native acrylamide gels as previously described [9]. The decomposition of H2O2 was followed directly by a decrease in absorbance at A240 nm and used as a measure of total catalase activity (U/mg protein).

Western Blot Analysis

Western blot analysis was used to (1) detect transgene expression and (2) monitor the induction of iNOS following cytokine exposure. The housekeeping proteins α-actin and GAPDH served as loading controls. RIN cells were plated at 3 × 105 cells/well in 6-well plates. About 24 h later, cells were transfected or transduced as appropriate and allowed to recover for 48 h and 96 h, respectively, before exposure to IL-1β. Cells were harvested by centrifugation at 5,000 rpm for 5 min. Total protein was extracted from cell pellets following suspension in whole cell lysis buffer (5 mM Tris–HCl, 0.1 mM EDTA, 0.1% 10% (v/v) SDS, 1% IGEPAL, 10% (v/v) protease inhibitor cocktail, pH 7.5) and incubated on ice for 20 min. Supernatants were collected and total protein concentrations determined by the BCA protein assay. About 15 μg samples of protein were loaded on 4% stacking/10% separating SDS-PAGE. Resolved proteins were transferred to a 0.2 mm nitrocellulose membrane on a semidry electrophoretic transfer system (Bio-Rad) before blots were placed in blocking buffer (5% non-fat milk in phosphate buffer saline/0.05% Tween 20) overnight at 4°C. Membranes were incubated with the primary antibodies (i.e., anti-iNOS (1:5,000), anti-nNOS (1:500), anti-GAPDH/anti-α-actin (1:10,000), MnSOD/CuZnSOD (1:3,000), Catalase (1:5,000), or IκBα (1:2,000)) diluted in blocking buffer overnight at 4°C. After washing, the membranes were incubated with the appropriate secondary antibodies conjugated to HRP for 1 h at room temperature. Protein bands were detected using enhanced chemiluminescence (GE Healthcare) and autoradiography.

Cell Viability

Cell viability was assayed using a Cell Proliferation Kit 1 (MTT assay) (Roche Diagnostics, Mannheim, Germany). RIN cells were seeded in 48-well microtitre plates at a density of 2.5 × 104 cells per well in a humidified atmosphere of 5% CO2, 37°C. About 24 h after plating, the cells were transfected as appropriate. After 48 h (transient transfection experiments) or 96 h (adenoviral and lentiviral transduction experiments), IL-1β incubations commenced. About 0.5 mg/ml of MTT labelling reagent was then added to each well and incubated at 37°C for 4 h before addition of solubilization solution. After the cells were allowed to stand overnight at 37°C, the A550 nm was recorded using a microtitre plate reader (SpectraMax, Molecular Devices). MTT assay is also used to monitor vector-associated toxicity.

Griess Assay

Nitrite accumulation was monitored as a marker for the activation of iNOS in RIN cells exposed to IL-1β. Briefly, RIN cells (2.5 × 104 cells/well) were plated in 48-well microtitre plates and transfected or transduced as appropriate 24 h later. For plasmid transfection and viral transduction experiments, cells were allowed to recover for 48 h and 96 h, respectively, before IL-1β incubations commenced. About 50 μl samples of media were taken at appropriate time-points and incubated with Griess reagent (1:1) (0.1% naphtyl ethylenediamine and 1% sulphanilamide in 0.1 mol/l HCl, 1:1 vol/vol) for 15 min in the dark at 25°C. Cells were harvested, lysed, and assayed for protein using the BCA protein assay. The A540 nm was then measured and the sample nitrite concentrations were normalized to total protein (μM of accumulated nitrite/mg protein).

Statistical Analysis

The experimental data are expressed as means ± SD. Treatment groups were compared by means of two-tailed Student’s t-test, assuming unequal variances; significance was established at p < 0.05 unless otherwise stated.

Results

Cytokines Induce RIN Cell Death

In order to establish our model for cytokine-induced RIN cell death, cells were plated in 48-well plates and exposed to IL-1β (100 U/ml). IL-1β exposure induced significant decreases in RIN cell viability (Fig. 1A) in a time-dependent manner with 91.25 ± 0.7%, 77.84 ± 1.95%, 71.57 ± 0.68, and 48.25 ± 2.55% viability remaining following 24, 48, 72, and 120 h of IL-1β exposure. Notably, these decreases coincided with the induction of iNOS expression (Fig. 1B). iNOS expression was evident as early as 4 h after initial incubation with IL-1β, and levels were seen to increase at all time-points examined. Concomitantly, significant accumulations in nitrite (Fig. 1C) were also noted following IL-1β exposure. These results indicate that exposure of RIN cells IL-1β induces an iNOS-dependent decrease in viability that is associated with an accumulation of nitrite.
https://static-content.springer.com/image/art%3A10.1007%2Fs12033-007-0049-6/MediaObjects/12033_2007_49_Fig1_HTML.gif
Fig. 1

Establishing the model for cytokine-induced β cell damage. (A) RIN cell viability following IL-1β exposure over an array of time-points as depicted. IL-1β induced significant decreases in RIN cell viability versus non-exposed controls. Data are mean ± SD from three individual experiments at n = 3. *Significance established at p < 0.005. (B) Western blot depicting the IL-1β-mediated time-dependent induction of iNOS protein expression over a range of time-points (panel A). Actin probing (panel B) was used to confirm equal loading of the whole cell lysates. The first detectable levels of iNOS protein were observed at 4 h post-IL-1β exposure. The concentration of IL-1β employed to induce iNOS was 100 U/ml. This blot is a representative image from experiments performed in duplicate. (C) Nitrite accumulation (μM/mg total cell protein) in RIN cells exposed to 100 U/ml IL-1β over 0, 24, 48, 72, and 120 h and 7 days. All experiments were performed at n = 3 in triplicate. Data are expressed as means ± SD. *Significance established at p < 0.005. **Significance established at p < 0.001. All data are expressed relative to mock treated control cell samples analysed at each of the respective time-points

Gene Transfer to RIN Cells

To compare the efficiency, duration, and associated toxicity of transgene expression conferred in RIN cells using either lipofectamine 2000 or adenoviral and lentiviral vector gene delivery, marker gene expression (GFP) was monitored at 1, 4, 7, 14, and 21 days post-gene transfer (Fig. 2A). Lipofected RIN-r cells displayed an onset of GFP transgene expression at 24 h post-transfection, with expression levels of GFP declining steadily to near non-detectable levels at 14 and 21 days post-transfection. In contrast, adenoviral and lentiviral vector-transduced RIN cells exhibited prolonged transgene expression together with an overall higher level of gene transfer efficiency once maximal expression from the respective vectors was reached. Both adenoviral vectors and plasmids conferred maximal levels of transgene expression as early as 24 h post-gene transfer reaching efficiencies of 81.98 ± 1.39% and 46.28 ± 1.61%, respectively. Lentiviral vector-mediated gene transfer displayed a slower onset of expression with maximal efficiency of expression obtained at 96 h post-transduction (i.e., 82.56 ± 1.67%). Moreover, the duration of transgene expression was markedly prolonged, when compared to lipofection, using either adenoviral or lentiviral gene delivery system with the lentiviral system yielding the most prolonged marker gene expression with 70.88 ± 2.85% cells expressing GFP at 21 days post-transduction versus 1.25 ± 0.26% and 3.93 ± 0.96% of lipofectamine treated and adenoviral vector-transduced RIN cells expressing GFP at the same time-point. These results highlight the superior gene expression profiles achievable using viral vectors, in particular the lentiviral vectors, when compared to those obtained through lipofection.
https://static-content.springer.com/image/art%3A10.1007%2Fs12033-007-0049-6/MediaObjects/12033_2007_49_Fig2_HTML.gif
Fig. 2

Comparative analysis of the gene transfer profiles of plasmid, adenoviral and lentiviral vectors expressing the GFP marker gene using empirically predetermined optimal parameters (A). The graph depicts percentage transduced (GFP-positive) cells in control, lipofected, adenoviral and lentiviral vector transduced RIN-r cell populations at days 1, 4, 7, 14, and 21 as determined by flow cytometry. Experiments were performed at n = 3 in duplicate. Data are expressed as means ± SD. Gene transfer efficiencies are expressed as the percentage GFP-positive RIN cells at the respective time-points after gene transfer. (B) Vector-induced cytotoxicity of RIN cells transfected or transduced as appropriate with plasmid, adenoviral and lentiviral vectors. Non-transfected/non-transduced RIN cells were normalized to 100% viability. Experiments were performed at n = 3 in triplicate. Data are expressed as means ± SD. Both lipofection and transduction resulted in significant levels of cytotoxicity that were exacerbated overtime. No significant differences between the treatments were observed. *Significance was established at p < 0.05

Notably, however, viability studies conducted in parallel experiments revealed that both lipofection and viral-mediated gene delivery exerted a mild but significant impact (p < 0.05) on viability relative to unmodified control cells (Fig. 2B). For the purposes of these experiments, non-transfected/non-transduced RIN-r cells were taken to be 100% viable at the respective time-points. Briefly, it was observed that both lipofection and viral transduction were associated with significant decreases in viability at all time-points (p < 0.05). The cytotoxic effect of both lipofection and viral transduction was most marked at the later time-points. However, no significant cytotoxic differences between lipofection and lentiviral transduction were observed at any of the time-points examined.

For the purposes of evaluating the cytoprotective potential of adenoviral-mediated gene transfer of the candidate transgenes, overexpression was quantified by Western blots performed on appropriately transduced RIN cell populations (Fig. 3). Densitometry analysis of the protein bands indicated expression levels were upregulated as follows: catalase, 3.55 ± 0.25-fold; MnSOD, 2.53 ± 0.05-fold; CuZnSOD, 2.55 ± 0.11-fold; and IκBα, 9.03 ± 2.1-fold (all figures expressed relative to mock transduced control cell populations).
https://static-content.springer.com/image/art%3A10.1007%2Fs12033-007-0049-6/MediaObjects/12033_2007_49_Fig3_HTML.gif
Fig. 3

Western blot analysis of transgene overexpression. Adenoviral-mediated gene delivery was carried at using an MOI of 100. (A) Catalase overexpression at 58 kDa. Lane 1: Non-transduced RIN, Lane 2: AdNull-RIN, Lane 3: AdCatalase-RIN. (B) Adenoviral-mediated MnSOD overexpression at 25 kDa. Lane 1: Non-transduced RIN, Lane 2: AdNull-RIN, Lane 3: AdMnSOD-RIN. (C) Adenoviral-mediated CuZnSOD overexpression at 23 kDa (endogenous rat CuZnSOD observed at 19 kDa). Lane 1: Non-transduced RIN, Lane 2: AdNull-RIN, Lane 3: AdCuZnSOD-RIN. (D) Adenoviral-mediated IκBα overexpression at 40 kDa. Lane 1: Non-transduced RIN, Lane 2: AdNull-RIN, Lane 3: AdIκBα-RIN. (E) Confirmation that mutant IκBα is resistant to cytokine-induced degradation. Mutant IκBα levels remain unchanged in RIN cells following 1 h cytokine exposure. Lane 1: AdIκBα-RIN non-exposed, Lane 2: AdIκBα-RIN exposed to IL-1β, Lane 3: AdIκBα-RIN exposed to the cytokine cocktail. (F) In contrast to mutant IκBα, endogenous IκBα undergoes rapid degradation (<1 h) following cytokine exposure irrespective of the cytokine conditions tested. Blots are representative of experiments performed in triplicate

Assessing the Impact of Cytoprotective Gene Transfer to RIN Cells

Following viral vector-mediated gene transfer, cells were exposed to IL-1β for 48 h, after which cell viability was assessed using the MTT assay. The 48 h time-point was selected given that maximal iNOS expression has been reached and the majority of β cell destruction has already occurred by this time-point (Fig. 1). Adenoviral vector-mediated antioxidant overexpression (either alone or in combination) did not confer significant (p < 0.05) cytoprotection against IL-1β toxicity. In contrast, adenoviral vector-mediated mutant IκBα gene transfer and lentiviral vector-based iNOS RNAi improved RIN cell viability significantly (p = 0.0053 and p = 0.0025, respectively) with 92 ± 1.05% and 104.69 ± 4.29% viability remaining, respectively, following IL-1β exposure when compared to mock transduced controls (75 ± 1.96%), AdNull transduced controls (70.75 ± 3.35%), and lenti transduced controls (75.7 ± 2.81%). Control vectors had no significant impact on viability versus non-transduced cell population (p = 0.33 and p = 0.3 for Ad and Lenti control vectors, respectively) (Fig. 4). As a positive control for the suppression of iNOS, the chemical inhibitor L-NIO also demonstrated a comparable degree of protection to that conferred by mutant IκBα expression or iNOS RNAi (i.e. 108 ± 5.62%, p = 0.0053).
https://static-content.springer.com/image/art%3A10.1007%2Fs12033-007-0049-6/MediaObjects/12033_2007_49_Fig4_HTML.gif
Fig. 4

Cell viabilities of RIN cells engineered with each of our candidate viral vector-mediated gene transfer strategies of interest following exposure to IL-1β (100 U/ml) for 48 h. Relative to mock transduced and vector-transduced controls, and in contrast with antioxidant-based overexpression strategies, transduction with either AdIκBα or the lenti.shRNA-iNOS vector preserved viability significantly over all time-points examined. The degree of protection afforded was comparable to that afforded by administration of the specific iNOS inhibitor, L-NIO, at 1 μM. Cell damage was thus determined to arise predominantly from the activation of iNOS given that chemical iNOS inhibition conferred complete protection upon RIN cells exposed to cytokines. Data are expressed as means ± SD. All experiments were performed at n = 3 in triplicate. *Significance established at p < 0.05. **Significance established at p < 0.005

Similarly, IκBα gene transfer and lentiviral vector-based iNOS RNAi, in contrast to antioxidant gene transfer, also suppressed the induction of iNOS as detected by Western blot. Given these results, it is likely that the cytoprotective impact of these strategies against IL-1β lies in their ability to suppress iNOS-mediated cell toxicity. These data are consistent with Griess assay results which indicate that both mutant IκBα expression and iNOS RNAi significantly (p < 0.005) suppresses the accumulation of nitrite 2.32 ± 0.2-fold and 2.33 ± 0.52-fold, respectively, in the media of RIN cells exposed to IL-1β, whereas antioxidant gene transfer did not impact this phenomenon significantly. The inclusion of a specific iNOS inhibitor, L-NIO, suppressed the accumulation of nitrite 3.98 ± 0.96-fold when compared to non-transduced IL-1b exposed RIN cells (see Figs. 56).
https://static-content.springer.com/image/art%3A10.1007%2Fs12033-007-0049-6/MediaObjects/12033_2007_49_Fig5_HTML.gif
Fig. 5

Western blot depicting cytokine-induced (6 h exposure) iNOS expression. iNOS induction was observed in all antioxidant overexpressing RIN cell populations irrespective of the transgenes employed. In contrast, mutant IkBα overexpression and lentiviral vector-based iNOS specific RNAi in RIN cells suppressed the induction of iNOS. (A) Mock transduced RIN (B) AdNull transduced RIN (C) AdMnSOD transduced RIN (D) AdCuZnSOD transduced RIN (E) AdCatalase transduced RIN (F) AdCuZnSOD/Catalase transduced RIN (G) AdMnSOD/Catalase transduced RIN (H) AdMnSOD/CuZnSOD/Catalase transduced RIN (I) AdIκBα transduced RIN (J) Lenti.shRNA-iNOS. Lane 1—Non-exposed, Lane 2—IL-1β exposed

https://static-content.springer.com/image/art%3A10.1007%2Fs12033-007-0049-6/MediaObjects/12033_2007_49_Fig6_HTML.gif
Fig. 6

Nitrite accumulation in RIN-r cells exposed to IL-1β (100 U/ml for 48 h) and transduced as appropriate with the gene transfer strategies of interest. The bar chart provides a comparative analysis of the impact of potentially cytoprotective strategies on nitrite accumulation in RIN cells following 48 h cytokine exposure. IκBα-overexpressing RIN and iNOS shRNA expressing cells display significantly reduced nitrite accumulation versus mock transduced and AdNull transduced cells. In contrast, antioxidants failed to suppress the accumulation of nitrite. *Significance established at p > 0.005. Data are mean ± SD, experiment performed at n = 4 in triplicate

Discussion

Protecting the β cells from immune-mediated destruction remains an unresolved issue in the development of islet transplantation [27, 28]. For this reason, our group and others have been optimizing and applying vector-mediated gene delivery systems to validated β cell model systems with the aim of enhancing cellular resistance against proinflammatory mediators. We hypothesized that by using a predetermined optimal vector system engineered to augment the β cell redox status or inhibit the activation of iNOS, RIN cells could be protected from cytokine-mediated destruction. Studies such as this provide an invaluable insight into the relative pathophysiological roles of NF-κB-dependent iNOS expression and ROS in β cells exposed to cytokines, given the ambiguity surrounding this issue to date.

We first established a robust model for cytokine-induced RIN cell death using treatment with IL-1β. IL-1β represents a well described β cell death mediator and potent inducer of iNOS. IL-1β exposure resulted in significantly decreased β cell viability, the induction of iNOS, and an accumulation of nitrite. These data are consistent with other reports [13, 29] in validated β cell model systems and primary β cells, which describe cytokine-induced activation of iNOS expression as a central element in cell toxicity. However, the relative roles of NO and ROS in determining β cell fate remain poorly understood.

Before conducting an evaluation of our library of candidate cytoprotective genes, we endeavoured to compare the three routinely used vector systems for β cell gene transfer pilot studies. Preliminary gene transfer studies confirmed the superior gene transfer profile of viral vectors compared to plasmid vectors with respect to duration and efficiency. Lentiviral vectors resulted in the most prolonged transgene expression. Prolonged exposure to all vector systems resulted in comparable levels of cytotoxicity.

Following a screen of the candidate cytoprotective molecules in our established model of IL-1β-induced β cell pathophysiology, our results indicate that antioxidant gene transfer does not represent a promising cytoprotective strategy. In corroboration with our findings, a recent report by Chen et al. [30] describes the inability of antioxidants to confer β cell protection against cytokine toxicity. The report cited the inability of antioxidants to suppress NF-κB activation and thus iNOS expression as the proposed reason for the lack of protection observed. The report further postulates that, given this fact, ROS have no role in cytokine toxicity in β cells. In keeping with this hypothesis is our finding that gene-based and RNAi-mediated cytoprotective strategies that suppress the activation of IL-1β-induced NF-κB activation and thus iNOS expression is sufficient to provide complete protection against IL-1β-induced cytotoxicity. The central role of iNOS-associated toxicity may be due to the ability of NO to inhibit insulin secretion, induce mitochondrial dysfunction and DNA damage, and also react with ROS to generate peroxynitrite, a potent β cell specific toxin [31]. As such, if this is the case, the role of ROS in β cell fate is not redundant.

This work serves to underscore the importance of iNOS activation in β cell pathophysiology following proinflammatory cytokine exposure. In addition, this report provides proof-of-principle that strategies designed to block the activation of iNOS harbour potential in promoting β cell viability following cytokine exposure. It may, however, be argued that the knock-on effects of NF-κB inhibition, given its broad-ranging role as a nuclear transcription factor capable of modulating the expression of an array of genes such as MnSOD and Fas, may well be ultimately detrimental to a cell in an in vivo scenario. In this regard, more exacting high-efficiency strategies, such as the lentiviral vector-based shRNA strategy targeting iNOS described in this report, may also have therapeutic potential.

In this regard, the development of alternative vector systems other than adenoviral vectors is central to the clinical application of gene therapy strategies, given that adenoviral vectors do not confer long-term, stable gene expression and the cytoprotective impact of candidate transgenes will therefore be short-lived. Although this report demonstrates the efficiency of lentiviral vector-mediated β cell gene transfer, the associated toxicity, although minimal, will likely make the application of this vector system to routine clinical use problematic. As such, our group and others are currently investigating novel capsid pseudotyping strategies aimed at augmenting the transduction efficiencies of VSV-g pseudotyped HIV-1-based vector particles. In this regard, our group has endeavoured to incorporate a truncated version of the β cell-specific LDL receptor ligand, Apolipoprotein E, onto the VSV-g pseudotyped vector capsid so as to enhance viral particle binding and entry (unpublished data). Strategies such as this, however, will serve to minimize the number of viral particles required to mediate efficient gene transfer and thereby reduce vector associated toxicity and immunogenicity.

To date, one report has demonstrated the utility of RNAi-mediated iNOS knockdown in insulin-producing cells [32]. While this report provided useful information regarding proof of feasibility of RNAi strategies to silence iNOS expression in a β cell line, the study had a number of limitations. First, the siRNA transfection strategy employed provided only short-term, low-level protection over a relatively short period of exposure to cytokines. Second, the clinical applicability of siRNA transfection by non-viral means to confer iNOS knockdown in primary β cells is questionable given the reported poor transfection efficiencies of this cell type (largely attributable to their relatively slow-proliferation rate and their centralized location within the islet as a whole) [33, 34]. Lentiviral vector-based shRNA delivery provided long-term knockdown of iNOS expression following prolonged exposure to IL-1β. The degree of knockdown conferred was such that iNOS levels were below the levels of detection at 6 h by Western blot in the lenti.shRNA-iNOS transduced cells. Concomitantly, nitrite levels were significantly reduced after 48 h of IL-1β exposure. We believe the degree of suppression afforded underscores the efficiency of vector-based RNAi as a tool to silence deleterious iNOS gene expression, given the rapid transcription rate of iNOS expression following IL-1β exposure. Previous data from our group has shown that the p65 subunit of the cytosolic NF-κB complex translocates to the nucleus <45 min after cytokine exposure, whereupon it binds to upstream elements in the iNOS gene to induce its transcription [9]. Notably, the suppression of iNOS expression was sufficient to confer prolonged significant preservation of RIN-r cell viability and was comparable to that mediated by the specific iNOS-inhibitor L-NIO.

The development of more efficient vector-based shRNA strategies designed to mediate iNOS knockdown such as that described herein will undoubtedly prove to be a useful tool in the study of β cell physiology and preservation. This investigation provides novel evidence for lentiviral vector-based RNAi-mediated long-term and efficient knockdown of iNOS expression in a β cell line exposed to the proinflammatory cytokine, IL-1β. Moreover, the data herein provide proof-of-principle that vector-based strategies targeting the NF-κB-dependent iNOS expression are sufficient to preserve the deleterious expression of iNOS in insulin-producing cells for prolonged periods. Strategies such as that outlined in this study, which aim to modulate the accumulation of cytotoxic levels of NO, may be useful in preserving β cell viability in the face of proinflammatory cytokine exposure. Similarly, such strategies may hold therapeutic utility in disease models where NO toxicity represents a key determinant of cell fate.

Acknowledgments

The technical assistance from all NCBES and REMEDI is gratefully acknowledged. This research was supported by a grant form the Health Research Board of Ireland. CMcC is funded by the IRCSET Embark Initiative. TOB is funded by an SFI CSET, and the HEA, and the Juvenile Diabetes Foundation International.

Copyright information

© Humana Press Inc. 2007