Combined Small Interfering RNA Therapy and In Vivo Magnetic Resonance Imaging in Islet Transplantation

The double-stranded siRNAs targeting human caspase-3 (GenBank accession numbers NM_004346, NM_032991) and scrambled siRNA were designed and synthesized by Dharmacon (Lafatette, CO). The siRNA was modified to incorporate a thiol group on the 5′ end of the sense strand. There was no modification of the antisense strand.

The probe consisted of magnetic nanoparticles (for MRI, synthesized as described in Ref. 20) conjugated to the near-infrared (NIRF) Cy5.5 dye (for correlative microscopy) and to siRNA (for gene silencing) (Fig. 1A). The synthesis of the probe involved four distinct steps: synthesis of dextran-coated magnetic nanoparticles (MN), conjugation of the fluorescent dye Cy5.5-N-hydroxy succinimide ester to MN, conjugation of the N-succinimidyl 3-(2-pyridyldithio) propionate (SPDP) to MN-Cy5.5, and conjugation of siRNA to MN through SPDP linker. In brief, a solution of monoactivated Cy5.5 succinimide ester (Amersham Biosciences, Piscataway, NJ) in 20 mmol/L sodium citrate and 0.15 mol/L NaCl was allowed to react with the previously dialyzed immunopure, amino-derivatized dextran-coated iron oxide (aminated iron oxide, MN-NH₂) at pH 8.5, with constant agitation over a period of 12 h at room temperature. The Cy5.5-labeled aminated iron oxide (MN-NIRF) was purified from unreacted dye using a Sephadex G-25, PD-10 column (Amersham Biosciences). A ratio of 2 Cy5.5 molecules per nanoparticle was obtained. MN-NIRF was then conjugated to the heterobifunctional cross-linker SPDP (Pierce Biotechnology, Rockford, IL) by means of the N-hydroxy succinimide ester, followed by purification using a Sephadex G-25, PD-10 column in phosphate-buffered saline (PBS)/EDTA, pH 7.5. The labeling of SPDP per crystal was determined based on the release of pyridine-2-thione at 343 nm (e = 8.08 × 10³ M⁻¹cm⁻¹) after the addition of the reducing agent, tris-(2-carboxyethyl) phosphine hydrochloride (TCEP; 35 mmol/L in DMSO). A ratio of 31 SPDP molecules per nanoparticle was obtained.

The caspase-3-targeting and scrambled siRNA duplexes were then conjugated to MN-Cy5.5-SPDP through its 5′-sense thiol group. Before conjugation, the disulfide protecting group on 5′-disulfide bond was deprotected using TCEP according to the manufacturer’s instructions. The double-stranded RNA (dsRNA) was then reacted overnight (4°C) with the previously activated MN-Cy5.5-SPDP product via the SPDP cross-linker in PBS/EDTA, pH 8, followed by purification using a Quick Spin Column G-50 Sephadex Column (Roche Applied Science, Indianapolis, IN). The probes designated MN-siCaspase-3/MN-siSCR (Fig. 1A) were next purified using magnetic separation columns as described by the manufacturer (Miltenyi Biotec, Auburn CA). The amount of conjugated siRNA was assayed using agarose gel electrophoresis. To assess siRNA dissociation from the nanoparticles under reducing conditions, the probe was pretreated with 15 mmol/L TCEP for 30 min. The siRNA standard probes, untreated probes, and probes treated with reducing agents were applied on a 2% agarose gel in Tris borate-EDTA buffer (Invitrogen, Carlsbad, CA) at 145 V for 1 h. After electrophoresis, the gel was stained with 0.5 μg/ml ethidium bromide for 30 min and visualized using a Molecular Imager FX scanner (Bio-Rad Laboratories, Hercules, CA). The image was quantitated using the software ImageJ, version 1.43u.

Human islets were obtained from the Islet Cell Resource Centers (National Institutes of Health and Juvenile Diabetes Research Foundation). The viability and purity of the islets were both higher than 90%. On arrival at our facility, islets were cultured in CMRL-1066 medium (GIBCO, Grand Island, NY) supplemented with 10% FBS and 100 μg/mL penicillin-streptomycin.

For labeling experiments, 1,000 islet equivalents were counted and incubated for 48 h with either MN-siCaspase-3 or MN-siSCR probe in the same medium (0.025 mg Fe, 105 pmol siRNA, in 1 mL medium). To mimic the apoptotic condition, we cultured a group of islets in CMRL-1066 medium supplemented with 100 μg/mL penicillin-streptomycin without serum (serum starvation conditions). After incubation, islets were washed three times in culture medium and used for in vitro and in vivo experiments.

Glucose-stimulated insulin secretion was evaluated using static incubation of experimental MN-siCaspase- and control MN-siSCR–labeled islets at low (1.7 mmol/L) and high (20 mmol/L) glucose concentrations. Insulin was measured in supernatants and islet extracts using a human insulin ELISA kit (Mercodia, Uppsala, Sweden). A stimulation index was calculated as the ratio of stimulated to basal insulin secretion normalized by the insulin content.

Islet cell viability was determined after labeling by colorimetric (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay according to the manufacturer’s protocol (Promega, Madison, WI). To determine silencing effect after islet treatment with the probes, we performed real-time RT-PCR (TaqMan protocol).

Total RNA was isolated from the experimental and control islets with the RNeasy Mini kit, according to the manufacturer’s protocol (QIAGEN, Valencia, CA). The PCR primers and TaqMan probe specific for caspase-3 were designed with Primer Express software 1.5. Primer and probe sequences were as follows: Forward primer, 5′-TGTTCCATGAAGGCAGAGCC-3′; Reverse primer, 5′-TGCGTATGGAGAAATGGGC-3′; and TaqMan Probe 5′-TGGACCACGCAGGAAGGGCCT-3′.

Eukaryotic 18S rRNA TaqMan PDAR Endogenous Control reagent mixture (PE Applied Biosystems, Foster City, CA) was used to amplify 18S rRNA as an internal control, according to the manufacturer’s protocol. All samples were run in duplicate.

Caspase-3 protein levels in pancreatic islets were determined by Western blot with rabbit monoclonal antibody to caspase-3 (Abcam, Cambridge, MA) followed by horseradish peroxidase-conjugated goat anti-rabbit antibodies (Cell Signaling Technology, Danvers, MA). Staining of β-actin was used as an internal control. In situ apoptosis detection was performed using transferase-mediated dUTP nick-end labeling (TUNEL) assay (Chemicon, Temecula, CA) according to the manufacturer’s protocol. The percentage of cells undergoing apoptosis was calculated from the number of nuclei positive for DNA fragmentation (fluorescein isothiocyanate [FITC] channel) versus the total number of cells present (diaminido phenylindol [DAPI] channel).

Fluorescence microscopy of labeled islets was performed with anticleaved caspase-3 rabbit monoclonal IgG antibody (clone 269518; R&D, Minneapolis, MN) followed by FITC-labeled anti-rabbit IgG (H+L) secondary antibody (Vector Laboratories, Burlingame, CA). Slides were mounted with DAPI-containing mounting medium (Vector Laboratories) and examined under Nikon Eclipse 50i microscope.

For iron staining, we used Prussian Blue stain as described previously (14). All images were acquired using a CCD camera (SPOT 7.4 Slider RTKE; Diagnostic Instruments, Sterling Heights, MI) and analyzed using iVision 4.015 version software.

All animal experiments were performed in compliance with institutional guidelines and approved by the Subcommittee on Research Animal Care at Massachusetts General Hospital. MN-siCaspase-3–labeled human pancreatic islets were implanted under the left kidney capsule (1,000 islet equivalents/kidney) of 6-week-old NOD.scid mice (n = 6). The same number of islets labeled with parental nanoparticles (MN) was transplanted under the right kidney capsule of the same animal.

Islet phantoms were prepared by fixing islet pellets in 2% paraformaldehyde and sedimenting them in 1% agarose gel. Imaging was performed using a 9.4T AVANCE scanner (Bruker BioSpin, Billerica, MA) equipped with ParaVision 3.0.1 software. The imaging protocol consisted of coronal T2 weighted spin echo (SE) pulse sequences with the following parameters: repetition time (TR)/echo time (TE) = 3,000/8, 16, 24, 32, 40, 48, 56, 64 ms, field of view (FOV) = 3.2 cm², matrix size 128, resolution 250 μm², and slice thickness = 0.5 mm. Image reconstruction and analysis were performed using Marevisi 3.5 software (Institute for Biodiagnostics, National Research Council, Canada). T2 relaxation times were determined by T2 map analysis of regions of interest drawn around the islet pellets.

In vivo MRI was performed on a 9.4T scanner 1, 3, 5, 7, and 14 days after islet transplantation. The imaging protocols consisted of Multi-Slice Multi Echo T2-weighted map (for volume and T2 relaxivity measurement). Parameters: TR = 2,000 ms and multiecho TE = 8, 16, 24, 32, 40, 48, 56, and 64; number of averages (NA) = 4; rapid acquisition with relaxation enhancement factor = 8; FOV = 4.4 cm², matrix size 128, spatial resolution 312 μm², and slice thickness = 0.5 mm. We calculated graft volumes by counting the area in each slice of the region of interest (ROI) outlining the graft and multiplying by the number of slices.

After the last imaging time point, kidneys were removed, fixed in 4% formaldehyde, and embedded in paraffin. Sections were stained for insulin with anti-human insulin antibody (Santa Cruz Biotechnology, Santa Cruz, CA) and for caspase-3 with anticleaved caspase-3 rabbit monoclonal IgG antibody (269518; R&D, Minneapolis, MN). After incubation with Alexa Fluor 594-conjugated secondary goat anti-mouse IgG (H+L; Invitrogen) and FITC-labeled secondary goat anti-rabbit IgG (Vector Laboratories) and counterstaining with DAPI, the sections were used for microscopy as described above. In situ apoptosis detection of kidney sections was performed using TUNEL assay similar to the procedure described above.

Data are presented as means ± SD. Statistical differences were analyzed by Student t test (SigmaStat 3.0; Systat Software, Richmond, CA). A value of P < 0.05 was considered statistically significant.

To evaluate the number of siRNA molecules associated with nanoparticles, we performed gel electrophoresis with MN-siCaspase-3 under reducing conditions (Fig. 1B). As shown in Fig. 1C, semiquantitative analysis indicated that 105 pmol siRNA was conjugated to MN-siCaspase-3 probe containing 25 μg of iron.

The uptake of the probes by human islets was confirmed by staining for iron with Prussian Blue stain. As shown in Fig. 2, A and B, approximately 95% of islet cells were labeled with MN-siCaspase-3 or MN-SCR probes, which was confirmed with fluorescence microscopy (Fig. 2C).

At the same time, insulin secretion of human islets following siRNA probe treatment was unaffected in both the experimental and control groups compared with nonlabeled (sham) islets. Glucose stimulation index was not significantly different in all treatment groups (Fig. 2Dand E; P > 0.05). MTT assay performed on labeled islets showed that the treatment with the probes did not have any significant effect on islet viability (Fig. 2F; P > 0.05).

To evaluate the protective effect of siRNA probes, we created treatment conditions where caspase-3 level in the islets was upregulated. We incubated islets without serum (serum starvation), treated them with the probes, and measured caspase-3 mRNA expression by RT-PCR. As shown in Fig. 3A, caspase-3 mRNA expression was significantly reduced by the protective action of the MN-siCaspase-3 probe. Although the islets treated with MN-siSCR probes seemed to show a similar trend, the difference between this treatment and treatment with MN-siCaspase-3 probe was highly significant. It is possible that off-target effects (18) played a role in this phenomenon. To determine the caspase-3 protein expression in islets after treatment with the probes, Western blot of the islet samples was performed. As shown in Fig. 3B, the expression of caspase-3 protein decreased in islets treated with MN-siCaspase-3 probe compared with control islet group. Fluorescence microscopy confirmed lower expression of caspase-3 in serum-starved islets treated with MN-siCaspase-3 probe compared with the starved nonlabeled islets or islets labeled with the control MN-siSCR probe (Fig. 3C). As a consequence, the level of apoptosis in the serum-starved islets after the treatment with the MN-siCaspase-3 probe was significantly lower than in control islets (Fig. 4A and B). Together, these results demonstrated that there was a significant protective effect of the siRNA probe in islets after incubation under serum starvation conditions.

Finally, we performed MRI of islet phantoms to prove that incubation with magnetic nanoparticles leads to islet labeling, which would allow for following transplanted islets in vivo. Indeed, labeled islets revealed a significant drop in the T2 relaxation times after incubation with MN-siCaspase-3 or parental nanoparticles compared with unlabeled islets (P < 0.05, Fig. 5Aand B).

To test our hypothesis that prelabeling with the magnetic siRNA-nanoparticle probe would lead to improved protection and better outcome of islet grafts, we compared the survival properties of human islet grafts labeled with MN-siCaspase-3 and control parental nanoparticles (MN) using in vivo MRI. MN-siCaspase-3–labeled islets were transplanted under the left kidney capsule, and MN-labeled islets were transplanted under the right kidney capsule, so each mouse served as its own control. As shown in Fig. 6A, on T2-weighted MR images, transplanted islet grafts underneath the renal capsule appeared as pockets of signal loss disrupting the contour of both kidneys. ROI analysis performed over time revealed that the graft volume decreased both in experimental and control groups. However, this decrease was significantly lower in the animals transplanted with MN-siCaspase-3–labeled islets and was especially pronounced on days 7 and 14 after transplantation (Fig. 6B, P < 0.05). These results correlated with ex vivo histology, which showed reduced expression of caspase-3 in MN-siCaspase-3–treated grafts (Fig. 6C). On the other hand, insulin expression was visibly increased in these grafts compared with controls. This result, in turn, was in accordance with TUNEL staining, which revealed higher apoptosis levels in control grafts (Fig. 6D). These data attest to the feasibility of the MN-siCaspase-3 probe to create a protective effect for pancreatic islets after transplantation.

Islet transplantation can provide type 1 diabetic patients with improved glycemic control and insulin independence. However, a substantial proportion of the transplanted islet mass fails to engraft because of death by apoptosis. A variety of approaches have been explored to prevent apoptotic destruction of islets in the experimental setting. They include a series of pan-caspase inhibitors and virus mediated anticaspase gene therapy (10,11). Major concerns regarding these approaches deal with poor selectivity and specificity of the inhibitors and the safety of viral vectors. In this study, we demonstrate for the first time that combined siRNA gene therapy and noninvasive imaging can create protection for the graft, reduce apoptosis, and improve the survival properties of the islets in vitro and in vivo.

Programmed cell death is a cascade of events leading to the upregulation of the caspase family. Hypoxia, oxidative damage, serum starvation, inflammatory cytokines (tumor necrosis factor-α, interleukin-1β, γ-interferon, etc.), and other chemical factors (staurosprine, etoposide, cytochrome c, etc.) can induce apoptosis in the islet (19,20). Here we showed that serum starvation-induced apoptosis leads to increased caspase activity in the islets. Because activation of the precursor caspase-3 is the converging point of apoptosis, we used the MN-siCaspase-3 probe to label islets to determine whether caspase-3 gene silencing can block apoptosis induced by serum starvation in vitro. Our data demonstrated that transfection of islets with MN-NIRF-siCaspase-3 reduced caspase-3 mRNA transcripts by 50% under serum starvation. Caspase-3 precursor protein was also significantly suppressed as determined by Western blotting 48 h after incubation.

Our in vivo MRI studies showed that the area of the graft treated with the MN-siCaspase-3 probe was significantly higher than the control islet graft under the renal capsule 7 days and 14 days after transplantation. These data also demonstrate that during the first week after transplantation, there were no differences in reduction of the graft volume between the control and MN-siCaspase-3 groups. Analysis of T2 maps showed that there was no significant change in T2 relaxation times between MN-siCaspase-3 and control groups within 2 weeks. This observation indicated a relative stability of the label within the graft in both experimental and control kidneys during the post-transplant period, which was consistent with our previous report (12).

Because immune rejection could induce apoptosis in the grafts and interfere with the results, all of the in vivo experiments were performed in immunocompromised mice. We propose experiments in diabetic immunocompetent mice transplanted with MN-siRNA–labeled pancreatic islets in the future.

In summary, our data show the feasibility of two-in-one combination of siRNA gene protection therapy and in vivo MRI-noninvasive monitoring of transplanted human islets in mice. We believe that our dual purposes of nanoparticles in addition to delivering siRNA to pancreatic islets offer the added benefit of monitoring grafts in early post-transplant period.

This study was supported by National Institutes of Health [Grant 5R01-DK080784] (to A.M.).

No potential conflicts of interest relevant to this article were reported.

P.W. performed the in vitro experiments, animal surgery, and MRI scanning and participated in the writing of the manuscript. M.V.Y. synthesized the probe. Z.M. participated in the experimental design. L.W. and C.S. participated in the animal studies. G.D. participated in the MRI scanning. A.M. conceived the idea of the project and wrote the manuscript.