Partial correction of cystinuria type A in mice via kidney-targeted transposon delivery

Abstract

We used kidney-targeted, non-viral, transposon-mediated gene delivery to express the mouse Slc3a1 transgene in one kidney of cystinuria type A (Slc3a1^−/−) mice. We found a 44% reduction in urinary cystine concentration at 154 days post-gene transfer, although there was no significant effect on cystine stone formation. Our results indicate that it is possible to achieve kidney-targeted gene transfer, resulting in reduction of cystine concentration in the urine of a cystinuria type A animal model. This proof of concept lays the foundation for future studies directed at gene therapy for cystinuria and other kidney diseases.

Graphical abstract

Wilson and colleagues use a kidney-targeted non-viral gene delivery system to achieve partial correction of an inherited kidney disease, cystinuria type A, in a mouse model. They show proof of concept that kidney-targeted injection of transposon constructs could be used for gene therapy of rare human diseases.

Introduction

Cystinuria is an autosomal recessive genetic kidney disease affecting 1 in 7,000 worldwide.¹ Patients with cystinuria develop kidney stones, leading to morbidity and mortality.¹ Effective treatments for cystinuria are lacking and those used can have severe side effects.² Cystinuria type A results from mutation of the SLC3A1 gene in humans affecting expression of the neutral and basic amino acid transport protein (rBAT) that is a component of the heteromeric amino acid transporter protein complex expressed in proximal tubule cells in vivo.^3,4 The piggyBac transposon system is a non-viral cut-and-paste system capable of integrating transgenes into kidney cell genomes in vivo following hydrodynamic renal pelvis injection of transposon plasmid DNA.^5,6 To test a non-viral gene therapy for cystinuria, we used transposon-mediated gene transfer to deliver the murine Slc3a1 transgene to one kidney in a mouse model of cystinuria type A.

Results

We previously found that hydrodynamic renal pelvis injection of plasmid DNA enables gene delivery in mice that is kidney-specific, as other organs lacked transgene expression.⁶ We found that the elongation factor 1 alpha (EF1α) promoter was capable of expressing transgenes in proximal tubular cells in mouse kidney in vivo after hydrodynamic renal pelvis injection of plasmid DNA.⁶ Therefore, we engineered a piggyBac transposon plasmid containing the mouse Slc3a1 transgene driven by the EF1α promoter with the goal of expression of the rBAT transporter in tubular cells (Figure 1A). We also engineered a transposon plasmid containing the mouse Slc7a9 transgene to express b^0,+AT protein, which is the partner protein of the heteromeric amino acid transporter (Figure 1A). To evaluate for stability of expression with and without piggyBac transposase, we transfected human proximal tubular (HK-2) cells with both pT-EF1α-Slc3a1 and pT-EF1α-Slc7a9, encoding the cotransporter required for rBAT function, and performed western blot analysis of rBAT expression over time. We found that piggyBac enabled greater long-term expression of rBAT in human proximal tubule cells (Figures 1B and 1C). We next evaluated for rBAT association with b^0,+AT protein forming functional cystine transporters in HK-2 cells. We confirmed cystine uptake in cells co-transfected with Slc3a1 and Slc7a9, indicating that our Slc3a1 transgene cassette resulted in functional transporter protein complexes in transfected cells (Figure 1D).

Figure 1.

A Slc3a1 transposon system to express the rBAT transporter

(A) Schematic of plasmid vectors. m7piggyBac, hyperactive piggyBac transposase; SV40 pA, polyadenylation sequence; Slc3a1, mouse Slc3a1 cDNA transgene; Slc7a9, mouse Slc7a9 cDNA transgene; blue arrows, piggyBac transposon inverted repeats. (B) Western blot analysis demonstrating expression of rBAT over time with and without transposase in transfected HK-2 proximal tubule cells. Cells were co-transfected with pT-EF1α-Slc3a1 and pT-EF1α-Slc7a9 with or without pEF1α-m7pB and analyzed for rBAT expression over time using β-actin as a loading control. PC, positive control mouse kidney lysate; NC, untransfected HK-2 cells. (C) Image quantitation of rBAT expression levels (∼100 kDa band) over time taken from images in (B), mean (SD). (D) Cystine-FITC uptake in transfected HK-2 cells. Statistical comparison was performed using Student’s t test from three independent experiments, mean (SD).

We next performed hydrodynamic renal pelvis injection of transposon DNA for kidney-targeted transgene expression.⁶ A unilateral flank incision was used to expose one kidney that was injected with transposon and transposase plasmids using hydrodynamic renal pelvis injection.^6,7 We evaluated rBAT expression in transfected mouse kidneys using immunofluorescence microscopy. We observed no detectable rBAT expression in Slc3a1^−/− mice injected with buffer alone or in immunoglobulin G (IgG)-stained controls. We observed patchy rescued expression of rBAT in mouse kidney after gene delivery to Slc3a1^−/− mice (Figure 2A). rBAT expression co-localized with proximal tubule cells as rBAT immunofluorescence colocalized with lotus tetragonolobus lectin-fluorescein (LTL) signal (Supplementary Information). We used droplet digital PCR (ddPCR) to quantitate mRNA expression of Slc3a1 after gene delivery. We found approximately 7% at 48 h and 4.8% at 6 weeks post-gene delivery in knockout animals when compared to the level of Slc3a1 RNA in wild-type mice (Figure 2B). We observed no Slc3a1 RNA in uninjected Slc3a1^−/− mice or in the liver, an off-target organ, of injected mice (Figure 2B).

Figure 2.

Confirmation of rBAT expression after hydrodynamic renal pelvis injection

A unilateral flank incision was made to expose one kidney, and the renal pelvis was hydrodynamically injected with the transposon system as outlined in the materials and methods. (A) Immunofluorescence microscopy demonstrates rBAT expression 48 h after gene delivery to Slc3a1^−/− mice. DAPI was used for nuclear staining. Staining with IgG control indicates no rBAT staining. (B) ddPCR quantitation of mRNA levels of Slc3a1 in Slc3a1^−/− mice 48 h (left) or 6 weeks (right) after gene delivery compared to uninjected. Statistical analysis was performed using one-way ANOVA followed by Turkey’s multiple comparisons test, mean (SD).

Slc3a1^−/− mice develop bladder stones over time that can be visualized using X-ray analysis.^8,9 We performed gene transfer experiments in randomly assigned male mice only because female mice do not reliably form stones.^8,9 We did not observe a statistically significant effect on stone formation over time in the Slc3a1 gene delivery as compared to mice receiving buffer only (Figure 3A). Of note, some of the mice had stones prior to administration of the DNA as the experiment was performed in adult mice. Exclusion of the mice with stones prior to gene delivery did not achieve statistical significance (p = 0.66 to 0.2 after exclusion via log rank Mantel-Cox test). We also evaluated the area of bladder stones comparing with and without gene transfer at 10 weeks post-gene transfer and did not observe a statistically significant difference (Figure 3B).

Figure 3.

Targeted kidney injection of rBAT-expressing transposons lowers urinary cystine in Slc3a1^−/− mice

(A) Evaluation of the % of mice with stones detected by X-ray analysis over time comparing naive to Slc3a1 gene transfer in Slc3a1^−/− animals. (B) The area of bladder stones calculated from X-ray taken 10 weeks after gene transfer, mean (SD). (C) Evaluation of urinary cystine concentration normalized to creatinine. The green arrow indicates gene delivery into 8-week-old mice. Student’s t test analysis revealed a difference at 154 days between groups. (D) Plot of urinary cystine levels over time for individual mice.

We next quantified cystine levels in the urine of mice over time post-gene transfer. We found a 44% ± 9% reduction over time that was significant at 154 days post-gene transfer in mice receiving the pT-EF1α-Slc3a1 transposon with transposase when compared to the animals that did not receive the transposon (Figure 3C). Individual animals had decreased urinary cystine levels over time comparing transposon-injected Slc3a1^−/− mice to uninjected Slc3a1^−/− mice (Figure 3D). There was a loss of naive animals over time due to death (Figure 3C).

Discussion

Long-term phenotypic correction of an inherited kidney disease in an animal model has remained a challenge, despite correction of other diseases in other organs. The primary roadblock to advancing gene therapy for kidney disease remains efficient delivery to the kidney in vivo.^10,11

We undertook these experiments as proof-of-principle studies for gene therapy for kidney disease. Cystinuria remains a targetable disease for gene therapy.¹² We chose cystinuria type A mice as they phenocopy the human disease. The recessive inheritance pattern suggested that low levels of rBAT expression in proximal tubule cells could affect the phenotype.

We engineered transposons to express functional rBAT in heterologous cells and in vivo. We previously found that our targeted single-kidney delivery surgical method had high survival with low morbidity^6,7; therefore, we proceeded with our single kidney delivery method. Although we did not affect stone formation over time, this was not surprising as the method was unable to lower urinary cystine concentration below the 1 mM threshold needed to reduce stone formation.¹³ We observed a significant reduction in urinary cystine concentration at 154 days, even though we stably transfected only one kidney. Our results suggest that more efficient delivery to both kidneys could further reduce urinary cystine concentration for possible therapy. Catheter-based kidney-targeted approaches in larger animals are an area of investigation that should be considered.

Both viral and non-viral approaches are being developed for kidney-targeted gene delivery. Viral-gene-delivery-based approaches including adeno-associated virus (AAV) remain a possibility. Gene editing or delivery of Cre-recombinase to the kidney in reporter mice has been achieved.^14,15 Ikeda et al. demonstrated that AAV delivered Cre-mediated cleavage of floxed Gli2 inhibited kidney fibrosis in a UUO model.¹⁶ However, delivery of a recombinase to cleave genomic DNA is a different challenge than integrating new sequences into the genome to treat a genetic disease. Recent studies have demonstrated the potential for therapy in mouse models targeting glomerular disease.^17,18 Liao et al. reported AAV9 delivery of Slc7a9 for cystinuria type B in mice.¹⁹ Although AAV9 has limited delivery to tubules of the kidney in mice,¹⁴ Liao et al. reported decreased urinary cystine concentration and stone formation over time in a mouse model generated by CRISPR/Cas9-mediated knockout of Slc7a9. Therefore, their results are also consistent with limited gene delivery mediating a therapeutic effect for cystinuria. Future studies to generate novel engineered viral capsids could enable organ-enhanced delivery of therapeutic transgenes with viruses such as AAV.²⁰ Recent studies have also developed novel AAV capsids targeting the proximal tubule, which may prove advantageous for cystinuria.²¹

Our studies extend the possibility of gene delivery for cystinuria type A using a nonviral delivery method. Nonviral delivery offers advantages over viral delivery by having less restricted packaging capacity, less immune response, and much decreased production cost. Therefore, nonviral delivery should be pursued for diseases such as cystinuria.

The piggyBac transposon system is currently being used in preclinical and clinical studies for CAR-T cells for hematologic and solid malignancies.²² Additionally, transposon-based approaches using lipid nanoparticles have enabled gene transfer to liver in mouse and non-human primate models.²³ The efficiency of transposon delivery and integration can be improved using minicircle vectors and hyperactive piggyBac elements.^24,25,26 The discovery of the piggyBac transposon structure may finally enable structure-based redesign for targeted transposon integration improving vector safety.²⁷ Therefore, additional refinements could improve transposon-based kidney gene delivery for pre-clinical and clinical application.

Nonviral delivery may be improved with kidney-targeted nanoparticles.^28,29 Retroureteral-catheter-based techniques already employed in patients could be harnessed for transposon delivery offering a non-viral therapeutic option for patients. Given the widespread morbidity and mortality of kidney disease along with dearth of transplant organs for end-stage renal disease patients, developing gene therapy for kidney disease is a high-priority area of investigation.

Materials and methods

Plasmids and gene delivery

pT-EF1α-Slc3a1 and pT-EF1α-Slc7a9 were generated by subcloning the mouse Slc3a1 and Slc7a9 cDNAs obtained from Origene (Rockville, MD) in place of the luciferase transgene in pT-EeL driven by the human EF1α promoter.⁶ pEF1α-m7pB has been described previously.³⁰ All plasmids were confirmed via DNA sequencing and prepared for in vivo use using endotoxin-free Maxipreps from Qiagen resuspended in Buffer QR (Mirus Bio, Madison, WI).

Hydrodynamic renal pelvis injection was performed as described previously.^6,7 A solution of 50 μg of pT-EF1α-Slc3a1 and 10 μg of pEF1α-m7pB resuspended in 100 μL of Buffer QR was loaded into an insulin syringe without a plastic safety guard and rapidly injected into the renal pelvis after surgical exposure of the kidney via flank incision.^6,7 The cystinuria type A mouse model and X-ray procedure to evaluate stone formation have been described previously.⁹ Mice were fed breeder chow as this increases the rate of stone formation.^8,9 All animal experiments were performed according to the Institutional Animal Care and Use Committee of Vanderbilt University Medical Center.

Detection of transgene expression and function

HK-2 (ATCC CRL2190) cells were seeded into 6-well plates at 1.5 × 10⁵ cells per well 1 day before transfection and were maintained in DMEM (Corning, Glendale, AZ) supplemented with 10% bovine growth serum (Cytiva, Chicago, IL), insulin-transferrin-selenium (Thermo Fisher, Waltham, MA), hydrocortisone (Sigma, St. Louis, MO), and penicillin-streptomycin (Corning, Glendale, AZ). Cells were co-transfected with 1 μg each of pT-EF1α-Slc3a1 and pT-EF1α-Slc7a9 with and without 0.5 μg pEF1α-m7pB using Lipofectamine LTX plus reagent (Invitrogen, Carlsbad, CA) according to manufacturer’s instructions. Cells were harvested on days 1, 3, 7, and 14 post-transfection for protein extraction and western blot analysis. Cells were resuspended in Pierce RIPA buffer (Thermo Fisher, Waltham, MA) supplemented with protease (Sigma, St. Louis, MO) and phosphatase (Roche, Indianapolis, IN) inhibitor cocktail. Protein concentration was determined by Pierce BCA protein assay kit (Thermo Fisher, Waltham, MA). Proteins were preheated at 70°C for 10 min with NuPage LDS Sample Buffer (Invitrogen, Carlsbad, CA) and NuPage Sample Reducing agent (Invitrogen, Carlsbad, CA). Each lane of a NuPage 4–12% Bis-Tris gel (Invitrogen) was loaded with 20 μg of protein. Proteins were transferred to nitrocellulose membrane using an iBlot 2 transfer system (Invitrogen, Carlsbad, CA). After transfer, immunoblots were blocked in Odessey Intercept blocking buffer TBS (LI-COR, Lincoln, NB) for 1 h at room temperature. Blots were incubated with rabbit α-rBAT primary antibody 16343-1-AP (Proteintech, Rosemont, IL) at 1:800 dilution and mouse anti β-Actin primary antibody (Invitrogen, Carlsbad, CA) at 1:1000 dilution at 4°C overnight. Antibodies were diluted in Odessey Intercept blocking buffer TBS (LI-COR, Lincoln, NB) + 0.2% Tween 20. The next day, the membranes were washed in TBS (Corning, Glendale, AZ) + 0.1% Tween 20. Secondary antibodies (IRDye-800 donkey anti-rabbit and IRDye-680LT donkey anti-mouse) (LI-COR, Lincoln, NB) diluted at 1:15,000 in Odessey Intercept blocking buffer TBS (LI-COR, Lincoln, NB) + 0.2% Tween 20 were added and incubated rocking at room temperature for 1 h. After additional washing in TBS +0.1% Tween 20 and final PBS (Corning, Glendale, AZ), immunoblots were imaged and quantified on the Odessey Imaging System (LI-COR, Lincoln, NB). β-actin (1:1,000, and mouse anti-βactin, Novus Biologicals, Centennial, CO, # NB600-501) was used for normalization of the western blot.

For microscopic analysis, paraffin slides were stained for the rabbit α-rBAT antibody 16343-1-AP (Proteintech, Rosemont, IL) diluted 1:200, secondary antibody AlexaFluor donkey anti-rabbit 594 (Thermo Fisher) diluted 1:500, and mounted in ProLong Gold Antifade Reagent with DAPI (Thermo Fisher) as previously described.⁶

For quantitation of RNA levels post-gene delivery, total RNA was prepared using the NucleoSpin/RNA/Protein kit (Macherey-Nagel, Allentown, PA). RNA was treated with DNase I (RNase-free) (New England Biolabs, Ipswich, MA) to remove the plasmid DNA. cDNA was synthesized from RNA (1,500 ng) using ProtoScript II reverse transcriptase (New England Biolabs) using an Slc3a1-specific primer (5′-CTAACACGAGCTATAGAGGATG-3′), purified using AMPure XP magnetic beads, and amplified using ddPCR OX200 (Bio-Rad, Hercules, CA) using the following primers: forward 5′-TCATTGGAGCCACCATAGCC-3’; reverse 5′-GATGGCAGCAACCAAATTCT-3′; and FAM-labeled probe 5′-GACTTCAGATACGCTGTTGA-3’.

For cystine uptake analysis, HK-2 cells were seeded into 6-well plates as described above. Cells were co-transfected with 1 μg each of pT-EF1α-Slc3a1 and pT-EF1α-Slc7a9 using Lipofectamine LTX plus reagent (Invitrogen, Carlsbad, CA) according to manufacturer’s instructions. One day after transfection cells were incubated with 30 μM or 60 μM BioTracker cystine-FITC live cell dye (Sigma, St. Louis, MO) for 1 h at 37°C. After incubation, cells were washed with HBSS. Imaging and flow cytometry were done immediately. Cystine-FITC uptake was first confirmed using a ZOE microscope (Biorad, Irvine, CA). The mean fluorescence intensity (MFI) of cystine-FITC was analyzed using Cytek Aurora flow cytometer (Cytek, Fremont, CA).

Phenotypic evaluation

To monitor stone formation over time, mice were imaged in a Faxitron 2000 X-ray machine at setting 35 for an exposure time of 4 s as described previously.⁸ Stone area was measured using ImageJ. Urine was collected from animals by placing them on a 96-well plate and covering them with a beaker until urine was collected. Urine amino acid concentrations were determined by reverse phase HPLC using a modified version of the methods of Bidlingmeyer et al.^8,31 The University of Alabama-Birmingham O’Brien Center Core C Biomarkers Laboratory measured the urine creatinine by LC-MS/MS.⁸

Data availability

The authors will make the original data presented herein available to the reader upon request.

Acknowledgments

M.H.W. was supported by the Ruth King Scoville Chair in Medicine at Vanderbilt University School of Medicine, the Department of Veterans Affairs (BX002190 and BX004285), and National Institutes of Health grants DK093660 and EB033676. L.E.W. was supported by the Department of Veterans Affairs (BX002797 and BX004845) and the Vanderbilt O’Brien Kidney Center (VCKD) (5P30DK114809-02). J.L.P. was supported by National Institutes of Health grants F30 DK134046 and T32 GM007347. This material is the result of work supported with resources and use of facilities at the VA Tennessee Valley Healthcare System.

Author contributions

L.E.W. and M.H.W. were responsible for conceptualization, formal analysis, and funding acquisition. T.A.I and M.H.W. provided supervision and resources. All authors contributed to investigation and methodology. M.H.W. conducted project administration, visualization, and writing of the original manuscript. L.E.W. and R.A.V. revised the manuscript.

Declaration of interests

L.E.W. and M.H.W. received funding for other projects from SalioGen Therapeutics and Bayer AG. M.H.W. has received funding for other projects from Torque Bio.