Site-specific kidney-targeted plasmid DNA transfer using nonviral techniques

A. Liposomes 1 HVJ-liposome

Tomita et al, (1992) first reported that the HVJ-liposome is efficient for glomerulus-targeted gene transfer. Its expression is transient. Four days after the injection of HVJ-liposome, SV40 large T antigen is detected immunohistochemically in 15% of the glomerular cells in the kidney. Isaka et al, (1993) demonstrated a 35% transfection efficiency of the glomerulus with plasmids carrying the gene for chloramphenicol acetyltransferase (CAT), transforming growth factor (TGF)-β, and platelet-derived growth factor (PDGF)-B genes, driven by the chicken β actin promoter. The blood urea nitrogen and plasma creatinine concentrations are not changed significantly by this procedure (Tomita et al, 1992; Isaka et al, 1993). Proteinuria is not observed in the rats treated with the HVJ-liposome method (Tomita et al, 1992; Isaka et al, 1993). Isaka et al (1993) demonstrated that the introduction of either the TGF-β1 or PDGF-B gene into the kidney induces glomerulosclerosis. Akami et al, (1994) reported the HVJ-liposome mediated gene transfer of a human CD59 expression plasmid, pSRαCD59, into the canine kidney. Human CD59 was detected immunohistochemically in the canine glomerular cells for 7 days with a maximum on day 2. Yamada et al (1995) identified a silencer in the mouse renin gene promoter region by transferring renin gene promoter-CAT fusion constructs into the kidney using the HVJ-liposome technique. Arai et al, (1995) demonstrated human renin expression in approximately 30% of the glomeruli on day 3 after the introduction of the human renin and angiotensinogen genes. In addition, on day 7 after the transfection of the renin and angiotensinogen genes, the extracellular matrix had expanded in the glomeruli and αsmooth muscle was expressed in the mesangial cells. These findings suggest that the locally activated renin angiotensin system induces glomerular sclerosis and a phenotypic change in mesangial cells.

The mechanism of the HVJ-liposome technique that results in glomerulus-specific gene expression is unknown. A possible explanation is that the intrarenal hydrodynamic state facilitates the contact of HVJ-liposomes with glomerular cell membranes, and another is that the phagocytotic activity of glomerular mesangial cells facilitates the uptake of HVJ-liposomes (Tomita et al, 1992).

B. Cationic polymer/DNA complex

Cationic polymer/DNA complexes have been used as nonviral vectors for gene transfer.

1. 25-kD Polyethyleneimine (PEI)

Boletta et al, (1997) demonstrated that efficient reporter gene expression was achieved by injecting DNA complexed to the branched 25-kD cationic polymer polyethylenimine (PEI 25k) into the kidney. The product of the reporter gene, β-galactosidase, was localized almost exclusively to the proximal tubules. The transfection efficiency of the branched PEI 25k was significantly higher than that of the linear form of PEI 22k, PEI 800k. A critical parameter in nonviral DNA delivery is the overall charge of the DNA-containing particles. Whereas a net positive charge is desired for particle interaction with the cellular plasma membrane and its entry into the cell, an excessive positive charge might favor entrapment of the DNA complexes by the extracellular matrix (Boletta et al, 1997).

To investigate whether the specific transfection of the proximal tubules involves glomerular filtration of the DNA-containing particles, Foglieni et al (2000) prepared fluorescent PEI 25k polyplexes containing fluoresceinated poly-L-lysine (mean diameter, 93 nm). This allowed visualization of the route of the particles into the kidney. Foglieni et al (2000) demonstrated that polyplexes that can transfect proximal tubular cells have access to these cells through glomerular filtration. Conversely, fluorescent lipoplexes containing the cationic lipid DOTAP (mean diameter, 160 nm) were never observed in tubular cells. The size of the transfecting particles is a key parameter in this process, as expected by the constraints imposed by the glomerular filtration barrier. To date, PEI 25k has been the most useful polymer for kidney-targeted gene transfer. Cationic polymer/DNA complexes injected via the left renal artery are not confined to the left kidney and are also entrapped by other organs (Boletta et al, 1997).

C. Electroporation

Tsujie et al, (2001b) first demonstrated that in vivo electroporation provides an efficient approach for glomerulus-targeted gene transfer. Expression vectors in

0.5 ml of balanced salt solution were injected into the left renal artery via a catheter in a one-shot manner and the left renal vein was clamped immediately after the injection. The left kidney was then sandwiched between a pair of oval-shaped tweezer-type electrodes and electroporated. Four days after the transfection of pCAGGS-lacZ, βgalactosidase expression was observed in 75% of glomeruli from the injected kidney. Thus, in vivo electroporation with intrarenal-arterial DNA injection was more effective than the HVJ-liposome method.

For this study, Tsujie et al, (2001b) used a luciferase expression plasmid, pEBAct-Luc, and transferred it by electroporation with 25, 50, 75, or 100 V. Luciferase activities in glomeruli did not change significantly at voltages of 25 to 100 V. The authors observed few harmful effects on the treated kidneys, except for small burns on the surface of the kidney in contact with electrodes. No histologic damage was seen in the glomeruli or tubular epithelial cells. A possible explanation of the glomerulus-specific gene expression is that, as the glomerular capillaries have abundant fenestrations, a large volume of DNA solution could also enter the mesangial area through them (Tsujie et al, 2001b). Electroporation-mediated gene transfer can deliver naked plasmid DNA into the entire area in contact with the DNA within the electric field. Therefore, the exact mechanism of the glomerulus-specific gene expression requires further study. We speculate that a hydrodynamics-based transfection by the injection of a large volume of DNA solution in a one-shot manner could achieve the desired gene transfer.

D. Ultrasound-microbubble (Optison)

Lan et al, (2003) first demonstrated that the ultrasound-mediated disruption of gas-filled microbubbles could be used effectively to transfer naked plasmid DNA into the kidney. The mechanism by which ultrasound microbubbles enhance transgene expression in all cell types within the kidney may largely be attributed to ultrasound-mediated microbubble cavitation. It is possible that the cavitation may largely increase the permeability of capillary and tubular basement membranes, which allows the locally released DNA to cross through these basement membranes and enter cells, including glomerular, interstitial, and tubular epithelial cells.

The ultrasound-Optison-mediated m2Smad7/Tet-on plasmid transfer resulted in transgene expression in more than 90% of glomerular, tubular, and interstitial cells. Gene transfer of inducible Smad7 using the ultrasoundmicrobubble system inhibited renal fibrosis in the rat unilateral ureteral obstruction model. The ultrasound treatment did not cause any abnormal histologic or functional changes, as evidenced by normal urinary protein excretion, normal glomerular and tubulointerstitial morphology, and the lack of cellular and interstitial edema and of local inflammation.

E. Naked plasmid DNA

It has not been possible to express transgenes in the kidney by injecting naked plasmid DNA via the renal artery (Lai et al, 1997, Boletta et al, 1997). Lai et al (1997) and Boletta et al (1997) performed intrarenal arterial injection of naked plasmid DNA while the renal arterial blood flow was interrupted but the renal venous blood flow was not. The hydrodynamics-based transfection mechanism did not work in either of these studies.

F. AVE-type HVJ-liposome ex vivo

The AVE-type HVJ-liposome method effectively induced the heat shock protein (HSP) 70 or the bcl-2 gene in kidney grafts. This was effective even when the HVJ-liposome vector was mixed with a cold-preservation solution (Ringer’s lactated or University of Wisconsin solution) and infused into the renal artery just prior to storage of the kidney at 4°C for 24 to 48 h in the same preservation solution (Kita et al, 2003). The induction of the HSP70 or bcl-2 gene reduced the occurrence of primary non-function of grafted rat kidneys after long-term preservation. The transgene expression in the kidney was limited to the tubules. In contrast, Tsujie et al (2001a), as described above, observed expression exclusively in the glomeruli when they introduced plasmid DNA using the AVE-type HVJ-liposome by renal arterial injection. The discrepancy between the two studies may be due to the difference in plasmid vectors used, or for other, as yet unknown, reasons.

A. Hydrodynamics-based transfection

Maruyama et al, (2002a) have developed a technique for transferring naked plasmid DNA into the kidney of normal rats by hydrodynamics-based transfection by retrograde injection of the DNA into the renal vein. When this technique was performed using a lacZ expression plasmid as the reporter gene, lacZ expression was detected exclusively in the interstitial fibroblasts near the peritubular capillaries (PTC) of the injected kidney, as assessed by immunoelectron microscopic analysis. No nephrotoxicity attributable to gene transfer was apparent either by histological or functional examinations of the injected kidney. These authors also used this method to transfect rat kidney with erythropoietin (Epo) gene. Maximal Epo expression was obtained when the vector, pCAGGS-Epo, was injected in Ringer’s solution within 5 sec, and with a volume of 1.0 ml. The transgene-derived Epo mRNA was detected by RT-PCR only in the targeted kidneys and without aberrant expression in nontarget organs. After an injection of 100 µg of pCAGGS-Epo, the serum Epo levels peaked at 208.3 ± 71.8 mU/ml at week 5, and gradually decreased to 116.2 ± 38.7 mU/ml at week

24. Transgene-derived Epo secretion resulted in significant erythropoiesis. The preparation of naked plasmid DNA is simple, compared with the preparation involved for other nonviral techniques. Moreover, the most long-term stable gene expression in the kidney has been obtained using this technique.

The mechanism underlying the transfer of the naked plasmid DNA into fibroblasts is unclear. The negatively charged (Dworkin et al, 2000) PTC could be refractory to the transfer of negatively charged naked plasmid DNA into the endothelium.

However, the PTC wall consists of an extremely thin endothelium, and fifty percent of the PTC endothelium is fenestrated, and thus highly permeable to water and small solutes (Lemley and Kriz, 1994). No incubation time is required for this technique; therefore, hydrostatic pressure (Liu et al, 1999) may be the mechanism underlying the transfer of the naked plasmid DNA. Efficient expression may depend on the elevated intravascular hydrostatic pressure caused by the rapid injection of a sufficient volume of fluid, which leads to the transfer of the naked plasmid DNA through the PTC endothelium, despite the negative charge. The sharp vascular resistance gradient between the efferent arteriole and the PTC (Lemley and Kriz, 1994) probably plays a major role in the blockade of the retrograde stream of injected naked plasmid DNA solution and in the dilation of the PTC, which are the most expandable sites in this route, resulting in transgene expression in the fibroblasts. Moreover, these cells may have the ability to take up the naked plasmid DNA.

Recently, Shimizu et al, (2003) demonstrated that the transfer of a kidney-targeted naked plasmid encoding 7ND (anti-monocyte chemoattractant protein-1) into the kidney interstitial cells using the technique of retrograde injection into the renal vein (Maruyama et al 2002a) attenuates the Table 2. Gene transfer via renal vein

Epo, erythropoietin; MCP-1, Anti-monocyte chemoattractant

tubulointerstitial renal injury induced by protein-overload proteinuria.

IV. Kidney-targeted gene transfer via the urinary tract Table 3 provides an overview of studies of kidney-targeted plasmid DNA transfer in a retrograde manner via the urinary tract. Two kinds of liposomes have been used in retrograde gene transfer via the pelvis or ureter. Using

them, investigators can target tubules and other interstitial fibroblasts.

A. Lipofectin via the pelvis

After the intrarenal pelvic injection of pCMV-β-gal-Lipofectin, lacZ was expressed mainly in the cytoplasm of renal tubular cells in the outer medulla with some staining in the cortex (Lai et al, 1997). The transgene expression site accessed by this route was similar to that accessed by the intrarenal arterial route, as described above. Lai et al (1997) speculated that the difference in the microenvironments of the inner and outer medulla might affect the interactions between the DNA and the liposome, therefore preventing gene uptake by the papillary cells, and resulting in a lack of staining in the inner medulla. Alternatively, the transgene taken up by the papillary cells may not have been expressed properly, owing to factors such as limited promoter efficiency and specificity, early degradation of the DNA, and altered half-life of the mRNA and/or protein synthesized from transgene. Intrarenal pelvic injection is more feasible than intra-renalarterial injection in mice, because it is difficult to inject an adequate amount of plasmid DNA-Lipofectin complex into the renal artery, and because intrarenal-arterial injection has a high incidence of renal ischemic injury (Lai et al, 1997).

Lai et al, (1998) also demonstrated a transient physiological effect of gene delivery into the kidney by the retrograde injection of a cationic liposome complexed with a carbonic anhydrase (CA) II chimeric gene, pCMV-CAII-Lipofectin into the renal pelvis of CAII-deficient mice. The CAII-deficiency was produced by introducing a point mutation into the CAII gene, and it manifested as renal tubular acidosis. The delivery of the CAII gene corrected the renal tubular acidosis of the CAII-deficient mouse model. Immunohistochemistry studies using anti-CAII antibodies showed that CAII was expressed in the tubular cells of the outer medulla and corticomedullary junction. The gene therapy was not associated with nephrotoxicity, as assessed by evaluating the blood urea nitrogen levels and the renal histology. This is the first successful gene therapy of a genetic renal disease.

B. Naked plasmid DNA via pelvis

It has not been possible to express transgenes in the kidney by injecting naked plasmid DNA via the pelvis (Lai et al, 1997).

C. AVE-type HVJ-liposome via

To examine the pattern of transgene-expressing cells, Tsujie et al (2000) used the N-lacF fragment containing the SV40 nuclear transport signal at the N terminus of the lacZ gene, which limits the β-galactosidase protein localization to the nucleus of transfected cells and can be distinguished from endogeneous β-galactosidase. pEBAct-NlacF was introduced into the kidney of normal rats retrogradely via the ureter using the AVE-type HVJ-liposome method (Tsujie et al, 2000). Nuclear βgalactosidase activity was observed in interstitial fibroblasts for 2 weeks. After the introduction of the pEBActN-lacF expression vector into kidneys with an obstructed ureter, Tsujie et al, (2000) observed a greatly expanded interstitium with marked cell proliferation on days 7 and 14. In contrast, no pathological changes nor increased numbers of infiltrated cells were observed in the treated kidneys 1 day after the transfection, suggesting that there was no interstitial damage from the HVJ-liposome solution, but that the damage was caused by ureteral obstruction. It has been speculated that the AVE-type HVJ-liposome may reach interstitial fibroblasts by slipping between either tubular epithelial cells or papilla epithelial cells; gene transfer could then occur via the fusion activity of the HVJ-liposome. Gene transfer targeting interstitial fibroblasts might be explained by the possibility that glycol-type sialic acids, the receptors for HVJ, are abundant in fibroblasts but rare in endothelial or epithelial cells (Tsujie et al, 2000). However, the precise mechanism remains to be elucidated.

Table 3 Gene transfer via urinary tract

A. In vivo catheter-mediated gene therapy

Putting the catheter technique to practical use, the above-mentioned gene transfer techniques via three different routes, renal artery anterogradely, renal vein retrogradely, the urinary tract (ureter or pelvis) retrogradely, can deliver therapeutic genes to the kidney less invasively than techniques that do not use catheters and therefore require an abdominal incision. Therefore, the development of catheters that are exclusively for gene transfer is key to the progress of catheter-mediated gene therapy. Similarly, ureterorenoscopes can be used to deliver therapeutic genes and specific adaptations to optimize their use should be undertaken (Lai et al, 1998).

B. Ex vivo gene therapy for kidney transplantation

To obtain successful outcomes following kidney transplantation, we need to overcome the common problems of post-transplant ischemic injury and rejection. Some of the above-described techniques (Tsujie et al, 2001b; Maruyama et al, 2002a; Lan et al, 2003; Kita et al, 2003) could be useful for ex vivo plasmid DNA delivery during kidney transplantation. The kidney to be used for grafting is readily accessible for gene transfer. In addition, the transgene expression in the graft that is induced by ex vivo nonviral techniques is, because of the nature of ex vivo transfection, confined to the injected kidney.