Serum amyloid A protein (SAA), a prominent component of the acute-phase response, is strongly expressed in developing and repairing kidneys and promotes tubulogenesis. Accordingly, we reprogrammed relatively undifferentiated NRK52E cells with the mouse SAA1.1 gene and transplanted SAA-positive and -negative cells into rats with acute renal failure. We found that SAA-positive cells accelerated renal recovery in three models of acute renal failure: gentamicin nephrotoxicity, cisplatin-mediated renal injury, and ischemia-reperfusion renal injury. The dramatic improvement of renal failure was demonstrable within 2 days, consistent with an early paracrine effect. However, abundant donor cells were also found integrated in the healing tubular architecture after 7 days. We conclude that infusions of SAA-positive cells promote renal recovery after acute renal failure and offer a potentially powerful and novel therapy of renal failure.
- acute-phase response
acute kidney injury (AKI) is widely unpredictable in occurrence, clinical course, and magnitude. AKI could just as easily resolve in part or completely, or it could be devastating, causing enough organ damage to terminate all kidney function. Moreover, in most cases the course of established AKI is largely uncontrollable and its medical management remains supportive (18). Work on developing specific and targeted therapies has continued over the years, and relevant scientific discoveries have been applied in experimental attempts to nullify the aggressiveness of AKI. In general, these applications involve replacement of elements thought to be needed for renal repair and presumed missing in the damaged kidney. For example, several growth factors have been tested with some success, and epidermal growth factor, hepatocyte growth factor, and insulin-like growth factor 1 are part of an earlier group of interventions. The main goal of this therapy was to promote intrinsic renal cell repair and growth and to limit cell death (9). More recent comparative work on AKI and renal ontogeny has uncovered shared mechanisms for reparative tubulogenesis and embryonic tubulogenesis (5, 27). Accordingly, factors that participate in kidney development have also being tested on AKI, including bone morphogenetic protein (28), glial cell line-derived neutrophic growth factor (25), and notch pathway ligand DLL4 (8). Current work has demonstrated a strong renal inflammatory component following AKI (14), and some have reported that mediators of inflammation could also be beneficial in AKI (4). The notion that components of inflammation can be protective is supported by our work showing that serum amyloid A protein (SAA), a classic component of the acute-phase reaction linked to inflammation, also promotes tubulogenesis (15). With the availability of bone marrow-derived stem cells, infusion of exogenous potential renal cell precursors became possible, and attempts have been made to replace and renew the damaged renal cell populations. Cytotherapy with injected bone marrow-derived stem cells limited AKI and were initially reported to assist recovery by direct integration into the damaged tubules (11, 23). However, while clear renal benefit was confirmed by other investigators, the claim for direct integration of stem cells into renal tissue was not verified, as stem cells could not be easily identified postinjection (7). Thus the benefit from stem cell infusions was thought to be secondary to a paracrine action rather than structural replacement by a cell population of exogenous origin (2, 26). In this work, we report that NRK52E cells reprogrammed to express tubulogenic mouse SAA1.1 (15) formed tubules in vitro and promoted recovery when these reprogrammed cells were injected into rats with AKI.
MATERIALS AND METHODS
NRK-52E cells were acquired from American Type Culture Collection (CRL-1571, Manassas, VA). Cells were stably cotransfected and then cultured on glass bottom culture dishes. The culture medium was DMEM (GIBCO Invitrogen, Carlsbad CA) with 5% FBS, in an atmosphere of 5% CO2-95% air at 37°C. NRK52E cells were transfected with Lipofectamine 2000 (Invitrogen) according to the manufacturer's protocol. In short, each plasmid DNA was dissolved separately at a concentration of 1 μg/μl in DMEM without serum or antibiotics, to which Lipofectamine was added at a proportion 1 μg DNA/2 μl Lipofectamine. The DNA and Lipofectamine complexes were applied to cells; the transfection medium was removed after 6 h and replaced with fresh DMEM with 5% FBS. The three plasmids used for cotransfection were pcDNA3.1-SAA1.1, a vector expressing mouse SAA1.1; pGFP1, a vector expressing green fluorescent protein (GFP; Ex 475 nM and EM 505 nM, Clontech, Mountain View, CA), and pNF-κB-SEAP (Clontech) a reporter plasmid that expresses secreted alkaline phosphatase (SEAP) upon nuclear binding to the NF-κB-regulatory sequences located upstream of the SEAP gene. SEAP released into the culture medium was measured by chemiluminescence using a Great Escape kit also from Clontech. The expression vector pcDNA3.1-SAA1.1 was manufactured in our laboratory. The murine SAA1.1 cDNA sequence, spanning ATG start and TGA stop sites, was amplified in two steps from a previously characterized clone (29). Primers F1 and R1/2 were used for the initial PCR. The product of that reaction was then amplified with primers F2 and R1/2:
PCR with the second primer set added NheI and BamHI restriction sequences (underlined above) at the 5′- and 3′-ends of the SAA sequence, respectively. The amplified fragment was then inserted between NheI and BamHI cloning sites in pcDNA3.1 (Invitrogen). The recombinant pcDNA-SAA1.1 was propagated in XL-10 Gold Escherichia coli (Stratagene, La Jolla, CA) and purified using a Qiagen Plasmid Purification Kit (Qiagen, Valencia, CA). Nucleotide sequencing of purified pcDNA-SAA1.1 confirmed its identity (ABI 3100 genetic analyzer, Applied Biosystems, Foster City CA). PCR amplification of transplanted SAA1.1 mRNA in homogenized kidneys was done using forward primer 5′- GAGTCTGGGCTGCTGAGAA-3′ and reverse primer 5′- TGTCTGTTGGCTTCCTCGT3′.
The cotransfected cells were selected for their resistance to G418 (200 μg/ml selection dose and 50 μg/ml maintenance dose, Sigma, St. Louis, MO) 24 h posttransfection. We selected four GFP-expressing cell lines after five passages in G418 (200 μg/ml). NRK52E-C1 and C2 cell lines lacked SAA.1.1 expression and did not form tubules. We retained C2, which had the strongest GFP expression, and used this line as control cells. Cell lines NRK52E-S1 and S2 expressed GFP and SAA1.1 and formed tubules in vitro after 15 days in culture. Tubule formation in vitro was more prominent in cell line S1, and it was used as the experimental cell line. The selected cotransfected cells also released SEAP, which was measured in the urine postinfusion (below).
The cotransfected and wild-type cells were cultured for 15 days and then were either fixed with 4% paraformaldehyde in preparation for immunohistochemistry (below) or subjected to fluorescein-uptake studies (below). Paraformaldehyde-fixed cells were incubated overnight with primary antibodies: rabbit anti-SAA1.1 (1:1,000) manufactured in our laboratory (15), rabbit anti-organic anion transporter 1 (OAT1; affinity-purified IgG) used at a concentration of 6.5 μg/ml (catalog no. OAT11-A; Alpha Diagnostic International, San Antonio, TX), and rabbit anti-rat sodium/hydrogen exchanger 3 (NHE3; affinity-purified IgG) used at a concentration of 20 μg/ml (catalog no. NHE31-A; Alpha Diagnostic International). Cells were rinsed in PBS and labeled with Texas red-conjugated goat anti-rabbit IgG (1:200, Jackson ImmunoResearch, West Grove, PA) for 1 h and the nuclear dye 4,6-diamidino-2-phenylindole (Molecular Probes, Eugene, OR) for 20 min. The cells were rinsed with PBS and imaged with a Zeiss LSM 510 confocal microscope equipped with UV, argon, and helium lasers (15). Apoptotic cells in culture and in kidney sections were identified as those with condensed, fragmented nuclei and expressed as the fraction of total nuclei in the image (14).
For functional studies on tubules formed in vitro, cultures and living cells were first incubated with fluorescein (2 μM) added to the culture media overnight. After rinsing of the cells in PBS five times, Hoescht 33342 was added (0.1 mg/ml, Molecular Probes) for 10 min to the media, cells were again rinsed in PBS, and imaged with the confocal microscope as described above. To assess cell proliferation, total DNA in transfected cells was quantified using the nucleic acid binding fluorophor PicoGreen (Invitrogen) according the supplier's instructions.
Cells were homogenized in 25 mM Tris, pH 7.6, 150 mM NaCl, 1% deoxycholate, 1% NP-40, 0.1% SDS, and 2× Halt Protease Inhibitor Cocktail (Thermo Scientific, Rockford, IL) and adjusted to a protein concentration of 2 mg/ml. The homogenates (20 μg) were fractionated by electrophoresis through 16.5% polyacrylamide Tris-Tricine gels. After electrophoresis, proteins were transferred to a nitrocellulose filter. Blocking was carried out in 1% casein, 1× PBS for 1 h. Incubation with primary antibodies diluted in 1× PBS was for 1 h; primary antibodies were rabbit anti-mouse SAA (1:2,000) generated in this laboratory (15) and mouse anti-actin (1:1,000, clone AC-40, Sigma-Aldrich). The filter was then washed in 1× PBS and incubated with secondary antibodies diluted in 1× PBS for 1 h. Secondary antibodies were IRDye 680 goat anti-rabbit IgG (1:15,000, Li-Cor Biosciences, Lincoln, NE) and IRDye 800 CW goat anti-mouse IgG (1:20,000, Li-Cor Biosciences). After washing in 1× PBS, the filter was scanned using the Odyssey Infrared Imaging System (Li-Cor Biosciences) for visualization of immunoreactive proteins. All steps were carried out at room temperature. SAA protein was also detected via ELISA (Immunology Consultant Laboratory, Newberg, OR) according to the supplier's instructions.
The experiments were conducted with 250-g Sprague-Dawley rats in conformity with the Guiding Principles for Research Involving Animals and Human Beings. The investigations were approved by the Institutional Animal Care and Use Committee of Indiana University School of Medicine. The rats were anesthetized with intraperitoneal (ip) pentobarbital sodium (50 mg/kg) and placed on a homeothermic table to maintain core body temperature at ∼37°C. After adequate anesthesia was ensured, renal ischemia was induced by occluding both renal pedicles for 30 min with microaneurysm clamps as described (14). In two other separate groups of rats, acute renal failure was caused with either gentamicin (100 mg·kg−1·day−1 twice daily for 7 days) or cisplatin (7.5 mg/kg ip, once). Creatinine was measured in sera via standard picric acid reaction. For intravital imaging (below), a small flank incision was made to expose the left kidney. For cell infusion studies, 106 stably cotransfected or wild-type NRK52E cells were trypsinized, washed, suspended in PBS, and infused via the tail vein. The number of cells infused was based on preliminary studies which showed that 3 × 105 cells resulted in minimal protection (mean serum creatinine of 0.77 ± 0.09 vs. 1.0 ± 0.15 mg/dl in controls) 48 h postischemia. An aliquot of cells prepared for infusion and passed through the same needle was always recultured to verify viability of the infused cells. The cell viability and growth tests always showed strong cell growth within 24 h of culture postinjection. The cells were infused after unequivocal rises in serum creatinine were observed following the renal injury: 5 days after the start of daily gentamicin injections, 4 days after the single injection of cisplatin, and 1 day after ischemia-reperfusion injury. Separate rats also subjected to the procedure of ischemia-reperfusion injury received either 5 × 106.
Intravital multiphoton fluorescence microscopy.
Intravital imaging was performed with a Bio-Rad MRC-1024MP confocal/multiphoton microscope (Hercules, CA) equipped with a titanium-sapphire laser (Spectraphysics, Mountain View, CA). Imaging was performed during cell infusion or 4 h later. The rats were placed on the heated (37°C) microscope stage and covered with a temperature-controlled pad. General anesthesia was accomplished with pentobarbital sodium (50 mg/kg ip) or thiobarbital sodium (80 mg/kg ip). The left kidney was surgically exposed, placed in a cell culture dish with a glass bottom (Warner Instruments, Hamden, CT), and bathed in warm 0.9% NaCl (14). Hoeschst 33342 (250 μg in 0.5 ml 0.9% NaCl, Molecular Probes) was injected intravenously (iv) immediately before imaging to identify nuclei and the focal plane. Renal microvascular flow was visualized using Texas red-conjugated large (100-kDa) dextran (400 μg in 0.5 ml of 0.9% NaCl iv) immediately before imaging. Excitation wavelength 800 nm, laser output (∼30%), and photomultiplier settings were chosen on the basis of prior studies (14), so that the fluorescence intensity of nuclei was ∼50% of maximum across different animals observed at different times.
Kidney sections were fixed in 4% paraformaldehyde and preserved in 30% sucrose before 10-μm sections were obtained. Fixed tissue sections were incubated with anti-mouse SAA1.1 and anti-mouse OAT1 or anti-mouse NHE3 followed by a Texas red-conjugated donkey anti-rabbit IgG (Jackson ImmunoResearch) and the nuclear dye 4,6-diamidino-2-phenylindole (Molecular Probes). Images were collected with a Zeiss LSM 510 confocal microscope and analyzed with Zeiss LSM software and MetaMorph (Universal Imaging, Downingtown, PA). Standard periodic acid-Schiff staining of kidney sections was also performed, and histology was graded on coded sections as described (17).
All image quantification was performed on coded images. Data are expressed as means ± SE. Analysis of variance was used to determine whether differences among mean values reached statistical significance. Tukey's test was used to correct for multiple comparisons. The null hypothesis was rejected at P < 0.05.
Phenotype of cotransfected NRK52E cells.
Initial experiments were performed to verify that stably cotransfected NRK52E cells expressing GFP also expressed SAA1.1. We employed rat cells because rats do not express the specific SAA1.1 protein isotype (22). The initial step was to determine whether cells expressing GFP and SAA were capable of functional tubule formation similarly to nontransfected NRK52E cells exposed to SAA (15). We selected and then maintained with G418 four stably cotransfected NRK52E cell lines that expressed GFP and either did or did not form tubule-like structures spontaneously. Cell lines C1 and C2 did not form tubules at all, and a light microscopic image of live C2 cells is shown in Fig. 1A. Although C2 cells expressed GFP, as shown by fluorescence microscopy in Fig. 1B, C2 cells lacked SAA expression (Fig. 1C). On the other hand, cell lines S1 and S2 readily formed tubule-like structures in vitro (Fig. 1, D and G). The average number of tubules counted per ×60 field for S1 and S2 cells was 14.4 ± 1.1 and 11.9 ± 1.2 tubules/field, n = 10. S1 cells expressed GFP (Fig. 1E) and also exhibited strong SAA expression, as shown by immunofluorescence in Fig. 1F. Moreover, the expression of SAA in cotransfected NRK52E cells was also confirmed by Western blotting (Fig. 1H). This panel shows that SAA was undetectable in lysates from cell lines C1 and C2. On the other hand, strong SAA immunoreactivity was easily detectable in lysates from cell lines S1 and S2. The two SAA-immunoreactive bands corresponded to approximate molecular weights of the SAA monomer (12 kDa) and SAA dimer (24 kDa) (15). SAA was not detected by immunoblotting or ELISA in the medium of SAA-positive or SAA-negative cells (data not shown).
Functional tubule-like structures in vitro.
One critical readout was the generation in vitro of tubule-like structures by cotransfected SAA-positive cells. We found that the SAA-positive NRK52E-S1 cell line was more prolific at forming tubules in culture, and it was used in subsequent experiments. These well-organized tubule-like structures also transported weak acids from media to lumens, as shown in the fluorescein transport experiments illustrated in Fig. 2, B–D. In contrast, the SAA-negative cell line NRK52E-C2 did not form tubules at all, and the cells lacked fluorescein uptake (Fig. 2A). NRK52E-S1 cells also manifested strong expression of the proximal transporters NHE3 (Fig. 2F) and OAT1 (Fig. 2, H–K). In contrast, control cells NRK52E-C2 did not express detectable amounts of NHE3 (Fig. 2E) or OAT 1 (Fig. 2G).
Renal cell transplantation.
Earlier experiments (15) revealed that SAA induction is a property of developing nephrons and recovering injured tubules, and we suggested that SAA, or other SAA isotypes, had a role in promoting renal tubule formation and repair (15). Accordingly, we set out to test this hypothesis by infusing SAA-positive or SAA-negative NRK52E renal tubular cells into rats in three separate models of tubular injury: gentamicin nephrotoxicity (6), cisplatin-induced renal injury (16), and renal ischemia-reperfusion (14). The subgroups included four rats, each infused with 106 control SAA-negative cells (NRK52E-C2), with SAA-positive cells (NRK52E-S1), or with wild-type NRK52E cells. The cells were infused after the rats developed acute renal failure, as indicated by unequivocally rising serum creatinine levels (Fig. 3). The gentamicin groups received gentamicin (100 mg/kg twice daily for 7 days), and renal cells were infused 5 days after gentamicin was started. The cisplatin groups received a single injection of cisplatin (7.5 mg/kg), and cells were infused after 4 days. The renal ischemia-reperfusion groups had their renal pedicles clamped for 30 min, and cells were infused 1 day postclamping. Acute renal failure was manifested in all three injury models, although progression and recovery patterns were different among the three injury groups (Fig. 3). While increases in serum creatinine levels exhibited a lag time in gentamicin nephrotoxicity, established acute renal failure did not recover within the period of observation. Serum creatinine in cisplatin injury was clearly higher at day 3 and then continued to climb before a subsequent incomplete recovery. The time course of renal failure in ischemia-reperfusion was more acute, in that serum creatinine increased rapidly 1 day postischemia. There was then a decline in the serum level followed by a plateau which remained higher than preischemia values. In contrast to cell infusions with SAA-negative cells, infusion of SAA-expressing NRK52E-S1 cells caused an earlier and sustained recovery in all three examples of renal failure (Fig. 3). This effect was noticeable by 24 h postinjection. In the gentamicin toxicity group, the sustained rise in serum creatinine was abrogated and leveled off for the remaining days of observation. In cisplatin toxicity, infusion of SAA-positive NRK52E-S1 cells was followed by a fall in serum creatinine which occurred earlier than in the group injected with SAA-negative cells. The effect of NRK52E-S1 cells on ischemia-reperfusion injury was dramatic and led to a near normalization of serum creatinine (Fig. 3). Infusion of a greater number of NRK52E-S1 cells (5 × 106) produced a similar improvement in renal function (mean creatinine 0.38 ± 0.07 mg/dl 48 h postischemia). Remarkable protection of renal histology in rats treated with NRK52E-S1 cells was also found (Fig. 3B and Table 1). In particular, SAA-positive cell transplantation was very effective in limiting the extent of tubular necrosis and decay 7 days after infusion. We examined the possibility that early effects of cell infusion were dependent on acute changes in microvascular renal blood flow, which was estimated from direct measurements of red blood cell velocity (Table 2). These measurements remained relatively constant and uniform before and after cell infusions.
Transplanted cells and recipient kidneys.
Transplanted NRK52E-C2 and NRK52-S1 cells harboring GFP and SAA genes were also cotransfected with the gene encoding SEAP. SEAP release is enhanced by a short NF-κB-activating sequence (materials and methods). Hence, we searched for and measured urinary SEAP in rats subjected to the three types of injury (Fig. 4). We compared SEAP levels in urine from rats injected with cells cotransfected with the SEAP plasmid (NRK52E-C2 and NRK52-S1 cells) and SEAP levels in the urines from rats injected with 106 wild-type (nontransfected) NRK52E cells. There was a rapid rise in SEAP urine levels within the 4 h of observation postinjection of NRK52E-C2 and NRK52-S1 cells. In marked contrast, injection of wild-type NRK52E cells did not increase urinary SEAP in any of the models of renal injury. SAA was not detected by either ELISA or immunoblotting in urine obtained 4 h after cell infusion or serum 24 h following cell infusion. The in vivo transit of transplanted NRK52-S1 SAA- and GFP-positive cells was also monitored in recipient kidneys with the two-photon microscope. In Fig. 5 is shown a remarkable sequence of images obtained after the intravenous injection of Texas red-conjugated dextran (which delineates the vascular space), followed 2 min later by injection of 106 NRK52-S1 cells. The injected GFP-labeled NRK52E-S1 cells are easily visible in abundant numbers circulating inside the peritubular capillaries 10 min postinjection, and 2 h postinjection they appeared to migrate out of the peritubular capillaries.
Renal engraftment of transplanted GFP-positive NRK52-S1 and NRK52-C2 cells was also examined in kidneys removed from rats 7 days following cell injections in the three models of acute renal failure. In Fig. 6 it is shown that 7 days postinfusion, transplanted GFP-positive NRK52-S1 cells and NRK52E-C2 cells had incorporated into the tubular architecture or were found adjacent to the luminal surfaces. There were occasional GFP-positive cells in the tubular lumina. Quantification of GFP-positive tubules and GFP-positive cells/tubules shows significantly greater engraftment of NRK52E-S1 cells. The number of transplanted GFP-positive NRK52-S1 cells and NRK52E-C2 cells encountered in the kidney, lung, and spleen is shown in Fig. 7A. In Fig. 7, A and B, is shown that GFP-positive NRK52-S1 cells and NRK52E-C2 cells were more abundant in the kidney than in either of the two other organs. In marked contrast, in the absence of renal failure, GFP-positive cells were rare (<1 GFP-positive cell/×63 field) in the kidney, lung and spleen. Immunostaining for proliferating cell nuclear antigen (PCNA) indicates increased renal mitoses in kidneys from rats treated with NRK52E-S1 cells (Fig. 8), consistent with increased proliferation of cultured NRK52E-S1 cells.
In earlier work it was shown that SAA induced the formation of functional tubules that expressed the renal transporters OAT1 and NHE3 in cultured wild-type NRK52E cells (15). Accordingly, we examined the expression of OAT1, NHE3, as well as SAA in rat kidneys following cell transplantation (Fig. 9 and Table 3). Immunoreactive OAT1 (red) was strongly expressed in renal tubular cells of rats injected with NRK52E-S1 cells in all three models of acute renal failure/acute kidney injury, but not in the kidneys of rats injected with control NRK52E-C2 cells. Immunoreactive SAA (red) was also found in renal tubular cells of rats treated with NRK52E-S1 cells but not in those that received NRK52E-C2 cells. The targeted expression of SAA in recipient kidneys was confirmed by PCR (Fig. 9B). NHE3 (red) was also expressed in renal tubular cells following injection of NRK52E-S1 but not NRK52E-C2 cells. Fluorescence intensity corresponding to OAT1, SAA, and NHE3 expression as well as GFP (indicating proportion of engrafted cells) in recipient kidney sections is shown in Table 3. The GFP signal was comparable in kidneys from rats that received NRK52E-S1 and those that received NRK52E-C2 cells. In contrast, the fluorescent signal of OAT1, SAA, and NHE3 were markedly (P < 0.05) higher in the kidneys from rats that received NRK52E-S1 cells compared with NRK52E-C2 cells.
Renal apoptosis largely reflects the magnitude of renal damage in acute renal failure and can be measured directly in the affected kidneys. Accordingly, we examined all available cell nuclei for apoptotic changes 7 days posttransplantation of cells (Fig. 10). The data are expressed as a percentage of total renal nuclei visualized. The fractional number of apoptotic renal nuclei was always higher in the kidneys transplanted with SAA-negative NRK52E-C2 cells than in SAA-positive NRK52E-S1 cells. The percentage of apoptotic renal nuclei averaged ∼7% in all three renal injury groups transplanted with NRK52E-C2 cells and ∼5% in all three injury groups transplanted with NRK52E-S1 cells. This is consistent with improved histology in the rats that received NRK52E-S1 compared with the rats that received NRK52E-C2 cells. We also found a significantly lower number of apopototic nuclei in NRK52E-S1 cells compared with NRK52E-C2 cells after 8 days in culture.
We tested the effect of SAA-expressing renal tubular cell transplantation in three rat models of acute renal failure: gentamicin nephrotoxicity (6), cisplatin-induced renal injury (16), and renal ischemia-reperfusion (14). The transplanted cells were rat proximal tubule NRK52E cells, either unmodified or reprogrammed to express GFP and without or with concomitant SAA expression. Infused NRK52E SAA-positive cells were chosen for transplantation because of our earlier finding that SAA is a powerful inducer of functional renal tubule formation in rat and mouse renal epithelial cells (15). Moreover, renal SAA is strongly expressed in forming mouse fetal tubules and it is induced in injured mouse renal tubules, suggesting that SAA, or a related SAA isotype, is involved in tubule formation and repair (15). Hence we reprogrammed NRK52E cells by cotransfection of plasmids expressing GFP, SAA, and with a third plasmid expressing SEAP under the control of the NF-κB-regulatory sequence. GFP and SEAP guided us in tracking the cells following transplantation by intravenous infusion. We found that transplantation with SAA-expressing tubular cells uniformly limited the degree of established renal injury in all three models, as indicated by sequential measurements of serum creatinine, assessment of renal histology, and degree of renal apoptosis, a valuable indicator of the severity of acute renal failure (13). Moreover, the significance in the degrees of apoptosis is likely to be considerable, considering that apoptotic cells are rapidly cleared from tissues (12).
We were able to track GFP-positive cells shortly after infusion and found them to be primarily localized to injured kidneys and in much lesser numbers in lung and spleen. The degree of renal cell implantation was also estimated by measurements of urinary SEAP and with determinations of renal cell GFP fluorescence. Both of these parameters were comparable in SAA-positive and SAA-negative cells. The renal protection afforded by transplanted SAA-positive renal epithelial cells was remarkable and in contrast to the lack of effect of SAA-negative renal epithelial cells which were infused in equal numbers. Thus it may be inferred that SAA cellular expression was directly or indirectly critical to the benefit reached by infused SAA-positive renal cell transplantation. However, our data are insufficient in revealing the detailed mechanisms renal protection achieved by SAA-positive cell transfer. The SAA-expressing cells constituted ∼10% of renal tubular cells and yet were associated with markedly improved function. The counterpart to our data are reports that seemingly small fractions of damaged renal cells can lead to severe renal functional deficits (17, 19, 21).
The infused cells acted relatively fast, and their effects appeared to last several days. The early improvement in renal function has been reported in other forms of cytotherapy of AKI (23, 24), and this early effect has been cited as reflecting a paracrine action (10). The early activity achieved by cell transfer is clearly desirable, for it is more likely to limit renal damage and subsequent progression to chronic renal failure (8, 18). The improved renal function was not derived from improved microvascular flow, which remained unmodified by cell transplantation, in agreement with prior reports derived in another cell transfer system (26). We are then left with the alternative that the early role played by SAA-positive cell infusion was a manifestation of paracrine or endocrine modulation of renal function. This broad conclusion leaves unanswered the basic question as to how SAA-positive cells, and other seemingly unrelated cells, improve renal function shortly after infusion (10). Reportedly, soluble factors released by donor cells (2) or transmission of mRNA contained in donor microvesicles (3) are potential mediators of renal repair. It is noteworthy that in contrast to other forms of cytotherapy (2, 7), transfer of SAA-positive renal cells was followed by robust integration of donor cells into the recipient tubular structure. We used two independent readouts for this purpose. First, our confocal microscope imaged SAA and GFP fluorescence on the same plane as the nuclei. Thus we were able to discern the fate of SAA- and GFP-positive tubular cells imbedded in the recipient renal tubules. The specific SAA readout we employed is also valuable because rats do not express SAA1.1, the presumed agent protein (22). Second, we were able to measure rising levels of donor-secreted alkaline phosphatase in the urine of the recipients. Accordingly, the mechanism of renal protection, although not fully addressed by these studies, seems to involve both early paracrine effects as well as enhanced proliferation of tubule cells following cytotherapy with SAA-expressing cells. However, we are not claiming that the relatively early integration seen at 7 days posttransplantation was required for eventual renal healing and repair. Rather, we propose that studies of longer duration and with a variety of readouts may shed more light on this important issue.
The mechanism of action of SAA on renal cellular reprogramming and redifferentiation is unknown. We previously suggested that a tissue-based form of the acute-phase reaction, with SAA as a major component, may participate in renal development in utero and in renal cell repair in adult life (15). This notion presupposes an exquisitely timed interplay between SAA and heretofore unrecognized pattern recognition molecules (15). This view has progressively gained experimental support, and newly described signaling mechanisms for SAA may advance this important topic (1). In any event, the core point of our report is that reprogramming relatively undifferentiated renal cells with the SAA gene committed those cells to a functional tubular phenotype in vitro. Moreover, these reprogrammed and committed cells integrated into recipient renal tubules and promoted renal healing in acute renal failure. Thus, while both SAA-negative and -positive infused renal tubular cells honed in on the damaged kidneys, only the SAA-positive cells improved renal failure parameters. We suggest that SAA by virtue of its expression in renal development and repair (15), and its role in promoting renal recovery after AKI, may constitute a novel therapeutic tool of AKI.
This work was supported by Clarian Health Values (K. J. Kelly) and the Department of Veterans Affairs (J. H. Dominguez).
No conflicts of interest, financial or otherwise, are declared by the authors.
The images were obtained at the Indiana Center for Biological Microscopy.