Bone Marrow Transplantation Can Attenuate the Progression of Mesangial Sclerosis



Bone marrow (BM) transplantation has been shown to provide beneficial effects in injured organs, including heart, liver, and kidney. We explored the therapeutic potential of BM transplantation (BMT) in Wilms' tumor suppressor 1 (Wt1) heterozygous mice, which represent a model of mesangial sclerosis. After transplantation of wild-type BM, there is statistically significantly lower urinary albumin and increased survival in Wt1+/ recipients. Control BMT using Wt1+/ donors showed no significant beneficial effects. The long-term beneficial effect of BMT was dependent on the dose of irradiation applied to the recipients before BMT. At a lethal dose of 1,000 cGy, the decrease in albuminuria and prolongation of lifespan in Wt1+/ mice were transient, with maximal amelioration at 12 weeks and resumption of albuminuria by 24 weeks after BMT. This was, at least in part, due to irradiation and not Wt1 heterozygosity because wild-type recipients also developed albuminuria within 24 weeks of BMT with 1,000 cGy. In contrast, Wt1+/ mice transplanted after 400 cGy showed long-term improvement in albuminuria and lifespan. Approximately 0.4% of podocytes were marrow derived, a level that is unlikely to be responsible for the therapeutic effects. In addition, donor BM cells formed rings surrounding the glomeruli, and approximately one third of the cells in these rings were macrophages. In conclusion, transplantation of wild-type BM attenuates progression of mesangial sclerosis in the Wt1+/ model of renal disease, and the mechanism by which this occurs may involve engraftment of BM-derived cells in the renal parenchyma.


The ultrafiltration barrier of the renal glomerulus is comprised of the fenestrated endothelial cell layer, the glomerular basement membrane (GBM), and the podocyte foot processes. Lesions in any part of this ultrafiltration unit can cause accumulation of mesangial matrix, proteinuria, secondary tubulointerstitial inflammation, and fibrosis [1]. The podocyte is perhaps the most complex cell type in the glomerulus; its foot processes act to counteract the elastic force of the GBM, and the contractile state of these foot processes, as regulated by vasoactive hormones, modulates ultrafiltration.

In mesangial sclerosis, the glomerular tuft becomes sclerotic within a dilated urinary space. There is usually a corticomedullary gradient of involvement, with the deepest glomeruli being the least affected. Tubules are severely damaged, especially in the deep cortex, where they are markedly dilated and often contain hyaline casts. Diffuse mesangial sclerosis (DMS) is the second most common cause of infantile nephrotic syndrome and is associated with glomerular injury and rapid progression to end-stage renal failure [2]. The same glomerular lesion is observed in Denys-Drash syndrome, which is characterized by a combination of nephropathy, male pseudohermaphroditism, and Wilms' tumor [3]. Mesangial sclerosis can occur due to mutations of the WT1 gene [4]. The zinc-finger transcription factor WT1, originally identified as Wilms' tumor suppressor 1, is crucial for development and maintenance of normal podocytes [510]. A primary defect involving epithelial cells or one of the components of the glomerular extracellular matrix has been proposed as the underlying cause in some cases of DMS. Recent findings in humans and mice have shown that mutations of WT1 are involved not only in Wilms' tumor but also in WAGR syndrome (characterized by Wilms' tumor, aniridia, genitourinary abnormalities, and mental retardation) [11], Frasier syndrome [12], Denys-Drash syndrome [13, 14], diffuse mesangial sclerosis [4, 15], and several types of tumors [16]. In addition to inherited diseases, podocyte injury also occurs in acquired diseases of the kidney, in which it leads to alterations in the glomerulus that again result in proteinuria. For instance, in membranous glomerulonephritis, a major cause of nephrotic syndrome, the accumulation of immune deposits in close contact with podocyte foot processes disrupts podocyte function [17].

Recovery of renal function after severe injury depends on the replacement of damaged epithelial cells. New epithelial cells may originate from kidney-resident and/or extrarenal cells, such as bone marrow (BM)-derived cells (BMDCs). Several studies indicated that BMDCs can participate in renal parenchymal repair both in mice and humans [18], as well as in the turnover of renal cells, such as tubular regeneration after acute ischemia/reperfusion injury [1820], albeit at a lower level than was previously reported [21]. BM-derived glomerular mesangial cells and endothelial cells have also been reported [2224]. Moreover, mesenchymal stem cells isolated from mouse BM may be renotropic and capable of contributing to renal repair after acute renal failure triggered by toxin administration [25, 26].

To date, mesangial sclerosis has been considered an irreversible disease leading rapidly to end-stage renal failure. Given the proposed high plasticity of BM cells, we explored the possibility of ameliorating the course of mesangial sclerosis by BM transplantation (BMT). The results obtained demonstrate that BMT can be of therapeutic benefit in mesangial sclerosis caused by loss of one copy of the Wt1 gene.

Materials and Methods

Animal Care and Bone Marrow Transplantation

Friend virus B NIH (FVB/N) wild-type mice were purchased from Charles River Laboratories (Wilmington, MA, Green fluorescent protein (GFP) donor mice expressing enhanced GFP from the constitutively active chicken beta-actin promoter and cytomegalovirus enhancer [C57BL/6-Tg(ACTbEGFP)1Osb/J] were purchased from Jackson Laboratory (Bar Harbor, ME, [27]. Wt1 heterozygous mice (K-mice) were genotyped as described previously [8]. Mice were maintained under approved housing conditions in accordance with federal and local regulations.

Four different combinations of BM donors and recipients were used: wild-type into wild-type (WT to WT), wild-type into K-mice (WT to K), K-mice into wild-type (K to WT), and K-mice into K-mice (K to K). There was no sex bias in the severity of kidney phenotype in K-mice. In most cases, the controls were littermates. Approximately half of the BMT used GFP+ donors. For the sake of simplicity in the text, GFP+ donors are categorized as wild-type in terms of the Wt1 locus. Four- to 8-week-old K-mice and their wild-type littermates on a pure FVB/N background were used in BMT throughout this study unless otherwise specified.

Donor BM isolation and lineage depletion by magnetic-activated cell sorting and BMT were carried out as described [28, 29]. Mice were irradiated using a Cs-137 source (∼150 cGy/min) at doses of 400, 600, 750, or 1,000 cGy. BM cells were transplanted via tail vein injection. Engraftment was verified in the blood 1 month after BMT and in the BM at the time of euthanasia by fluorescence-activated cell sorter analysis of peripheral blood when GFP donor was used and by Y-chromosome fluorescence in situ hybridization (Y-FISH) on cytospins of blood and BM when male-to-female transplantation was carried out.

Urine Analysis

Mouse albumin concentration was quantified using an albumin ELISA kit (Bethyl Laboratories, Montgomery, TX, following the manufacturer's protocol. Creatinine concentration was measured following standard protocol using picric acid under alkaline conditions to produce an orange product that was quantified by absorption spectrophotometry at 505 nm. All samples were measured in duplicate. ELISA data of urine albumin were standardized to creatinine levels. The standard deviation is relatively high due to the high variation in urinary albumin concentration between individuals. Protein gel analysis of urine samples was performed as described [8].

Histology, Y-FISH, and Immunofluorescence

Mice were anesthetized with an overdose of ketamine followed by systemic perfusion of 10 ml of 1× phosphate-buffered saline with 1 U/ml heparin. Kidneys were then cut into two pieces and fixed in buffered formalin for 4 hours or overnight at 4°C and embedded in paraffin, sectioned at 3 microns, and periodic acid-Schiff stained. To display the relative degree of renal pathology, we used a semiquantitative system as follows: 0, +, ++, +++, ++++, +++++, with the best (score = 0) being completely normal histology and the worst (+++++) being completely sclerotic, which was converted to a 0 to 5 scale for histogram presentation (Fig. 3). Y-FISH was performed as previously described [28].

Immunostaining on paraffin-embedded tissue was performed as described [30] using rabbit polyclonal antibodies against Wt1 (C-19, Santa Cruz Biotechnology, Santa Cruz, CA,, α smooth muscle actin (αSMA) (Abcam, Cambridge, MA,, and pancytokeratin (DAKO, Carpinteria, CA, Biotin-Lotus tetragonolobus agglutinin (LTA) (Sigma-Aldrich, St. Louis,, F4/80 antibody (eBioscience, San Diego,, and rat anti-mouse CD45 (BD Biosciences) were also used. Rabbit polyclonal anti-nephrin antibody was a kind gift from Dr. Larry Holzman (Ann Arbor, MI). Goat anti-GFP polyclonal antibody, rabbit polyclonal anti-desmin, and anti-von Willebrand factor (vWF) were purchased from Abcam (Cambridge, MA, All secondary antibodies were from Molecular Probes (Eugene, OR, Antigen retrieval was carried out in BD Retrievagen A in a steamer for 15 minutes, except for Nephrin, wWF, and desmin antibodies, for which the antigens were retrieved in Signet retrieval 2 buffer (Signet Laboratories, Inc, Cambridge, MA, at 92°C for 2 hours.

Apoptosis was detected on formalin-fixed, paraffin-embedded tissue using the in situ Cell Death kit from Roche Diagnostics (Mannheim, Germany,, following the manufacturer's recommended protocol.

Statistical Analysis

Data were analyzed using Student's t-test, with p < .05 regarded as significantly different.


Inbred Wt1-Heterozygous Mice (K-Mice) Represent an Excellent Model of Mesangial Sclerosis

Homozygous deletion of Wt1 abolishes development of the urogenital system, adrenal glands, and spleen and leads to abnormalities in epicardial development, resulting in embryonic lethality [6, 7, 31]. In contrast, Wt1 heterozygous mice, which were originally reported to be viable and healthy on a mixed genetic background [6], have since been shown to represent an excellent model for mesangial sclerosis even on a mixed background of CBA, SWR, 129/ola, C57Bl/6, and Balb/C [8, 32]. After being backcrossed more than 15 generations to FVB/N mice, Wt1 heterozygous mice (K-mice hereafter) develop renal abnormalities even earlier than those on a mixed background; they start to develop albuminuria by 3 weeks of age, at which time histological changes are still very subtle or undetectable in the kidneys. Therefore, urinary albumin loss is used as an early and sensitive marker to monitor mesangial sclerosis in Wt1+/ mice. Mesangial sclerosis progresses rapidly such that more than half of the mice develop complete glomerulosclerosis in less than 24 weeks and only 44% of FVB/N K-mice survive to 200 days of age when left untreated (Fig. 1, upplemental Table 1, Supplemental Fig. 1). This phenotype is in sharp contrast to the later onset and slower progression of renal disease when Wt1+/ mice are maintained on a mixed background (Supplemental Table 1). Therefore, inbred FVB/N K-mice represent an excellent genetic model for studying potential therapeutic strategies for mesangial sclerosis and are used in this study.

Bone Marrow Transplantation After High-Dose Total Body Irradiation Improves Survival and Transiently Ameliorates the Progression of Mesangial Sclerosis

BM Transplantation Improves the Survival of K-Mice

To determine the effects of BMT on the survival of K-mice, we transplanted either 1 × 105 lineage-depleted marrow cells from β-actin-GFP C57Bl/6 donors [27] or 1 × 106 unfractionated BM cells from wild-type FVB/N male donors into male and female recipients, respectively. Recipients included K-mice and wild-type control mice that were first exposed to a lethal dose (1,000 cGy) of total body irradiation (TBI). Two months after BMT, the recipients had approximately 90% donor-derived nucleated blood cells, indicating that hematopoietic engraftment had occurred (Table 1). A Kaplan-Meier plot of the survival of mice from each group is shown in Figure 1. Compared with the 65% 24-week survival rate of nonirradiated, untransplanted K-mice, K-mice that underwent BMT with wild-type BM had a statistically significant improvement in survival to 87% (p = .03), and at 28 weeks, compared with the 44% survival rate of nonirradiated, untransplanted K-mice, K-mice transplanted with wild-type BM had a survival rate of 76% (Fig. 1). In contrast, K-mice transplanted with BM from K-mice did not have a statistically significant difference in survival from untransplanted K-mice (p = .95, data not shown). As controls, we also assessed wild-type recipients of BM from wild-type and K donors. These two groups had nearly identical survival at 24 weeks, which was not statistically significantly different from that of K recipients that received WT BM (data not shown). These data indicate that BMT is therapeutically beneficial for mesangial sclerosis due to low Wt1 levels.

BMT Decreases Urinary Albumin Loss

Damage to podocytes affects ultrafiltration in the glomerulus and leads to proteinuria/albuminuria. To monitor the status of mesangial sclerosis, we assessed the ratio of albumin to creatinine in urine over time. K recipients of wild-type BM after lethal irradiation started to show a decrease in the albumin/creatinine ratio by 2 weeks after BMT, and this statistically significant beneficial effect extended for 12 to 16 weeks but was then followed by an increase in albuminuria (Fig. 2A). Each K recipient of WT BM displayed the same tendency of having the albuminuria levels decreased for the first 12 to 16 weeks after BMT, followed by an increase. This late increase was likely due to radiation nephropathy secondary to TBI because wild-type recipients of wild-type BM also had progressive albuminuria starting approximately 12 weeks after BMT (Fig. 2A).

To monitor renal status over time, urine samples were also analyzed by protein electrophoresis to compare whole protein composition before and after BMT at various time points. As expected, K recipients of wild-type BM showed a dramatic reduction in proteinuria during the first 12 weeks after BMT, before the adverse effects of TBI reversed the course (see representative data in Fig. 2B).

BMT Attenuates Progression of Mesangial Sclerosis Pathology

To determine whether the decrease in proteinuria was due to decreased pathology, the renal histology of the BMT recipients was assessed at different times after BMT. To display the degree of pathology, we used a semiquantitative system, as follows: 0, +, ++, +++, ++++, +++++, with the best (score = 0) being completely normal histology, and the worst (+++++) being completely sclerotic, consistent with end-stage renal disease. The score was converted to a 0 to 5 scale for graphic display. Normally, adult K-mice (6 to 12 weeks old) on the FVB/N background develop mesangial sclerosis with mild tubular atrophy, protein casts, and occasional interstitial inflammation. After undergoing BMT with wild-type marrow, K recipients showed significant improvement in histology as evidenced by the absence of tubular atrophy and no or fewer microcysts, as well as decreased interstitial inflammation compared with untransplanted littermate controls (8 to 12 weeks, p < .005; 13 to 20 weeks, p < .05; Fig. 3). This tubulointerstitial improvement lasted as long as 4 months after BMT. There was mild accumulation of extracellular matrix in the glomeruli of K recipients throughout the time course (Fig. 3). Compared with their untransplanted K littermates, K-mice transplanted with wild-type BM had a far better histology score (p < .05). In the two control groups (WT to WT and K to WT), mesangial sclerosis started to develop approximately 4 months after BMT (Supplemental Fig. 2b shows the WT to WT group), for which it is likely that the lethal dose (1,000 cGy) of irradiation was the cause. Taken together, these data suggest that BMT of K-mice can eliminate or slow the progression of mesangial sclerosis for 4 months, after which detrimental side effects of TBI occur. To further analyze glomerular integrity, we studied the staining pattern of the slit-diaphragm protein nephrin [33], which is a sensitive indicator for the status of the podocyte foot processes. Not surprisingly, wild-type control mice had a normal, global staining pattern, and K-mice had a less complex pattern, indicating retraction of podocyte foot processes. K-mice transplanted with wild-type marrow showed normalization of the nephrin staining pattern, although still less intense than that of wild-type controls (Fig. 4A).

Impact of Different Doses of TBI on Renal Damage

Based on the development of glomerulopathy in the wild-type mice transplanted after 1,000 cGy of irradiation, we tested the effects of BMT after different doses of TBI on wild-type FVB/N mice. In contrast to the progressive development of renal pathology starting 12 to 16 weeks after BMT with 1,000 cGy, mice that received 400 cGy (Table 1), 600 cGy (data not shown), or 750 cGy had normal histology (Table 1). Moreover, analysis of the apoptosis rate in the 1,000-cGy group using an in situ cell death kit (Roche Diagnostics) revealed that there is an obvious increase in apoptosis compared with lower doses of irradiation. The apoptotic cells were scattered throughout the kidney and were not restricted to any particular structure or cell type (data not shown). This indicated that programmed cell death caused by high-dose radiation could be the primary reason for radiation nephropathy.

As shown in Table 1, high levels of hematopoietic engraftment were obtained with irradiation doses as low as 400 cGy. Therefore, we compared the extent of renal pathology in mice undergoing BMT after different doses of irradiation. Whereas 100% of mice transplanted after 1,000 cGy developed renal pathology, none of the mice transplanted with 400 cGy showed any abnormality and 750 cGy gave intermediate results. Therefore, we next tested whether BMT can have long-term beneficial effects in K-mice after irradiation with just 400 cGy.

BMT with Sublethal Irradiation Significantly Attenuates the Progression of Mesangial Sclerosis and Improves Long-Term Survival Without the Adverse Effects of TBI

We applied 400-cGy TBI to 4-week-old wild-type and K-mice and transplanted 17 × 106 unfractionated wild-type BM cells. BM engraftment of 64% ± 11% was achieved (Table 1). None of the wild-type recipients developed albuminuria throughout the 24 weeks of monitoring after BMT, and all had normal renal histology at the end of the 24-week period, validating that sublethal TBI at 400 cGy does not have a deleterious effect on the kidney. Of the six K-mice transplanted after 400 cGy, three were killed 8 weeks after BMT for analysis, at which time all three had low to mild levels of albuminuria, with a histology score of + to ++. The remaining three all survived for more than 29 weeks. In contrast, 75% (n = 8) of unmanipulated littermate control K-mice died with an average lifespan of 22 ± 3 weeks. Only two of these eight untransplanted K-mice littermates survived over 29 weeks, and both had severe albuminuria and a histology score of +++. In contrast, the transplanted K-mice had significant improvement in histology (p < .05), with two mice having normal histology (score of 0) and one having only minimal glomerulosclerosis (histology score of +). The urinary albumin/creatinine levels also reflected the long-term improvement in K-mice after BMT with 400 cGy. Whereas unmanipulated K-mice developed severe albuminuria within 12 weeks, K-mice recipients of wild-type marrow had low levels of albuminuria (Fig. 2C). It is also noteworthy that sublethal irradiation alone does not improve the renal pathology of K-mice (data not shown). Thus, in contrast to the data with lethal TBI, sublethal TBI followed by transplantation with a high dose of BM cells leads to long-term improvement.

Syngeneic BM Administration Without Irradiation Provides No Therapeutic Effect

To examine the necessity of hematopoietic reconstitution in the setting of BMT-mediated therapy of mesangial sclerosis, we carried out control experiments of BM administration without prior irradiation. We transplanted 5 × 106 sex-mismatched BM cells from wild-type FVB/N donors into 3-week-old K-mice and wild-type littermate controls. Although 80% of the wild-type recipients (n = 5) survived longer than 200 days, none of the K-mice recipients (n = 4) survived for longer than 200 days, nor was there any improvement in albuminuria compared with untransplanted K control mice (data not shown). Graft-versus-host disease (GvHD) can be excluded as the cause of death because the donor and recipients were all FVB/N and histologic analysis of organs showed no sign of GvHD. Moreover, without irradiation, the BM cells administered did not engraft the hematopoietic system. No male cells were identified in the kidney or BM using Y-FISH (data not shown). Based on these data, the therapeutic effect of BMT on renal diseases likely requires BM engraftment.

Donor Bone Marrow Contributes to Renal Regeneration

To better understand how BMT exerts its beneficial effects, kidney sections from recipient mice obtained at different time points after transplantation with 1,000 cGy or 400 cGy were analyzed by Y-FISH and immunofluorescent staining to identify any marrow-derived epithelial cells. In the absence of severe inflammation, 0.5%–4% of the cells in the kidney of female recipients of male BM were positive for Y-FISH, most of which (>95%) were in the interstitium and at least 30% of which were hematopoietic cells based on CD45 expression. To identify marrow-derived podocytes, a combination of Y-FISH and immunofluorescence for Wt1 was performed. It should be noted that only podocytes express high levels of the nuclear protein Wt1 that can be detected in the kidney by immunofluorescence. Even though Wt1 can be detected by reverse transcription–polymerase chain reaction from total nucleated peripheral blood cells from both healthy humans and wild-type mice (our own unpublished data and [34, 35]), as well as from mobilized peripheral blood CD34+ cells and BM, these cells do not have adequate levels of Wt1 protein to be detectable by immunofluorescence using the polyclonal antibody (WT1, C-19) used in this study (our own unpublished data and [35]). Therefore, in the kidney, all cells staining for nuclear Wt1 protein are podocytes, which is supported by the fact that all are in the glomeruli and that this staining is significantly weaker in wild-type mice than in K-mice (Wt1+/−) (Fig. 4Bb). Two weeks after BMT, donor-derived podocytes (double-positive for Y-FISH and Wt1) were detected by Y-FISH and Wt1 colocalization at a level of 0.3%–0.4% of total podocytes (Table 2). This rate increased such that by 8 to 16 weeks after BMT, the highest level reached 0.75% and the marrow-derived podocytes (Y+Wt1+) observed in the WT-to-K group appeared to have higher expression levels of Wt1 compared with the endogenous podocytes within the same glomerulus, as judged by immunostaining (Figs. 4Bc, 4Bd). The presence of marrow-derived podocytes was also confirmed by combining anti-GFP and anti-Wt1 staining. Cells double-positive for GFP and Wt1 comprised approximately 0.6% of total podocytes (Table 2, Supplemental Fig. 3), which is consistent with the degree of engraftment of marrow-derived podocytes detected with Wt1 and Y-FISH. Donor-derived podocytes ranged from one to six cells per 100 glomeruli on the same tissue section. Bearing in mind that technically Y-FISH will not stain every male nucleus visible on the slide due to the small size of the Y-chromosome and the 3-micron thickness of the section, one needs to normalize the result by a factor of 1.25 to correct for sampling in a male control (80% of Y+ podocytes on a section of male kidney). Because the 100 glomeruli studied have approximately 1,000 podocytes in the plane of section, approximately 0.1%–0.75% podocytes were donor marrow derived.

Fifteen female recipients that had received 1,000-cGy radiation were examined intensively for evidence of donor-derived tubular epithelial cells by combining Y-FISH with staining for the proximal tubule markers wide-spectrum cytokeratin and/or lectin from Tetragonobulys purpurea. Anti-mouse CD45 immunostaining was included to verify the nonhematopoietic phenotype of the donor cells. No Y+Wt1+ cells or Y+LTA+ cells were CD45+. Donor-derived renal tubular epithelial cells were very rare. From serial kidney sections covering 12 microns in depth (4 × 3 μm thickness) from each mouse, only 36 tubular cells in 7 out of 15 mice were double-positive for Y-FISH and LTA, which accounts for fewer than 0.03% of the proximal tubular epithelial cells (Fig. 4Bf), and there were no double-positive tubular epithelial cells detected using Y-FISH and pancytokeratin staining. In contrast, approximately 10% of interstitial cells were donor derived, as discussed below.

Attempts to Identify Nonepithelial Cells of Donor Origin

It is of great importance to determine whether donor BM contributes to regeneration of glomerular mesangial and/or endothelial cells in our model system. We attempted to costain sections with Y-FISH plus the mesangial cell marker desmin and with Y-FISH plus the endothelial marker vWF. However, the immunostaining results did not provide definitive results because desmin and vWF staining are cytoplasmic, and, given the long thin shape of these cells, it was not possible to definitively determine whether an adjacent Y+ nucleus was from the same cell even with confocal analysis of 8-micron sections (data not shown). We were also unable to perform colocalization staining in combination with GFP because the GFP transgene driven by the chicken beta-actin promoter, although claimed to be universally expressed, is not expressed in mesangial and endothelial cells of the glomerulus even in GFP+ donor mice.

Because the very low levels of marrow-derived podocytes are not likely to be adequate to achieve the therapeutic benefits obtained with BMT, we also performed analyses of the marrow-derived interstitial cells in the kidney. Myofibroblasts are ubiquitous cells with features of both fibroblasts and smooth muscle cells. Direkze et al. [36] recently reported that BM can contribute to myofibroblast populations in a variety of tissues and that this is enhanced by injury. To assess for marrow-derived myofibroblasts in our animal model, we examined the kidney sections by costaining for Y-FISH and αSMA, a marker for myofibroblasts. Although many cells were positive for Y-FISH or αSMA, double-positive cells were not found. From analysis of more than 100 sections from multiple different time points (2, 4, 8, 12, 16, 20, and 24 weeks after BMT), there were no Y+SMA+ interstitial cells detected in K-mice transplanted with wild-type BM or in K-mice transplanted with K BM (data not shown); thus, marrow-derived myofibroblasts, at least in this study, do not seem to play a role in the ameliorative effect of BMT.

Many donor-derived macrophages surrounded the glomeruli in recipient mice at distinct times after BMT. Two weeks after BMT, approximately 0.5% of cells in the kidneys were F4/80-positive macrophages, with more than 90% of them being of recipient (GFP-) origin. In contrast, untreated wild-type and Wt1+/− mice had at least 100 times fewer macrophages in the kidney. Although the number of F4/80-positive cells decreased in the kidney by 4 weeks after BMT, approximately 10% of the glomeruli were surrounded by GFP+ cells, and of these GFP+ cells, one third were F4/80 positive (Fig. 5). The ring structures disappeared by 8 weeks after BMT. It will be important to determine what controls the formation of these cells surrounding individual glomeruli and whether the appearance and disappearance of these interstitial macrophages play a significant role in the process of BMT-mediated attenuation of mesangial sclerosis.


In this study we show that BMT has a beneficial effect on the mesangial sclerosis caused by Wt1 deficiency. BMT attenuates the progression of mesangial sclerosis for at least 12 to 16 weeks if recipients undergo 1,000-cGy TBI and for at least 24 weeks if recipients undergo 400 cGy TBI, as evidenced by reduced urinary albumin loss, improvement of renal histology, and prolonged survival compared with untreated littermates. In the 1,000-cGy TBI group, the beneficial effect of BMT is overridden by the side effects of radiation nephropathy. The mechanisms underlying this amelioration are not clear. Although up to 0.75% of podocytes are donor derived and have high levels of Wt1 expression, this level of engraftment is unlikely to be adequate to achieve the therapeutic benefit of BMT in K-mice. It remains to be shown whether the amelioration is due to paracrine factors produced by donor-derived macrophages. This proof-of-principle study strongly suggests that total BM contains a cell population that is capable of ameliorating the progression of mesangial sclerosis and that BM-derived renal cells, including possibly podocytes, tubular epithelial cells, and interstitial cells, may account for the positive effects of BMT. Note that the very low level of engraftment of renal tubular cells (<.03% detected by lectin + Y-FISH but not by cytokeratin + Y-FISH) is unlikely to be related to the therapeutic benefits of BMT. Further studies are needed to determine whether the facilitation of resident renal stem cells by donor BM cells is involved in the renal improvements observed.

Our findings are consistent with studies by other investigators. By analyzing female kidneys that had been transplanted into male patients, Poulsom et al. [18] provided evidence that circulating Y chromosome–positive cells could differentiate into renal tubular epithelial cells. However, the frequency of BM-derived tubule cells was reported to be low. Kale et al. [19] and Lin et al. [20] observed that BM-derived cells were capable of tubular repair after ischemia/reperfusion injury. Although they initially reported a high frequency of donor-derived renal tubular epithelial cells (∼20%) using a chemical staining approach (X-gal staining to identify LacZ+ cells), much lower frequency of marrow-derived tubular epithelial cells (∼0.1%) was found by Y-FISH in the follow-up studies by the same laboratory [21]. Lin et al. [37] recently reported that resident renal tubular epithelial cells contribute more significantly than marrow-derived cells in renal repair after ischemia/reperfusion (I/R) injury. Also using the I/R model, Duffield et al. [38] revealed that BM-derived cells do not contribute significantly to the restoration of epithelial integrity after injury. Moreover, Szczypka et al. [39] recently reported extremely rare incorporation of BM-derived cells into the kidney after folic acid-induced injury (none by staining of kidney sections; a cluster of 7 out of 4 million cells in culture). Although a different model of injury was used, these data indicate that donor BM contributes to the healing process of renal injury, which likely occurs via pathways other than the transformation of donor cells into renal epithelium.

Most prior studies in which BM-derived cells have been identified were performed using C57Bl/6 mice, an irradiation dose range of 350 to 1,100 cGy, and analysis of tissues within 4 to 12 weeks after BMT [19, 20, 24], except Szczypka et al. [39], who monitored for 9 months. It is likely that none of these studies showed the adverse effects of radiation nephropathy in their mice due to the fact that the C57Bl/6 strain is less sensitive to radiation than is the FVB/N strain and because the time of examination after BMT was less than 16 weeks when a dose of 1,000 cGy was used. Radiation nephropathy in mice has been reported to occur at least 16 weeks after exposure to 800 cGy administered at a high dose rate of 71 cGy/min, and the severity increases with higher doses [40, 41].

Compared with littermates and age-matched untransplanted controls, K recipients of wild-type BM had normal or nearly normal tubular histology. Although donor-derived tubular epithelial cells were extremely rare, it is possible that the tubules could return to a normal status without generation of tubular epithelium from extrarenal sources when the upstream leakage of protein is decreased due to improved ultrafiltration. This could also occur due to an increased ability of the remaining renal tubule cells to take up albumin, perhaps with the help of BM-derived interstitial cells.

A role for the migration of resident renal progenitor cells from papilla or any other potential niches in renal epithelial repair was not assessed in our study. Oliver et al. [42] reported that the renal papilla provide a niche for renal progenitor/stem cells. Bussolati et al. [43] recently isolated renal progenitors, which are capable of differentiating into epithelial and endothelial lineages, from adult human kidneys, suggesting that the adult kidney may have its own resident repair machinery. Togel et al. [44] revealed that transplanted MSCs can protect kidneys from acute ischemic injury, but probably through paracrine actions, rather than by differentiation into renal cells such as tubular epithelial cells. Interestingly, we detect donor-derived cells surrounding glomeruli starting as early as 2 weeks after BMT (Supplemental Figs. 4, 5). Although these cells do not stain as epithelial cells (pan cytokeratin), mesangial/smooth muscle cells (desmin), endothelial cells (vWF), or myofibroblast/smooth muscle cells (αSMA), up to one third of them do stain positive with the macrophage-specific anti-F4/80 antibody. More interestingly, before the formation of donor-derived rings in the interstitium, endogenous macrophages accumulated in a scattered pattern throughout the entire renal interstitium and formed clusters in arches or ring-like structures outside of the glomeruli. We speculate that endogenous macrophages are attracted to local tissue injury via specific signaling pathways after lethal irradiation and that the cells then recruit donor-derived cells to the damaged region to provide a therapeutic effect. Further investigation will be needed to determine the underlying mechanisms.

The plasticity of BM cells to become nonhematopoietic cells has been questioned due to findings suggesting that cell fusion, i.e., BM cells fusing with tissue specific cells, could be responsible for the appearance of marrow-derived epithelial cells. A recent study from our laboratory demonstrates that cell fusion is absent or very rare under normal BMT conditions but occurs, at least in skeletal muscle and liver, when additional damage is induced [45]. Future studies will assess whether the mechanism of podocyte regeneration requires cell-cell fusion.

Taken together, our data provide strong evidence that transplanted BM contributes, to a small extent, to the regeneration of glomerular and tubular epithelial cells and that BMT has an ameliorative effect in mesangial sclerosis, as indicated by a dramatic decrease of urinary albumin loss, attenuation of renal pathology, and extended lifespan.

Table Table 1.. Relationship between total body irradiation dose, BMT, and renal outcome
original image
Table Table 2.. Summary of donor-derived podocytes at different time points
original image
Figure Figure 1..

Bone marrow transplantation extends the lifespan of K-mice. (A): Percentage of animals surviving at 200 days in each of six experimental groups. The white and black bars represent wild-type (WT) and K-mice, respectively. Control group: untransplanted mice, n = 6 for wild-type, n = 16 for K; K-donor group, n = 13 for WT, n = 8 for K; WT donor group, n = 10 for WT, n = 21 for K. (B): Kaplan-Meier plot of survival up to 24 weeks of age. Left, Comparison of untransplanted WT (dotted line) and K (solid line) mice (p < .05). Right, Comparison of untransplanted K-mice (solid line) and K-mice that were transplanted with Wild-type bone marrow after 1,000 cGy (dotted line) irradiation (p < .05).

Figure Figure 2..

Albuminuria is improved in K-mice after bone marrow transplantation (BMT). (A): Plotted is the mean ± standard deviation of the albumin concentration normalized to creatinine for urine samples from different experimental groups (X-axis) over time after BMT as indicated in the figure. The value shown in the Y-axis is albumin (mg/dl)/creatinine (mg/dl). Student's t-test between age-matched groups (K vs. WT to K, and K vs. K to K) revealed that only 4-week and 8- to 12-week post-BMT groups (K vs. WT to K) differ significantly (**p < .005). No other statistical significance was found; K group, n = 64; WT to K, n = 14; WT to WT, n = 9; K to K, n = 5; K to WT, n = 16. #Standard deviation is 14. (B): Representative example of the changes of urine composition after BMT examined by Coomassie brilliant blue-stained protein gels. Legends indicate the age in weeks of the mouse when the sample was collected. The wild-type mouse shows a trace amount of albumin within the normal range that was undetectable using albumin diagnostic dip-sticks (Bayer). Untransplanted K-mice show a gradual increase in the albumin level over time. The seven lanes on the right are from the same K-mouse that had undergone BMT at 4 weeks of age. They show a decreased level of albumin secretion within the first 12 to 16 weeks, with a subsequent increase in urine albumin. (C): Urinary albumin analysis of mice from the 400-cGy group. Abbreviations: WT to WT, WT recipients of WT marrow, killed 24 weeks after BMT, n = 6; WT, untransplanted wild-type control, n = 3; K, untransplanted K-mice control (only three time points are included in this figure, n = 8 [0 weeks after BMT], n = 7 [4 weeks after BMT], and n = 5 [8 weeks after BMT] because the case number dropped over time); WT to K, K-recipients of Wild-type BM, n = 6 (0 to 12 weeks after BMT), n = 3 (16 to 24 weeks after BMT).

Figure Figure 3..

Bone marrow transplantation (BMT) attenuates the pathological progression of mesangial sclerosis. (A): Histology of four littermates killed at the age of 20 weeks. (a): Untreated wild type with normal histology; (b): untreated K-mouse with numerous dilated tubules filled with protein casts (*) and showing a mild elevation of glomerular extracellular matrix (ECM); (c, d): mice that underwent BMT with wild-type donors at the age of 4 weeks and were killed 16 weeks after BMT for analysis; (c): wild-type recipient; (d): K-mouse recipient. Note that the kidneys shown in (c) and (d) have mild ECM accumulation. Although it is not known whether the sclerotic tufts in (d) are caused by irradiation or the Wt1 genetic defect, it is clear that mouse (c) developed glomerulosclerosis from total body irradiation. Insets show higher-power images of individual glomeruli. Original magnification ×200. (B): Overall histology of the different experimental groups. For untransplanted K-mice, n = 16, 6, and 13 for subgroups of 8 to 12, 13 to 20, and >20 weeks, respectively. WT to K group: n = 5, 6, and 16 for age subgroups of 8 to 12, 13 to 20, and >20 weeks, respectively; WT to WT group: n = 4, 4, and 12 for subgroups of 8 to 12, 13 to 20 and > 20 weeks, respectively; K to K group: only one age group > 20 weeks was scored, n = 8; K to WT: only one age group > 20 weeks was evaluated, n = 11. There is a significant decrease in severity of mesangial sclerosis in the WT to K group compared with the untransplanted K control group (**p < .005 and *p < .05 for 8- to 12-week and 13- to 20-week groups, respectively). However, the effect of BMT is overridden by the side effects of TBI over time, as evidenced in the WT to WT group. #, not available.

Figure Figure 4..

Characterization of kidneys by immunostaining. (A): Representative patterns of nephrin staining for wild-type control (a), K-mouse control (b), wild-type recipient of wild-type bone marrow transplantation (BMT) (c), K-recipient of wild-type bone marrow (d). Note that although (c) and (d) have a less complicated nephrin staining pattern than (a), they have more nephrin staining than (b), the K-mouse control. Original magnification ×400. (B): (a–d): Y-chromosome fluorescence in situ hybridization (Y-FISH) (red) and Wt1 immunostaining (green) to identify donor-derived podocytes; (e, f): Y-FISH and Lotus tetragonolobus agglutinin (LTA) staining (green) to identify donor-derived tubular epithelial cells. (a): Male wild-type control; (b): male K-mouse. Not all of the cells in male controls show a Y-chromosome due to partial nuclear sampling in these 3-μm sections. Note that the wild-type shows higher levels of Wt1 expression (green nuclear stain) than does the K-mouse (ad were littermates, processed together under identical conditions). (c, d): Female wild-type and K-mouse recipients of wild-type male bone marrow, respectively, 4 months after BMT. (c): Glomerulus with a donor-derived podocyte (white arrow) by Y-FISH and Wt1 staining. (d): One donor-derived podocyte (white arrow). Boxed regions are shown in the individual channel images. Top inset: 4′,6′-diamidino-2-phenylindole staining of nuclei, blue; middle inset: Y-FISH, Texas red; bottom inset, Wt1, AlexaFluo-488, green. (e): Male wild-type control showing Y+/LTA+ proximal tubules. The arrow indicates representative Y+/LTA+ tubular epithelial cells. (f): A female recipient of male wild-type bone marrow 16 weeks after BMT showing a Y+/LTA+ tubular epithelial cell (arrow). Original magnification ×400. Three-micrometer paraffin sections, original magnification ×400.

Figure Figure 5..

Macrophage infiltration in bone marrow transplantation (BMT) recipients. Green fluorescent protein staining in red, F4/80 staining in green, and nuclear staining of 4′,6′-diamidino-2-phenylin-dole in blue. (a): Wild-type control showing minimal detection of macrophages by F4/80 antibody (the same is also true of the K-mice); arrow indicates a macrophage. (b): Two weeks after BMT, recipients have a dramatic increase in the number of macrophages in the kidney, and these cells tend to form ring structures just outside of the glomeruli. Boxed area is split into three channels in (b); (c): 4 weeks after BMT, recipients still have high number of resident macrophage in the kidney, and the rings are composed of mainly donor-derived cells. Boxed area is split into three channels in (c); (d): 8 weeks after BMT, the ring structures decrease in size and eventually disappear (arrow points to a donor-derived macrophage). (b, c): Channel splits of images shown in (c) and (d), respectively (original magnification ×400). Arrows indicate double-positive cells, with the color of the arrows representing the individual channels (red, green, or blue).


We thank Stephanie Donaldson and the Yale University YARC facility for excellent mouse care and David Tuck for assistance with Kaplan-Meier plots. We also thank Dr. Michael Kashgarian for his invaluable advice. We are grateful for the rabbit polyclonal anti-nephrin antibody kindly provided by Dr. Larry Holzman (University of Michigan).


The authors indicate no potential conflicts of interest.