In this paper we demonstrate a protective role of hAFSC when injected directly into nu/nu mouse kidneys with glycerol-induced ATN. This model of ATN involves a complex sequence of events wherein myoglobin, released from damaged muscle, damages the epithelial cells of the proximal tubules, producing cast formation, vasoconstriction and decreased glomerular filtration. Peak of the damage represented by high levels of apoptotic cells, increased levels of creatinine and BUN and by histological analysis was confirmed to be between 48–72 after induction of ATN.
The ability for hAFSC to home to injured organs was confirmed using luciferase positive hAFSC detected their bioluminescence after injection. As shown in our previous paper 
the amount of hAFSC detected by bioluminescence in nu/nu
mice, wherein the acute injury was performed in the lung instead of the kidney, decreased over time, showing a strong signal in the lung at the beginning that fades over the next few days. Nevertheless herein following intrarenal injection after ATN kidney damage, the bioluminescent signal is still present in the kidney after 21 days and hAFSC were found specifically located among the tubules. In order to determine if the injected cells can differentiate into tubular epithelial cells we performed both immunohistochemistry and RT-PCR using human specific antibodies and human primers. Genes such as PAX2
and lectins like Dolicholus Biflorus and Peanut Agglutinin are expressed by injected cells, indicating that at least some of the integrated cells are able to commit toward renal differentiation. Furthermore, in some rare instances, injected hAFSC cells could express human Glial Derived Neurotrophic Factor (GDNF
), which is expressed during very early kidney development; GDNF
is not usually detectable in the adult kidney, indicating that injected hAFSC can express also embryonic renal markers, retaining the ability of going through the nephrogenesis. We have previously shown that hAFSC have the potential to integrate into embryonic kidneys and can participate in key steps of nephrogenesis, indicating that hAFSC can be induced toward to a renal fate when placed in an appropriate environment 
. However, both in this study as well as in this previous study, the efficiency of integration into kidney lineages was not strikingly high and we did not detect significant clonal expansion of integrated hAFSC. We, therefore, postulated that perhaps another benefit could be afforded the injured kidney than simply a structural one by the presence of hAFSC during the acute phase of injury.
We determined that hAFSC have the ability to modulate kidney function during ATN, as reflected in serum creatinine and BUN levels. However, the timing of hAFSC is critical: hAFSC injected during the acute phase of ATN (between 48–72 hours after the IM injection of glycerol), had no effect on creatinine and BUN levels (data not shown), and implies that if injury is already established the damage is not attenuated. In contrast, hAFSC injected into the kidney on the same day of glycerol injection resulted in no observed peaks in creatinine or BUN. This beneficial effect with hAFSC was also correlated with significant increases in proliferative activity of tubular epithelial cells, decreased cast formation, and decreased apoptosis of tubular epithelial cells.
We therefore speculate that hAFSC can, when injected early enough (in this study contemporaneously with the time of injury) attenuate acute renal damage, underscoring the potential protective effects of hAFSC. Moreover, even 14 days after injury hAFSC injection into the kidney still correlated with increased tubular cell proliferation and less tubular damage. Thus, we hypothesize that hAFSC might accelerate the proliferation of epithelial tubular cells that were only partially damaged, while in addition preventing apoptosis. This mechanism of protection therefore appears to lead to better maintenance of tubular structure, as seen in PAS staining, thus avoiding the increase in BUN and creatinine typically seen in IM-glycerol-induced ATN.
During acute renal injury the immune response plays a key role. Damaged kidney endothelial cells attract leukocytes, vasomediators are released with injury, and epithelial cells of the tubule produce pro-inflammatory and chemotactic cytokines 
. Bonventre et al. 
and Lin et al. 
have shown recently that the mechanism by which bone marrow stem cells contribute to renal repair was by attenuating the immune response, rather than through integration or differentiation of the stem cells into the cells of the damaged organ. They also speculated that the protection in these animals was not through integration and differentiation of the injected MSC, because of the very short period of time with which a protective response was observed 
The animal model used in these experiments (Athymic Nude-Foxn1nu
) was chosen in order to be able inject human derived stem cells and evaluate their effect over time, avoiding outright rejection. Nu/nu
mice are immunodeficient, lacking activated T Lymphocytes but not their precursors. They have normal B Lymphocytes and they have evidence of increase numbers of NK cells. Thus, even though they are immunodeficient, they do possess some ability to mount and modulate a partial immune response when stimulated 
. In addition to the animals that underwent only ATN and mice with no treatment, we introduced mice that were injected with a vehicle saline solution (PBS) after glycerol damage and compared their respective cytokine levels. Therefore, to further investigate the potential mechanisms by which hAFSC enhance renal protection, we examined intrarenal cytokines, to determine whether there is a general change in inflammatory cytokine pattern in mice that were treated with hAFSC compared to our controls. We decided to evaluate the immune response in a time period extending up to two weeks, instead of only the first 3 days, because even if this is an acute injury model, the response of nu/nu
mice with glycerol induced ATN to human stem cells after kidney injection has never been demonstrated before, and the response could diverge significantly from wild type mice injected with mouse-specific stem cells.
The kidneys of the ATN mice treated with hAFSC presented a profile of mouse cytokines that indicated a lower expression of pro-inflammatory molecules, as compared to mice treated with either ATN plus PBS or that underwent only ATN, versus the basal level in mice before any treatment. One possible explanation for this result may be the increased presence of IL-10 and IL1Ra, both of which are anti-inflammatory cytokines, in the mice treated with hAFSC cells. Moreover, the presence of both of these two anti-inflammatory interleukins appears to have stimulated the production of IL-6, which also possess anti-inflammatory properties in the presence of increased levels of IL-10 and IL1Ra 
. In mice with ATN and treated with PBS or only with ATN, there is no change in expression of IL-10 and IL-6. Therefore we conclude that inflammation can persist over a longer period of time in these non-hAFSC treated mice. Furthermore, mice treated with hAFSC did not show increased levels of important pro-inflammatory cytokines, such as IL27, a Natural Killer cell attractor 
or IL12p70, which is known to be inhibited by IL-10 
. Furthermore, the injection of the cells prevented an increase in SDF-1 (Stromal cell-Derived Factor-1) a potent B chemoattractor, which is produced by several cell types 
, thus indicating that many different immune functions can be slowed down or controlled by injection of hAFSC. Specific chemoattractants such as KC (Keratinocyte Chemoattractant), JE/MCP-1 (Monocyte Chemoattractant Protein 1), MCP-5 (Monocyte Chemoattractant Protein 5) and MIP-2 (Macrophage Inflammatory Protein 2) are also diminished in mice treated with hAFSC.
It is beyond the scope of the current study to define the specific role of each individual cytokine involved in the immune response in kidney injury and repair. Nevertheless, the total effect of the cytokines and chemokines expressed in the ATN kidneys of mice treated with hAFSC evidently lead to a combined action helping to ameliorate the acute phase of injury.
It is also important to note that in ATN mice that were injected with hAFSC, all cytokines (pro and anti-inflammatory) had returned to normal by 14 days, particularly when compared with ATN mice injected with PBS or with just ATN, i.e., in the latter mice, inflammation persists longer. This is also in accordance with the presence of tubular damage in the latter groups at 14 days. This suggests that an important function of hAFSC may be to actually prevent the acute injury. We further speculate that, in this particular nu/nu
model of ATN, the beneficial down-modulation of the most important immune cytokines may actually be relatively delayed, because nu/nu
mice, lacking activated T lymphocytes, can induce alternative pathways of immune responses. One great practical advantage of studying the impact of hAFSC in the nu/nu
model of ATN is that we could measure and compare the cytokines expressed by the human cells versus those the recipient mouse kidney. Interestingly, human cytokine levels (compared with the basal level secreted in the hAFSC supernatant in vitro
before injection) are significantly increased early in the course of ATN (as shown Table S3
), while the mouse cytokines are only significantly increased after 1 or 2 weeks. Thus the human cells are able to produce cytokines that can modulate the nu/nu
mouse response to ATN.
With a few exceptions, most human cytokines are also active on mouse cells 
, so both the combined effects of human and mouse cytokines may have affected the kidney inflammation and tissue homeostasis milieu. We think that this may be an important concept, and therefore consider that maybe cytokines derived from both the injected hAFSC and the endogenous mouse cytokines are responsible for the observed protective effects. In addition, it is relevant to underline the decrease in number of injected hAFSC found in the kidney over time, as shown both by the bioluminescence images and by the immunohistochemistry data, as well as by the absence of human cytokine expression at 14 days. What is essential to underline in this work and in this mouse model of ATN is that, in nu/nu
mice treated with hAFSC, there is over all improvement in the maintenance of tissue homeostasis, which prevented progression of the acute phase in glycerol-induced ATN, most likely through stimulating proliferation of tubular epithelial cells and by cytokine-mediated paracrine mechanisms. Furthermore, after the acute phase, the decrease of creatinine levels, as well as less cast formation and the increase of tubular cell proliferation is still statistically significant at two weeks, at which time damage still persists in the ATN mice treated only with PBS or not treated at all. This is also in accordance with the data on cytokine levels that are still significantly elevated at 1 and 2 weeks in ATN mice treated with PBS only or not treated at all, as compared to ATN mice treated with hAFSC in which cytokine levels had resolved by 1–2 weeks.
To rule out whether hAFSC alone can cause damage to the kidney and/or produce cytokines once injected in vivo, we injected them into nu/nu mice without ATN induction. In these mice, injected only with hAFSC there was no alteration of creatinine or BUN, no increase in apoptosis or proliferation or tubular damage; thus indicating that hAFSC per se do not alter the normal physiology and morphology of the kidney. However injection of hAFSC alone did stimulate the production of some cytokines by nu/nu mouse kidney, but their levels were relatively low and only detectable predominantly during the first few time points, and rapidly resolved back to normal before 14 days, as compared with ATN mice injected with PBS. In addition, in these mice, there was an increase of levels in IL-10 and IL-1Ra and IL-6 following hAFSC injection (similar to that noted in the mice injected with hAFSC following ATN), there was no increase in SDF-1 and also they expressed lower levels of NK attractors as well as macrophage attractors. Thus, hAFSC do stimulate a modest tissue reaction by nu/nu mice, but this did not cause any measurable perturbation of kidney function. In addition, the human cytokines produced by the injected cells in mice that did not undergo ATN disappear quickly and with no apparent adverse consequences.
In conclusion, we have demonstrated that early direct injection of hAFSC into the kidney strongly ameliorates ATN injury, as reflected by more rapid resolution of tubular structural damage, by tubular cell proliferation and by normalized creatinine and BUN levels. In addition, our data show evidence of potent immunomodulatory effects of hAFSC that appear to control the local immune response in favor of a tissue cytokine and cellular milieu that promotes prevention or resolution of tissue damage. Taken together these findings suggest that hAFSC may have therapeutic potential in ATN and, by extrapolation, perhaps in other kidney diseases.