Sphingosine Kinase 2 Deficiency Attenuates Kidney Fibrosis via IFN-γ

Abstract

Maladaptive repair after AKI may lead to progressive fibrosis and decline in kidney function. Sphingosine 1-phosphate has an important role in kidney injury and pleiotropic effects in fibrosis. We investigated the involvement of sphingosine kinase 1 and 2 (SphK1 and SphK2), which phosphorylate sphingosine to produce sphingosine 1-phosphate, in kidney fibrosis induced by folic acid (FA) or unilateral ischemia-reperfusion injury. Analysis of Masson trichrome staining and fibrotic marker protein and mRNA expression 14 days after AKI revealed that wild-type (WT) and Sphk1^−/− mice exhibited more kidney fibrosis than Sphk2^−/− mice. Furthermore, kidneys of FA-treated WT and Sphk1^−/− mice had greater immune cell infiltration and expression of fibrotic and inflammatory markers than kidneys of FA-treated Sphk2^−/− mice. In contrast, kidneys of Sphk2^−/− mice exhibited greater expression of Ifng and IFN-γ–responsive genes (Cxcl9 and Cxcl10) than kidneys of WT or Sphk1^−/− mice did at this time point. Splenic T cells from untreated Sphk2^−/− mice were hyperproliferative and produced more IFN-γ than did those of WT or Sphk1^−/− mice. IFN-γ blocking antibody administered to Sphk2^−/− mice or deletion of Ifng (Sphk2^−/−Ifng^−/− mice) blocked the protective effect of SphK2 deficiency in fibrosis. Moreover, adoptive transfer of Sphk2^−/− (but not Sphk2^−/−Ifng^−/−) CD4 T cells into WT mice blocked FA-induced fibrosis. Finally, a selective SphK2 inhibitor blocked FA-induced kidney fibrosis in WT mice. These studies demonstrate that SphK2 inhibition may serve as a novel therapeutic approach for attenuating kidney fibrosis.

AKI is one of several initiating events that contribute to chronic progressive kidney disease.¹ Some patients with AKI fully recover renal function, whereas in others the development of CKD is accompanied by a progressive decline in kidney function, ultimately leading to ESRD. Maladaptive repair in response to severe injury or multiple insults after AKI contributes to the progression to CKD and ESRD.^1–3 Regardless of the cause of CKD (nephrotoxic kidney injury, ischemia, infection, genetics, paraneoplastic syndromes, immunologic processes), there is a stereotypical response leading to interstitial fibrosis, tubular atrophy, and peritubular rarefaction and inflammation.⁴ A key feature is the activation of extracellular matrix–producing myofibroblasts.⁵ Other factors important in CKD progression include endothelial cell and vascular damage in AKI,⁶ hypoxia-HIF,⁷ innate and adaptive immunity,⁸ cell cycle arrest,³ and epigenetic mechanisms.⁹ Although some injured tubules may undergo repair and regeneration, injury may be accompanied by inflammation, maturation, and proliferation of fibroblasts, and extracellular matrix deposition as part of the process of fibrosis. The source of fibroblasts in the injured kidney (fibrocytes, epithelial cells through EMT, intrinsic fibroblasts, and pericytes) remains controversial.^10,11

Sphingosine 1-phosphate (S1P), a pleiotropic lysophospholipid that is involved in diverse functions such as cell growth and survival, lymphocyte trafficking, and vascular stability,^12,13 is the product of sphingosine phosphorylation by two sphingosine kinase (SphK) isotypes. SphK1 and SphK2 have different subcellular localizations^14,15 and may serve different functions. Localization of SphK1 near the plasma membrane provides S1P that is exported out of the cell (inside-out signaling)¹⁶ by spinster homolog 2 (SPNS2),¹⁷ and other less well characterized mechanisms where it can bind to its five different G-protein coupled receptors (S1P1–5).¹⁸ SphK1 promotes cell survival and proliferation and regulates cell transformation.¹⁹ Less is known about SphK2; it may serve proapoptotic functions,^20,21 although some studies show prosurvival properties.²² SphK2 in the nucleus,¹⁵ mitochondria, and endoplasmic reticulum points to an intracellular signaling role for locally produced S1P.²³ Nuclear S1P is a histone deacetylase (HDAC) inhibitor,²⁴ which allows gene expression to be induced, however no direct targets have yet been discovered. SphK2 also phosphorylates FTY720,²⁵ which is an immunomodulatory prodrug used in the treatment of multiple sclerosis.²⁶

We found that Sphk2^−/− but not Sphk1^−/− mice are more susceptible to kidney ischemia-reperfusion injury (IRI)²⁷; a more modest but opposite effect was found by others,²⁸ whereas both studies showed upregulated Sphk1 mRNA after IRI. In addition to acute injury, S1P also plays a role in fibrosis.^29–32 The involvement of S1P in many cell processes, and hence its potential for involvement in disease, has prompted interest in the development of SphK1 and SphK2 inhibitors for treatment of cancer and inflammatory diseases.³³ Given the importance of S1P in kidney injury and inflammation and the central role of inflammation in fibrosis,⁸ we sought to determine the roles of SphK1 and SphK2 in kidney fibrosis and their potential as therapeutic targets.

Results

Sphk2^−/− Mice Are Protected from Folic Acid-Induced Fibrosis

In our previous AKI studies,²⁷ susceptibility to bilateral IRI was similar in Sphk1^−/− and wild-type (WT) mice but greater in Sphk2^tr/tr mice²⁷ (another Sphk2^−/− mouse strain with the same phenotype as the Sphk2^−/− mouse used in the current study). The subsequent availability of Sphk2^−/− mice allowed comparison to WT and Sphk1^−/− mice (all on a B6 background). Additional experiments replicated the earlier finding that the increase in plasma creatinine 24 hours after IRI is greater in Sphk2^−/− mice than in WT or Sphk1^−/− mice (Supplemental Figure 1A). Furthermore, the greater susceptibility to acute injury in Sphk2^−/− mice is likely because of increased Th1 responses, as deletion of Ifng blunts the injury (Sphk2^−/− Ifng^−/−) (Supplemental Figure 1B). IRI increased expression of Sphk1 but not Sphk2 mRNA in WT mouse kidneys for up to 96 hours after IRI,²⁷ and expression of S1p1 and S1p3 mRNA 4–6 hours after injury.³⁴ In the current study we found that mRNAs encoding all five S1P receptor subtypes and both SphK isotypes are expressed in the kidney (although the expression of S1P4 and S1P5 was negligible). In our chronic injury models, i.e., fibrosis in kidneys induced either by unilateral IRI or by administration of folic acid (FA), there was a small increase in Sphk2 mRNA expression (1.4–2.3-fold, relative to control) and somewhat larger (2–9-fold, relative to control) increase in expression of S1p1–3 and Sphk1 mRNAs in kidneys 14 days after unilateral IRI (compared with the contralateral uninjured kidney) or FA (Supplemental Figure 2).

To investigate the role of SphKs in progressive fibrosis, WT, Sphk1^−/−, and Sphk2^−/− mice were subjected to FA-induced kidney injury (experimental timeline in Supplemental Figure 3). WT and Sphk1^−/− mice displayed a significant rise in plasma creatinine that was greater at day 3 than at day 14 after FA, while creatinine in Sphk2^−/− mice was elevated at day 3 but not at day 14 compared with its vehicle control (Figure 1A). Similarly, kidneys of WT and Sphk1^−/− mice displayed a marked increase compared with vehicle in extracellular collagen deposition 14 days after FA, as revealed by Masson trichrome staining (Figure 1B) and stereological analysis of tubulointerstitial fibrosis (Figure 1C). Densely packed cells could be seen in the interstitial space, likely because of increased infiltration of immune cells or proliferation of fibroblasts, in fibrotic areas of kidneys from FA-treated mice (Figure 1D). Paradoxically, Sphk2^−/− mice, which are more susceptible to AKI than WT mice, had less kidney fibrosis after FA (Figure 1, B and C).

Figure 1.

FA treatment induces fibrosis in kidneys of WT and Sphk1^−/− mice but Sphk2^−/− mice are protected. (A) Plasma creatinine in WT, Sphk1^−/−, or Sphk2^-/ mice measured 3 and 14 days after vehicle (0.3 M sodium bicarbonate) or FA (250 mg/ml, intraperitoneal) treatment (see Supplemental Figure 3 for timeline). (B) Deposition of extracellular collagen indicated by Masson trichrome staining (blue) in kidney sections of WT and Sphk1^−/− is much greater than in Sphk2^−/− mice euthanized 14 days after FA treatment. Scale bar, 1 mm. (C) Extent of fibrosis determined by quantitative stereological analysis of Masson trichrome-stained sections (expressed as percentage of total surface area of kidney section occupied by interstitial fibrosis). (D) Higher magnification of cortex in Masson trichrome-stained sections from WT mouse kidney showing interstitial cell infiltration (arrows). Scale bar, 100 μm. n=3–4. *P<0.05; **P<0.01; ***P<0.001.

Kidneys of vehicle-treated mice exhibited low levels of fibronectin, vimentin, and α-smooth muscle actin (α-SMA) protein (immunofluorescence) and mRNA (RT-PCR), which increased significantly after FA treatment in WT and Sphk1^−/− but not Sphk2^−/− mice (Figure 2, A–C; corresponding Western blots in Supplemental Figure 4).

Figure 2.

FA induces increased expression of fibrotic markers in kidneys of WT and Sphk1^−/− but not Sphk2^−/− mice. Tissue samples from same mice as Figure 1 (euthanized on day 14). Expression of (A) α-SMA (Acta2), (B) fibronectin (Fn1), and (C) vimentin (Vim) measured by real-time quantitative RT-PCR (relative to glyceraldehyde 3-phosphate dehydrogenase; n=3–4) and immunofluorescence labeling (representative photomicrographs from the same groups of mice). Scale bar, 100 μm. *P<0.05.

FA Induces Leukocyte Infiltration in Kidneys of WT and Sphk1^−/− but Not Sphk2^−/− Mice

We investigated the profile of leukocyte infiltration in kidneys of FA-treated mice. By flow cytometry, the number of neutrophils (CD11b⁺ Ly6G^high), T cells (CD3⁺), and macrophages (CD11b⁺F4/80^low) increased significantly in kidneys of WT and Sphk1^−/− mice 14 days after FA treatment compared with kidneys of Sphk2^−/− mice (Figure 3A). Immunofluorescence labeling of kidney sections also revealed that FA-treated WT and Sphk1^−/− mice had greater infiltration of neutrophils (7/4), macrophages (F4/80), and T cells (CD3) in fibrotic areas (similar to the unlabeled cells observed in the interstitium in Figure 1D) compared with FA-treated Sphk2^−/− mice (Figure 3B).

Figure 3.

FA induces infiltration of immune cells in kidneys of WT and Sphk1^−/− but not Sphk2^−/− mice. (A) Macrophage (CD11b⁺F4/80^low), neutrophil (CD11b⁺GR-1^high [Ly6G]), and T cell (CD3⁺) number in kidneys determined by flow cytometry 14 days after treatment with vehicle or FA. Tissue samples from same mice as Figure 1 (euthanized on day 14). n=3–4. *P<0.05; **P<0.01; ***P<0.001 relative to corresponding vehicle control. (B) Immunofluorescent labeling of neutrophils (7/4), monocytes (F4/80), and T cells (CD3) in kidney sections from mice 14 days after vehicle or FA. Blue 4′,6-diamidino-2-phenylindole–labeled nuclei. Green autofluorescence reveals architecture of tubules. Scale bar, 100 μm.

Expression of Proinflammatory and Fibrotic Markers Associated with Fibrosis Is Significantly Higher in Kidneys of Sphk1^−/− Mice than in Sphk2^−/− Mice after FA Treatment

Kidneys of FA-treated WT or Sphk1^−/− mice had significantly higher mRNA levels of profibrotic and inflammatory markers. FA administration led to an increase in Tgfb (Figure 4), Cxcl2, Tlr4, Mcp1, Pcna, and Col3a1 in WT and Sphk1^−/− but not in Sphk2^−/− mice, but the levels of IL-10 in WT, Sphk1^−/−, and Sphk2^−/− mice were comparable (Table 1).

Figure 4.

FA induces increased expression of Tgfb in kidneys of WT and Sphk1^−/− but not Sphk2^−/− mice, and increased Ifng in Sphk2^−/− but not WT and Sphk1^−/− mice. RNA was extracted from kidneys harvested 14 days after administration of vehicle (sodium bicarbonate buffer) or FA (treatments as in Figure 1). mRNA levels were measured by quantitative RT-PCR and expressed relative to glyceraldehyde 3-phosphate dehydrogenase. n=3–4. *P<0.05; **P<0.01.

Table 1.

FA-induced increases in mRNA expression of proinflammatory and profibrotic markers on day 14 in WT and Sphk1^−/− but not Sphk2^−/− mice

Gene	Treatment	WT	Sphk1−/−	Sphk2−/−
CXCL1 (×10−3)	Vehicle	0.3±0.08	0.7±0.16	0.9±0.29
	FA	29.8±15.9a	11.9±4.4b	1.7±0.4c
CXCL2 (×10−5)	Vehicle	0 (n.d.)	0.05±0.01	0.13±0.03
	FA	4.16±0.01	4.53±2.27a	0.19±0.05
TLR-4 (×10−3)	Vehicle	7.4±2.3	5.5±0.7	6.9±0.6
	FA	24.8±5.5b	13.1±2.7c	4.5±0.4d
MCP-1 (×10−6)	Vehicle	0.8±0.002	1.5±0.3	2.9±0.4
	FA	4.7±1.2e	6.1±0.8d	2.0±0.6d
PCNA (×10−2)	Vehicle	5.7±0.1	4.9±0.3	3.6±0.2
	FA	19.6±3.5e	8.1±1.4d	3.1±0.3d
Collagen (Col3a1) (×10−1)	Vehicle	0.1±0.01	0.09±0.03	0.1±0.06
	FA	6.5±2.6a	4.6±0.6a	1.1±0.4
IL-10 (×10−4)	Vehicle	2.0±0.01	1.5±0.3	1.7±0.4
	FA	4.2±0.8	2.9±0.6	2.4±0.7

RNA extracted from kidneys harvested 14 days after administration of vehicle (sodium bicarbonate buffer) or FA (treatments as in Figures 1 and 2, Supplemental Figure 3). mRNA levels measured by quantitative RT-PCR and expressed relative to glyceraldehyde 3-phosphate dehydrogenase. n=3–4. n.d., not detectable.

^aP<0.05 compared with respective vehicle.

^bP<0.01 compared with respective vehicle.

^cP<0.05 compared with FA-treated WT mice.

^dP<0.001 compared with FA-treated WT mice.

^eP<0.001 compared with respective vehicle.

In summary, while there may be subtle differences in the changes in various markers and measures of fibrosis between fibrosis-susceptible WT and Sphk1^−/− mice, with some responses appearing less robust in Sphk1^−/− than in WT mice (e.g., fibronectin and vimentin expression, Figure 2, B and C; neutrophil and T cell infiltration, Figure 3A; Cxcl1, Tlr4, and Pcna expression, Table 1), we focused our attention on the protective phenotype of Sphk2^−/− mice.

Mice with Sphk2 Deletion in Either Bone Marrow or Nonbone Marrow Cells Are Resistant to Fibrosis

To examine the contribution of SphK2 in hematopoietic or parenchymal cells to kidney fibrosis, we generated bone marrow chimeric mice. Fibrosis increased after FA administration to WT mice reconstituted with WT bone marrow (WT → WT) but was blunted in Sphk2^−/− → WT or WT → Sphk2^−/− mice to about the same extent as in global Sphk2^−/− mice (Supplemental Figure 5), thus demonstrating that deletion of Sphk2 from either bone marrow or nonbone marrow compartments is sufficient to protect mice from fibrosis. A pattern of increased expression of Ifng was detected (but not significant) in kidneys after FA in Sphk2^−/−→WT mice that is consistent with prior observations in global Sphk2^−/− mice.

Reduced Fibrosis in Sphk2^−/− Mice Is Dependent on IFN-γ

In contrast to the above findings, mRNA expression of Ifng was markedly enhanced in kidneys of Sphk2^−/− but not WT or Sphk1^−/− mice after FA, although basal levels were not different between groups (Figure 4). T cells are one potential source of IFN-γ, and IFN-γ function is modulated by S1P³⁵ and Sphk2.³⁶ Confirming prior findings by others who used a different Sphk2^−/− mouse (Balb/c background),³⁶ FACS analysis revealed that splenic T cells from untreated Sphk2^−/− mice were hyperproliferative (Figure 5, A–C), more of these T cells produced IFN-γ (intracellular measurement by FACS; Figure 5, D and E), and they released more IFN-γ (measured by ELISA; Figure 5F) when stimulated in vitro compared with cells from WT or Sphk1^−/− mice.

Figure 5.

T cells from Sphk2^−/− mice proliferate faster and make more IFN-γ. (A–C) T cell proliferation measured in Cell Trace Violet dilution studies. T cells were isolated from spleens of untreated mice (control conditions), stimulated in culture with anti-CD3- and anti-CD28-conjugated beads, incubated with Cell Trace Violet for 10 minutes and allowed to proliferate for 72 hours. (A) Cell Trace Violet measured by flow cytometry. Numbers in each panel are percentage of the total population of Cell Trace Violet-labeled cells that has divided over the course of 3 days. Results are from a representative experiment; similar results were achieved in a replicate experiment by labeling cells with CFSE. (B) Overlay of proliferation histograms for splenic T cells from the four types of mice. (C) Proliferation measured in triplicate wells of cells (representative experiment from two mice per group) and expressed as percentage of the population of cells that divided at least once during the 3-day incubation period (mean±SEM). (D) FACS analysis of intracellular IFN-γ in proliferating T cells. (E) Percentage of proliferating, stimulated T cells that are IFN-γ–positive; from triplicate wells analyzed by FACS as shown in panel D. (F) IFN-γ concentration (measured by ELISA) in supernatant of triplicate wells of T cell cultures (representative experiment from three mice per group, repeated at least three times) 4 and 7 days after stimulation with vehicle or anti-CD3- and anti-CD28-conjugated beads. **P<0.01 compared with values on day 7 for WT and Sphk1^−/−. n.d., not determined in mice lacking IFN-γ.

Prior evidence that IFN-γ may be antifibrotic in the kidney³⁷ and other organs (reviewed by Wynn and Ramalingam³⁸) and our observation of the altered phenotype of Sphk2^−/− T cells, which could be directly IFN-γ–dependent and/or could lead to downstream protective mechanisms, prompted us to explore the role of IFN-γ in the resistance of Sphk2^−/− mice to FA-induced fibrosis. Kidney fibrosis was exacerbated in WT and Sphk1^−/− mice treated with IFN-γ blocking antibody (150 μg per mouse, beginning 1 day before FA and every 3 days thereafter) (Supplemental Figure 6) and there was >60% mortality in WT and Sphk1^−/− mice treated with IFN-γ blocking antibody in the FA model (data not shown).

Because of the high mortality associated with administration of IFN-γ blocking antibody in the FA model, we used unilateral IRI as an alternate model of fibrosis for assessing the role of IFN-γ. In the unilateral IRI model, the contralateral uninjured kidney maintains function, and plasma creatinine does not change. Impaired function can be revealed by removing the contralateral uninjured kidney on day 13, and assessing kidney function (plasma creatinine) 1 day later. Plasma creatinine, fibrosis (Masson trichrome), and expression of fibrotic markers increased markedly in WT, Sphk1^−/−, and Ifng^−/− mice 14 days after unilateral IRI and contralateral nephrectomy, but as in the FA model, Sphk2^−/− mice had less fibrosis and better renal function (Figure 6, A–D). The degree of protection in Sphk2^−/− mice in unilateral IRI-induced fibrosis was somewhat less than in the FA model, perhaps because the initial injury and therefore the extent of fibrosis is greater after IRI than with FA in Sphk2^−/− mice (Figure 1, Supplemental Figure 1). To determine the role of IFN-γ in mediating the protective effect of SphK2 deficiency, we generated Sphk2^−/−Ifng^−/− mice (Supplemental Figure 7). The protection observed in Sphk2^−/− mice was lost when Ifng was deleted (Sphk2^−/−Ifng^−/− mice; Figure 6, A–D) but the hyperproliferative phenotype of Sphk2^−/− T cells was not altered (Figure 5, B and C), and therefore their hyperproliferation is not dependent on IFN-γ.

Figure 6.

Protection from unilateral kidney IRI-induced fibrosis in Sphk2^−/− mice is dependent on expression of IFN-γ. (A–D) Mouse kidneys were exposed to sham surgery or 26 minutes unilateral ischemia followed by reperfusion. To reveal kidney function, the contralateral unclamped kidney was removed (and saved as the control) on day 13. (A) Plasma creatinine was measured 1 day after nephrectomy on day 14 after unilateral IRI. (B) Masson trichrome staining (blue) in kidney sections 13 (contralateral control kidney) or 14 days (IRI kidney) after IRI. Scale bar, 1 mm. (C) Extent of fibrosis determined by quantitative stereological analysis of Masson trichrome-stained sections. n=4–5. (D) mRNA levels of fibrotic markers, chemokines, and cytokines in IRI and contralateral control kidneys (Cntrl) measured by quantitative RT-PCR and expressed relative to glyceraldehyde 3-phosphate dehydrogenase. Two-way ANOVA with post hoc analysis revealed statistical significance between Sphk2^−/− and WT or Sphk1^−/− mice in level of expression of various markers that was reversed in Sphk2^−/− Ifng^−/− mice, but comparisons have been omitted from the figure for simplicity. For CXCL9 and CXCL10, ***P<0.001 relative to all other groups. (E) Mice were subjected to unilateral IRI and contralateral nephrectomy as in A–D, and were treated with IgG1 isotype control antibody or anti–IFN-γ blocking antibody (150 μg per mouse, administered intraperitoneally) on days 2, 5, 8, 11, and 13 after IRI (see timeline in Supplemental Figure 3). Plasma creatinine was measured on day 14. n=4–5. *P<0.05; **P<0.01; ***P<0.001.

We further characterized the Sphk2^−/− protective phenotype in the unilateral IRI model, including studies on localization of intracellular IFN-γ. Additional studies to reveal IFN-γ–positive cells by immunofluorescence and flow cytometry showed reduced infiltration of leukocytes in Sphk2^−/− mice after unilateral IRI, and no evidence of cells with increased IFN-γ in kidneys at day 14 (Supplemental Figure 8).

Although we were unable to identify a cellular source of IFN-γ that would be consistent with increased mRNA expression at 14 days, two additional findings further substantiate the importance of IFN-γ: (1) expression of the IFN-γ–responsive genes, CXCL9 and CXCL10, was considerably higher in Sphk2^−/− mice than WT, Sphk1^−/−, or Sphk2^−/−Ifng^−/− mice (Figure 6D); and (2) administration of IFN-γ blocking antibody (150 μg per mouse, 2 days after FA to avoid effects on acute ischemic injury and every 3 days thereafter) restored susceptibility to kidney fibrosis in Sphk2^−/− mice but these mice remained resistant to IRI-induced fibrosis after treatment with isotype control antibody (Figure 6E). Thus, an increase in levels of IFN-γ in Sphk2^−/− mice at some point in time during the period after FA contributed to inhibition of progression to fibrosis.

Adoptive Transfer of Sphk2^−/− T Cells Protects WT Mouse Kidneys from FA-Induced Fibrosis and Requires IFN-γ

We evaluated the immunomodulatory function of SphK2 in T cells in adoptive transfer studies. Adoptive transfer of Sphk2^−/− T cells (1.5×10⁶ CD4 cells) to WT mice 7 days before administration of FA attenuated kidney fibrosis (Masson trichrome) and blocked the increase in Acta2 and Col3a1 in kidneys 14 days after FA compared with adoptive transfer of WT or Sphk1^−/− T cells (Figure 7; T cell transfer produced similar changes in Acta2 and Col3a1 in kidneys 14 days after unilateral IRI [data not shown]). Adoptively transferred Sphk1^−/− T cells also had a small protective effect that was not significantly different from WT or Sphk2^−/− T cells. The protective effect of adoptively transferred Sphk2^−/− T cells to WT mice was lost if the transferred cells were also Ifng-deficient (T cells from Sphk2^−/−Ifng^−/− mice). These studies point to a dependence on IFN-γ in these cells; although they could be releasing more IFN-γ in vivo after FA, the presence of adoptively transferred, phenotypically altered Sphk2^−/− T cells with additional protective characteristics could also stimulate a local or distant antifibrotic or prorecovery mechanism.

Figure 7.

Adoptive transfer of splenic CD4⁺ T cells from Sphk2^−/− mice 7 days before FA protects kidneys of recipient WT mice from injury. CD4⁺ T cells were isolated from spleens of WT, Sphk1^−/− Sphk2^−/−, or Sphk2^−/−Ifng^−/− mice, and 1.5×10⁶ cells were intravenously injected via the tail vein into WT mice 7 days before treating with vehicle or FA (as in Figure 1). (A) Masson trichrome staining (blue) in kidney sections 14 days after vehicle or FA. Scale bar, 1 mm. (B) Extent of fibrosis determined by quantitative stereological analysis of Masson trichrome-stained sections. (C) Relative expression of fibrotic markers in kidneys measured by quantitative RT-PCR (relative to glyceraldehyde 3-phosphate dehydrogenase). n=3 for vehicle, 4–5 for FA groups. *P<0.05. NC, no cells.

Inhibition of SphK Activity in WT Mice Reproduces the Protective Phenotype of Sphk2^−/− Mice

To further examine the role of SphK2 in mice and to demonstrate therapeutic efficacy, we tested a SphK inhibitor (SLP120701) with 10-fold selectivity for SphK2 over SphK1.³⁹ Administration of SLP120701 in WT mice was delayed until 1 day after FA administration to avoid potential inhibition of AKI and enable analysis of effects on the chronic phase of injury by subsequent daily administration. SLP120701 (1, 3, 10, and 30 mg/kg, administered intraperitoneally) produced a dose-responsive decrease in fibrosis/injury (data not shown). Using a dose of 10 mg/kg, we found that SLP120701 increased blood levels of S1P 14 days after FA (data not shown), consistent with prior acute studies of the inhibitor alone and indicative, as a biochemical marker, of SphK2 engagement in vivo; blood levels of SLP120701 after a single dose of 10 mg/kg were 7–8 μM, which is well above the Ki of SLP120701 at recombinant mouse SphK2.³⁹ Daily administration of SLP120701 (10 mg/kg) yielded marked reduction of fibrotic markers (mRNAs for Acta2, Col3a1, Fn1, Vim, Ctgf, and Tgfb) and fibrosis compared with FA alone (Figure 8) (39.1%±7.6% versus 25.4%±1.4% of total surface area of kidney section occupied by fibrosis for FA versus FA+SLP120701, respectively, as assessed by trichrome stain; P<0.05). Consistent with increased IFN-γ in kidneys after FA treatment of Sphk2^−/− mice, increased expression of Cxcl9 and Cxcl10 (target genes for IFN-γ) was found in kidneys of WT mice treated with FA+SLP120701 (19.89±8.12 and 0.64±0.25, relative mRNA levels for Cxcl9 and Cxcl10) relative to FA alone (8.63±2.73 and 0.23±0.04; P<0.07 and P<0.08, respectively, compared with FA+SLP120701).

Figure 8.

SphK2 inhibitor reduces FA-induced fibrosis in kidneys. (A) Relative expression of fibrotic markers measured by quantitative RT-PCR (relative to glyceraldehyde 3-phosphate dehydrogenase) 14 days after vehicle (sodium bicarbonate) or FA in kidneys of WT mice that were treated once daily with the SphK2 inhibitor SLP120701 (10 mg/kg, administered intraperitoneally) beginning 1 day after FA. n=3 for vehicle, 4–5 for FA or FA+SLP120701. *P<0.05; **P<0.01; ***P<0.001 relative to vehicle or comparisons as shown. (B) Masson trichrome staining (blue) in kidney sections. Scale bar, 1 mm. Acta2, actin α2 smooth muscle aorta; Col3a1, collagen type III α1; Ctgf, connective tissue growth factor; Fn1, fibronectin; Tgfb, transforming growth factor β1; Vim, vimentin.

Discussion

Sphk2^−/− mice are more susceptible to IRI,²⁷ however despite early injury, we now show that Sphk2^−/− mice have less kidney fibrosis 14 days after AKI in either unilateral IRI- or FA-induced fibrotic models. The reduced susceptibility to fibrosis (either from lack of progression to fibrosis or enhanced recovery from injury) is due, at least in part, to IFN-γ, which although proinflammatory in AKI,⁴⁰ also has antifibrotic properties in chronic stages of disease.³⁸ Sphk2^−/− kidneys have higher levels of expression of Ifng mRNA after FA than WT or Sphk1^−/− kidneys; although a cellular source of IFN-γ was not identified, increases in IFN-γ–responsive genes were found in Sphk2^−/− kidneys, demonstrating enhanced IFN-γ functional activity during the 14-day period after injury. Adoptive transfer of Sphk2^−/− T cells, which are hyperproliferative and make more IFN-γ in vitro but may have other characteristics that afford a protective phenotype or response in vivo, confers protection from fibrosis to WT recipient mice. IFN-γ blocking antibodies or Ifng deletion (Sphk2^−/−Ifng^−/− mice) reversed the protection from unilateral IRI- or FA-induced fibrosis in Sphk2^−/− mice, and adoptively transferred Sphk2^−/−Ifng^−/− T cells were not protective. Although the contribution of endogenous Sphk2^−/− T cells to protection in Sphk2^−/− mice is unknown, our results strongly support a role for increased IFN-γ (perhaps from multiple cell sources over a discrete period of time after FA administration) in protecting Sphk2^−/− kidneys from fibrosis; other mechanisms, such as reduced transformation of fibroblasts to myofibroblasts and perhaps enhanced recovery processes, likely also contribute to protection (Figure 9). The clinical potential for targeting SphK2 in kidney disease is suggested by our finding that a moderately selective SphK2 inhibitor protected WT mice from FA-induced injury. Delineating the cellular source of IFN-γ and the time(s) at which its concentration increases will help future studies to further understand the mechanism(s) of protection or enhanced recovery in Sphk2^−/− mice.

Figure 9.

Mechanisms of antifibrotic effects of SphK2 deficiency. In WT mice or SphK1-deficient mice, FA or IRI leads to early epithelial injury followed by progressive fibrosis. Fibroblasts and pericytes transform into myofibroblasts leading to extracellular matrix deposition. Deletion of Sphk2 leads to enhanced early injury to an extent greater than WT or Sphk1^−/− mice following IRI (Supplemental Figure 1A), an effect that is dependent on the proinflammatory effects of IFN-γ (Supplemental Figure 1B). Despite the increase in initial injury in Sphk2^−/− mice, there is reduced fibrosis at 14 days in both unilateral IRI and FA-induced injury. T cells from SphK2-deficient mice are hyperproliferative and produce more IFN-γ in vitro. Although T cells may be one source of IFN-γ, other hematopoietic and nonhematopoietic cells likely to contribute during the period of time following initial injury to the kidney IFN-γ levels and IFN-γ response genes that lead to reduced fibrosis. Developing drugs that block SphK2 may lead to therapeutic agents that attenuate kidney fibrosis.

S1P in Kidney Injury and Fibrosis

S1P is an important target in kidney injury and disease.^34,41,42 S1P has pleiotropic effects in fibrosis.³¹ Prolonged S1P1 activation exacerbates fibrosis in the lungs,⁴³ whereas S1P3 deletion attenuates pulmonary fibrosis.⁴⁴ FTY720 blocks fibrosis in the kidney (UUO model)^29,30 and other organs.

Functional Role of SphKs and Consequences of Deletion

Considerable attention has been focused on the role of SphK1 in cancer but less is known of the specific roles of Sphk2.²³ SphK2 plays a proapoptotic role,²¹ and increased migration and proliferation of mouse fibroblasts and mesangial cells was found with SphK2 deletion.⁴⁵ Mice born without functional Sphk2 alleles are viable and fertile.^46,47 Mouse embryos with germline deletion of alleles for both isotypes are not viable,⁴⁷ but mice can survive Cre-mediated conditional deletion of both kinases after birth.⁴⁸

The antifibrotic effect of SphK2 deletion could be attributed to a change in S1P levels. Deletion of Sphk1 or administration of a SphK1 inhibitor to WT mice leads to reduced plasma S1P levels⁴⁹; unexpectedly, deletion of Sphk2 or administration of a SphK2 inhibitor to WT mice yields higher S1P levels (2–4-fold, relative to controls),⁵⁰ as also found in our current and prior studies.³⁹ SphK2 may have a role in clearance of S1P from blood, independent of its enzymatic role in generating S1P, that may be important for maintaining the steep gradient of S1P between blood (high) and tissues (low),⁵¹ whereas different sources of S1P may be important for lymphocyte trafficking.⁴⁸ S1P chaperones that bind circulating S1P could impart specificity to the various biologic actions of S1P, including regulation of immune function.⁵² In our experiments, we do not yet know if reduced fibrosis in Sphk2^−/− mice is due to altered concentrations of S1P, S1P receptor-mediated signaling, or intracellular effects of SphK2 and S1P. Our study was limited by the use a single SphK inhibitor, with moderate selectivity for SphK2 over Sphk1, although the increased blood levels of S1P, the biochemical marker of SphK target engagement in vivo, strongly support the predicted effect. Future studies will be needed when new SphK inhibitors are available, in order to define the relative contribution of the two enzymes.

SphKs in Injury and Fibrosis

The kidneys of Sphk2^−/− mice are more susceptible to IRI²⁷ but they do not progress to fibrosis (as shown in the current study). SphK1 expression increases in lungs of humans and mice with pulmonary fibrosis, SphK inhibition attenuates pulmonary fibrosis, and Sphk1^−/− but not Sphk2^−/− mice are protected from bleomycin-induced fibrosis.³² However, antifibrotic effects are not predicted by susceptibility to acute injury, similar to our findings on AKI and kidney fibrosis, as deletion of SphK1 exacerbated injury in the LPS-induced model of acute lung injury.⁵³ SphK1 upregulation/activation is antifibrotic in mesangial cells⁵⁴ and renal fibrosis⁵⁵ but may contribute to kidney fibrosis in diabetic nephropathy,⁵⁶ and questions remain on the role of intracellular versus inside-out extracellular signaling of S1P generated by SphK1 in fibrosis. Although we found subtle differences in various measures of fibrosis between WT and Sphk1^−/− mice, the overall finding was that their susceptibility to fibrosis was nearly equivalent and vastly different from the protective effect observed in Sphk2^−/− mice. The observation that both enzymes phosphorylate sphingosine yet produce remarkably different phenotypes suggests that the nuclear compartmentation of SphK2 could play a critical role in kidney fibrosis.

Role of IFN-γ in Fibrosis

We found that the protection from fibrosis in Sphk2^−/− mice is dependent on IFN-γ. Sphk2^−/− mice have increased expression of IFN-γ in the kidneys, and activated Sphk2^−/− T cells make and secrete more IFN-γ in vitro. Perhaps appearing to be at odds with increased Ifng mRNA expression in fibrotic kidneys, as immune cells are believed to be the main producers of IFN-γ, leukocyte infiltration in the kidneys of Sphk2^−/− mice, as expected for tissues that are protected from fibrosis, did not increase 14 days after FA. In addition, bone marrow chimera studies showed that hematopoietic Sphk2^−/− cells are not the only cells essential for protection. This suggests that nonhematopoietic cells could be a source of IFN-γ, as shown previously,⁵⁷ or that other phenotypic changes in hematopoietic and nonhematopoietic cells in response to Sphk2 deletion contribute to protection. Our intracellular cytokine studies did not identify a source of increased IFN-γ. It is possible that dynamic expression of IFN-γ during the course of injury and repair leads to protein levels that are higher at earlier time points, which would be consistent with our IFN-γ blocking antibody studies showing that increased IFN-γ at some point during the 14-day period is necessary for protection from kidney fibrosis. Another possibility for the observed discrepancy between mRNA and protein is that IFN-γ can induce its own mRNA expression.⁵⁷ Nevertheless, increased expression of the Th1 target genes, Cxcl9 and Cxcl10, is consistent with increased expression of IFN-γ, albeit from an as yet unknown source and time point, and there could be elevated levels of other IFN-γ–responsive genes that are important for the observed antifibrotic effects.

IFN-γ is typically viewed as a proinflammatory cytokine, particularly in acute injury or inflammation; Ifng^−/− mice are less susceptible to kidney IRI⁴⁰ and the increase in acute injury in Sphk2^−/− mice (24 hours after IRI), likely because of increased Th1 response, is reduced in the absence of Ifng in Sphk2^−/− Ifng^−/− mice (shown in the current study). However, IFN-γ also has antifibrotic properties³⁸ mediated by a number of mechanisms, and that has made it an attractive though as yet unsuccessful therapeutic approach for treatment of a variety of disorders. The conclusion that IFN-γ is necessary for the protective effect of Sphk2 deletion does not rule out the possibility that other factors or pathways contribute to the protective mechanism in Sphk2^−/− mice.

S1P and T Cells

We focused our initial studies on T cells because they play a significant role in fibrosis, various S1P-mediated T cell properties are important in inflammation, and because IFN-γ is produced primarily by natural killer cells, CD4, CD8, and natural killer T cells. S1P inhibits stimulated T cell proliferation and increases activated T cell-mediated IL-2 or IFN-γ secretion.³⁵ T cells from Sphk2^−/− mice are hyperactive and hyperproliferative, and in contrast to our findings in an acute setting involving innate immunity, adoptive transfer of Sphk2^−/− T cells exacerbated irritable bowel disease (an autoimmune disease), perhaps via increased IL-2 in Sphk2^−/− T cells.³⁶ Sphk2^−/− T cells have increased secretion of cytokines (IL-2, IL-4, IL-10, and IFN-γ) in response to stimulation, are less sensitive to T regulatory cell-mediated suppression, and show increased IL-2 responsiveness. We confirmed some of these findings in our study.

Does Less Fibrosis Have Anything To Do with the Hyperproliferative Properties of Sphk2^−/− Cells?

Increased proliferation of Sphk2^−/− fibroblasts⁴⁵ appears at odds with reduced susceptibility to fibrosis, however their proliferative behavior, migration, and phenotype in vivo is unknown, and proliferation may be self-limited by increased IFN-γ production.⁵⁸ Proliferative properties of Sphk2^−/− tubule epithelial cells have not been examined. Whereas repair and recovery after acute injury depends on the proliferative capacity of tubule epithelial cells,³ hyperproliferative tubule cells are more susceptible to acute injury, and AKI can be attenuated by cell-cycle inhibitors⁵⁹; this is consistent with our data that Sphk2^−/− mice are more susceptible to AKI but are resistant to fibrosis, and may have enhanced repair and/or recovery. Future studies will examine the effect of SphK2 deletion on cell cycle pathways.

Our finding that IFN-γ is necessary for the protective effect in Sphk2^−/− mice but not the hyperproliferative phenotype in T cells, clearly separates the role of SphK2 in these two biologic phenomena and indicates that Sphk2 deletion in T cells alters another pathway that regulates proliferation besides IFN-γ.

Protective Effect of Adoptively Transferred Sphk2^−/− T Cells

In addition to protection in global Sphk2^−/− mice, adoptive transfer of Sphk2^−/− T cells was sufficient to protect recipient WT mice from kidney fibrosis, and the effect was long-lasting (at least 7 days). Furthermore, it was necessary that these donor Sphk2^−/− T cells retain the ability to make IFN-γ, as Sphk2^−/−Ifng^−/− T cells were not protective. Although increased release of IFN-γ may or may not be necessary and a cellular mechanism has not yet been identified, Sphk2^−/− T cells appear to have a regulatory phenotype that confers protection from fibrosis when they are adoptively transferred; the contribution of endogenous Sphk2^−/− T cells to the protection from fibrosis in Sphk2^−/− mice has not yet been examined, but our bone marrow chimera studies suggest that both hematopoietic and nonhematopoietic SphK2-deficient cells contribute to protection. Translation of this potential therapy to humans would require testing of selective pharmacologic inhibition of SphK2 to produce a phenotype ex vivo in donor T cells that yields antifibrotic properties upon transfer to a recipient, or that can be used in vivo to expand an antifibrotic T cell phenotype.

Therapeutic Potential of SphK2 as a Target in Kidney Disease

The clinical relevance of our studies in Sphk2^−/− mice is bolstered by our findings that a selective SphK2 inhibitor protected WT mice from FA-induced injury. New SphK inhibitors in development^33,39 could modulate the sphingolipid rheostat and have broad implications for tissue fibrosis and other diseases. Production of S1P in the nucleus by SphK2 has broad epigenetic effects through inhibition of HDAC,²⁴ which plays a role in T cell responses,⁶⁰ and therefore may be important in the protective phenotype of Sphk2^−/− T cells. HDAC is implicated in the progression of kidney fibrosis,⁶¹ and its inhibition by S1P induces p21 transcription,²⁴ which could reduce cell proliferation and attenuate fibrosis. HDAC inhibitors reduce renal fibrosis^62,63 and may promote recovery from early insults by reducing the epithelial G2/M cell-cycle arrest⁶⁴ that occurs in fibrosis.³ Inhibition of this nuclear pathway by SphK2 inhibitors could be advantageous for treatment of fibrosis, as transcriptional regulation through this mechanism likely yields a complex array of altered expression of many genes, and the paradoxic increase in S1P levels in Sphk2^−/− mice and after the administration of SphK2 inhibitors could drive the pathway.

In summary, we have shown that SphK2 plays an important role in kidney fibrosis through mechanisms involving IFN-γ and perhaps other pathways, and that deletion or pharmacologic inhibition of SphK2 protects kidneys from fibrosis.

Concise Methods

Animals and Models of Fibrosis

Two models of fibrosis were used (see Supplemental Figure 3 for timeline of treatments). We used C57BL/6 (WT) mice (approximately 20–24 g, 8–12 week of age, National Cancer Institute [NCI], Frederick, MD) and WT littermate controls; Sphk1^−/− and Sphk2^−/− mice (congenic on C57BL/6) were provided by Dr. Richard L. Proia (National Institutes of Health/National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, MD). In pilot studies, the response to either FA or IRI was the same in WT littermate controls and purchased or house bred NCI WT mice, therefore WT NCI C57BL/6 mice were used as controls throughout. Previous studies employed the Sphk2^tr/tr mice,²⁷ which have the same phenotype as Sphk2^−/− mice. Double knockouts were generated by breeding Ifng^−/− mice (The Jackson Laboratory, Bar Harbor, ME) and Sphk2^−/− mice to produce Sphk2^−/−Ifng^−/− mice (genotyping in Supplemental Figure 7). In each experiment, animals cohoused in groups of five per cage (on average) were randomized to the treatment design.

Fibrosis was induced in WT, Sphk1^−/−, Sphk2^−/−, Ifng^−/−, and Sphk2^−/−Ifng^−/− mice (12 weeks of age) by a single injection of FA (250 mg/kg, administered intraperitoneally) in 0.3 M sodium bicarbonate (vehicle); the optimal dose of FA was determined by preliminary studies and was similar to previous reports.^65,66

In some experiments, mice (8–12 weeks of age) were exposed to unilateral IRI, and fibrosis in the ischemic kidney was assessed at day 14. Mice were anesthetized with an intraperitoneal mixture of ketamine (120 mg/kg), xylazine (12 mg/kg), and atropine (0.324 mg/kg), and were subjected to unilateral flank incisions. The right kidney pedicle was exposed and crossclamped for 26 minutes, then clamps were released (reperfusion) as described before.⁶⁷ The kidney pedicle was exposed but not clamped in sham-operated mice. During the surgery, mouse core temperature was maintained at 34°C–36°C with a TR-200 Heating Pad System (Fine Science Tools); during the recovery and reperfusion period (18–24 hours), mice were housed in a warming incubator with ambient temperature at 30°C–32°C. To assess kidney function in the chronic model after unilateral IRI, when renal reserve ordinarily would mask changes in the injured kidney, the contralateral kidney was removed in some experiments (nephrectomy) 13 days after IRI and saved as the control kidney. Mice were euthanized 24 hours later, and plasma creatinine was determined.

Surgeries or drug treatments were always performed in a dedicated surgical/animal procedure room between the hours of 10:00 a.m. and 2:00 p.m. so that blood collection and euthanasia could also occur during this same time of day on the desired number of days after treatment. Between procedures animals were housed in the vivarium in intervening periods of time >24 hours with free access to food and water and with a standard light/dark cycle.

Blood was collected under isoflurane anesthesia by retro-orbital bleeding at 3 and 14 days after FA or vehicle treatment, or on day 14 after unilateral IRI (1 day after contralateral nephrectomy). On day 14, all mice were euthanized (by cervical dislocation under anesthesia immediately after collecting blood) and kidneys were harvested.

The SphK2 inhibitor, SLP120701,³⁹ was prepared in a vehicle of 2% hydroxypropyl-β-cyclodextrin in PBS and administered intraperitoneally once daily (1, 3, 10, or 30 mg/kg) for 12 days, beginning 1 day after FA.

Rat anti-mouse IFN-γ (XMG1.2; BioXcell, West Lebanon, NH) or rat IgG1 isotype control antibody (BioXcell) was administered (150 μg per mouse) 1 day before FA and every 3 days thereafter in the FA model, or 2 days after IRI (to avoid possible effects on the acute phase of injury) and on days 5, 8, 11, and 13 (before nephrectomy) in the unilateral IRI model (see Supplemental Figure 3).

In experiments to detect intracellular IFN-γ by immunofluorescence or flow cytometry, kidneys were harvested 5 hours after treatment with the Golgi inhibitor brefeldin A (250 μg per mouse, administered intraperitoneally; Sigma-Aldrich, St. Louis, MO), which enables accumulation, and therefore detection, of intracellular cytokines.

Generation of Bone Marrow Chimeric Mice

Sphk2^−/− mice, littermate progeny controls, and B6 Cd45.1 (B6.SJL-Ptprc^a Pepc^b/BoyJ; The Jackson Laboratory) were used to generate bone marrow chimeras as previously described.⁶⁸ Briefly, recipient mice were lethally irradiated and were immediately reconstituted with 10⁶ donor bone marrow cells (injected intravenously). The resulting chimeric mice were maintained for 6–8 weeks before experimentation. Cd45.1 mice were used as donors to generate Cd45.1 WT → Sphk2^−/− chimeras; detection of Cd45.1⁺ donor cells by flow cytometry confirmed that reconstitution efficiency in chimeric mice was >90%.

Adoptive Transfer Studies

Total CD4⁺ cells were isolated from single cell suspensions of WT, Sphk1^−/−, or Sphk2^−/− or Sphk2^−/−Ifng^−/− mouse spleens using the Dynal mouse CD4 negative isolation kit (Invitrogen, Life Technologies, Grand Island, NY). WT mice (12 weeks old) were intravenously injected via the tail vein with 1.5×10⁶ freshly isolated CD4⁺ cells, or an equal volume (200 μl) of PBS as a control, 7 days before FA administration.

Assessment of Kidney Function and Histology

Plasma creatinine was measured using a colorimetric assay according to the manufacturer’s protocol (Sigma-Aldrich). Kidneys were fixed in 0.2% sodium periodate/1.4% DL-lysine/4% paraformaldehyde (PLP) in 0.1 M phosphate buffer, pH 7.4 (4% PLP), and embedded in paraffin. Paraffin-embedded kidney sections (4 μm) were stained with Masson trichrome and viewed by light microscopy (Zeiss AxioSkop; Carl Zeiss GmbH, Jena, Germany), and images were taken using a SPOT-RT Camera (software version 3.3; Diagnostic Instruments, Sterling Heights, MI).

To assess collagen deposition in tissue sections, kidneys were fixed in 1% PLP and embedded and frozen in Optimal Cutting Temperature compound (Ted Pella, Redding, CA). Frozen sections (5 µm) were washed with PBS, fixed with 4% paraformaldehyde/0.1 M phosphate buffer, pH 7.0, for 10 minutes, washed twice with PBS, stained with picrosirius red solution (Polysciences) for 1 hour at room temperature, washed in 0.5% acetic acid (in water), and dehydrated through a graded series of alcohols. Coverslips were applied, and photomicrographs of picrosirius red birefringence were captured using a Zeiss AxioImager microscope with polarizing filter (Carl Zeiss GmbH) and Stereo Investigator software (version 9; MBF Bioscience, Williston, VT). The percentage of the kidney section occupied by red biorefringent pixels was calculated using Adobe Photoshop.

The extent of kidney tubulointerstitial fibrosis revealed by trichrome staining was assessed in an unbiased, systematic manner using design-based stereology to achieve statistically accurate random sampling of kidney sections. The investigator was blinded to the experimental identity of the sections. Sections were imaged by using a Zeiss Axio Imager Z2/Apotome Microscope fitted with motorized focus drives and motorized XYZ microscope stage (MBF Bioscience) and integrated to a workstation running Stereo Investigator software (version 11.06.2, MBF Bioscience). The area fraction fractionator probe (Stereo Investigator) was used for stereological analysis of the fractional area (percentage of total surface area) of the section occupied by tubulointerstitial collagen deposition (Masson trichrome stain). The following parameters were defined: counting frame, 400×400 μm; sample grid, 800×800 μm; and grid spacing, 85 μm. These values were determined empirically such that adequate numbers of sample sites were visited and adequate numbers of markers (indicating extracellular deposition of collagen) were acquired, in keeping with accepted counting rules for stereology. A total of 625±20 (mean±SEM) grid sites were evaluated per section; the sampling fraction was 25% of a total average area of 2.85±0.38 ×10⁶ μm² for each kidney section.

Using the same Zeiss AxioImager Z1 and MBF Bioscience system, virtual tissue section images were acquired using Stereo Investigator software. Parameters were defined in this application of the software so that images of kidney sections acquired with a 5× or 10× objective were automatically stitched together to produce a full view of the kidney section (e.g., as shown in Figures 1 and 6–9, Supplemental Figure 6) that can be magnified from the image with preservation of the original digital resolution.

Immunofluorescence Labeling of Kidney Sections

Immunofluorescence labeling for markers of fibrosis (fibronectin, vimentin, and α-SMA) and inflammatory cells (T cells [CD3], neutrophils [7/4], and macrophages/monocytes [F4/80]) was performed on frozen sections. Kidneys were fixed in 1% PLP and embedded and frozen in Optimal Cutting Temperature compound (Ted Pella). Frozen sections (5 µm) were permeabilized with 0.3% Triton X-100, and nonspecific binding was blocked with 10% horse serum and anti-mouse CD16/32 (10 μg/ml; clone 2.4G2; StemCell Technologies, Vancouver, BC, Canada). Sections were labeled with (1) rabbit anti-fibronectin (20 μg/ml; Abcam, Inc., Cambridge, MA), goat anti-vimentin (5 μg/ml; Santa Cruz Biotechnology, Santa Cruz, CA), or rabbit anti-Ki-67 (1:200, VP-RM04; Vector Laboratories, Burlingame, CA) followed by Cy3-tagged donkey anti-rabbit antibody for detection of fibronectin and Ki67, or Cy3-tagged donkey anti-goat secondary antibody for detection of vimentin; or with (2) Cy3-conjugated mouse anti–α-SMA (5 μg/ml, clone 1A4; Sigma-Aldrich; unlabeled antibody was conjugated with Cy3 using Apex Antibody Labeling Kit; Invitrogen, Carlsbad, CA); or with (3) A647-conjugated rat anti-PDGFRβ (1:50 from 0.2 mg/ml stock, clone APB5; Biolegend, San Diego, CA). For detecting leukocytes, neutrophils, monocytes, and T cells, sections were labeled with FITC-conjugated mouse anti-CD45 (7 μg/ml, clone 30-F11; eBioscience, San Diego, CA), FITC-conjugated rabbit anti-7/4 (5 μg/ml; Invitrogen), mouse anti–F4/80-PE (5 μg/ml), or mouse anti–CD3-FITC (10 μg/ml), respectively, for 1 hour at room temperature. Intracellular IFN-γ was detected in kidney sections from brefeldin A-treated mice (as described above) by labeling with APC-conjugated anti–IFN-γ (1:50 from 0.2 mg/ml stock, XMG1.2; eBioscience). After washing, specimens were covered with ProLong Gold Antifade reagent with 4′,6-diamidino-2-phenylindole (Molecular Probes, Life Technologies), to label nuclei, and coverslips were applied. Images were acquired using a Zeiss Axiovert 200 microscopy system with ApoTome imaging (Carl Zeiss GmbH) and processed using Zeiss AxioVision 4.8.2 software.

Western Blot Analysis

Kidney samples were homogenized in RIPA lysis buffer (Thermo Fisher Scientific, Vernon Hills, IL) enriched with 1% protease and phosphatase inhibitor cocktail (formulation: sodium fluoride, sodium orthovanadate, β-glycerophosphate, sodium pyrophosphate, aprotinin, bestatin, E64, leupeptin, EDTA; Thermo Fisher Scientific) by shaking with metal beads (50 s⁻¹ for 10 minutes) using a TissueLyser (Qiagen, Germantown, MD) followed by sonication to improve protein extraction. Homogenates were centrifuged (12,000 rpm for 15 minutes at 4°C), and the protein concentration in the supernatant was measured by using the BCA protein assay kit (Pierce, Rockford, IL) or Bradford protein assay (Bio-Rad, Hercules, CA). Equal volumes of the remaining cell or tissue lysate supernatants were boiled with Laemmli buffer and β-mercaptoethanol (10 minutes at 100°C); proteins were separated using a 10% (fibronectin) or 7.5% (α-SMA or vimentin) SDS-PAGE gel and transferred to PVDF membranes (EMD Millipore, Billerica, MA). PVDF membranes were incubated with primary antibody mouse anti–α-SMA (1:500; Sigma-Aldrich), rabbit anti-fibronectin (1:250; Santa Cruz Biotechnology), goat anti-vimentin (1:500, sc7557; Santa Cruz Biotechnology), or mouse anti-tubulin (1:5000, sc53140; Santa Cruz Biotechnology). Blots were then washed and incubated at 1:1000 for 1 hour with horseradish peroxidase-conjugated secondary antibodies (Santa Cruz Biotechnology). Bands were visualized by chemiluminescence according to the manufacturer’s protocol with SuperSignal West Pico chemiluminescent substrate (Thermo Fisher Scientific) and densitometric analysis was performed.

T Cell Proliferation Assay and T Cell Response

Total CD4⁺ cells were isolated from single cell suspensions of WT, Sphk1^−/−, Sphk2^−/−, or Ifng^−/−Sphk2^−/− mouse spleens using the Dynal mouse CD4 negative isolation kit (Invitrogen, Life Technologies) and were counted using a Vision Cellometer (Nexelcom Bioscience LLC). Freshly isolated cells were cultured in RPMI 1640 (Gibco, Life Technologies) culture medium (10% FBS [Invitrogen], 1% glutamine, 0.1% β-mercaptoethanol, 1% P/S) in 96-well round bottom plates (100,000 cells per well), stimulated with anti-CD3- and anti-CD28-conjugated beads (Invitrogen; 1:1 beads to cells) for 72 hours then incubated with Cell Trace CFSE or Cell Trace Violet (Invitrogen; 5 μM for 10 minutes at 37°C), cultured and allowed to proliferate for 72 hours. Flow cytometry was used to measure the rate of proliferation (by analyzing dye-positive cell populations) and IFN-γ–positive cells (see section FACS Analysis). Supernatants were collected for analysis of IFN-γ by ELISA (eBioscience).

Real-Time RT-PCR

Kidneys were homogenized in TRI Reagent, and total RNA was isolated according to the manufacturer’s protocol (Molecular Research Center, Inc., Cincinnati, OH). cDNA was generated by reverse transcription using the iScript cDNA synthesis kit (Bio-Rad). The mRNA expression of various inflammatory markers (TGFβ, TNFα, CXCL1, CXCL2, TLR-4, MCP-1, proliferating cell nuclear antigen [PCNA], IL-10, IL-4, and IFN-γ), markers of fibrosis (fibronectin, vimentin, Col3a1, CTGF, and α-SMA), S1P receptors (S1P1–S1P5), and SphK1 and SphK2 was measured by performing quantitative RT-PCR using the iScript one-step RT-PCR kit with SYBR Green (Bio-Rad), according to the manufacturer's protocol and as described previously.³⁴ Primer sequences, designed using PrimerQuest (Integrated DNA Technologies; http://www.idtdna.com), are shown in Supplemental Table 1. Sample values were calculated with normalization to glyceraldehyde 3-phosphate dehydrogenase.

FACS Analysis

Flow cytometry was used to analyze leukocyte infiltration of kidneys 14 days after FA treatment. In brief, kidneys were extracted, minced, digested, and then passed through a filter and a cotton column, as described previously.⁶⁹ Nonspecific Fc binding was blocked with anti-mouse CD16/32 (2.4G2), and fresh kidney suspensions were then incubated with anti-mouse CD45-FITC (30-F11) to determine total leukocyte cell numbers. CD45-labeled samples were then labeled with different combinations of fluorophore-conjugated anti-mouse F4/80 (BM8), GR-1 (Ly6G), CD11b, CD3, CD4, and CD8 antibodies. 7-AAD (BD Biosciences, San Jose, CA) was added 15 minutes before analyzing the sample, to separate live cells from dead cells. The BD Biosciences Fix/Perm buffer set was used according to the manufacturer’s protocol for intracellular staining of IFN-γ (anti-mouse IFN-γ, 1 μg/ml; eBioscience,) and as described previously,⁴⁰ in cells isolated from kidneys of mice treated with brefeldin A, as described above. Flow cytometry data acquisition was performed on FACS Calibur (Becton Dickinson, San Jose, CA) with Cytek 8-color flow cytometry upgrade (Cytek Development, Inc., Fremont, CA). Data were analyzed by FlowJo software 9.0 (TreeStar, Inc., Ashland, OR) using a gating strategy as described previously.^70,71 All antibodies (except as noted) were purchased from eBioscience and were used at a concentration of 5 μg/ml.

Statistical Analyses

Comparisons between treatment groups were examined by one-way or two-way ANOVA by using Sigma Plot 11.0 (Systat, Chicago, IL) with post hoc multiple comparisons using Tukey test. Data are expressed as mean±SEM. Statistical significance was identified at P<0.05.

Study Approval

All experiments were performed in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals, and all procedures were approved by the University of Virginia Animal Care and Use Committee.

Disclosures

K.R.L. is one of the founders of SphynKx Therapeutics LLC (Charlottesville, VA), which was created to commercialize S1P-related discoveries, including sphingosine kinase inhibitors. The inhibitor featured in this paper, SLP120701, is included in a patent application that is licensed currently to SphynKx.