Targeting mTOR Signaling Can Prevent the Progression of FSGS

FSGS Is Associated with mTORC1 Activation in Humans and Mice

To study the potential association of mTORC1 activation and podocyte injury in human FSGS, we analyzed glomerular gene expression of microdissected glomeruli from patients with primary FSGS or minimal change disease (MCD) in comparison to healthy controls (living donors). Increased expression of major mTORC1 target genes responsible for mitochondrial biogenesis, fatty acid synthesis, and angiogenesis as well as mTOR mRNA itself indicated the activation of mTORC1 in primary FSGS, whereas transcription levels were not influenced in MCD (Figure 1A). Activation of mTORC1 could be confirmed on protein level, as phosphorylation of glomerular ribosomal protein S6 (pS6), a marker of mTORC1- dependent S6K activation, was significantly enhanced in glomeruli and especially podocytes of patients with FSGS compared with glomeruli of patients with MCD (Figure 1B). To further dissect the role of mTORC1 in the development of FSGS lesions, we investigated a commonly used FSGS mouse model, where the primary podocyte injury is induced by the injection of Adriamycin (ADR) in susceptible mouse strains. ADR injection resulted in significant phosphorylation of S6, highlighting mTORC1 activation as a common feature of human- and animal-associated FSGS-like lesions (Figure 1, C and D). Increasing proteinuria (Figure 1E) and glomerulosclerosis (Figure 1F) correlated well with the progression of podocyte loss (Figure 1, G–I).

Interestingly, however, plasma from patients with recurrent primary (idiopathic) FSGS after transplantation was not associated with activation of mTORC1 (Supplemental Figure 1). Cultured human podocytes were treated for 24 hours with media containing 5% serum from patients with recurrent FSGS. Compared with treatment with serum from healthy controls, phosphorylation of S6 was not significantly altered in recurrent FSGS serum–treated podocytes. This suggests that mTORC1 is not the primary target of FSGS permeability factors, but rather might represent a chronic, potentially maladaptive response after a primary podocyte injury.

Constitutive mTORC1 Activation Causes FSGS-Like Lesions

Chronic mTORC1 activation in the podocyte was previously described to lead to podocyte hypertrophy and progressive glomerular disease.¹² Because we identified mTORC1 activation as a hallmark of FSGS, we used this model to study the time course of FSGS development in response to mTOR hyperactivation. Therefore, we generated a podocyte-specific knockout of Tsc1 (Supplemental Figure 2A). Loss of Tsc1 results in constitutive hyperactivation of mTORC1 (Supplemental Figure 2D). In fact, Tsc1^Δpodocyte mice already developed FSGS-like histologic lesions and nephrotic-range proteinuria at 3 weeks of age, confirming mTORC1 activation as a pathogenic factor driving the development of FSGS (Supplemental Figure 2, B, C, and E). At the cellular level, podocyte numbers were significantly reduced but the glomerular tuft volume was increased compared with controls, suggesting a hypertrophy of the remaining podocytes (Supplemental Figure 2, F–H).

Complete Genetic Abrogation of mTORC1 Activity Facilitates FSGS-Like Disease

Next, we tested the phenotypic consequences of suppressing ADR-induced mTOR activation by the genetic deletion of mTORC1 activity. Previously, we showed that deleting Raptor can efficiently diminish the function of mTORC1 in podocytes. The constitutive podocyte-specific deletion of Raptor during embryonic development (Raptor flox/flox; Nphs2.Cre mice) resulted in progressive proteinuria associated with growth retardation and increased mortality.¹³ Raptor deletion induced in adult mice (doxycycline-treated Raptor flox/flox; Nphs2.rtTA; tetO.Cre mice), however, caused neither proteinuria nor alteration of glomerular histology.¹³ Therefore, we combined the ADR-induced FSGS model with a timed inducible deletion of Raptor (Figure 2, A and B). Unexpectedly, mTORC1 abrogation did not ameliorate the course of FSGS-like lesions, but resulted in a dramatically increased vulnerability toward ADR-induced FSGS-like lesions. Compared with control mice, mice with induced homozygous Raptor deletion developed massively increased proteinuria (Figure 2C), showed much more severe signs of glomerulosclerosis (Figure 2, D and E), and exhibited a significantly increased loss of podocytes (Figure 2, F and G). The total glomerular tuft volume was not different between Raptor-deficient and controls (Figure 2H).

To corroborate these surprising findings, we performed a similar series of experiments by combining a subtotal nephrectomy (Nx) model—known to cause FSGS—with induced deletion of Raptor in adult animals (doxycycline-treated Raptor flox/flox; Nphs2.rtTA; tetO.Cre mice). Again, compared with Nx wild-type mice, induced Raptor-deficient Nx mice developed more severe proteinuria, renal failure, and uremia (Figure 2, I and J). Histologic examination and transmission electron microscopy documented increased glomerular lesions, podocyte foot process effacement, and tubular casts of Nx Raptor-deficient versus control mice (Figure 2K, Supplemental Figure 3). In agreement with the ADR model with an induced homozygous Raptor deletion, a significant loss of podocytes was detected after Nx whereas no glomerular hypertrophy was seen, underlining the need of mTOR activation for podocyte and subsequent glomerular hypertrophy (Figure 2, L–N).

Partial Genetic Reduction of mTORC1 Activity Ameliorates FSGS-Like Disease

The fact that constitutive mTORC1 activation by itself can cause FSGS-like lesions, but complete loss of mTORC1 activity aggravates underlying FSGS models, suggested a nonlinear role of mTOR in FSGS-related glomerular disease. Although some mTOR activation is likely required for a compensatory adaptive hypertrophy of remaining podocytes, persistent mTOR hyperactivation presumably propagates podocyte dedifferentiation and progressive podocyte loss. Conceptually, this model suggested a modest mTOR inhibition as a novel therapeutic strategy. To genetically test this hypothesis, we generated mice lacking just one Raptor allele in podocytes (Raptor^{Het podocyte}, Figure 3A). Reduced RAPTOR expression was documented by western blot analysis (Figure 3B) and was associated with significantly reduced mTOR activation in response to ADR-induced FSGS (Figure 3C). Strikingly, progression of proteinuria in ADR nephropathy was substantially ameliorated in Raptor^{Het podocyte} mice (Figure 3D). Furthermore, quantification of glomerular lesions demonstrated unambiguously less glomerulosclerosis in Raptor^{Het podocyte} compared with control mice (Figure 3, E and G). In-depth quantitative stereologic analyses using the dissector principle 3 weeks after ADR injection allowed us to directly determine podocyte cell volumes and revealed that podocyte-specific lowering of mTORC1 activity in ADR nephropathy completely prevented podocyte hypertrophy, whereas podocyte number and glomerular tuft volume were not different (Figure 3, F and H). In summary, these data demonstrate that modest genetic mTOR inhibition can ameliorate FSGS progression.

Metabolic Podocyte Reprogramming by mTOR Signaling

To gain more insights into the molecular consequences of mTOR signaling in podocytes, we analyzed the mTOR-dependent metabolic profile in immortalized human podocyte cells. Key features of cellular metabolism are reflected by mitochondrial respiration and ATP generation. Interestingly, mTOR signaling exhibited a dose-dependent effect on mitochondrial function. High-dose treatment with Torin 1 massively reduced mitochondrial respiration and ATP synthesis due to lower levels of the oxygen consumption rate, whereas low-dose treatment did not affect the oxidative phosphorylation (Figure 4, A and B) or the total ATP content (Figure 4C). In agreement with these results, high-dose treatment with Torin 1 lowered the expression levels of SDHB, a key component of the respiratory chain (Figure 4D, Supplemental Figure 4A), thereby directly linking mTOR downstream signaling events to oxidative ATP production. In addition, functional mitochondrial impairment in response to high-dose treatment with Torin 1 led to high levels of oxidative stress (Figure 4E). ADR itself had no effect on the mTOR signaling pathway in immortalized human podocytes (Figure 4F), but mitochondrial respiration and oxidative phosphorylation were impaired (Figure 4, G and H). Low-dose treatment with Torin 1, however, ameliorated ADR-induced mitochondrial dysfunction (Figure 4, I and J) and maintained the capacity for mitochondrial respiration and oxidative ATP synthesis. In contrast, high-dose Torin 1 treatment aggravates ADR-induced mitochondrial dysfunction (Figure 4, K and L).

These data are in line with the clinical observations in mice and unravel that complete inhibition of mTOR causes dramatic metabolic changes in podocytes, being associated with mitochondrial dysfunction and increased ROS production. Strikingly, however, incomplete mTOR inhibition preserves mitochondrial function even in the presence of an experimental FSGS model.

Rapamycin Treatment Prevents FSGS-Like Lesions in ADR Nephropathy

To test whether the genetic and metabolic data can be transferred to a new therapeutic rationale, we employed a rapamycin treatment protocol on mice with ADR-induced FSGS. Because a partial reduction of mTORC1 activity should be sufficient, we chose a low-dose protocol (0.5 µg/g mouse body wt three times weekly intraperitoneally) to avoid adverse effects of high-dose rapamycin. Low-dose rapamycin (first dose of rapamycin together with ADR) reduced mTORC1-dependent phosphorylation of S6 (Figure 5A) and significantly ameliorated proteinuria (Figure 5B). In agreement, podocyte numbers were preserved in rapamycin-treated animals (Figure 5, D and E), whereas glomerular volume did not differ significantly between the groups (Figure 5F). Histologic analysis evidenced that low-dose rapamycin prevents progressive glomerulosclerosis (Figure 5, C and G). Strikingly, the beneficial effect of pharmacologic mTORC1 inhibition in ADR nephropathy was fully preserved even when rapamycin treatment was started 2 weeks after ADR injection, suggesting that prolonged mTOR activation after a podocyte injury is the critical step for FSGS disease progression (Supplemental Figure 5A).

High-dose rapamycin (4 µg/g rapamycin three times weekly) protocols did not differ in their efficiency to prevent FSGS-like lesions from low-dose rapamycin applications in ADR nephropathy (Supplemental Figure 5, A and B). We did not observe any statistically significant effects of either low- or high-dose rapamycin on serum markers of glomerular function (Supplemental Figure 5B), indicating that our high-dose therapy protocol did not reach the same levels of mTOR inhibition as caused by genetic deletion of both Raptor alleles.

mTORC1 Activation Propagates FSGS Progression

Here, we identified mTORC1 activation as a common response to primary or secondary podocyte loss in human FSGS as well as in FSGS-like animal models. Genetically induced constitutive mTORC1 activation proved that persistent mTORC1 activation is self-sufficient to cause progressive glomerulosclerosis. Recently, the critical role of mTORC1 for glomerular development has been demonstrated and it appears to be a general theme that podocyte injury leads to the reactivation of developmental programs such as Notch, Wnt, and mTOR.^11,29 At the cellular level it was shown that activation of mTORC1 in diabetic animal models resulted in podocyte hypertrophy, dedifferentiation, and redistribution of slit diaphragm proteins, ultimately leading to podocyte loss and glomerulosclerosis.^12,13 Furthermore, it was revealed that mTORC1 activation results in increased levels of Nox4, Nox1, and NADPH oxidase activity in podocytes,³⁰ resulting in mTORC1-dependent ROS and ER stress signaling pathways known to instigate glomerular disease progression.¹² Thus, persistent mTORC1 activation presents a cellular stress pathway in podocytes that specifically propagates glomerular disease and FSGS progression (Figure 6).

Adaptive Functions of mTORC1 Provide a Model To Understand Rapamycin-Induced Proteinuria

Although we identified mTORC1 activation as an underlying cause of glomerular disease progression, a timed inducible knockout of the mTORC1 complex did not prevent, but dramatically facilitated, glomerular disease with FSGS-like lesions. These data clearly point to a nonlinear dual role of mTORC1 function for glomerular maintenance versus disease progression. Recent findings indicated mTOR as a regulator of podocyte size, allowing podocytes to adjust for podocyte loss or an expanding glomerular surface.^{10–13,31,32} Our data now further underline the importance of mTOR-dependent adaptive podocyte functions. In agreement, rapamycin treatment has been reported to cause proteinuria and FSGS in patients with underlying glomerular disease or after renal transplantation.^33–35 Hence, our data may help to better predict situations where mTORC1 sustains an adaptive response and rapamycin-induced glomerulopathy is a more likely scenario to occur. Recently, a role for the mTORC2-Akt2 survival axis was also documented in podocytes. Although rapamycin predominantly inhibits mTORC1, mTORC2 signaling appeared to be as well abolished in patients receiving prolonged rapamycin therapy.³⁶ Similar to the deletion of mTORC1 demonstrated here, the deletion of mTORC2^12,13 sensitized mice to proteinuria and foot process effacement in models of nephron reduction.³⁶ However, mTORC1-deficient models even developed nephrotic-range proteinuria and renal failure, underlining the critical contribution of impaired mTORC1 signaling to rapamycin-induced glomerulopathy (Figure 6). In any case, mTORC1 and mTORC2 seem to synergistically promote podocyte adaptation and maintenance, as simultaneous deletion of both mTORC complexes resulted in the most severe phenotype.¹³ Thus, it will be important to test whether novel mTORC1-specific inhibitors or dual inhibitors with simultaneous, but less efficient, inhibitory function on both mTOR complexes will cause less glomerular side effects and proteinuria than currently used mTOR inhibitors.

mTOR-Dependent Metabolic Podocyte Programming

Clarification of downstream events of mTOR signaling should lead to a better understanding of both the etiology and consequences of aberrant podocyte size regulation. In general, cell hypertrophy and an increased protein synthesis is an energy-dependent process that involves ATP hydrolyzation.³⁷ Although the balance of cell size regulation and energy supply is usually tightly controlled,³⁸ pathologic conditions can shift the metabolic profile and mitochondrial efficiency as has been extensively demonstrated in physiologic versus pathologic cardiac hypertrophy.^39,40 In podocytes, mTOR seems to be the main driving force linking both cell size and energy supply. mTOR inhibition appears to not only affect mitochondrial energy supply, but also mitochondrial efficiency and functionality, resulting in significantly increased ROS generation in a dose-dependent manner that likely contributes to podocyte injury. This is in line with previous work, demonstrating that TGF-β1 mTOR dependently increased both mitochondrial respiration and ROS production in podocytes.⁴¹ Mechanistically, increased ROS production is likely explained by the excessive demands of electron carriers in the presence of persistent mTOR activation,⁴² whereas the decreased expression of the distal subunits of complex II, SDHB, is a sufficient explanation for the increase in ROS production caused by high-dose mTOR inhibition.⁴³

Surprisingly, the cellular effects of mTOR inhibition are highly dose-dependent and vary from shut-down of oxidative phosphorylation accompanied with high levels of ROS production to a distinct modulation of cellular energy homeostasis, strengthening podocytes against toxic stimuli like ADR or FSGS-inducing factors in human disease. The underlying mechanisms of how low-dose mTOR inhibition preserves mitochondrial function in FSGS are not completely clear yet. In general, we predict a metabolic shift in podocytes to a more resting-like state with increased stress resistance.

Partial mTOR Inhibition as a Novel Therapeutic Concept To Treat FSGS and Progressive Glomerular Disease

Our genetic, as well as metabolic, data clearly implied a rationale for inhibition of mTOR as a therapeutic intervention. Excitingly, rapamycin, even when used at very low doses, could significantly reduce proteinuria and prevent podocyte loss and glomerulosclerosis. In contrast to previously published work, high-dose rapamycin treatment also exerted positive effects on halting the development of FSGS-like lesions in our ADR model, suggesting that the potential podocyte-toxic effects of higher doses of rapamycin vary depending on the mouse background, the pharmacodynamics, and the dose of ADR.⁴⁴ Nonetheless, there is need for further studies to define the subpopulation of human patients with FSGS who may benefit from pharmacologic inhibition of mTORC1, as well as the dose dependency of renal adverse events.

These data are now shedding new light on an interesting contradiction concerning the effects of mTOR inhibition in glomerular diseases. Although some human as well as animal studies evidenced beneficial effects, other studies targeting the same disease entities documented a worse outcome in response to mTOR inhibition. Supplemental Table 1 summarizes some of the recent human mTOR-inhibitor–based studies targeting steroid-resistant FSGS, ranging from massive worsening of proteinuria and renal function to complete remission of disease. Although this cannot be tested, our results may suggest that different treatment responses might correlate with the underlying severity of podocyte injury, levels of adaptive mTOR activation, and interindividual efficiency of mTOR inhibition. In any case, mTOR inhibitors used at high, fully immunosuppressive dosages appear to always bear the risk of causing a worsening of an underlying podocyte disease.

In summary, our data indicate clear experimental evidence for a novel treatment regimen on the basis of partial mTOR inhibition to prevent progressive glomerulosclerosis and FSGS (Figure 6B). Clinical studies will be required to confirm as well as to titrate the exact levels of an mTOR inhibitor therapy. Ideally, treatment protocols should also be directly adapted to the individual mTOR activity in biopsy specimens. The required serum levels of an mTOR-inhibitor–based therapy are likely very well below the current immunosuppression dosing regimens, which would also significantly reduce side effects of mTOR pharmacotherapy. Thus, these data suggest a novel targeted, better tolerated, and individualized mTOR-based therapy to prevent FSGS and progressive renal disease.

Human Kidney Samples, Study Approval, and Microarray Analysis

Microarray analysis was performed with human renal biopsy specimens collected within the framework of the European Renal cDNA Bank–Kröner-Fresenius Biopsy Bank. Microdissection of glomeruli, RNA hybridization, and analysis were performed as reported previously.⁴⁵ Please see our Supplemental Material for details.

Patients with recurrent FSGS were recruited at Charité-Universitätsmedizin Berlin. The Ethics Committee of the Charité-Universitätsmedizin Berlin approved the study. Please see our Supplemental Material for details.

Formalin-fixed, paraffin-embedded renal tissue specimens were obtained from the Hôpital Européen Georges Pompidou, Assistance Publique-Hôpitaux de Paris, Paris, France. Human tissue was used after approval from, and following the guidelines of, the local Ethics Committee (IRB00003888, FWA00005831). We chose representative images at high magnification. Samples from three patients each with FSGS and MCD were analyzed.

Mouse Experiments

For all models of acquired FSGS (ADR nephropathy and subtotal nephrectomy), mice were on or backcrossed to an ICR/CD1 background because of its known sensitivity toward the development of FSGS-like lesions.

We initially purchased CD1 (Crl: CD1 [ICR]) from Charles River (Europe) for ADR nephropathy experiments on a wild-type background. Later on, we used mice on a cognate ICR background (IcrTac: ICR; Taconic) for ADR injection and backcrossing from ADR-resistant genetic backgrounds (C57BL/6J). Doxorubicin (ADR) was aseptically prepared by our hospital pharmacy. A single dose of 12 µg/g ADR was intravenously injected once in the retro-orbital plexus under isoflurane anesthesia.

Sirolimus was injected at a dose of 0.5 µg (low dose) or 4 µg (high dose) (Pfizer, Germany) per gram of mouse body weight intraperitoneally, in a TWEEN 20 and PEG-400 solution (Sigma-Aldrich, Germany), three times a week for a follow-up course of 6 (delayed treatment) or 8 weeks (simultaneous treatment). Trough levels were obtained 24 hours after the last injection using liquid chromatography–mass spectrometry.

Raptor^{Het Podocyte} mice were generated by crossing Raptor^flox/flox with NPHS2.Cre mice after backcrossing for six generations on an ICR background (IcrTac: ICR; Taconic) as previously described.¹³ Raptor^flox/flox; NPHS2.rtTA; tetO.CreNPHS2.rtTA mice have been backcrossed on an ICR background as previously described.¹³ Induction of Raptor deletion in NPHS2.rtTA; tetO.Cre mice was performed as previously described.^13,46–48 In general, we used littermates as controls in experimental groups. Tsc1^Δpodocyte (Tsc1^flox/flox; NPHS2.Cre) mice were generated by backcrossing of Tsc1^fl/+ (129S4/SvJae; C57BL/6J) mice for 6 generations on a C57BL/6J background as previously described.¹² These animal studies were approved by the Committee on Research Animal Care, Regierungspräsidium Freiburg.

Subtotal nephrectomy (Nx) was performed as previously described.³⁶ After surgery, mice were fed a defined diet containing 30% casein and 0.5% sodium. Twenty-four hours before euthanasia, blood was collected from the tail veins of overnight-fasted mice for determination of urea concentrations. Urine was collected at week 5 of the experiment for determination of proteinuria and albuminuria. Animal procedures for subtotal nephrectomy were approved by the Departmental Director of Services Vétérinaires de la Préfecture de Police de Paris and the ethical committee of Paris Descartes University. Plasma urea and urinary protein, albumin, and creatinine concentrations were measured using an Olympus multiparametric analyzer (Instrumentation Laboratory).

Quantification of Podocytes and Glomerular Tuft Volume

To determine the number of podocytes per glomerulus as well as the glomerular tuft volume (6 weeks after ADR injection in wild-type animals [Figure 1], and 8 weeks after ADR injection or nephrectomy for other experiments; Tsc1^Δpodocyte and control animals at 6 weeks of age), we modified the approach of Hodgin et al.⁴⁹ using frozen kidney sections, which were stained for WT1, Nidogen, and Hoechst 33342. Please see our Supplemental Material for a detailed description of data analysis.

Quantification of Immunofluorescence Results

Glomeruli were selected as region of interest and a macro (consisting of a color threshold procedure, followed by filtering and Danielsson algorithm; AxioVision software; Carl Zeiss SpA) was applied to select stained areas and calculate them as percentage of the region of interest area. Please see our Supplemental Material for details.

Histologic Analysis

Kidneys were fixed in 4% paraformaldehyde and embedded in paraffin and further processed for PAS staining. Sclerosis index was done as described previously by el Nahas et al.⁵⁰ Please see our Supplemental Material for a detailed description of data analysis.

Quantitative Stereologic Analysis

Quantitative stereologic analysis of kidney sections was performed as described previously.⁵¹ Briefly, the mean glomerular volume (_[Glom]) was determined from the mean glomerular profile area (_Glom),⁵² which was obtained by measuring 100 systematically sampled glomerular profiles per animal. The physical dissector principle was applied for counting podocytes (Q⁻) as described, using semithin sections.^53,54 The numerical density of podocytes in glomeruli (N_V[P/Glom]) was calculated as the quotient of the sum of Q⁻ divided by the dissector volume. The mean podocyte volume, , was calculated by dividing Vv_(P/Glom) by N_V(P/Glom). All results were corrected for embedding shrinkage.

Cell Culture

For in vitro experiments, immortalized human podocytes were kindly provided by Moin Saleem (Bristol, UK), and directly isolated primary murine podocytes were used. Primary cells were isolated as previously described.⁵⁵ Please see our Supplemental Material for a detailed description.

Assessment of Oxidative Stress

Oxidative stress was assessed using CellROX Deep Red Flow Cytometry Assay Kit (#C10491, Molecular Probes; Life technologies, Eugene) following the manufacturer’s instructions. Please see our Supplemental Material for a detailed description.

ATP Measurement

ATP was measured using ADP/ATP Ratio Assay Kit (bioluminescent) (ab 65313; Abcam, Cambridge, UK) according to the manufacturer’s instructions. Please see our Supplemental Material for a detailed description.

Mitochondrial Analysis

Mitochondrial analysis was performed using Seahorse XFp (Seahorse Bioscience Billerica, MA) according to the user’s manual provided by the manufacturer. Presented data were normalized for cellular counts. Please see our Supplemental Material for a detailed description.

Urine and Serum Analyses

Urinary albumin and urinary or serum creatinine, respectively, were measured as previously described.¹³ Proteinuria was expressed as milligram albumin per milligram creatinine.

Statistical Analyses

All data are expressed as the mean±SEM if not stated otherwise. All shown data reflect a minimum of three subjects per test. Statistical comparisons were performed using two-tailed paired t test if not stated otherwise. P=0.05 was set as statistical significance level for rejection of the null hypothesis if not stated otherwise.