Renal Physiology

The mTOR-inhibitor rapamycin mediates proteinuria in nephrotoxic serum nephritis by activating the innate immune response

A. H. Kirsch, V. Riegelbauer, A. Tagwerker, M. Rudnicki, A. R. Rosenkranz, K. Eller


Rapamycin (Rapa) is an immunosuppressant used to prevent rejection in recipients of renal transplants. Its clinical use is limited by de novo onset or exacerbation of preexisting proteinuria. In the present study, Rapa administration was started 14 days after induction of murine nephrotoxic serum nephritis (NTS) to study glomerular effects of this mammalian target of rapamycin (mTOR) inhibitor. Glomeruli were laser-microdissected, and real-time PCR was performed to assess effects on glomerular cells and the expression of inflammatory cytokines. Immunohistochemical stainings were performed to confirm mRNA data on the protein level. Compared with nephritic control animals, Rapa-treated mice developed significantly increased albuminuria. This was accompanied by a more prominent glomerular infiltration by CD4+ T cells and macrophages. Glomerular mRNA expression profiling revealed increased levels of the proinflammatory cytokines interleukin-6 and tumor necrosis factor-α, and the chemokines monocyte chemoattractant protein-1 and macrophage inflammatory protein-1β and their cognate macrophage-associated receptors CCR2 and CCR5 in the Rapa-treated animals. Furthermore, there were elevated glomerular transcription levels of the regulatory T cell phenotype transcription factor Foxp3. No differences in the glomerular expression of the podocyte marker nephrin or the endothelial cell marker CD31 were observed on the mRNA or protein level. In conclusion, our data indicate that Rapa-induced proteinuria in NTS is a result of the activation of the innate immune system rather than a direct toxicity to podocytes or glomerular endothelial cells.

  • mammalian target of rapamycin inhibition
  • rapamycin
  • glomerulonephritis
  • proteinuria

the search for optimal immunosuppressive regimens for recipients of renal transplants has lead to the introduction of various agents, among them mammalian target of rapamycin (mTOR) inhibitors. Compared with current standard immunosuppressive agents, mTOR inhibitors have a variety of potentially crucial advantages, most notably a lack of inherent nephrotoxicity, a smaller incidence of CMV-infections and neoplasms, as well as antiatherogenic effects (28). Even though one recent study has shown improvement of renal function after early conversion from a calcineurin inhibitor to an mTOR-inhibitor regimen (4), later conversion failed to show improvements in graft survival (28). This has reaffirmed worries about detrimental effects of mTOR inhibition in the setting of preexisting proteinuria and impairments in renal function (8, 28).

The pathogenesis of rapamycin (Rapa)-induced proteinuria is still incompletely understood. Previous studies on mTOR inhibition in experimental models of renal disease have speculated that mTOR inhibitors might harm endothelial cells (5, 15), impair glomerular repair mechanisms (5), or enhance the inflammatory response (5, 6). Furthermore, it has been presumed that vascular endothelial growth factor (VEGF)-A plays a crucial role in the pathogenesis of Rapa-induced proteinuria. VEGF-A, the expression of which is decreased by mTOR inhibition (3, 5, 34, 37), has been shown to be crucial to the glomerular integrity (10, 11). In a previous study, we have shown that mTOR inhibition can have both protective and detrimental effects in a model of nephrotoxic serum nephritis (NTS). While the early administration of Rapa, concomitant with disease induction, reduced disease severity in nephritic mice, the initiation of Rapa treatment 14 days after disease onset significantly increased the nephritic phenotype (15). Later administration of Rapa was associated with an increase in albuminuria, significantly more renal T cell and macrophage infiltration, and morphologic evidence of glomerular endothelial cell damage (15).

To provide additional insights into the pathogenesis of Rapa-induced deterioration of NTS, the present study was designed to evaluate the renal substructure most directly involved in this process: the glomerulus.


Study design, mice.

The respective Austrian authorities approved all animal experiments. Animals used in this study were 8- to 12-wk-old, male C57BL/6 mice (Charles River, Sulzfeld, Germany) that were maintained in a virus- and antibody-free environment at the central research animal facility (Medical University Innsbruck, Innsbruck, Austria).

NTS was induced as described elsewhere (27). Briefly, mice were preimmunized 3 days before administration of the nephrotoxic serum with 2 mg/ml whole rabbit IgG (Jackson ImmunoResearch Laboratories, West Grove, PA) dissolved in incomplete Freund's adjuvant (Sigma, St. Louis, MO) with nonviable, desiccated Mycobacterium tuberculosis H37a (Difco Laboratories, Detroit, MI). Three days thereafter, heat-inactivated rabbit-derived anti-mouse glomerular basement membrane serum was administered intravenously via the tail vein. Rapa (LC Labs, Woburn, MA) stock solution was prepared by dissolving Rapa in 100% ethanol, which was then dissolved in sterile saline for intraperitoneal injection at a dose of 0.5 mg/kg body wt. Daily intraperitoneal Rapa or vehicle administration was started on day 14 after disease induction. Rapa trough levels were assessed by HPLC-MS/MS as reported previously (13) and were found to be similar to previously published data.

Urine samples were obtained on days 0, 7, 14, 21, 28, and 35 in metabolic cages. Mice were killed, and samples were collected on day 35.

Assessment of urinary albumin/creatinine excretion.

As reported previously (15), urinary albumin excretion was quantified by means of a double-sandwich enzyme-linked immunosorbent assay (Abcam, Cambridge, MA), whereas urinary creatinine excretion was measured photometrically using a picric acid-based kit (Sigma).

Immunhistochemical stainings.

A three-layered immunoperoxidase staining technique was used for the assessment of infiltration by macrophages and T cells and the expression of glomerular cell markers. Frozen tissue sections (4 μm) were stained using rat-derived primary antibodies for F4/80 (Serotec, Oxford, UK), CD4 monoclonal antibody (clone YTS191; Serotec), CD68 (Abcam), VEGF-A (clone 1F07–2C01; BioLegend, San Diego, CA), and CD31 (Serotec) and a goat anti-mouse nephrin antibody for nephrin (Jackson ImmunoResearch Laboratories). A biotin-conjugated goat anti-rat IgG (Serotec) or donkey anti-goat IgG antibody (Jackson) was used as secondary antibody. Intraglomerular cells were counted assessing a minimum of 50 glomerular cross sections per mouse at a magnification of ×200 and counting the number of positive cells within the glomerulus. Glomerular cell marker and VEGF-A staining intensity was scored by a blinded examiner, as were all other histological assessments. Representative images were taken on a BX51 microscope using a DP71 digital camera and CellD software (all Olympus Austria, Vienna, Austria).

Isolation of glomerular tissue by laser-capture microdissection.

Renal cryosections were cut at 8 μm and transferred onto nonelectrostatic laser-capture microdissection (LCM) slides (Arcturus, Mountain View, CA). Slides were stained using the HistoGene LCM Frozen Section Staining Kit (Arcturus) according to the manufacturer's protocol. LCM of renal cryosections was performed on a PixCell IIs LCM system (Arcturus). A minimum of 50 dissected glomeruli were transferred onto CapSure LCM caps (Arcturus), and specificity for isolation of glomerular tissue was confirmed microscopically (Fig. 1).

Fig. 1.

Isolation of glomeruli by means of laser-capture microdissection. Renal cryosections were stained; glomeruli were then identified morphologically (A), and separated from surrounding tissue (B). Isolated glomeruli were assessed microscopically for contaminating nonglomerular tissue (C).

Isolation of glomerular tissue by sieving.

Glomeruli were isolated using a series of three sieves with serially declining mesh size (150, 90, and 45 μm) as described previously (29). Briefly, minced murine kidneys were passed over the sieves, and glomeruli were collected from the top of the middle sieve. Purity of sieved fractions was controlled by microscopy.

RNA isolation and reverse transcription real-time PCR.

RNA from sieved glomeruli was isolated using TRIzol (Sigma) with a standardized protocol. Thereafter, the Superscript III Transcription Kit (Invitrogen, Carlsbad, CA) and random primers (Roche, Basel, Switzerland) were used to reverse transcribe 2 μg of total RNA.

RNA from microdissected glomeruli was isolated using the PicoPure RNA isolation kit (Arcturus) according to the manufacturer's protocol. After reverse transcription and confirmation for linearity of preamplification for the genes of interest, cDNA obtained from microdissected glomeruli was preamplified using the TaqMan PreAmp MasterMix (Applied Biosystems) according to the manufacturer's protocol. To selectively preamplify genes of interest, we used TaqMan gene expression assays (Applied Biosystems) for hypoxanthine-guanine-phosphoribosyltransferase (HPRT; Mm01324427_m1), CD31 (Mm001242584_m1), nephrin (Mm00497828_m1), VEGF-A (Mm0001281449_m1), interleukin (IL)-6 (Mm00446190_m1), tumor necrosis factor (TNF)-α (Mm_00443258_m1), interferon (IFN)-γ (Mm00801778_m1), Foxp3 (Mm00475162_m1), RORγt (Mm01261022_m1), T-bet (Mm00450960_m1), CCR2 (Mm00438270_m1), CCR5 (Mm01216171_m1), macrophage inflammatory protein (MIP)-1β (Mm0044311_m1), and monocyte chemoattractant protein (MCP)-1 (Mm00441242_m1).

Real-time analysis was performed on a CFX96 real-time system (Bio-Rad, Vienna, Austria). For quantification of relative gene expression, TaqMan Universal PCR Mastermix (Applied Biosystems) and the respective gene expression assays were used. HPRT was used as a reference for the laser-dissected samples, whereas 18S RNA (Mm03928990_g1) was used in the case of sieved glomeruli. The relative glomerular gene expression level was then calculated using the ΔΔCt method.

Statistical analysis.

All statistical evaluations were performed using SPSS 13.0 for Windows (SPSS, Chicago, IL). Results are presented as medians or means ± SE. Normality was assessed using the Kolmogorov-Smirnov test with Lillefors correction. Treatment groups were compared using a nonparametric Mann Whitney U-test or unpaired t-test as appropriate, depending on the variable in question. A two-tailed P < 0.05 was considered statistically significant.


In the present study, mice previously subcutaneously immunized against rabbit IgG were subjected to NTS by means of rabbit-derived nephritogenic serum as described elsewhere (15) and were treated intraperitoneally with vehicle or Rapa at a dose of 0.5 mg/kg body wt on a daily basis starting on day 14 after immunization. Evaluation of the urinary albumin-to-creatinine ratio revealed a significantly higher ratio in Rapa-treated animals compared with vehicle-treated controls (Table 1).

View this table:
Table 1

Urinary albumin-to-creatinine ratio

To gain insights on the molecular background of Rapa-mediated proteinuria in this context, we performed LCM on renal cryosections and isolated a minimum of 50 glomeruli/sample (see concise methods). We then performed real-time PCR of preamplified reverse-transcribed glomerular cDNA to assess the expression of glomerular cell markers, VEGF-A, as well as inflammatory cytokines and transcription factors. Because Rapa has been proven to decrease VEGF-A (3, 5, 37), which is crucial for glomerular integrity (10, 11), we first evaluated whether glomerular VEGF-A mRNA expression is influenced by Rapa treatment 35 days after induction of NTS. To our surprise, no difference in the glomerular mRNA expression of VEGF-A in the two study groups was detected (Fig. 2A). These data are contradictory to our previous findings in whole kidney extracts showing decreased VEGF-A mRNA expression in Rapa-treated mice (15). In our previous study, we hypothesized that glomerular endothelial cell-derived VEGF-A rather than podocyte-derived VEGF-A is decreased by Rapa treatment in NTS (15). In the isolated glomerular fractions, we did not detect a difference between Rapa-treated and control mice in the viability of glomerular endothelial cells or podocytes: Rapa treatment starting 14 days after induction of NTS neither affected the glomerular mRNA expression of the endothelial cell marker CD31 nor the podocyte marker nephrin, which has been shown to be diminished in the setting of podocyte injury (22, 40), compared with vehicle control (Fig. 2A). These data were confirmed on the protein level by immunohistochemistry, where the glomerular staining score for VEGF-A (Fig. 2B), CD31 (Fig. 2C), and nephrin (Fig. 2D) revealed no differences (Fig. 3). VEGF-A has been shown to not only be vital to glomerular endothelial cell and podocyte survival (10, 11) but also to be of importance to extraglomerular endothelial cells and peritubular capillary integrity (17, 18). Because, contrary to our data gained from whole kidney (15), we did not detect any glomerular changes in VEGF-A and CD31 mRNA expression, we speculate that Rapa treatment might diminish extraglomerular, peritubular vessels rather than glomerular endothelial cells. Nevertheless, further studies are needed to prove this hypothesis.

Fig. 2.

Integrity of glomerular endothelial cells and podocytes. Glomerular vascular endothelial growth factor (VEGF)-A, CD31, and nephrin mRNA expression was assessed by performing real-time PCR from RNA obtained from laser-microdissected glomeruli (A) and was not different between vehicle-treated (gray bars; n = 7) and rapamycin-treated (white bars; n = 7) animals. Protein expression of VEGF-A (B), the endothelial cell marker CD31 (C), and the podocyte marker nephrin (D) was assessed by attributing a glomerular staining score to immunohistochemically stained slides. Not significant (ns), P > 0.05 vs. controls.

Fig. 3.

Representative glomerular stainings for VEGF-A (A), CD31 (B), and nephrin (C). Magnification in all panels is ×600, except the middle and right panels in B, where magnification is ×400. Glomeruli are marked by black circles.

Rapa has been proven to activate the innate immune system, namely monocyte/macrophages and dendritic cells in vitro and in vivo (14, 36). Previously, we provided evidence that Rapa might similarly activate the innate immune system in NTS, since whole kidney extracts from Rapa-treated mice showed increased IL-6 mRNA expression and there was increased renal macrophage infiltration compared with control mice (15). Therefore, we first evaluated the infiltration of macrophages into the glomeruli by performing immunohistochemistry for the macrophage markers CD68 and F4/80. There was a significant increase in the number of glomerular F4/80+ and CD68+ cells in Rapa-treated animals compared with controls (Fig. 4, A and B).

Fig. 4.

Glomerular macrophage invasion is increased significantly by treatment with rapamycin. Immunohistochemical evaluation of F4/80+ (A) and CD68+ (B) intraglomerular cells showed that rapamycin treatment (white bars; n = 7) significantly increased the no. of positive glomerular cells compared with vehicle treatment (gray bars; n = 7). Laser-capture microdissection of murine glomeruli with subsequent RNA isolation, reverse transcription, and real-time PCR showed that Rapa significantly increased the glomerular gene expression levels of interleukin (IL)-6 and tumor necrosis factor (TNF)-α (C). Sieved glomeruli were analyzed for mRNA expression of CCR2, monocyte chemoattractant protein (MCP)-1, CCR5, and macrophage inflammatory protein (MIP)-1β (D). *P < 0.05 vs. controls.

This was accompanied by increased inflammatory activity in Rapa-treated animals as evidenced by significantly higher transcription levels of the proinflammatory cytokines IL-6 and TNF α (Fig. 4C), both of which are known to be expressed by macrophages (21, 26).

To test whether the increase in glomerular macrophage infiltration was due to increased chemotactic migration associated with higher levels of chemokines, we assessed the gene expression levels of the chemokines MCP-1 and MIP-1β. Because of the relatively scarce expression level of these genes in RNA, which was isolated from laser-microdissected glomeruli, we then assessed glomerular RNA obtained from sieved glomeruli. The latter is a less specific method of isolating glomerular tissue, which, however, allows isolating RNA in higher abundance. Using this approach, we found glomeruli isolated from the Rapa-treated animal group to show significantly increased expression of MCP-1 in glomeruli compared with vehicle-treated controls (Fig. 4D). MIP-1β was also found to be increased in glomeruli of Rapa-treated mice, but significance was not reached (Fig. 4D). Furthermore, we determined the expression level of the respective chemokine receptors expressed on macrophages and found the MCP-1-binding partner CCR2 to be expressed at significantly higher levels (Fig. 4D). Again, the MIP-1β-binding partner CCR-5 was increasingly expressed in glomeruli of Rapa-treated mice, but significance was not reached (Fig. 4D).

To evaluate the T cell response in this setting, we stained for CD4+ and CD8+ T cells. CD8+ T cells were only found in the interstitium rather than in the glomeruli (data not shown), but CD4+ T cells did partly migrate into the glomeruli. There were significantly higher numbers of glomerular CD4+ T cells detectable in Rapa-treated mice compared with controls 35 days after induction of NTS (Fig. 5A). To identify the subset of CD4+ T cells accounting for the observed increase in CD4 positivity in Rapa-treated animals, we assessed for the expression of Th1, Th2, Th17, Th9, and regulatory T cell (Treg) markers. Using the glomerular sieving approach, we detected no difference in the glomerular mRNA expression of the Th1 cytokine IFNγ and its master regulator T-bet between the two groups (Fig. 5, B and C). The Th2 and Th9 master regulators GATA-3 and PU.1, respectively, were not detectable in the sieved glomerular fractions (data not shown). There was also no difference in the mRNA expression of the Th17 master regulator RORγt (Fig. 5C). Most interestingly, we detected a significant increase in the glomerular transcription level of the Treg marker Foxp3 in sieved glomerular fractions of Rapa-treated mice 35 days after NTS induction compared with respective controls (Fig. 5C).

Fig. 5.

The no. of invading CD4+ T cells was increased significantly in rapamycin-treated nephritic mice (white bars; n = 7) compared with vehicle-treated controls (gray bars; n = 7) (A). Sieved glomerular fractions were assessed for gene expression and did not reveal any differences in the expression of the Th1 cytokine interferon (IFN)-γ (B), whereas there was a significant increase in the expression of Foxp3 (C). *P < 0.05 vs. controls.

Taken together, our assessment of the inflammatory response suggests a significant increase in the number of infiltrating glomerular macrophages, since there were significantly more F4/80+ and CD68+ glomerular cells in animals treated with Rapa. Concurrently, glomerular mRNA levels of TNF α and IL-6, both secretion products of macrophages (21, 26), were increased. The observed increase and activation of macrophages in the glomeruli due to Rapa treatment seems to be responsible for disease aggravation in Rapa-treated animals subjected to NTS. Macrophages have repeatedly been shown to be key mediators of renal injury by induction of apoptosis and inhibition of mitosis of mesangial cells in a nitric oxide-dependent manner (7, 35).

Recent in vivo and in vitro studies have documented a proinflammatory effect of mTOR inhibition on macrophages. Weichhart et al. (36) provided in vitro and in vivo data proving that Rapa-treated macrophages and mononuclear cells are clearly driven toward a proinflammatory phenotype, as evidenced by increases in IL-6. In this study, Rapa significantly prolonged the survival of Listeria monocytogenes-infected mice in association with an increase in CD68 expression by splenic macrophages (36). Furthermore, in a study by Jagannath et al. (16), Rapa significantly upregulated the in vitro and in vivo capacity of macrophages and dendritic cells to take up and subsequently present antigens. In our model, the increased presence of glomerular macrophages in Rapa-treated animals can be attributed to either an increase in macrophage proliferation, a higher rate of macrophage trafficking to the glomeruli, or a combination of the two. The observed increase in glomerular expression levels of MCP-1 and MIP-1β as well as their cognate receptors CCR2 and CCR5, respectively, strongly supports the concept of a Rapa-mediated pathogenic augmentation in macrophage trafficking to the glomeruli. This is supported by the fact that antibody-mediated blockade of MCP-1/CCR2 interaction in the thioglycollate-induced peritonitis model almost completely abolished monocyte recruitment, whereas the influx of neutrophils and lymphocytes remained unaffected (19). In line, CCR2 blockade in a rat model of anti-glomerular basement glomerulonephritis has been shown to attenuate the disease (32), and the same holds true for the neutralization of MCP-1 in this model (12) or the selective depletion of CCR2-expressing cells (20). In these studies, the number of renal macrophages was reduced significantly, which clearly documents the essential role MCP-1 plays in macrophage recruitment to glomeruli. CCR5, which is expressed on macrophages differentiated from blood monocytes and Th1 cells activated in response to inflammatory stimuli (39), has been shown to act pathogenic in experimental models of rheumatoid arthritis (33) as well as allograft rejection (13). Furthermore, CCR5-expressing cells have repeatedly been shown to be recruited abundantly to the kidney and specifically to glomeruli in the setting of renal inflammation (19, 31), and the role of CCR5-ligands in the pathogenesis of experimental glomerulonephritis is documented (24, 31). However, more detailed studies are clearly warranted to answer the question of how Rapa influences the chemokine expression and thereby the migration behavior of macrophages in NTS.

Although Rapa treatment had a considerable effect on macrophages, we also detected an effect on cells of the adaptive immune system. We found an increased number of CD4+ T cells infiltrating the glomeruli of Rapa-treated mice, but these cells accounted neither for Th1, Th2, Th9, nor Th17 cells. On the basis of the available data, we believe these cells to be Foxp3+ Tregs. Considering the clear evidence of increased inflammatory activity in Rapa-treated-animals, an increase in Tregs may seem counterintuitive. Nevertheless, others and we have shown that Tregs can only exert their immunosuppressive capacity when located to the secondary lymphoid organs rather than to the kidney (9, 23, 25, 38). The observed increase in the Treg population might be explained by a direct Rapa-mediated effect, since Rapa has been proven to increase the Treg population in vitro and in vivo (1, 2). On the other hand, glomerular Treg infiltration might be a compensatory reaction to the increased inflammatory activity in Rapa-treated animals. Clearly, more studies will be needed to further elucidate the phenotype of these Foxp3-expressing cells (30).

Taken together, we here provide evidence that Rapa induces proteinuria in NTS by activating the innate immune response, namely an increase in the number of glomerular macrophages rather than by injuring endothelial cells or podocytes.


A. H. Kirsch is supported by the PhD program in molecular medicine at the Medical University of Graz. This work was sponsored by a Pfizer grant to A. R. Rosenkranz.


No conflicts of interest, financial or otherwise, are declared by the authors.


Author contributions: A.H.K., V.R., A.T., and K.E. performed experiments; A.H.K. and K.E. analyzed data; A.H.K. and K.E. interpreted results of experiments; A.H.K. prepared figures; A.H.K. and K.E. drafted manuscript; A.H.K., M.R., A.R.R., and K.E. approved final version of manuscript; M.R., A.R.R., and K.E. conception and design of research; M.R., A.R.R., and K.E. edited and revised manuscript.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
View Abstract