SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Objective

Toll-like receptor 9 (TLR-9), a receptor for CpG DNA, has been implicated in the activation of immune cells in lupus. We undertook this study to determine whether the expression of TLR-9 in resident renal cells in lupus nephritis is related to the development of tubulointerstitial injury.

Methods

TLR-9 was analyzed in selectively retrieved renal tissue from (NZB × NZW)F1 mice at different stages of disease by laser capture microdissection combined with real-time quantitative reverse transcriptase–polymerase chain reaction, and in renal biopsy specimens from lupus nephritis patients by immunohistochemistry. We investigated for the molecular component responsible for TLR-9 activation by cultured proximal tubular cells in serum from patients with lupus.

Results

Renal tissue from NZB × NZW mice displayed robust TLR-9 expression localized to proximal tubular cells. TLR-9 levels correlated with proteinuria and tubulointerstitial injury to the extent that a cyclin-dependent kinase inhibitor, while reducing proteinuria and renal structural damage, prevented tubular TLR-9 generation in lupus mice. Consistently, exaggerated TLR-9 staining was found in proximal tubular cells of lupus patients, which correlated with tubulointerstitial damage. DNA-containing immune complexes purified from sera of patients with lupus induced TLR-9 in cultured proximal tubular cells. This was prevented by CCGG-rich short oligonucleotides, specific antagonists of CpG DNA, indicating that the DNA component of immune complexes was required for TLR-9 stimulation.

Conclusion

These findings suggest that tubular TLR-9 activation has a pathogenetic role in tubulointerstitial inflammation and damage in experimental and human lupus nephritis, and they indicate a novel target for future therapies.

Lupus nephritis is a severe organ manifestation of systemic lupus erythematosus (SLE) (1) and may influence morbidity and mortality. In SLE, abnormalities of immune regulation lead to loss of self tolerance, which in turn triggers autoimmune responses. Those include T and B cell dysfunction and activation of autoreactive B cells that produce antibodies against nuclear antigens. Immune complexes are detected in the circulation, and their deposition, together with complement activation in glomeruli and along the tubular basement membrane, plays a major role in the pleomorphic histopathology of lupus nephropathy. The initiating event may be local binding of nuclear or other antigens to glomerular and tubular sites followed by in situ immune complex formation. Antigens contain nucleic acids, and antinuclear antibody (ANA) formation occurs as a consequence of stimulation of invariant receptors that recognize nucleic acid determinants (2).

Toll-like receptors (TLRs) are a family of transmembrane proteins that recognize conserved molecular patterns shared by a wide variety of microorganisms (3, 4). Among TLRs, TLR-9 specifically binds CpG DNA, a hypomethylated form of DNA typical of bacteria and virus (5). CpG DNA activates potentially autoreactive B cells and plasmacytoid dendritic cells to secrete Th1-like cytokines (6). Aberrant TLR-9 activation and release of cytokines and chemokines occurs in plasmacytoid dendritic cells after exposure to immune complexes containing DNA, rich in CpG motifs, isolated from the serum of patients with active lupus nephritis (7).

Lesions of lupus nephritis involve the renal glomerulus and tubulointerstitium. Whereas the glomerular lesions have been studied extensively, the pathophysiology of tubulointerstitial inflammation and damage remains ill-defined. Yet tubulointerstitial changes are prominent and contribute prominently to the unfavorable long-term prognosis (8). Proximal tubular epithelial cells play an active role in tubulointerstitial inflammation and fibrotic lesions, to the extent that when these cells are exposed to anti–double-stranded DNA (anti-dsDNA) antibodies from patients with active lupus, they produce cytokines that promote local recruitment of inflammatory cells (9). DNA-containing autoantibody complexes found in the serum of patients with lupus activate dendritic cells to generate cytokines and chemokines via TLR-9 (7).

Considering that proximal tubular cells have antigen-presenting capacity as dendritic cells under certain circumstances (10), we wondered whether TLR-9 could be involved in the development of tubular damage and interstitial inflammation in lupus nephritis. To this end, we investigated renal TLR-9 expression in (NZB × NZW)F1 lupus-prone mice at different stages of the disease, both when animals had elevated serum levels of anti-dsDNA antibodies and later on, when proteinuria and tubulointerstitial damage ensued. TLR-9 expression in resident renal cells was assessed by laser capture microdissection combined with TaqMan real-time quantitative reverse transcriptase–polymerase chain reaction (RT-PCR). The finding that increased tubular TLR-9 generation was associated with proteinuria provided the rationale for using seliciclib, a cyclin-dependent kinase (CDK) 2, 7, and 9 inhibitor, which effectively reduced proteinuria and ameliorated renal injury in this model (11). To assess the relevance of animal data to the pathophysiology of interstitial lesions in human lupus, we evaluated the expression of TLR-9 in renal biopsy specimens from patients with lupus nephritis. Based on the previously reported finding that endogenous DNA-containing autoantibody complexes in the serum of SLE patients activated dendritic cells through TLR-9 (7), we finally assessed whether DNA-containing immune complexes purified from SLE serum stimulated cultured human proximal tubular cells to express TLR-9.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Animal experiments.

Female (NZB × NZW)F1 mice (Harlan, Milan, Italy) and CD-1 mice (Charles River, Calco, Italy) were used. Animal care and treatment were conducted according to the institutional guidelines in compliance with national (Decreto Legislativo n.116, Gazzetta Ufficiale suppl 40, 18 febbraio 1992, Circolare n.8, Gazzetta Ufficiale 14 luglio 1994) and international (European Community [formerly, the European Economic Community] Council Directive 86/609, OJL358-1, December 1987; Guide for the Care and Use of Laboratory Animals, US National Research Council, 1996) laws and policies. NZB × NZW mice were divided into the following groups: group 1 (n = 5), killed at age 2 months, before the onset of renal disease (12, 13); group 2 (n = 5), killed at age 5 months, when immune complex deposition occurs but renal structure and function are not yet severely impaired (14); and group 3 (n = 10), killed at age 8 months. An additional group of NZB × NZW mice (n = 10) was treated with 200 mg/kg seliciclib (R-Roscovitine, CYC202; Cyclacel, Dundee, UK) (11), given daily by gavage from age 2 months to age 8 months. Eight-month-old normal CD-1 mice (n = 5) were used as controls for TLR-9 expression analyses.

At the end of the experimental period, urinary protein excretion and serum blood urea nitrogen (BUN) levels were determined as described previously (14), and renal tissue specimens were removed. Serum levels of anti-dsDNA autoantibodies were evaluated by enzyme immunoassay (Diastat anti-dsDNA kit; Bouty Laboratory, Milan, Italy) (15).

Renal morphology.

Duboscq-Brazil–fixed and paraffin-embedded renal cortex sections (3 μm, Microtome V; LKB, Bromma, Sweden) were stained with hematoxylin and eosin, Masson's trichrome, and periodic acid–Schiff reagent. Glomerular and tubulointerstitial lesions were assessed using light microscopy. Glomerular intracapillary hypercellularity was given a score ranging from 0 to 3 (0 = absent; 1 = mild; 2 = moderate; 3 = severe). Extracapillary proliferation was also graded (0 = absent; 1 = <25%; 2 = 25–50%; 3 = >50% of glomeruli involved). Deposits were graded from 0 to 3 (0 = no deposits; 1 = <25%; 2 = 25–50%; 3 = >50% of glomeruli involved). At least 100 glomeruli were examined for each biopsy specimen. Glomerular damage was expressed as a single mean value of the intracapillary hypercellularity, extracapillary proliferation, and deposits scores.

Tubular changes (atrophy, casts, and dilatation) and interstitial inflammation were graded from 0 to 3 (0 = no changes; 1 = <25%; 2 = 25–50%; 3 = changes affecting >50% of the sample). At least 10 fields per sample were examined at low magnification (10×) for scoring of the interstitium. All biopsy specimens were analyzed by the same pathologist in a blinded manner.

Immunohistochemical analysis.

Paraffin kidney sections (3 μm) were hydrated, retrieved, and immunostained with a monoclonal antibody against F4/80 antigen of mouse monocyte/macrophages (Caltag, Burlingame, CA) and a polyclonal antibody recognizing mouse and human TLR-9 (Santa Cruz Biotechnology, Santa Cruz, CA). Bound primary antibodies were visualized with diaminobenzidine staining by using ABC kits (Vector, Burlingame, CA) as previously described (14). Negative controls were obtained by omitting the primary antibody. F4/80-labeled cells were counted in at least 20 randomly selected high-power fields (at 400× magnification) per animal. A semiquantitative score was assigned to the tubular TLR-9 staining intensity according to the following pattern: 0 = absent; 1 = weak; 2 = moderate; 3 = strong staining.

Real-time quantitative RT-PCR.

Real-time RT-PCR analysis was performed as previously described (16). Messenger RNA (mRNA) expression was quantified by SYBR Green 2-step RT-PCR for TLR-9. The sequence of the primers for mouse TLR-9 was 5′-CAACATGGTTCTCCGTCGAA and 5′-TGTACCAGGAGGGACAAGGG, and that for human TLR-9 was 5′-CACCCTCAACTTCACCTTGGA and 5′-TGCACGGTCACCAGGTTGT. The expression of mouse and human TLR-9 was normalized to GAPDH and hypoxanthine guanine phosphoribosyltransferase 1 (HPRT-1) mRNA content, respectively, and calculated relative to controls.

Laser capture microdissection and real-time RT-PCR.

Frozen renal specimens were subjected to laser capture microdissection as described (17). Tubular structures and interstitial areas with infiltrating cells were dissected from 2 areas of identical proportions (mean area 52,386 μm2) on the same section. Microdissected tissue areas were collected for RNA extraction using the QIAamp DNA Micro kit (Qiagen, Hilden, Germany). Samples were used for TaqMan real-time RT-PCR analysis according to previously described protocols (18). TLR-9 and HPRT-1 quantitation was performed using primers and probes from Assay on Demand kits (Mm00446193_m1 and Mm00446968, respectively; Applied Biosystems, Foster City, CA). HPRT-1 was used for normalization. Messenger RNA levels were quantitatively analyzed by comparing experimental levels with standard curves generated with serial dilutions of the same positive sample.

Patient biopsy specimens.

Biopsy samples obtained for diagnostic purposes from patients admitted to the Unit of Nephrology and Dialysis, Azienda Ospedaliera, Ospedali Riuniti di Bergamo were studied. We selected 15 patients diagnosed as having class IV lupus nephritis according to the World Health Organization classification criteria (19). There were 13 women and 2 men, with a mean ± SEM age of 32.1 ± 3.2 years, proteinuria of 5.37 ± 1.21 gm/24 hours, and serum creatinine of 1.15 ± 0.13 mg/dl. At the time of renal biopsy, 5 patients were being treated with steroids. For control specimens, we used normal kidney tissue from 4 patients undergoing nephrectomy for kidney adenocarcinoma (proteinuria <0.5 gm/24 hours). Informed consent was obtained from all patients. Twenty-four–hour urinary protein excretion and serum creatinine concentration were measured using an autoanalyzer (CX5; Beckman Instruments, Fullerton, CA). Analysis of tubulointerstitial lesions and TLR-9 protein expression was performed on paraffin-embedded biopsy specimens as described above for murine sections.

In vitro studies.

Cell culture.

Human proximal tubular HK-2 cells (20) (American Type Culture Collection, Manassas, VA) were grown in Dulbecco's modified Eagle's medium (DMEM)/Ham's F-12 with 5% fetal calf serum supplemented as previously described (21). Confluent cells were maintained overnight in DMEM/Ham's F-12 without serum prior to use in experiments.

Serum samples and isolation of immune complexes from sera.

Serum samples were collected from ANA-positive/anti-dsDNA–positive lupus patients (n = 4), ANA-positive/anti-dsDNA–negative patients with undifferentiated connective tissue disease (UCTD) (n = 3), and ANA-negative/anti-dsDNA–negative healthy controls (n = 3) for comparison. Positivity for ANA and anti-dsDNS was defined as titers greater than 1:320 and 35 IU/ml, respectively. Anti-dsDNA antibodies were determined by Farr assay (Diagnostic Products, Los Angeles, CA). IgG was purified from sera by protein G affinity chromatography (HiTrap Protein G; Amersham Pharmacia Biotech, Uppsala, Sweden) as described (22). Immune complexes in the purified IgG fraction were separated using a 300,000-Da–cutoff concentrator (VIVASPIN; VivaScience, Hannover, Germany).

Experimental design.

HK-2 cells were exposed for 6 hours to 20% human sera, 100 ng/ml of immune complexes, or 20% IgG-depleted sera from SLE patients, UCTD patients, or healthy subjects. To block DNA-induced activation, experiments were performed in the presence of oligonucleotide (ODN) TCCTGGATGGGAAGT specific for TLR-9 (1 μg/ml) (23).

Statistical analysis.

Data are expressed as the mean ± SEM. Data for real-time RT-PCR were analyzed by parametric analysis of variance with Bonferroni adjustment. For all other parameters, the Kruskal-Wallis test was used. P values less than 0.05 were considered significant. Linear regression analysis of TLR-9 levels and tubulointerstitial damage scores or proteinuria levels was performed.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Systemic and renal function and structure parameters in NZB × NZW mice.

Serum levels of circulating anti-dsDNA antibodies in NZB × NZW mice are reported in Table 1. Anti-dsDNA antibody levels were significantly increased starting from age 5 months compared with those in 2-month-old mice (P < 0.01) and were further augmented at age 8 months. None of the NZB × NZW mice that were given vehicle developed proteinuria at ages 2 and 5 months. In contrast, urinary protein excretion was significantly increased in 8-month-old NZB × NZW mice compared with that in 2- and 5-month-old NZB × NZW mice (P < 0.01) (Table 1). Renal function, evaluated as serum BUN levels, was normal up to age 5 months but deteriorated at age 8 months (Table 1). Renal morphology assessed by light microscopy revealed no structural abnormalities at age 2 months in NZB × NZW mice (Table 1). Focal glomerular hypercellularity was occasionally found in the kidneys of 5-month-old NZB × NZW mice. At age 8 months, when the animals were proteinuric, pronounced morphologic changes were observed, with intracapillary hypercellularity associated with focal extracapillary proliferation. Immune-type deposits were detected in the mesangium and in the glomerular capillary walls with subendothelial distribution. Tubulointerstitial damage was also observed, along with a consistent interstitial accumulation of F4/80-positive monocyte/macrophages (Table 1).

Table 1. Systemic and renal function and structure parameters in NZB × NZW mice*
GroupSerum anti-dsDNA antibody, units/mlProteinuria, mg/daySerum BUN, mg/dlGlomerular damage scoreTubulointerstitial damage score, 0–3Interstitial F4/80-positive monocyte/macrophages, cells/hpf
  • *

    Values are the mean ± SEM. Anti-dsDNA = anti–double-stranded DNA; BUN = blood urea nitrogen; hpf = high-power field.

  • See Materials and Methods for description of groups.

  • See Materials and Methods for description of glomerular and tubulointerstitial damage scores.

  • §

    P < 0.01 versus 2-month-old NZB × NZW mice.

  • P < 0.01 versus 2-month-old and 5-month-old NZB × NZW mice.

  • #

    Seliciclib (200 mg/kg) was administered daily by gavage from age 2 months to age 8 months.

  • **

    P < 0.05 versus 8-month-old NZB × NZW mice.

  • ††

    P < 0.01 versus 8-month-old NZB × NZW mice.

Age 2 months10.29 ± 1.190.59 ± 0.0717.48 ± 1.220.07 ± 0.070 ± 00.25 ± 0.25
Age 5 months145.26 ± 59.58§0.89 ± 0.0622.42 ± 1.140.13 ± 0.080 ± 00.22 ± 0.18
Age 8 months335.06 ± 99.79§53.08 ± 13.93109.83 ± 16.231.73 ± 0.111.25 ± 0.2045.3 ± 6.90
Seliciclib-treated#62.37 ± 24.90**2.55 ± 1.72††29.02 ± 4.05††0.13 ± 0.07**0.10 ± 0.07**7.71 ± 1.98††

TLR-9 mRNA expression in the kidneys of NZB × NZW mice.

The expression of TLR-9 in kidneys from NZB × NZW mice during the evolution of the disease was analyzed by real-time quantitative RT-PCR. TLR-9 mRNA was detectable in the kidney of control mice, although at much lower levels than in the spleen. In 2- and 5-month-old NZB × NZW mice, levels of TLR-9 mRNA were comparable with those in controls (Figure 1A). In contrast, a 6-fold increase in the receptor expression was observed in kidneys of 8-month-old NZB × NZW mice (P < 0.01 versus 2- and 5-month-old NZB × NZW mice and versus control mice).

thumbnail image

Figure 1. Renal Toll-like receptor 9 (TLR-9) mRNA and protein expression in (NZB × NZW)F1 (NZB/W) mice. A, Expression of TLR-9 mRNA evaluated by real-time quantitative reverse transcriptase–polymerase chain reaction in kidneys of NZB/W mice at ages 2, 5, and 8 months and after treatment with seliciclib. Data are expressed relative to findings in normal mice used as controls. Values are the mean and SEM. ◊ = P < 0.01 versus NZB/W mice at ages 2 months and 5 months and versus controls; ∗ = P < 0.05 versus NZB/W mice at age 8 months. B, Semiquantitative evaluation of renal tubular TLR-9 protein expression assessed by immunostaining. Values are mean and SEM scores. ◊ = P < 0.01 versus NZB/W mice at ages 2 months and 5 months and versus controls; ∗∗ = P < 0.01 versus NZB/W mice at age 8 months. C, Linear regression analysis of TLR-9 protein expression and tubulointerstitial damage. Ct = threshold cycle.

Download figure to PowerPoint

Expression of TLR-9 protein in renal tubuli of NZB × NZW mice.

To localize the cellular source of renal TLR-9 mRNA expression, we performed immunohistochemistry experiments. In kidneys of lupus mice at ages 2 and 5 months, TLR-9 staining was weak (Figures 2A and B) and similar to that in controls. In contrast, renal sections from 8-month-old NZB × NZW mice exhibited a strong positive signal. Of interest, the staining was localized in tubular epithelial cells, predominantly in proximal tubuli (Figure 2C), but not in glomerular cells (Figure 2D). Inflammatory cell infiltrates were also markedly positive for the receptor (Figures 2C and E). No signal was observed when the primary antibody was omitted (results not shown). Consistent with data on TLR-9 mRNA expression, the semiquantitative evaluation of TLR-9 protein staining revealed an increase of the protein in NZB × NZW mice at age 8 months, which was statistically significant compared with levels in 2- and 5-month-old lupus mice and control mice (P < 0.01) (Figure 1B). Linear regression analysis of the relationship between scores for tubular TLR-9 protein expression and tubulointerstitial damage in NZB × NZW mice showed a highly significant correlation (r = 0.82, P < 0.01) (Figure 1C). The levels of TLR-9 protein expression in tubuli also significantly correlated with the degree of proteinuria (r = 0.65, P < 0.01).

thumbnail image

Figure 2. Representative photomicrographs of TLR-9 protein localization in kidneys of NZB/W mice. A and B, In kidneys from NZB/W mice at age 2 months (A) and age 5 months (B), TLR-9 protein was almost undetectable. C and D, In renal tissue from 8-month-old NZB/W mice, a strong TLR-9 signal was present in the tubuli (C) but not in glomeruli (D). C and E, TLR-9–positive inflammatory cells (arrows) localized around TLR-9–positive tubuli were observed. F, TLR-9 staining was dramatically decreased in seliciclib-treated NZB/W mice. (Original magnification × 250.) See Figure 1 for definitions.

Download figure to PowerPoint

TLR-9 expression in microdissected mouse renal specimens.

Laser capture microdissection was used to selectively retrieve proximal tubuli and inflammatory cell infiltrates from kidneys of 8-month-old NZB × NZW mice, followed by assessment of TLR-9 gene expression through real-time quantitative RT-PCR. After control of tissue preservation by light microscopy, kidney sections were used for microdissection. Captured proximal tubuli expressed high levels of TLR-9 mRNA compared with proximal tubuli from control mice, which showed low but detectable transcript values (Table 2). As expected, TLR-9 mRNA was markedly elevated in microdissected inflammatory cell infiltrates compared with the values measured in proximal tubuli.

Table 2. Toll-like receptor 9 mRNA expression as assessed by real-time reverse transcriptase–polymerase chain reaction analysis of microdissected proximal tubuli and inflammatory cell infiltrates*
GroupProximal tubuli, fg/52,386 μm2 DNACell infiltrates, fg/52,386 μm2 DNA
  • *

    Values are the mean ± SEM. NE = not evaluated.

  • See Materials and Methods for description of groups.

  • P < 0.01 versus control mice.

  • §

    Seliciclib (200 mg/kg) was administered daily by gavage from age 2 months to age 8 months.

  • P < 0.01 versus 8-month-old NZB × NZW mice.

NZB × NZW mice  
 Age 8 months4.13 ± 1.3627.46 ± 11.0
 Seliciclib-treated§0.01 ± 0.00NE
Control mice0.01 ± 0.00NE

Reduction of proteinuria and TLR-9 expression by seliciclib in the kidneys of NZB × NZW mice.

Treatment of NZB × NZW mice with seliciclib from age 2 months to age 8 months significantly reduced serum levels of anti-dsDNA antibodies (Table 1). Seliciclib prevented the development of proteinuria and limited renal function impairment, as documented by lower BUN levels compared with those in untreated mice (Table 1). Glomerular and tubulointerstitial changes as well as interstitial accumulation of F4/80-positive monocyte/macrophages were remarkably attenuated (Table 1). Seliciclib administration prevented up-regulation of both TLR-9 mRNA and protein in the kidney of lupus mice. Real-time RT-PCR showed that TLR-9 mRNA levels from whole kidney were halved by treatment with the drug compared with those in untreated NZB × NZW mice (P < 0.05) (Figure 1A). When quantitative evaluation of TLR-9 mRNA was performed in microdissected tubuli, a significant decrease in receptor expression was found after treatment (P < 0.01) (Table 2).

Consistent with data on TLR-9 mRNA, immunohistochemistry analysis of the corresponding protein revealed that in kidneys of seliciclib-treated NZB × NZW mice (Figure 2F), the intensity of tubular TLR-9 protein staining was markedly reduced compared with that in untreated lupus mice of the same age (Figure 2C). The reduction of the intensity of TLR-9 protein staining after treatment with seliciclib was statistically significant (P < 0.01) (Figure 1B).

Tubular TLR-9 expression in human lupus nephritis.

The pattern of TLR-9 protein expression in renal biopsy specimens from patients with lupus nephritis (Figure 3) was similar to that observed in NZB × NZW mice. Intense and diffuse staining for TLR-9 was detected in proximal tubuli from patient specimens (mean ± SEM score 1.35 ± 0.26) (Figures 3A and B), which contrasted with findings in normal kidney, in which receptor expression was almost undetectable (mean ± SEM score 0.06 ± 0.06) (Figure 3D). This difference was statistically significant (P < 0.05). TLR-9 protein expression correlated positively with tubulointerstitial injury (tubulointerstitial injury score 1.43 ± 0.25) (r = 0.70, P < 0.01). No staining was observed in glomeruli. Areas of cell infiltrates were markedly positive for TLR-9 (Figure 3C).

thumbnail image

Figure 3. Photomicrographs of Toll-like receptor 9 (TLR-9) protein immunostaining on human renal biopsy sections from patients with class IV lupus nephritis. A and B, TLR-9–positive tubuli (A) and both tubuli and interstitial inflammatory cells positive for TLR-9 (B) were found in kidneys of patients. C, Boxed area in B at higher magnification, showing positivity for TLR-9 in cell infiltrates. D, No TLR-9–positive tubuli were found in kidneys of controls. E, Negative control omitting the primary anti–TLR-9 antibody. (Original magnification × 250 in A, B, D, and E; × 650 in C.)

Download figure to PowerPoint

TLR-9 expression in tubular cells exposed to SLE serum.

The ability of sera from patients with SLE to stimulate TLR-9 expression in proximal tubular epithelial cells was investigated and compared with that of sera from patients with UCTD without anti-dsDNA antibodies or from healthy subjects. Anti-dsDNA–positive sera from SLE patients induced TLR-9 expression in proximal tubular epithelial cells with a 9-fold average increase over control sera, whereas anti-dsDNA–negative sera from UCTD patients did not (Figure 4A). The gene expression augmentation observed was statistically significant (P < 0.01). Furthermore, immune complexes purified from SLE sera, but not from UCTD sera, were able to induce a significant (P < 0.01) increase in TLR-9 expression (Figure 4B). The addition of short ODNs rich in CCGG motifs, which are specific antagonists of CpG DNA, prevented TLR-9 stimulation by immune complexes. IgG-depleted sera from SLE patients were still able to induce significant TLR-9 gene expression (P < 0.05 versus control) (Figure 4C). This increase was partially inhibited by short ODNs.

thumbnail image

Figure 4. Effect of serum, immune complexes (ICs), and IgG-depleted serum from patients with systemic lupus erythematosus (SLE) on TLR-9 mRNA expression in cultured human proximal tubular epithelial cells. A, HK-2 cells were incubated with 20% serum isolated from healthy subjects (control), patients with undifferentiated connective tissue disease (UCTD), or SLE patients. B and C, HK-2 cells were exposed to 100 ng/ml of ICs purified from sera of healthy subjects, UCTD patients, or SLE patients, the latter either alone or in the presence of 1 μg/ml of short oligonucleotides (ODNs) rich in CCGG motifs, specific antagonists of CpG DNA (B), or with 20% IgG-depleted sera from healthy subjects or SLE patients, the latter either alone or in the presence of 1 μg/ml of short ODNs (C). After 6 hours of incubation, cells were harvested and TLR-9 mRNA expression was evaluated by real-time quantitative reverse transcriptase–polymerase chain reaction. Results were normalized to hypoxanthine guanine phosphoribosyltransferase 1 expression and expressed relative to control. Values are the mean and SEM. ∗ = P < 0.01 versus control and UCTD sera or ICs; ○ = P < 0.01 versus SLE ICs without ODNs; # = P < 0.05 versus control. See Figure 1 for other definitions.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Previous studies in the MRL-lpr/lpr mouse model of lupus nephritis demonstrated that injection of synthetic or bacterial DNA rich in CpG motifs activated TLR-9 in inflammatory cells infiltrating the kidney, leading to aggravation of renal inflammation, proteinuria, and tissue damage (24, 25). The present study demonstrated that proximal tubular epithelial cells were sites of robust expression of TLR-9 mRNA and protein in NZB × NZW mice with overt nephropathy. Up-regulation of tubular TLR-9 expression was concomitant with the development of proteinuria and correlated with tubulointerstitial damage.

To our knowledge, this is the first study to show TLR-9 expression in proximal tubuli in experimental lupus. Laser capture microdissection combined with real-time quantitative RT-PCR allowed us to selectively retrieve proximal tubuli and inflammatory cell infiltrates and quantify the expression of TLR-9, which was remarkably increased in proximal tubuli of NZB × NZW mice compared with that in normal mice. Despite a number of reports (including the present one) which suggested a pathogenetic role of TLR-9 in lupus nephritis (24–26), studies in TLR-9–deficient mice have shown that lack of TLR-9 worsened the course of lupus disease (27, 28). Data obtained in studies of sepsis syndrome are relevant for reconciling the contrasting observations on TLR-9 in lupus (29). A direct link between lipopolysaccharide (LPS) poisoning and sepsis acting on TLR-4 was derived from the observation that TLR-4 activation by LPS caused fever, shock, and death, whereas Sultzer's mutant mice, bearing nonfunctional TLR-4, were protected against shock when given LPS (30). However, when these mice were infected with gram-negative bacteria, which release LPS and pathogen-associated molecules, the manifestations of sepsis worsened and the rate of death increased. These paradoxical results can be explained if TLR-4 both protects the host by sequestering infectious organisms and causes manifestation of sepsis when the system is maximally activated and protection is overwhelmed.

A major finding of the present study is the positive correlation between tubular TLR-9 expression and the development of proteinuria and tubulointerstitial damage, which was further suggested by experiments using the CDK 2, 7, and 9 inhibitor, seliciclib, known to ameliorate autoimmune disease in NZB × NZW mice (11). Targeting aberrant renal cell proliferation by CDK inhibitors was found to be an effective therapeutic strategy in animal models of proliferative glomerulonephritis (31, 32). Here we report that the renoprotective effect of seliciclib in NZB × NZW mice was associated with inhibition of TLR-9 expression. To explain such an effect, one should consider the intracellular pathways leading to TLR-9 activation.

TLR-9 is localized in the endoplasmic reticulum and traffics to the endosomal–lysosomal compartment after cellular activation, becoming accessible to endocytosed foreign CpG DNA (33–35). It has been shown that CDK-2–cyclin E complexes control the vesicular fusion reaction of the endosomal apparatus in liver parenchyma cells (36). It is possible that under proteinuric conditions, seliciclib could have influenced endocytosis in proximal tubular cells, preventing TLR-9 activation at the endosomal sites as well as the downstream pathologic events. Seliciclib prevented tubular injury and exerted a remarkable antiinflammatory effect by limiting the interstitial accumulation of monocyte/macrophages, a major source of TLR-9. Taken together, these observations implicate TLR-9 as a trigger of tubulointerstitial damage in lupus nephritis.

We then sought to determine whether animal data were relevant to the pathophysiology of interstitial lesions in human lupus nephritis. We found intense staining for TLR-9 localized at proximal tubular cells in biopsy specimens from patients with active lupus nephritis, which would extend the significance of our observation to humans. In immune-mediated glomerular diseases, which include lupus nephritis, alteration of size-selective properties of the glomerular capillary wall is associated with abnormal filtration of plasma proteins, which conceivably include immune complexes (37) containing DNA enriched in hypomethylated CG motifs (38, 39). Ultrafiltered macromolecules and proteins are then actively reabsorbed by receptor-mediated endocytosis in proximal tubular cells, giving rise to endosomes that, upon acidification, progress to lysosomes (40). In vitro studies have shown that internalization of SLE immune complexes into subcellular lysosomes containing TLR-9 induces a signaling cascade leading to activation of plasmacytoid dendritic cells to produce cytokines and chemokines (7).

It is conceivable that a mechanism of TLR-9 activation by DNA-containing immune complexes similar to that described in dendritic cells occurs in proximal tubular cells. Proximal tubular cells possess Fc neonatal receptors on the plasma membrane, which could favor the uptake of immune complexes containing DNA (41), followed by DNA-induced activation of TLR-9. Our in vitro experiments showing that sera from patients with lupus nephritis, but not from patients with another autoimmune disease, stimulated cultured proximal tubular epithelial cells to produce TLR-9 would confirm such a possibility.

We found that immune complexes present in sera from patients with lupus induced up-regulation of TLR-9 in proximal tubular cells. Short ODNs rich in CCGG motifs, specific antagonists of CpG DNA, prevented the TLR-9 mRNA increase induced by immune complexes from lupus patient sera, suggesting that the DNA component of immune complexes is required for the stimulatory activity. IgG-depleted sera from patients with lupus still retained the ability to stimulate TLR-9 in proximal tubular cells, but this was prevented by the inhibitory short ODNs. This indicates that additional molecules in lupus sera participate in the engagement of the receptor. In this context, nucleosomes that bear hypomethylated DNA might act as endogenous ligands responsible for TLR-9 up-regulation (23).

In conclusion, we provide evidence that resident proximal tubular epithelial cells are activated to express TLR-9 in experimental and human lupus nephritis. Tubular TLR-9 expression correlates with tubulointerstitial damage. Cultured proximal tubular cells produce TLR-9 after exposure to immune complexes from lupus patient sera. DNA complexed with IgG is required for TLR-9 stimulatory activity. These findings suggest that tubular TLR-9 activation has a pathogenetic role in tubulointerstitial inflammation and damage in lupus nephritis, and they indicate a novel target for future therapies.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Dr. Benigni had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Benigni, Zoja, Remuzzi.

Acquisition of data. Caroli, Longaretti, Gagliardini, Galbusera, Moioli, Romagnani.

Analysis and interpretation of data. Benigni, Caroli, Longaretti, Gagliardini, Zoja, Galbusera, Moioli, Remuzzi.

Manuscript preparation. Benigni, Caroli, Zoja, Remuzzi.

Statistical analysis. Caroli, Longaretti, Gagliardini, Galbusera.

Provision of human sera and purified IgG. Tincani, Andreoli.

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

We thank Daniela Corna, Sara Conti, Fabio Sangalli, and Daniela Rottoli for excellent technical assistance. Seliciclib was kindly provided by Cyclacel (Dundee, UK).

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES