Protective outcomes of low-dose doxycycline on renal function of Wistar rats subjected to acute ischemia/reperfusion injury

Abstract

Renal ischemia-reperfusion injury (IRI) is a major cause of acute renal failure. Doxycycline (Dc) belongs to the tetracycline-class of antibiotics with demonstrated beneficial molecular effects in the brain and heart, mainly through matrix metalloproteinases inhibition (MMP). However, Dc protection of renal function has not been demonstrated. We determined whether low doses of Dc would prevent decreases in glomerular filtration rate (GFR) and maintain tubular Na⁺ handling in Wistar rats subjected to kidney I/R. Male Wistar rats underwent bilateral kidney ischemia for 30 min followed by 24 h reperfusion (I/R). Doxycycline (1, 3, and 10 mg/kg, i.p.) was administered 2 h before surgery. Untreated I/R rats showed a 250% increase in urine volume and proteinuria, a 60% reduction in GFR, accumulation of urea-nitrogen in the blood, and a 60% decrease in the fractional Na⁺ excretion due to unbalanced Na⁺ transporter activity. Treatment with Dc 3 mg/kg maintained control levels of urine volume, proteinuria, GFR, blood urea-nitrogen, fractional Na⁺ excretion, and equilibrated Na⁺ transporter activities. The Dc protection effects on renal function were associated with kidney structure preservation and prevention of TGFβ and fibronectin deposition. In vitro, total MMP activity was augmented in I/R and inhibited by 25 and 50 μM Dc. In vivo, I/R augmented MMP-2 and -9 protein content without changing their activities. Doxycycline treatment downregulated total MMP activity and MMP-2 and -9 protein content. Our results suggest that treatment with low dose Dc protects from IRI, thereby preserving kidney function.

1. Introduction

Acute kidney injury (AKI) is an important global healthcare issue, whose incidence has increased over the years [1,2]. Currently, AKI constitutes 15–20% of all hospital stays, affects 30–40% of patients admitted to a critical care unit, and accelerates progression to end stage renal disease and premature death [3,4]. Renal transient ischemia is a multifactorial pathology, causing an impairment in oxygen, nutrient supply, and product waste removal from kidney cells. Subsequent reperfusion promotes endothelial dysfunction and is a major cause of kidney injury [3]. Consequently, ischemia-reperfusion injury (IRI) results in inflammatory processes and oxidative stress, leading to apoptosis and tubular necrosis: eliciting a loss in kidney function [5–7].

A new paradigm of matrix metalloproteinases (MMPs) action has arisen in ischemic organs, such as heart and brain, undergoing injury. MMPs regulate inflammation, epithelial-mesenchymal transition, cell proliferation, angiogenesis, and apoptosis [8,9]. MMPs belong to a family of 28 structurally related enzymes containing a signal peptide, catalytic Zn²⁺ binding, a carboxy-terminal domain, and an amino-terminal propeptide, that covers the catalytic site [10]. In addition to their classical function in extracellular matrix remodeling, MMPs also localize to subcellular organelles, promoting proteolysis of intracellular proteins, a process that is also susceptible to ischemic activity [11]. MMP-2 (gelatinase A) and MMP-9 (gelatinase B) are the main MMPs present in the kidney. Both enzymes are localized within cells where they can cleave substrates inside and outside of the cells [12–14]. The involvement of MMPs in kidney damage depends on the stage of the disease and the underlying cause. In general, MMP-2 and -9 are up-regulated after I/R and their activation modulates renal microvascular permeability, and enhances inflammation and repair [15–17]. MMP-9 activates TGFβ1, and MMP-2 and -9 cleave and activate the precursor of IL-1; both actions further amplify inflammation. However, MMP-2 is also activated during the tubular repair phase after IRI [18]. In a diabetic nephropathy model, MMP-2 may be an anti-fibrotic agent regulating the severity of renal fibrosis [19]. In renal transplantation models of acute allograft rejection, the ratio between MMPs and its endogenous inhibitor TIMPs (tissue inhibitors of matrix metalloproteinases) is an important molecular mechanism for the development of tubulo-interstitial fibrosis that culminates in end-stage renal disease (ESRD) [20]. Indeed, using a well-documented model of fibrosis, the unilateral urethral obstruction, MMP-9 leads to myofibroblastic activation and MMP-9/TIMP-1 interaction is reduced in patients with resistant albuminuria [21].

Doxycycline (Dc) is of special interest because it is the most potent MMP inhibitor of the tetracycline-class antibiotics. Its mechanism of action involves Zn²⁺ chelation, a property distinct from the antibacterial action [22]. Many studies have demonstrated the beneficial use of tetracycline for several inflammatory diseases in which MMPs play a pathological role [23–25]. In the kidney, Dc treatment attenuates IRI by reducing oxidative stress, inflammation, and apoptosis as well as facilitating repair [14, 26–30]. To the best of our knowledge, other than the molecular benefits of Dc, the outcomes of renal function protection by I/R have not been demonstrated. Another issue involves the appropriate Dc dose needed to improve renal function without causing bacterial resistance. We hypothesized that intraperitoneal treatment with low doses of Dc prevents glomerular filtration rate (GFR) decrease and the disruption in tubular Na⁺ handling. To test this hypothesis, we used a model of bilateral kidney ischemia for 30 min followed by 24 h reperfusion. The protection provided by treatment with low doses of Dc 2 h before induction of ischemia was evaluated by measuring GFR and the activity of the primary Na⁺ transporters, as well as the expression levels of fibronectin, TGFβ, and MMPs.

2. Materials and Methods

2.1. Animal model

Forty male Wistar rats (200–250 g) were purchased from Centro de Criação de Animais de Laboratório (CECAL, Fundação Oswaldo Cruz, Rio de Janeiro, Brazil). The rats were maintained in the vivarium with constant temperature (23 ± 2 °C), a standard dark/light cycle (12/12 h), fed regular chow, and allowed to drink water ad libitum as recommended by the good practice standards in research and approved by the Ethics Committee for the Use of Animals (CEUA) of the Federal University of Rio de Janeiro, Brazil (protocol number 083/15).

2.2. Experimental protocol

Rats were subjected to bilateral ischemia for 30 min followed by 24 h reperfusion (I/R) iparallel to sham surgery for the control animals. (control, n = 8) [31]. Briefly, rats were anesthetized with ketamine (50 mg/kg, i.p., Cristalia, São Paulo, Brazil) and xylazine (5 mg/kg, i.p., Syntec, São Paulo, Brazil). In the I/R group (n = 27), ischemia was induced under aseptic conditions by applying a non-traumatic vascular clamp in both renal pedicles. In the Dc-treated groups (n = 19), 2 h before ischemia (to guarantee plasma steady state concentration [32]) Dc (doxycycline hyclate, D9891, Sigma-Aldrich, Saint Louis, MO) was administered in doses of 1, 3, or 10 mg/kg (i.p.) resulting in 3 groups: I/R + Dc1 (n = 4); I/R + Dc3 (n = 9); and I/R + Dc10 (n = 6), respectively. For internal control, we evaluated the effects of 3 mg/kg Dc treatment on sham-operated rats (control + Dc3; n = 5). Next, during the 24 h reperfusion interval, rats were individually housed in metabolic cages for 24 h-urine and 24 h-water intake measurements. Afterward, rats were euthanized by conscious decapitation. Blood samples were collected in chilled tubes containing 5 mM EDTA, and centrifuged at 3,000 g (10 min at 4°C) for plasma fraction separation to measure Na⁺, creatinine, and urea-nitrogen concentration. Immediately after kidney harvesting, five poles of the left kidney were sectioned from the: 1) control; 2) control + Dc3; 3) I/R; and 4) I/R + Dc3 groups and immersed in 4% paraformaldehyde for histological studies [33]. For the primary Na⁺ transporter activities, Western blot, and gelatin zymography; the kidney cortexes were dissected and processed as described below.

2.3. Renal functional responses to I/R

The concentration of Na⁺ in urine and blood was measured by flame spectrometry (Analyzer 910 MS, Analyzer, São Paulo, Brazil) and urine osmolality was measured by the freezing point depression technique using a cryoscopic osmometer (Osmomat 030, Gonotec, Berlin, Germany). GFR was calculated by the creatinine clearance and expressed as μl/min. Urinary and plasma creatinine, blood urea nitrogen, and proteinuria were measured by spectrophotometry using specific colorimetric kits (Gold Analisa, Belo Horizonte, Brazil).

2.4. Renal histology and immunofluorescence

The fixed left kidney pole tissue was processed, embedded in paraffin blocks, and stained with Periodic acid-Schiff (PAS) and Picrosirius red, as described by Beiral et al. [31]. The pathologist (L.S.R.) analyzed the sections in a blind manner. In PAS stained tissue we evaluated glomerulus diameter and tubular injury. The sizes of the individual glomeruli in the cortex were calculated as the average of glomerular diameters within a field of view; the calculations involved at least 18 glomeruli per rat kidney (n = 5 for each group; total of 90 glomeruli per group). The tubular injury score was evaluated by tubular dilation, tubular atrophy, vacuolization, and the degeneration and sloughing of tubular epithelial cells, or thickening of the tubular basement membrane. The scoring system was: 0 = no tubular injury; 1 = < 10%; 2 = 10%–25%; 3 = 26%–50%; 4 = 51%–75%; and 5 = > 75% tubular injury (n = 5, each group). Immunofluorescence was performed as previously described [34] using anti-rabbit TGF-β1 antibody (1:100) (sc-146; Santa Cruz Biotechnology, Santa Cruz, CA) or anti-rabbit fibronectin antibody (1:250) (ab2413; Abcam, Cambridge, UK) at 4°C, followed by incubation with specific secondary antibody (1:1,000, AlexaFluor 488- labeled secondary antibody; Invitrogen, Carlsbad, CA). Digital images were captured by a DS-U2/L2 USB camera attached to a Nikon Eclipse 50i microscope.

2.5. Primary Na⁺ transporters activity

First, homogenate preparation from kidney cortex was obtained as described previously [35]. Total protein concentration was assayed according to Lowry et al. [36], using bovine serum albumin (BSA) as standard. Samples were maintained at −20°C until use. The activities of the primary Na⁺ transporters: (Na⁺ + K⁺)-ATPase and ouabain-resistant, furosemide-sensitive Na⁺-ATPase were measured as previously described [35].

2.6. Western blot

One hundred micrograms of cortex homogenates were separated by electrophoresis in a polyacrylamide gel (SDS PAGE 10% for MMP-9 and 7.5% for MMP-2 detection) and transferred to a nitrocellulose membrane (10600003; GE Healthcare Life Sciences, Freiburg, Germany) [36]. After blocking, the membrane was incubated with the corresponding primary antibodies (1:200): human MMP-9 monoclonal antibody (MAB911; R&D Systems, Minneapolis, MN), mouse MMP-2 monoclonal antibody (sc-13594; Santa Cruz Biotechnology), and secondary fluorescent antibodies (anti-mouse, Li-Cor, IRDye 680RD, 1:4000). The immunofluorescence was detected using the Odyssey System (Li-Cor Bioscience, Lincoln, NE) for infrared imaging recording, and band intensities at the same molecular band identified by zymography were quantified using Image J software (National Institutes of Health, Bethesda, MD). Blots were striped and re-probed using β-actin monoclonal antibodies (A5316, Sigma-Aldrich, Saint Louis, MO). The quantification of MMPs immune signal was normalized to the β-actin quantification.

2.7. MMP activity

For MMPs activity, kidney cortex extracts were obtained by repeated freeze/thaw cycles in a buffer, pH 7,6 (50 mM Tris-HCl, 150 mMNaCl, 5 mM CaCl₂, 0,05% Brij35, and 1% Triton X-100). Then, the cellular extracts were centrifuged at 10,000 g for 30 min at 4 °C, and the supernatants were used immediately to determine proteolytic activity. Enzyme activity was determined using the fluorogenic peptide substrate Z-Arg-Arg-AMC (C5429, Sigma-Aldrich). The cleavage of substrate was monitored continuously in a spectrofluorometer (SpectraMax Gemini XPS, Molecular Devices, CA) using an excitation wavelength of 380 nm and an emission wavelength of 460 nm. Reactions were initiated by addition of substrate (10 μM) to the extract (30 μg protein) in a final volume of 60 μL of the same buffer used for kidney lysis, in the absence or presence of Dc (25 and 50 μM), metalloproteinase inhibitors at 1 mM (1,10-phenanthroline [PHEN], ethylene glycol-bis(β-aminoethyl ether)-N,N,N′,N′-tetraacetic acid [EGTA] and ethylenediaminetetraacetic acid [EDTA]), cysteine peptidase inhibitor trans-epoxysuccinyl L-leucylamido-(4-guanidino) butane (E-64) at 10 μM, serine peptidase inhibitor phenylmethylsulfonyl fluoride (PMSF) at 5 mM, and aspartic peptidase inhibitor pepstatin A at 5 μM. The reaction mixture was incubated at 37 °C for 1 h. The assays were controlled for self-liberation of the fluorophore over the same time interval [37].

2.8. Gelatin Zymography

For analysis by sodium dodecyl sulfate-containing polyacrylamide gel electrophoresis (SDS-PAGE), 40 μg of protein from the samples, obtained as described above, were collected and then sample buffer was added to SDS-PAGE (125 mM Tris, pH 6.8, 4% SDS, 20% glycerol and 0.002% bromophenol blue) in the ratio of 1: 1. Proteins were electrophoresed on 7.5% SDS-PAGE with 0.1% gelatin incorporated as substrate [38]. Electrophoresis was performed at 200 V and 20 mA per gel for 2 h at 4°C. Afterward, the gels were incubated in 2.5% Triton X-100 (1 h at room temperature under constant stirring) for removal of SDS. The gels were then incubated for 2 h at 37°C in pH 7.6 buffer (50 mM Tris-HCl supplemented or not with unspecific MMP inhibitors: 1,10 phenanthroline, EGTA, or EDTA). The gels were stained with 0.2% Coomassie Blue R-250 in methanol-acetic acid-water (50:40:10) and decolorized in a solution containing the same solvents in the ratio 85:5:10 [39]. The gels were photographed using LPix photo documenter and quantified by the IMAGE J program.

2.9. Statistical Analysis

Except where otherwise indicated, the data are presented as the mean ± SEM. Multiple comparisons were made by one-way analysis of variance (ANOVA) followed by Tukey’s multiple comparison test. In Figure 7B, unpaired t-test was applied. Statistical tests and graphs used GraphPad Prism 6.0 software (GraphPad Inc., La Jolla, CA).

Figure 7. Effect of 3 mg/kg Dc treatment on TGF-β1 expression.

Immunofluorescence was performed as described in Methods. Paraffin embedded kidney poles from control, I/R and I/R + Dc3 rats were incubated overnight with anti-rabbit TGF-β1 antibody (1:100) at 4°C, followed incubation with specific secondary antibody (1:1,000, AlexaFluor 488- labeled secondary antibody). Representative photomicrographs of kidney poles in glomeruli (×400) and tubulointerstitial areas (×400).

3. Results

3.1. Doxycycline treatment prevents reduction of renal function promoted by I/R

To evaluate the influence of the Dc treatment in the reduction of renal function promoted by I/R, rats were subjected to an i.p. injection 2 h before a bilateral I/R procedure, using 3 different doses of the drug (1, 3, and 10 mg/Kg). Table 1 shows that water intake did not change under any of the conditions. We also observed that Dc prevented the ~250% increase in urine volume and proteinuria of rats subjected to I/R, even at the lowest dose of 1 mg/kg. Interestingly, 3 mg/kg Dc blocked the reduction in urine osmolality seen in I/R (in mOsm/kg H₂O: control = 1.5 ± 0.2; I/R = 0.77 ± 0.14; I/R + Dc3 = 1.84 ± 0.18; P < 0.05). Sham-operated rats treated with Dc 3 mg/kg presented physiological parameters similar to control rats.

Table 1.

The impact of Dc treatment in water intake, urine volume and proteinuria in the I/R rat

	control (n=8)	I/R (n=8)	I/R + Dc (mg/kg)			control + Dc 3 (n=5)
			1 (n=4)	3 (n=5)	10 (n=5)
Water intake (ml/100 g BW in 24 h)	9.4±1.8a	10.8±1.8a	11.2±0.2a	7.6±1.6a	10±1.1a	8.2±0.4a
Urine volume (ml/100 g BW in 24 h)	3.7±0.4a	9.4±1.1b	4.3±0.6a	2±0.3a	4.1±0.3a	2.3±0.4a
Proteinuria (mg/100 g BW in 24 h)	3.6±0.6a	9.3±0.9b	2.9±0.7a	2.2 ±0.1a	3.9±0.5a	3.9±0.5a

Rats were subjected to bilateral ischemia for 30 min followed by 24 h reperfusion (I/R) or to sham surgery. Two hours before the surgery, 3 different doses of doxycycline (1, 3, and 10 mg/kg) were administered to the I/R rats (I/R + Dc1, I/R + Dc3, and I/R + Dc10). Dc 3 mg/kg was administered to sham-operated rats as an internal control. Urine and plasma collections were taken at the end of the 24 h reperfusion, corresponding to those in which renal function parameters were measured. Values are means ± SEM. Different superscripted lower-case letters indicate statistically significant differences (P < 0.05; one-way ANOVA followed by Tukey’s post-test). BW: body weight.

In addition, the ~60% GFR reduction observed in I/R rats was prevented in the I/R + Dc3. GFR was 435 ± 77μl/min, similar to the control rat (Figure 1A). Blood urea-nitrogen (BUN) reflected the GFR data. There was an increase in BUN (~250%) in the I/R rats; 3 mg/kg Dc treatment completely prevented this effect (Figure 1B). The same profile was observed for creatinine accumulation in plasma (inset to Figure 1B). Tubular Na⁺ handling is presented in Figure 2. Plasma Na⁺ concentration did not change in any condition (~137 mM; data not shown) and Na⁺ filtered load (FL_Na) paralleled GFR, with 3 mg/kg Dc being the lowest dose needed to return FL_Na to control values (Figure 2A). One of the characteristics of IRI is the decrease in GFR and an increase in tubular Na⁺ reabsorption [40]. In this model, we observed a decrease in the fractional excretion of Na⁺ (FE_Na) from 0.55 ± 0.06 (control) to 0.20 ± 0.02% in I/R rats. Doxycycline treatment with 3 mg/kg avoided FE_Na decrease. In the I/R + Dc1 rats, the FE_Na is even greater than in I/R rats. We ascribe this effect to the inverse relationship to the FL_Na, which is decreased at this condition. None of these parameters was modified when Dc 3 mg/kg was administered to sham-operated rats. Because 3 mg/kg Dc was the lowest dose to protect renal function, we measured the influence of this drug at the primary Na⁺ transporters: (Na⁺ + K⁺)-ATPase and ouabain-resistant, furosemide-sensitive Na⁺-ATPase (Figure 3). In the I/R rats, cortical (Na⁺ + K⁺)-ATPase activity was augmented (165%) (Figure 3A) whereas cortical Na⁺-ATPase was decreased (85%) (Figure 3B). In the medulla, both Na⁺ transporter activities were reduced by ~50% (Figure 3C and 3D). I/R rats treated with 3 mg/kg Dc maintained (Na⁺ + K⁺)-ATPase activity similar to the control animals in both cortex and medulla (Figure 3A and 3C). Dc 3 mg/kg administered to control rats did not affect the enzyme activity. However, 3 mg/kg Dc treatment either partially or did not recover Na⁺-ATPase activity (Figure 3B and 3D). This effect could be because Dc administered to control rats inhibits Na⁺-ATPase activity in both cortex and medulla (Figure 3B and 3D).

Figure 1. Impact of doxycycline (Dc) treatment on renal filtration function of rats subjected to I/R.

Rats were subjected to bilateral ischemia for 30 min followed by 24 h reperfusion (I/R) or to sham surgery (control). Two hours before the surgery, 3 different doses of doxycycline (1, 3, and 10 mg/kg) were administered to I/R rats (I/R + Dc1, I/R + Dc3, and I/R + Dc10, respectively). Doxycycline at 3 mg/kg was administered to sham-operated rats as an internal control. Creatinine and blood urea-nitrogen (BUN) were measured after 24 h of reperfusion. (A) Glomerular filtration rate (GFR) was calculated from the clearance of endogenous creatinine: Cl_cr = U_cr× V/P_cr, where V is the urinary volume (in ml/24 h) and U_cr and P_cr are the urinary and plasma creatinine concentrations, respectively (in mg/dl). (B) Blood urea-nitrogen (BUN) in mg/dl. Inset to (B) Plasma Creatinine (mg/dl). Values are means ± SEM. Different superscripted lower-case letters indicate statistically significant differences (P < 0.05; one-way ANOVA followed by Tukey’s post-test).

Figure 2. Effect of Dc treatment on Na⁺ handling of rats subjected to I/R.

Rats were subjected to bilateral ischemia for 30 min followed by 24 h reperfusion (I/R) or to sham surgery (control). Two hours before the surgery, 3 different doses of Dc (1, 3, and 10 mg/kg) were administered to the I/R rats (I/R + Dc1, I/R + Dc3, and I/R + Dc10, respectively). Doxycycline, at 3 mg/kg, was administered to sham-operated rats as an internal control. (A) Na⁺ filtered load (in μmol/min) = GFR (in μl/min) ×P_Na (Na⁺ plasma concentration in μM). (B) Fractional excretion of Na⁺ represents the percentage of the Na⁺ filtered by the kidney, which is excreted in the urine. FE_Na = UE_Na/FL_{Na ×} 100; where FE_Na is the fractional excretion of Na⁺(in %), FL_Na is the Na⁺ filtered load (in μmol/min) and UE_Na is the urinary Na⁺ excretion (in mmol in 24 h). UE_Na = U_{Na ×} V,U_Na is the urine concentration of Na⁺ (in mM) and V is the urinary volume (in ml/24 h). Values are means ± SEM. Different superscripted lower-case letters indicate statistically significant differences (P < 0.05; one-way ANOVA followed by Tukey’s post-test).

Figure 3. Influence of 3 mg/kg Dc treatment on the renal cortex primary Na⁺ transporters and medullary (Na⁺ + K⁺)ATPase activity.

Experimental groups were designed as in Methods. ATPase activity was measured in the homogenate obtained from kidney cortex (A–B) and medulla (C–D). (A, C) Ouabain-sensitive (Na⁺+ K⁺)ATPase activity and (B, D) Ouabain-resistant, furosemide sensitive Na⁺-ATPase. Values are expressed as percentage of control. Different superscripted lower-case letters indicate statistically significant differences (P < 0.05; one-way ANOVA followed by Tukey’s post-test). I/R: rats subjected to ischemia/reperfusion procedure; I/R + Dc3: I/R rats treated with 3 mg/kg Dc and control + Dc3: sham-operated rats treated with 3 mg/kg Dc.

3.2. Doxycycline treatment protects kidney from fibrotic injury

Figure 4 shows representative cortical and medullary images of PAS staining in all experimental groups. In kidney cortex of I/R rats (Figure 4D and 4D.I), we observed space between the tubules (black circles), congestive glomerular capillaries tufts (black triangles), detached necrotic tubular cells, granular casts with necrotic cell debris (black squares) and denuded basement membranes (stars). Moreover, the glomerulus diameters were decreased (Figure 4M). In the medulla (Figure 4E–F and 4 E.I–F.I), tubular dilatation and flattened tubular epithelium (black pentagon) was observed. This rat model of I/R promoted a mild tubular injury – the score for kidney cortex was 2 (Figure 4N) and in the outer medulla it was 2.5 (Figure 4O). The glomerular structure and the tubular lesions provoked by the I/R were prevented by 3 mg/kg Dc treatment (Figure 4G–I and 4 G.I–I.I). Administration of 3 mg/kg Dc to sham-operated rats did not alter kidney structure (Figure 4J–L and 4 J.I–L.I).

Figure 4. Effect of 3 mg/kg Dc treatment on IRI.

Representative photomicrographs of PAS staining of kidney cortex, outer medulla, and medulla. (A–C) control; (D–F) I/R; (G–I) I/R + Dc3 and (J–L) control + Dc3. The squares drawn in each image and labelled as (I) are shown in subsequent panels at higher magnification (×400). For example, (A.I) represents the square marked in the upper image (A). Black circles: space between the tubules, black triangles: congestive glomerular capillaries tufts, black squares: detached necrotic tubular cells and granular casts with necrotic cell debris, stars: denuded basement membrane and black pentagon: tubular dilatation and flattened tubular epithelium. (M) Glomerular diameter: sizes of the individual glomeruli located in the cortex were calculated as the average of glomerular diameters within a field of view. (N) Cortical tubular injury score and (O) Outer medulla tubular injury score. The score was evaluated by tubular dilation, tubular atrophy, vacuolization, the degeneration and sloughing of tubular epithelial cells, or thickening of the tubular basement membrane. The scoring system applied was 0 = no tubular injury; 1 = < 10% of tubules injured; 2 = 10%–25% of tubules injured; 3 = 26%–50% of tubules injured; 4 = 51% 75% of tubules injured; and 5 = > 75% of tubules injured.

Extracellular matrix collagen was detected with Pricrosirius red staining (Figure 5). I/R provoked a vast red staining in the cortex and principally in the medulla. Treatment with Dc 3 mg/kg blocked the interstitial collagen accumulation. Doxycycline itself did not change collagen accumulation.

Figure 5. Effect of 3 mg/kg Dc treatment on interstitial collagen accumulation induced by I/R.

Representative photomicrographs of Picrosirius red staining of kidney cortex, outer medulla, and medulla. (A–C) control; (D–F) I/R; (G–I) I/R + Dc3 and (J–L) control + Dc3.

The profile of restoration is displayed in Figures 6 and 7. IRI is marked by fibronectin deposition around glomeruli and tubules in the interstitial space in both cortex and outer medulla (Figure 6) and high TGFβ immunoexpression in the glomerulus and tubules of the outer medulla (Figure 7, bottom lane). Doxycycline treatment (3 mg/kg) prevented an increase in fibronectin deposition in the interstitial space and TGFβ tubular expression in the outer medulla (Figures 6 and 7). Under these conditions, TGFβ immunoexpression in glomeruli was completely abolished (Figure 7, upper lane).

Figure 6. Effect of 3 mg/kg Dc treatment on fibronectin deposition.

Immunofluorescence was performed as described in Methods. Paraffin embedded kidney poles from control, I/R, and I/R + Dc3 rats were incubated overnight with anti-rabbit fibronectin antibody (1:250) at 4°C, followed incubation with specific secondary antibody (1:1,000, AlexaFluor 488-labeled secondary antibody). Representative photomicrographs of kidney poles in glomeruli (×400) and tubulointerstitial areas (×400). Arrows demonstrates fibronectin deposition.

3.3. Doxycycline treatment prevents increase in total MMPs activities and restores MMP-2 and -9 protein contents

The well-known non-antibiotic mechanism of action of Dc is the inhibition of MMPs by chelating the co-factor Zn²⁺ at the catalytic site [22]. To determine whether the inhibition of MMPs is related to the beneficial effect of Dc on kidney function, we evaluated MMP activity. Figure 8A shows that I/R rats presented significantly increased (~60%) total MMP activity compared with control. In addition, the treatment with 3 mg/kg Dc restored the total MMP activity level in the kidney cortex. The addition of Dc (25 and 50 μM) and classical unspecific MMP inhibitors (1,10-phenanthroline, EDTA, and EGTA) to kidney cortex homogenates blocked the MMP activity independent of the rat condition (Figure 8B). The protease inhibitors, non-MMP (E-64, PMSF, and PepA) did not change the fluorescence levels observed in the condition without inhibitors, confirming the specificity of the phenomena (Figure 8B). Finally, MMP-2 and -9 (the major kidney MMPs) protein contents and its activities are presented in Figure 9. Unexpectedly, MMP-2 and -9 activities did not change under any condition (Figure 9A–C). We observed that I/R provoked augmentation of both MMP-2 and -9 protein content. Dc (3 mg/Kg) treatment restored the protein contents to control levels (Figure 9D–E).

Figure 8. Effect of 3 mg/kg Dc treatment on MMPs activity.

(A) Kinetics of the MMP activity. The cortex kidney extract (30 μg protein) was incubated with the fluorogenic substrate for 60 min at 37 °C, and the peptidase activity was assessed by measuring the hydrolysis of the peptide fluorogenic substrate Z-Arg-Arg-AMC. The inset represents the endpoint after 60 min of hydrolysis. The asterisks highlight the significant differences between I/R or I/R+Dc3 and control (P < 0.05; one-way ANOVA followed by Tukey’s post-test; n = 5 rats/group). (B) Alternatively, the kidney cortex extracts were incubated with the fluorogenic substrate for 60 min at 37 °C in the absence or presence of 1 mM 1,10-phenanthroline (Phen), 1 mM EGTA, 1 mM EDTA, 25 μM doxycycline (Dc 25), 50 μM doxycycline (Dc 50), 10 μM E-64, 5 mM PMSF, and 5 μM pepstatin A (PepA). The graphic represents the endpoint after 60 min of hydrolysis. The asterisks highlight the significant differences between the respectively control and its correspondent peptidase-treated (P < 0.05, t-test). The results were expressed as arbitrary fluorescence units (AFU). The values represent the mean ± standard deviation of three independent experiments performed in triplicate.

Figure 9. Effect of 3 mg/kg Dc treatment on MMPs in kidney cortex.

(A) Representative images of a MMP activities by zymography. The first panel is the loading control (A.1) followed by the zymography developed in the absence (A.2) and presence of different MMPs inhibitors: phenanthroline (A.3), EGTA (A.4), and EDTA (A.5). The bands corresponding to MMP-9 (~180 kDa) and MMP-2 (~72 kDa) were quantified in (B) and (C), respectively. MMP-9 (D) and MMP-2 (E) protein content in the kidney cortex from control, I/R and I/R+Dc3 rats. Upper panels: representative image of the immunoblots. Lower panels: densitometric measurement of the immunobands. β-actin was used as the loading control. Values are expressed as arbitrary units (a.u.). Different superscripted lower-case letters indicate statistically significant differences (P < 0.05; one-way ANOVA followed by Tukey’s post-test; n = 5 rats/group). I/R: rats subjected to ischemia/reperfusion procedure; I/R + Dc3: I/R rats treated with 3 mg/kg doxycycline.

4. Discussion

We provide evidence that intraperitoneal administration of low-dose Dc (3 mg/kg) protects kidney function (glomerular filtration and epithelial Na⁺ transport) from IRI. To our knowledge, there are no previous studies comparing different low doses of Dc on the renal function impairment provoked by I/R, although its beneficial effects on kidney structure and signaling pathways associated with kidney injury have been reported [14, 26–30]. In our study, the usual Dc dose was 10 mg/kg, administered either orally or by intraperitoneal injection. We have demonstrated that treatment with 3 mg/kg Dc attenuates total MMP activities, restores MMP-2 and -9 protein contents, fibronectin deposition, and tubular TGF-β expression. All of these changes appear to be associated with preservation of GFR and tubular Na⁺ transport. Notably, Dc 3 mg/kg, the dose with beneficial effects in renal function in I/R rats, did not affect kidney function and structure when administered to sham-operated rats.

The transient ischemic episode (30 min) followed by 24 h reperfusion induced polyuria and proteinuria, reduced GFR, increased urea-nitrogen accumulation in the blood, and diminished FE_Na. Moreover, the alteration in kidney morphology (score of 2 in the cortex and 2.5 in the medulla) corresponded to a mild tubular injury which is similar to that found in a previous rat model of AKI [41–46]. This disrupted renal function in an early stage of IRI can be related to rapid alterations in the renal cortex structures, interstitial fibronectin accumulation, and augmented tubular TGFβ1 immunoexpression. Although fibrosis has been generally accepted as a marker of chronic kidney injury, it was also demonstrated in IRI. Yang et al., [47] linked fibrosis to cell cycle arrest in five different models of acute tubular injury. The authors elegantly demonstrated that the development of fibrosis in each AKI model correlated with the arrest of proximal tubule cells cycle in G2/M. Injured tubular cells drive damage and inflammation by releasing profibrotic factors, in particular TGF-β, identifying proximal tubule cells as a major player in injury and fibrosis [48]. In the kidney cortex, we observed interstitial fibronectin accumulation and increase in tubular TGF-β1immunoexpression. Treatment with 3 mg/Kg Dc hinders these events, thus we propose that Dc may impede acute kidney injury from progressing to chronic kidney disease.

The immunoexpression of TGF-β in the glomeruli of the control rat could be related to the regulation of the extracellular matrix mediated by mesangial cells, since these cells express and respond to TGF-β1 signaling [49,50]. The same profile was observed in patients with established structural kidney injury that presented a significant increase in serum creatinine and decreased GFR [51].

Although all the three Dc doses prevented the excretion of diluted urine enriched by proteins, 1 mg/kg Dc did not prevent the reduction of GFR nor BUN accumulation. Instead, because I/R + Dc1 rats presented lower GFR values compared with I/R, the FE_Na (the percentage of the Na⁺ filtered by the kidney that is excreted in the urine) was enhanced, showing the inability of these rats to conserve Na⁺. It seems that 1 mg/kg Dc is effective for tubular damage (evaluated by urine volume and proteinuria) but not completely effective to modify glomerular filtration rate (GFR is similar to I/R and BUN is partially prevented); thus, invalidating the continued use of this dose in the study. The total beneficial effect of Dc was observed at the dose of 3 and 10 mg/kg, protecting GFR, and reducing BUN and FE_Na.

Guimarães et al., [52] demonstrated a dose-dependent effect of Dc in a renovascular hypertension model. Doxycycline treatment, at doses lower than 10 mg/kg/day, attenuated hypertension but not vascular alterations found in the 2K1C hypertensive rat. However, 30 mg/kg/day Dc treatment prevented vascular alterations in addition to attenuating hypertension. The authors correlated the reduced efficacy of the lower Dc doses to the lack of MMP-2 inhibition. Most previous kidney studies used 10 mg/kg (oral or intraperitoneal) to prevented kidney injury by attenuating reactive oxygen species (ROS), MMPs, TGF-β1, and nephritis [14, 26–30]. We used intraperitoneal Dc administration at 3 different low doses (1, 3, and 10 mg/kg) before inducing IRI. We observed that both 3 and 10 mg/kg prevented the kidney function loss, suggesting the effectiveness of Dc in the kidney compared with the vasculature. Indeed, it has been shown that the levels of Dc measured in renal tissue averaged twice the concentration found in serum [53]. According to Prall et al. [54], 10, 50, and 100 mg/kg Dc administered to mice produce a dose-dependent increase in Dc serum concentration of 2.7, 5.3, and 23.2 μM, respectively. Assuming that bioavailability is similar to rats, in our study serum Dc concentration should be equal to or lower than 2.7 μM and around 5.4 μM in the kidney. A concentration higher than 4 μM has been shown to be effective in reducing MMP activity [54]. This may partly explain the low dose effects of Dc in the kidney but not in the vasculature. The observation that low Dc doses inhibit MMP content and activity in the kidney, but not in the vasculature, reinforce this explanation.

The future use of Dc treatment as a rescue therapy is unsettled. Antonio et al., [55] showed that although Dc promoted MMP inhibition associated with antioxidant and antihypertensive effects, there was no reversal of the vascular remodeling induced by hypertension. However, Kholmukhamedov et al. [56] demonstrated that minocycline or Dc were similarly protective in attenuating kidney injury when given before or after resuscitated hemorrhage. These divergent effects could be related to the ability of the kidney to accumulate Dc.

In the present rat model of IRI, total MMPs activities were highly augmented as shown by the in vitro analysis. Cortical kidney extract of I/R rats incubated with both 25 and 50 μM Dc blocked augmentation of total MMPs activity on its specific fluorogenic substrate. Our observation differs from in vitro studies by Guimarães et al., 2011 [52] that showed Dc concentrations below 50 μM did not inhibit human recombinant MMP-2 activity, suggesting that rat MMP was more susceptible to Dc treatment. In vivo, the dual effect of MMP, whether protecting or promoting kidney injury, depends upon the temporal MMP pattern expression/activity. Among all MMPs, MMP-2 and -9 are emerging as key MMPs in kidney diseases. Kaneko et al. [17] showed that MMP-2 and -9 expression/activity are upregulated in 24 h of reperfusion. MMPs levels are even higher up to 14 days. MMP-2 is related with tubular repair phase after IRI and as an anti-fibrotic agent, regulating the severity of renal fibrosis [18, 19]. We observed that in I/R rats, there was an increase in MMP-2 and -9 protein content compared with control or I/R+Dc3, yet their activities did not change, suggesting that the augmented protein content sustained the enzyme activity among all three experimental groups. Indeed, the MMP-2 or -9 protein activity/protein content was reduced by 50 % and 30 %, respectively. Considering that MMP-2 is more related with protection and MMP-9 with tissue damage, it may be that the loss of the protective component (MMP-2) contributes to kidney injury in I/R. Moreover, it has been shown that Dc decreases MMPs mRNA stability, thereby explaining how MMPs decreases its protein content to control levels [57, 58]. We cannot rule out the involvement of other MMP isoforms since we showed that I/R generally augmented MMPs activities. Therefore, we propose that during I/R, an imbalance exists between the MMPs isoforms that culminates in an increase of total MMP activity. Doxycycline prevents this imbalance, maintains control levels of MMP activity, including MMP-2 and -9 activities and MMP-2 and -9 protein contents.

Doxycycline action on other MMPs and on TIMPs expression in recovering MMPs levels cannot be ruled out. TIMP-2 has been validated as a risk marker of AKI, predicting moderate to severe damage in patients [59] and augmented TIMP-1 levels promoted by Dc presented a protective effect on IRI by inhibiting the activity of MMP-2 in mice [29]. The preventive effect of Dc was accompanied by a bleached immunodetection of fibronectin in the interstitium and tubular TGFβ expression.

Our data did not reproduce MMPs inhibition using a daily dose of BB-94 (Batimast, a synthetic MMP inhibitor) 2 days before the I/R procedure. In this model, plasma creatinine levels of rats subjected to I/R levels remained high even when treated with this drug [60]. Contrary to what we observed, the augmentation of plasma creatinine was independent of MMPs expression and activity, since MMPs were unaltered. Considering these observations, the existence of alternative mechanisms mediated by Dc independent of MMP activation cannot be ruled out. Indeed, Dc prevented inflammation and ROS activation in the kidney [14, 26–30].

We focused in MMP changes in I/R and Dc treatment in the kidney cortex because: (1) the hallmark of acute renal failure is a decreased glomerular filtration rate due to vasoconstriction of afferent arterioles to glomeruli [61]; (2) MMP-2 and -9 are involved in endothelial cell dysfunction in ischemic acute renal failure [62]; (3) evidence in vivo has demonstrated that ischemic acute renal failure provokes a rapid disruption of proximal tubule function associated to cytoskeleton disruption [63]; and (4) the extent of proximal versus distal tubule injury is controversy in humans [62] and the detection of biomarkers of proximal tubule injury in the urine following ischemia have revealed significant damage in human proximal tubules [63, 64].

The proximal tubule is responsible for 70% of Na⁺ reabsorption from the ultrafiltrate; thus, any change in Na⁺ reabsorption in proximal tubule cells could impact the final urinary composition and volume even with compensatory reabsorption of Na⁺ in distal segments of the nephron. Transepithelial proximal tubule Na⁺ reabsorption is mainly driven by two sodium pumps: ouabain-sensitive (Na⁺ + K⁺)ATPase activity and the ouabain resistant, furosemide-sensitive Na⁺-ATPase. Located in basolateral membranes, these enzymes are involved in the genesis of the Na⁺ electrochemical gradient driving not only Na⁺ entry in the luminal membrane, but also unidirectional transport of solute across the epithelium [65, 66]. Na⁺-ATPase is also an efficient mechanism for Na⁺ extrusion without interference in intracellular K⁺ homeostasis to maintain membrane potential [67].

In the I/R rat, cortical (Na⁺ + K⁺)ATPase activity increased, which correlated with the observed decrease in FE_Na. This observation concurs with early stage AKI, diminished GFR associated with higher tubular reabsorption leading to acute tubular necrosis [40]. (Na⁺ + K⁺)ATPase is responsible for the massive Na⁺ reabsorption (high capacity and low affinity for Na⁺) across the epithelia as Na⁺-ATPase allows fine adjustment (low capacity and high Na⁺ affinity), thus acting together to maintain water and Na⁺ body homeostasis [65–68]. Altered regulation exclusively of the Na⁺-ATPase appears to underpin renovascular and cardiac dysfunctions, revealing the importance of this pump in the genesis of life-threatening diseases [69]. In the cortex, Na⁺-ATPase activity was diminished opposing the (Na⁺ + K⁺)ATPase activity. We speculate that the opposite reaction to IRI of Na⁺-ATPase in the kidney cortex is an unsuccessful attempted to compensate for Na⁺ loss by decreased (Na⁺ + K⁺)ATPase activity. In contrast, both Na⁺ transporters presented decreased activities in the medulla, clearly demonstrating the inability of the kidney to concentrate the urine. Doxycycline treatment (3 mg/kg) prevented the alteration of ATPase activity, although a few isolated areas of epithelium necrosis are still detected in kidney histology.

The correlation between (Na⁺ + K⁺)ATPase activity and MMPs has been described in the heart [70]. After a prolonged period of ischemia, cardiac tissue shows augmented MMP-2 activity, diminished (Na⁺ + K⁺)ATPase activity, and corresponding protein content changes. It has been shown that Dc protects against I/R-induced cardiac disease [71]. Our study did not evaluate the association of MMPs, (Na⁺ + K⁺)ATPase activity and IRI during prolonged periods of I/R.

Because Dc is an antibiotic, we tried to minimize gut microbiota disruption by using the lowest doses of Dc, doses below minimal inhibitory concentration (MIC) which caused little or no antimicrobial effects. Intraperitoneal administration also avoided local action in the intestine during absorption. Indeed, there was no alteration in kidney function of control rats treated with Dc compared with control. On the other hand, the influence of gut microbiota in IRI outcomes has been shown to protect the kidney from deteriorating structural and functional renal injury [71].

In addition to the issues of bacterial resistance and influence on gut microbiota, the use of the tetracycline antibiotic family, such as Dc and minocycline, in prevention of IRI is promising because it is well-tolerated, at least for 6 months by 92 % of patients [72]. The use of synthetic peptides that bind to Zn²⁺ at the active site of MMP, thus promoting enzyme inhibition, is clinically limited. Batismat (which is not available in an oral formulation) showed promise in decreasing tumor development, metastasis, and limiting aneurism expansion [72], but it did not inhibit MMP activity in the kidney nor renal damage provoked by I/R. Marismat (oral administration) caused significant musculoskeletal side effects in 30% of patients [72]. Recent studies using MMI270, a more specific inhibitor (MMP-2, -8, and -9), have shown similar side effects [72] including both plaque growth and plaque stability [72]. Figure 10 provides an illustration summarizing the mechanism that may be involved the protective effects of Dc on renal function.

Figure 10. Illustration of the proposed mechanism in the protective outcomes of low-dose Dc (3 mg/kg) on renal function of Wistar rats subjected to acute I/R injury.

After 24 h of I/R procedure, Wistar rats showed decreased glomerular filtration and impaired tubular function leading to diluted urine enriched in proteins, and creatinine and urea accumulation in the blood. These events were associated with altered structure of the glomerulus and tubules, interstitial accumulation of collagen and fibronectin, augmented tubular TGFβ expression, and a disrupted tubular Na⁺ transport. In the cortex, total MMP activity was enhanced, as also protein content of MMP-2 and MMP-9 were augmented. Arrows represent the proposed mechanism of action of low dose Dc in the kidney, administered i.p. 2 h prior to I/R procedure.

5. Conclusion

Our study demonstrates the non-classical mechanism of low dose Dc beyond its antimicrobial effects. Our findings shed light on the beneficial effects of low dose Dc in the preservation of IRI. The renal physiology alterations triggered by IRI are attenuated by Dc, which is correlated with the preservation of renal filtration and tubular Na⁺ reabsorption, thus opening new vistas to the pharmacological treatment of AKI.

Highlights.

• The doxycycline impact in ischemic-reperfusion kidney function is evaluated.

• Doxycycline preserves glomerular filtration and impedes proteinuria.

• Doxycycline restores (Na⁺ + K⁺) ATPase activity and fractional Na⁺ excretion.

• A correlation of kidney structure and function by doxycycline is proposed.