Noninvasive Assessment of Fibrosis Following Ischemia/Reperfusion Injury in Rodents Utilizing Na Magnetic Resonance Imaging

Fibrosis is often heterogeneously distributed, and classical biopsies do not reflect this. Noninvasive methods for renal fibrosis have been developed to follow chronic kidney diseases (CKD) and to monitor anti-fibrotic therapy. In this study, we combined two approaches to assess fibrosis regression following renal ischemia-reperfusion injury (IRI): magnetic resonance imaging (MRI) and noninvasive extracellular matrix (ECM) biomarkers. MRI was used to evaluate fibrosis in bilateral IRI in rats after reperfusion at 7, 14, and 21 days. This was performed with 1HT1 and T2* mapping, dynamic contrast-enhanced (DCE)-MRI, and chemical shift imaging (CSI)-23Na. The degradation of laminin gamma-1 chain (LG1M) and type III collagen (C3M) was measured in urine and plasma. Fibrosis was analyzed in tissue using fibronectin (FN) and alpha-smooth muscle actin (α-SMA) using quantitative polymerase chain reaction qPCR and western blotting. We found increased fibrosis 7 days after reperfusion, which dropped to sham levels after 21 days. Single kidney glomerular filtration rate (skGFR), perfusion (DCE-MRI), and total 23Na kidney content correlated positively with fibrotic markers FN and α-SMA as well as noninvasive LG1M and C3M. We showed that novel MRI protocols and ECM markers could track fibrogenic development. This could give rise to a multi-parametric practice to diagnose and assess fibrosis whilst treating kidney disease without using invasive methods.


Introduction
Chronic kidney diseases (CKD) lead to renal fibrosis and ultimately kidney failure. CKD is a leading cause of mortality, and its prevalence is steadily increasing globally, currently affecting about 10% of the population [1]. Renal fibrosis is a typical observed outcome in which the main characteristic is the increase of myofibroblasts activity in the renal interstitium producing extracellular matrix (ECM) deposition and accumulation [2].
Renal fibrosis assessment typically requires an invasive kidney biopsy causing several risks, e.g., bleeding. Since fibrosis is often heterogeneously distributed, a biopsy analysis is therefore inherently subjected to sampling bias [3]. Thus, there is a need for noninvasive methods to assess renal Pharmaceutics 2020, 12, 775 2 of 16 fibrosis in order to follow the disease development of CKD and to monitor anti-fibrotic therapy. In this study, we investigate two noninvasive approaches to evaluate renal fibrosis regression following ischemia/reperfusion injury, namely magnetic resonance imaging (MRI) and noninvasive ECM biomarkers.
Noninvasive methods such as diffusion-weighted (DW)-magnetic resonance imaging (MRI), blood oxygen level-dependent (BOLD)-MRI, T 1 mapping, and renal MRI elastography have been proposed as novel methods to provide insight into renal function and fibrosis severity [4][5][6]. Accumulation of ECM components is believed to cause reduced water movement, and DW-MRI allows detection of Brownian motion of water; i.e., a higher DW-MRI signal may reflect the extent of fibrosis [7,8]. Complementarily, T 1 mapping is a parametric plot where each pixel of a kidney image represents the 1 H T 1 spin-lattice of longitudinal relaxation time. 1 H T 1 relaxation time is mainly dependent on the amount of water molecules in the tissue. During renal fibrosis, a relaxation shift is likely to occur, which may be detected by T 1 mapping [9]. Concomitantly, CKD patients have been shown to suffer from chronic hypoxia in the kidneys [10] and the inverse relationship with blood oxygenation assessed with the BOLD imaging may detect fibrotic lesions [11,12]. Magnetic elastography is an MR technique which in its essence enables the user to measure tissue "stiffness". It is believed that kidney fibrosis elevates stiffness, leading to a higher signal [13].
Here, we aim to combine several MRI approaches, including not only the well-known sequences 1 H T 1 mapping and BOLD MRI but also dynamic contrast enhanced (DCE) MRI and 23 Na chemical shift imaging (CSI) to assess renal function and the progression of renal fibrosis in an ischemia-reperfusion injury (IRI) disease model and to compare these to the results from a unilateral ureteral obstruction (UUO) model. A dynamic contrast enhanced (DCE)-MRI sequence allows to calculate perfusion and single-kidney glomerular filtration rate (skGFR) [14]. Currently, estimated GFR (eGFR) is the most important parameter used to diagnose CKD [15]. For this reason, the ability to assess skGFR is thought to give relevant information on kidney function and the related fibrotic state. Moreover, combining 23 Na chemical shift imaging (CSI)-MRI might be an extra valuable tool to follow renal fibrosis [16]. A recent study has provided evidence that 23 Na-MRI might be a promising tool to assess skin fibrosis by quantifying Na + in the skin [16]. In this scenario, we utilize a similar approach to evaluate the signal intensity of 23 Na in the whole kidney. In addition, we will evaluate 23 Na signal intensity throughout segments in the kidney, allowing us to calculate the 23 Na signal slope reflecting the osmotic gradient in the kidney. The use of 23 Na mapping has not previously been evaluated for fibrosis quantification in the kidney, and this will, to our knowledge, be the first study to investigate this potential.
To complement our MRI data and to improve the understanding of pathophysiological alterations in tissue turnover following ischemia-reperfusion injury, we also measured novel noninvasive ECM markers in plasma and urine. Kidney fibrosis changes include type III collagen (COL III) accumulation in the tubulointerstitial matrix [17,18]. During disease progression, the altered microenvironment in the glomerular basement membrane (GBM) leads to an increased degradation of one of its most important components, namely laminin-11, which consists of the alpha-5, beta-2, and gamma-1 chain (LAMC1). The markers reflecting turnover of COL III (C3M) and LAMC1 (LG1M) measured in our study are both generated by matrix metalloproteinase-9 (MMP-9), a protease known to have a temporal and spatial expression during renal disease progression [19]. COL III degradation was assessed using the C3M enzyme-linked immunosorbent assay (ELISA), and LAMC1 degradation was assessed using the LG1M ELISA. The C3M biomarker has been previously evaluated in CKD animal models, where it correlated with the extent of renal fibrosis [17,20], while the LG1M, a LAMC1 biomarker, is still at early investigational stages [21].
Here, we compared different and innovative MRI techniques with C3M and LG1M measurements to improve the understanding of pathophysiological alterations in tissue turnover during kidney fibrosis development.

Animal Models
All experiments were performed on male Wistar rats divided into the following groups: sham n = 6; 240-290 g; 7 days IRI n = 6; 210-240 g; 14 days IRI n = 4; 235-260 g; 21 days n = 5; 300-330 g; and unilateral ureteral obstruction (UUO) n = 6; 200-230 g. These animals have been used in our previous paper published by Rasmussen et al. [22]. Some data was intentionally reused, and some are presented as originals, mainly the LG1M and MRI data. Two animals from two weeks IRI and one animal from three weeks IRI were euthanized prematurely. Animals had ad libitum access to standard rodent diet (Altromin, Lage, Germany) and water, and the 12 h:12 h light-dark cycle as well as temperature (21 ± 2 • C) and humidity (55 ± 5%) were controlled. The studies were carried out in accordance with the Danish National Guidelines for animal care and were approved by the Danish Animal Experiments Inspectorate under the Danish Veterinary and Food Administration (license no. 2015-15-0201-00658 for UUO animals and license no. 2013-15-2934-00810 for IRI animals).
The animals were anesthetized with sevoflurane (induction 6%, sustained 2.5%; Abbot Scandinavia, Solna, Sweden) mixed with air (2 L/min). Soon before the surgery began, Temgesic ® (Reckitt Berkshire, Slough, UK) was given subcutaneously as analgesics (0.05 mg/kg) and was then supplied in the drinking water (0.3 mg/mL) for 3 days post-surgery. During surgery, the animals were placed on a heating pad (CMA 450 temperature controller, Harvard apparatus, Cambourne, United Kingdom) to maintain a rectal temperature of 37 • C, while the respiration was visually monitored.
IRI induction: the abdomen was shaved and cleaned with ethanol, a surgical incision was made, and both the left and right renal arteries were carefully dissected. A nontraumatic clamp was placed on both arteries for 45 min to induce ischemia, where after the clamps were released, restoring blood flow and reperfusion visually were confirmed. To maintain postoperative water balance, 2 mL of isotonic saltwater was injected subcutaneously at the beginning of the IRI operation.
UUO induction: the abdomen was shaved and cleaned with ethanol, a surgical incision was made, and the left ureter was exposed and occluded with a 3-0 silk ligature. The incision was sutured separately in muscle tissue and skin. Sham animals underwent the same procedures, meaning dissection of either the renal artery or the left ureter but with no occlusion in both sham surgical cases.
After MR scanning, the IRI animals were euthanized. The UUO animals were scanned at both day 5 and day 7 and then euthanized. Arterial blood and urine were collected to estimate plasma creatinine and blood urea nitrogen (BUN). Creatinine and urea in plasma were measured using a Roche Cobas 6000 analyzer (Roche Diagnostics, Hvidovre, Denmark).

Fumarase Activity Assay
Fumarase activity was measured in plasma according to the manufacturer's instructions (Sigma Aldrich, Brøndby, Denmark) with minor alterations, i.e., the analysis was performed in a 384-well costar plate using a PHERAstar FS microplate reader (BMG Labtech, Birkerød, Denmark). Absorbance readout was performed at the highest peak change (654 nm). Plasma was distributed without pretreatment.

Western Blotting (WB)
Renal cortex proteins were lysated in a dissecting buffer (0.3 M sucrose, 25 mM imidazole, and 1 mM (EDTA), pH 7.2) including protease inhibitors Complete Mini Protease Inhibitor Cocktail Tablets (Roche Diagnostics, Hvidovre, Denmark) and Phosphatase Inhibitor Cocktail 2 and 3 (Sigma-Aldrich, Brøndby, Denmark) using a tissue homogenizer (Qiagen, Hilden, Germany) followed by centrifugation. The total protein concentration was determined using a Pierce BCA protein assay kit (Roche Diagnostics, Hvidovre, Denmark) following the manufacturer's instructions.
Proteins were separated on a 12% Criterion TGX Precast Gel (Bio-Rad Laboratories, Copenhagen, Denmark) and transferred to a Hybond ECL nitrocellulose membrane (GE Healthcare, Hatfield, UK). The membrane was then blocked in 5% non-fat dry milk in PBS-T (80 mM Na 2 HPO 4 , 20 mM NaH 2 PO 4 , Pharmaceutics 2020, 12, 775 4 of 16 100 mM NaCl, and 0.1 Tween 20, pH 7.4), washed in PBS-T, and incubated with primary antibodies overnight at 4 • C. Subsequently, the membrane was again washed and incubated with HRP-conjugated secondary antibody for one hour at room temperature. Antigen-antibody reactions were visualized using a chemiluminescence system (Amersham ECL Plus, GE Healthcare). All western blots were normalized to total protein content measured with Stain-Free technology [23]. Primary and secondary antibodies are listed in Table 1. A 2-mm kidney section was dissected from the IRI and sham kidneys when the rats were sacrificed. The sections were fixed in 4% paraformaldehyde for 2 hours and washed 3 times (10 min) with 0.01 M PBS buffer. The fixed kidney sections were then dehydrated, embedded in paraffin, and cut into 2-µM sections on a rotary microtome (Leica Microsystems, Herlev, Denmark). To assess the renal morphology, paraffin embedded kidney sections were stained with Hematoxylin and Eosin (HE). Renal fibrosis was highlighted by Masson's trichrome, in which collagenous fibers were stained blue and the cytoplasm was stained red. The interstitial fibrosis degree was assessed by a renal pathologist examining the sections in a blinded manor and scoring the sections with either no fibrosis, mild fibrosis (1-5%), incipient fibrosis (6-10%), profound fibrosis (11-25%), severe fibrosis (26-50%), or advanced fibrosis (51-100%).

RNA Extraction and Quantitative PCR (qPCR)
Total RNA was extracted from renal cortex using NucleoSpin RNA II mini kit according to the manufacturer's instructions (AH diagnostics, Aarhus, Denmark). Quantification of total RNA was performed by spectrophotometry and stored at −80 • C. cDNA synthesis was performed with RevertAid First strand cDNA synthesis kit following the manufacturer's protocol (MBI Fermentas, Burlington, ON, Canada). qPCR was carried out using Maxima SYBR Green qPCR Master Mix according to the manufacturer's instructions (AH diagnostics, Aarhus, Denmark) on an AriaMx QPCR reader (Agilent, Santa Clara, CA, USA). Briefly, 100 ng cDNA was utilized as template for PCR amplification. Specificity of products was confirmed by melting curve analysis and by gel electrophoresis. Primer sequences used are given in Table 2.

MRI Analysis
MR images were converted to DICOM. Regions of interest (ROI) were drawn and analyzed in the commercial software OsiriX (Pixmeo SARL, Bernex, Switzerland). The mean 1 H T 1 , 1 H T 2 *, or 23 Na T 2 * was calculated as the mean of all pixels included in the ROI. SkGFR calculated from DCE images was processed and analyzed in the commercial software Matlab (Mathworks, Natick, MA, USA). SkGFR was calculated using the Baumann-Rudin model. Perfusion changes from DCE wer calculated using the plugin UMMperfusion for Osirix, utilizing a fast deconvolution model. From the CSI images, a map of 23 Na signal intensity was reconstructed. An ROI was drawn around the whole kidney and was normalized to a back-muscle ROI, giving rise to the total 23 Na signal intensity kidney value. In order to reduce the user bias in the ROI placement, the novel twelve-layer concentric objects method (TLCO) [24] was adjusted to rodent MRI examinations, allowing correct renal assessment of the intrarenal modifications for all MRI examinations. In short, six equidistantly separated segment layers (12 layers in the original human version) were calculated based on a user-selected whole kidney ROI perimeter, which enabled us to calculate the 23 Na slope through the segments, reflecting the osmotic kidney gradient.

Noninvasive ECM Markers
Levels of the noninvasive degradation markers of COL III (C3M) and LAMC1 (LG1M) in serum and urine from the animals were measured using competitive enzyme-linked immunosorbent assays (ELISA) developed by Nordic Bioscience, Denmark. Both of the monoclonal antibodies employed in the C3M and LG1M ELISA specifically detect neoepitopes of 10 amino acids in COL III (610 .KNGETGPQGP 619) and LAMC1 (1232 LNRKYEQAKN. 1241), both generated by MMP-9. Intra-and inter-assay variations of the assays were <10% and <15%, respectively. The tests were carried out as follows. Briefly, a streptavidin-coated 96-well ELISA plate (cat. 11940279, Roche) was coated with a biotinylated peptide. The plate was washed in washing buffer five times and afterwards incubated with standard peptide or sample together with HRP-conjugated monoclonal antibody. Subsequently, the plate was again washed five times, followed by incubation of TMB (Kem-En-Tec, Taastrup, Denmark) in the dark. To end the reaction, a 1% sulfuric acid (H 2 SO 4 ) solution was added and the plate was analyzed on the ELISA reader at 450 nm with 650 nm as the reference. All incubation steps were performed with constant shaking at 300 rpm. Concentrations were corrected for dilution factor of the samples. To normalize for urine output, biomarker levels in urine were divided by urinary creatinine levels.

Statistics
All data are presented as mean ± SEM. Normality was assessed with quantile plots. Differences were analyzed using a one-way analysis of variance (ANOVA) with a multiple comparison test comparing all IRI and UUO groups with sham and using a Bonferroni correction. Associations between variables were assessed by Pearson's correlations analysis including both sham and IRI animals. Statistical analysis was performed in GraphPad PRISM 6 (GraphPad Software, San Diego, CA, USA) when comparing the different treatment groups. All p-values below 0.05 were considered statistically significant.

Bilateral Ischemia Reperfusion Injury (IRI) and UUO Induced Kidney Injury
Rats were exposed to 45 min of bilateral ischemia followed by 7, 14, and 21 days of reperfusion. Similarly, a group of animals was exposed to UUO for 7 days. Bilateral renal IRI-induced plasma creatinine and blood urea nitrogen (BUN) increase peaked at 7 days after reperfusion and then normalized at 21 days ( Figure 1A,B). The UUO rats indicated a tendency towards elevated values of plasma creatinine and BUN, although it did not reach significance. Fumarase activity, an indicator of necrosis development [25,26] was found to be elevated in the 7 days postischemic group and this elevation declined over time and normalized to sham levels in the 21 days postischemic group. No elevation of fumarase activity was found in the UUO group ( Figure 1C). To evaluate kidney injury, we measured kidney injury marker 1 (KIM-1) and neutrophil gelatinase-associated lipocalin (NGAL). KIM-1 mRNA expression was increased after 7 days bilateral IRI and in rats subjected to UUO, whereas no significant change was observed at day 14 and day 21 after IRI ( Figure 1D). NGAL mRNA expression was significantly increased in the UUO rats and elevated in the 7 days and 14 days postischemic kidney, but it did not reach statistical significance ( Figure 1E). elevation of fumarase activity was found in the UUO group ( Figure 1C). To evaluate kidney injury, we measured kidney injury marker 1 (KIM-1) and neutrophil gelatinase-associated lipocalin (NGAL). KIM-1 mRNA expression was increased after 7 days bilateral IRI and in rats subjected to UUO, whereas no significant change was observed at day 14 and day 21 after IRI ( Figure 1D). NGAL mRNA expression was significantly increased in the UUO rats and elevated in the 7 days and 14 days postischemic kidney, but it did not reach statistical significance ( Figure 1E).

Bilateral IRI and UUO Induced Renal Fibrosis
In order to evaluate progression of fibrosis after IRI and UUO, we measured the fibrosis markers alpha-smooth muscle actin (α-SMA) and fibronectin (FN) in kidney cortex. α-SMA and FN mRNA expression as well as protein levels were significantly increased 7 days after reperfusion in rats subjected to bilateral IRI and UUO (Figure 2A-D). In addition, we also observed a significant increase in FN mRNA expression after 14 days IRI ( Figure 2B). No change in both mRNA and protein levels of α-SMA and FN were observed after 21 days of reperfusion. Representative immunofluorescence staining of α-SMA and FN immunoperoxidase staining confirmed our mRNA and protein data, showing stronger staining intensity at 7 and 14 days after IRI ( Figure 2F,G,J,K respectively) compared to the sham group ( Figure 2E,I) and the IRI groups with a longer reperfusion time ( Figure 2G,H,K,L).

Bilateral IRI and UUO Induced Renal Fibrosis
In order to evaluate progression of fibrosis after IRI and UUO, we measured the fibrosis markers alpha-smooth muscle actin (α-SMA) and fibronectin (FN) in kidney cortex. α-SMA and FN mRNA expression as well as protein levels were significantly increased 7 days after reperfusion in rats subjected to bilateral IRI and UUO (Figure 2A-D). In addition, we also observed a significant increase in FN mRNA expression after 14 days IRI ( Figure 2B). No change in both mRNA and protein levels of α-SMA and FN were observed after 21 days of reperfusion. Representative immunofluorescence staining of α-SMA and FN immunoperoxidase staining confirmed our mRNA and protein data, showing stronger staining intensity at 7 and 14 days after IRI ( Figure 2F,G,J,K respectively) compared to the sham group ( Figure 2E,I) and the IRI groups with a longer reperfusion time ( Figure 2G,H,K,L). subjected to bilateral IRI and UUO (Figure 2A-D). In addition, we also observed a significant increase in FN mRNA expression after 14 days IRI ( Figure 2B). No change in both mRNA and protein levels of α-SMA and FN were observed after 21 days of reperfusion. Representative immunofluorescence staining of α-SMA and FN immunoperoxidase staining confirmed our mRNA and protein data, showing stronger staining intensity at 7 and 14 days after IRI ( Figure 2F,G,J,K respectively) compared to the sham group ( Figure 2E,I) and the IRI groups with a longer reperfusion time ( Figure 2G,H,K,L).

Changes in the Level of ECM Markers C3M and LG1M in Response to Bilateral IRI
The biomarker levels of C3M and LG1M were investigated in the urine and plasma fraction of the bilateral IRI groups. Levels of urinary C3M (U-C3M) closely mirrored the fibrotic response with increased presence of the biomarker at 7 and 14 days after reperfusion and returning to baseline levels at 21 days after reperfusion ( Figure 3A). We found no significant change in the plasma fraction of C3M (P-C3M) ( Figure 3B). LG1M in urine (U-LG1M) was increased at 14 days of reperfusion, whereas no significant change was observed in the plasma LG1M (P-LG1M) ( Figure 3C,D) although a trend towards increased plasma levels was observed at 14 days after reperfusion.
the bilateral IRI groups. Levels of urinary C3M (U-C3M) closely mirrored the fibrotic response with increased presence of the biomarker at 7 and 14 days after reperfusion and returning to baseline levels at 21 days after reperfusion ( Figure 3A). We found no significant change in the plasma fraction of C3M (P-C3M) ( Figure 3B). LG1M in urine (U-LG1M) was increased at 14 days of reperfusion, whereas no significant change was observed in the plasma LG1M (P-LG1M) ( Figure 3C,D) although a trend towards increased plasma levels was observed at 14 days after reperfusion.

MRI Images, Values, and Correlations with Fibrosis Markers and Noninvasive ECM Markers
The MRI protocols were adapted for a 9.4 T rodent scanner. The skGFR and perfusion calculated from the DCE images were significantly lower at 7 days of reperfusion and in UUO rats scanned at both day 5 and day 7 ( Figure 4A,B). The total 23 Na signal intensity/kidney was found to be significantly elevated at 7 days of reperfusion ( Figure 4C). In addition, the kidney 23 Na slope was significantly reduced at 7 days of UUO ( Figure 4D). Representative 23 Na-MRI images of the kidneys were obtained in all the animals ( Figure 4E-J). MRI with 1H T1 and T2* mapping did not show any significant change among the groups (data not shown).

MRI Images, Values, and Correlations with Fibrosis Markers and Noninvasive ECM Markers
The MRI protocols were adapted for a 9.4 T rodent scanner. The skGFR and perfusion calculated from the DCE images were significantly lower at 7 days of reperfusion and in UUO rats scanned at both day 5 and day 7 ( Figure 4A,B). The total 23 Na signal intensity/kidney was found to be significantly elevated at 7 days of reperfusion ( Figure 4C). In addition, the kidney 23 Na slope was significantly reduced at 7 days of UUO ( Figure 4D). Representative 23 Na-MRI images of the kidneys were obtained in all the animals ( Figure 4E-J). MRI with 1H T1 and T2* mapping did not show any significant change among the groups (data not shown).  To investigate the correlation between fibrosis and our MRI data, injury markers, and noninvasive ECM markers, we performed a Pearson's correlation analysis of all groups. Assessment of fibrosis was performed using HE ( Figure 5A-D) and Masson's trichrome staining ( Figure 5E-H) as previously published 21 and now used for further analyses related to our MRI data. The fibrotic score was significantly elevated after 7 and 14 days of reperfusion ( Figure 5I), and no change was seen after 21 days. We found a negative correlation between fibrosis score and skGFR ( Figure 5J) and a positive correlation between fibrosis score and total 23 Na signal ( Figure 5K). No significant correlation was found in relation to 23 Na signal slope ( Figure 5L). To investigate the correlation between fibrosis and our MRI data, injury markers, and noninvasive ECM markers, we performed a Pearson's correlation analysis of all groups. Assessment of fibrosis was performed using HE ( Figure 5A-D) and Masson's trichrome staining ( Figure 5E-H) as previously published 21 and now used for further analyses related to our MRI data. The fibrotic score was significantly elevated after 7 and 14 days of reperfusion ( Figure 5I), and no change was seen after 21 days. We found a negative correlation between fibrosis score and skGFR ( Figure 5J) and a positive correlation between fibrosis score and total 23 Na signal ( Figure 5K). No significant correlation was found in relation to 23 Na signal slope ( Figure 5L).
To investigate the correlation between fibrosis and our MRI data, injury markers, and noninvasive ECM markers, we performed a Pearson's correlation analysis of all groups. Assessment of fibrosis was performed using HE ( Figure 5A-D) and Masson's trichrome staining ( Figure 5E-H) as previously published 21 and now used for further analyses related to our MRI data. The fibrotic score was significantly elevated after 7 and 14 days of reperfusion ( Figure 5I), and no change was seen after 21 days. We found a negative correlation between fibrosis score and skGFR ( Figure 5J) and a positive correlation between fibrosis score and total 23 Na signal ( Figure 5K). No significant correlation was found in relation to 23 Na signal slope ( Figure 5L). ANOVA with Bonferroni multiple comparison test was performed between Sham and IRI groups. Each bar represents the mean ± SEM. * p < 0.05 compared to sham group. (J) Correlation between fibrosis scoring and single-kidney glomerular filtration rate (skGFR), (K) correlation between fibrosis scoring and total 23 Na signal, and (L) correlation between fibrosis scoring and 23 Na signal slope.
MRI measurements of skGFR, perfusion, and total 23 Na have a positive correlation with mRNA of α-SMA and FN ( Figure 6A-C). The 23 Na signal slope has a negative correlation with mRNA of α-SMA and FN ( Figure 6A,D). In addition, U-C3M, P-C3M, and P-LG1M were associated with increased mRNA levels of α-SMA ( Figure 6E-G). U-LG1M had no significant association with α-SMA and FN mRNA levels.
MRI measurements of skGFR, perfusion, and total 23 Na have a positive correlation with mRNA of α-SMA and FN ( Figure 6A-C). The 23 Na signal slope has a negative correlation with mRNA of α-SMA and FN ( Figure 6A,D). In addition, U-C3M, P-C3M, and P-LG1M were associated with increased mRNA levels of α-SMA ( Figure 6E-G). U-LG1M had no significant association with α-SMA and FN mRNA levels. After establishing possible fibrosis correlations, we further investigated correlations between the MR data (skGFR, perfusion, total 23 Na signal slope, and total 23 Na signal) and our noninvasive ECM markers ( Figure 7A). MRI measures of skGFR and total 23 Na signal slope as well as total 23 Na signal correlated inversely with U-C3M ( Figure 7B,D,E). No correlation was found for perfusion and U-C3M ( Figure 7C). P-C3M correlated with skGFR and perfusion as well as total 23 Na signal ( Figure 7F-I). No correlation was found for 23 Na slope ( Figure 7H). In addition, a correlation between 23 Na slope and P-LG1M was observed ( Figure 7A).
After establishing possible fibrosis correlations, we further investigated correlations between the MR data (skGFR, perfusion, total 23 Na signal slope, and total 23 Na signal) and our noninvasive ECM markers ( Figure 7A). MRI measures of skGFR and total 23 Na signal slope as well as total 23 Na signal correlated inversely with U-C3M ( Figure 7B,D,E). No correlation was found for perfusion and U-C3M ( Figure 7C). P-C3M correlated with skGFR and perfusion as well as total 23 Na signal ( Figure 7F-I). No correlation was found for 23 Na slope ( Figure 7H). In addition, a correlation between 23 Na slope and P-LG1M was observed ( Figure 7A).

Discussion
The main finding of this study is that skGFR, perfusion, 23 Na slope, and total 23 Na signal intensity as estimated by DCE imaging and CSI were correlated with renal function and fibrosis regression in both IRI and UUO models while MR relaxometry did not show any association. In addition, the noninvasive ECM urinary biomarker C3M correlated with the fibrogenic response as well as MRI. Measured skGFR and total 23 Na in the renal cortex were able to noninvasively reflect the fibrotic response following IRI in rats.
We used two different animal models, unilateral ureteral obstruction (UUO) and bilateral ischemia reperfusion injury (IRI), to follow fibrosis progression and recovery. The UUO model is a well-known fibrosis model in rodents [27]. The bilateral IRI model is a more complex one where strict  23 Na, and 23 Na slope. Linear correlations graphs between (B) skGFR and U-C3M, (C) perfusion and U-C3M, (D) 23 Na signal slope and U-C3M, (E) total 23 Na signal and U-C3M, (F) skGFR and P-C3M, (G) perfusion and P-C3M, (H) 23 Na signal slope and U-C3M, and (I) total 23 Na signal and P-C3M. * p < 0.05 was considered significant.

Discussion
The main finding of this study is that skGFR, perfusion, 23 Na slope, and total 23 Na signal intensity as estimated by DCE imaging and CSI were correlated with renal function and fibrosis regression in both IRI and UUO models while MR relaxometry did not show any association. In addition, the noninvasive ECM urinary biomarker C3M correlated with the fibrogenic response as well as MRI. Measured skGFR and total 23 Na in the renal cortex were able to noninvasively reflect the fibrotic response following IRI in rats.
We used two different animal models, unilateral ureteral obstruction (UUO) and bilateral ischemia reperfusion injury (IRI), to follow fibrosis progression and recovery. The UUO model is a well-known fibrosis model in rodents [27]. The bilateral IRI model is a more complex one where strict control of the severity of the induced ischemic renal injury is critical. If the induced ischemic insult is too severe, animals are likely to die within 48 h [28].
We used 45 min of ischemia to ensure that the kidneys could recover again; thus, we were able to test our noninvasive approach in the fibrogenic response and during recovery of the kidney. Interestingly, fibrosis markers and fibrosis scoring were highly elevated in the early reperfusion phase, followed by a drop towards sham levels after 21 days. The noninvasive extracellular matrix (ECM) biomarker reflecting urinary degradation fragments of COL III (U-C3M) closely followed the fibrosis scoring with the highest elevation after 7 days of reperfusion, remaining elevated at 14 days and dropping to sham levels after 21 days of reperfusion, supporting the restructuring of the ECM. Laminin gamma chain-1 (LG1M) levels in both plasma and urine peaked after 14 days of reperfusion and then dropped to near sham levels at 21 days of reperfusion. ECM turnover during fibrogenesis is a dynamic process, with different rates of formation and degradation of ECM proteins [29]. Our findings for both U-C3M and LG1M support this temporal turnover. Together, this indicates that the bilateral IRI model Pharmaceutics 2020, 12, 775 13 of 16 seems to be a quite useful injury model as it is reversible compared to the UUO. It is important to note that blood urea nitrogen (BUN), creatinine, and fumarase were not significantly elevated in UUO, which is likely caused by the reduced perfusion. Renal injury markers (KIM-1, NGAL, and fumarase) were similarly found to correlate with single-kidney filtration glomerular rate (skGFR), perfusion, and 23 Na-chemical shift imaging (CSI). It was not possible, though, to separate the effects on renal function imposed by fibrosis alone using MRI methods.
This study confirms previous findings that renal function is directly coupled to the level of renal damage and fibrosis [30]. This was particularly evident from the evaluation of renal function using two independent sequences, including dynamic contrast enhanced (DCE; skGFR) and CSI (total 23 Na signal and signal slope). This gives rise to a more reliable set of markers of kidney dysfunction than MR relaxometry, which is inherently based on the magnetic properties and molecular environment of water in the investigated tissue. This means that all factors affecting water balance will be detected in the MR sequence, including perfusion changes, blood flow, edema, kidney growth, blood volume, GFR, diuresis, and ECM composition [5]. Naturally, we cannot rule out the possibility that the quality, sensitivity, and resolution of our images is too poor to utilize relaxometry. However, the resolution of our DCE and CSI images show significant changes with good robustness in spite of a potentially reduced image quality. Interestingly, total 23 Na kidney content also reached a significant change and showed strong correlation with fibrosis regression in the kidney as well as correlation with the COL III degradation marker in the urine (U-C3M). Total 23 Na kidney content is potentially independent of kidney function, giving rise to an independent measurement of fibrosis compared to skGFR and 23 Na kidney slope.
Kidney fibrosis is often heterogeneously distributed, and in principle, MRI can differentiate the heterogeneous fibrosis; however, the current resolution was limited, and as such, the data presented are limited to rather large voxels. This is solely a limiting case in small animals, and as such, focal differentiation is possible in humans. All MRI methods can be applied in humans [31][32][33][34][35][36][37]. However, the use of gadolinium contrast agents is contradicted in severe CKD, and as such, alternative methods would likely be used to determine the hemodynamic nature of kidneys in humans [38].
The ability to noninvasively visualize kidney function and specifically pinpoint areas giving rise to the reduced kidney function makes the 23 Na content and/or slope an intriguing new MRI method that may be used to monitor renal fibrosis. These MRI measurements can either alone or in combination with noninvasive ECM markers, such as C3M, be a valuable tool to assess fibrosis regression in CKD patients where biopsy may not be recommended. This means that ECM markers and MR data potentially can give rise to a completely noninvasive procedure to evaluate fibrosis in a variety of renal diseases. Such noninvasive procedures may also be used in drug development to better assess the effect of reno-protective or anti-fibrotic drugs on structural aspects of the kidney. This would greatly improve clinical trials that currently mostly focus on change in renal function (serum creatinine) and/or albuminuria. Therefore, this study highlights the importance of incorporating multi-parametric information in the stratification of renal injury and fibrosis.

Patents
The patents for the C3M and LG1M ELISAs are owned by Nordic Bioscience.