Slice of Life: Porcine Kidney Slices for Testing Antiﬁbrotic Drugs in a Transplant Setting

: Circulatory death donor (DCD) kidneys are increasingly used to enlarge the donor pool. These kidneys undergo ischemia-reperfusion injury, frequently leading to renal ﬁbrosis. Transforming growth factor beta 1 (TGF-β 1) and matrix metalloproteases have been identiﬁed as central mediators of ﬁbrosis and inhibition of these targets could attenuate ﬁbrosis. We studied whether galunisertib, doxycycline, taurine, and febuxostat alleviated ﬁbrosis in precision-cut kidney slices (PCKS). PCKS were prepared from porcine kidneys that were exposed to 30 min of warm ischemia followed by 3 h of oxygenated hypothermic machine perfusion. We subsequently incubated PCKS for 48 h at 37 ◦ C with the described compounds. To further elucidate the antiﬁbrotic effects of galunisertib, we cultured PCKS with TGF-β 1. We ﬁrst screened the effects of the compounds without TGF-β 1. Most signiﬁcant effects were observed for galunisertib which lowered the expression of ACTA2 , TGFB1 , FN2 , and SERPINE1 . We then investigated the effects of galunisertib in ﬁbrotic PCKS incubated with TGF-β 1. TGF-β 1 signiﬁcantly increased expression of TGFB1 , FN1 , SERPINE1 , and SERPINH1 . Galunisertib, however, attenuated the expression of all ﬁbrosis-related genes. Galunisertib appears to be a promising antiﬁbrotic compound requiring further research in a preclinical model and may ultimately be administered during machine perfusion as an antiﬁbrotic treatment in a transplant setting.


Introduction
Chronic kidney disease is currently prevalent in 11% of the world's population and this number is still rising. It is characterized by a gradual loss of renal function due to interstitial fibrosis, often requiring a kidney transplant for survival [1,2]. As waiting lists keep growing, the need for additional donor kidneys is crucial. Consequently, suboptimal kidneys such as circulatory death donor (DCD) kidneys are increasingly used to enlarge the donor pool [3][4][5]. Unfortunately, these kidneys are deprived of nutrients and are subjected to the re-introduction of oxygen-possibly leading to ischemia-reperfusion injury, delayed graft function, interstitial fibrosis, and renal failure [6,7].
The most common cause for kidney allograft failure is chronic allograft nephropathy which is characterized by interstitial fibrosis and tubular atrophy [8]. Interstitial fibrosis is a common pathological process characterized by an imbalanced extracellular matrix (ECM). Multiple pathways have been shown to play a role in the formation and progression of interstitial fibrosis by inducing inflammation, endothelial-to-mesenchymal transition, epithelial-mesenchymal transition (EMT), activating fibroblasts, and ECM formation [9][10][11][12]. One of the most important cytokines involved in fibrogenesis is transforming growth factor beta 1 (TGF-β1) [13,14]. TGF-β1 primarily causes tissue scarring by activating its downstream small mother against the decapentaplegic (SMAD) signaling pathway [15]. Inhibiting TGF-β1 has shown significant reductions in renal fibrosis formation in rat and mice models yet this has not yet been explored in a transplant setting [16]. While the accumulation of ECM proteins certainly play a role in fibrosis and chronic rejection, recent studies show that excessive degradation of the ECM also plays a pathophysiological role [17][18][19]. One of the proteases responsible for this degradation-matrix metalloprotease 9 (MMP9)-is a key protein in the development of kidney fibrosis [20,21]. Furthermore, studies show that MMP2 and 9 contribute to fibrosis formation by promoting EMT and the activation of fibroblasts [16]. Therefore, TGF-β1 and MMP2/9 could be promising drug targets to attenuate fibrosis caused by transplant-related injury.
Galunisertib, taurine, febuxostat, and doxycycline represent compounds that could potentially attenuate renal fibrosis by modulating TGF-β1 signaling and MMP2/9 activity in a transplant setting. Galunisertib-originally developed as anti-cancer treatment-is an inhibitor of the TGF-β1 receptor kinase which lowers the phosphorylation of SMAD2/3 [22]. It has already been shown that galunisertib has antifibrotic potency in liver and renal fibrosis [23,24]. Taurine-an organic acid, and a derivative of the amino acid cysteine-is another compound that modulates TGF-β1 signaling. It significantly reduces TGF-β1 expression via inhibition of the MAPK pathway which plays an important role in TGF-β1 production in cells [25]. Furthermore, studies have shown that taurine has antioxidative, antiapoptotic, and anti-inflammatory capacities, and protective effects against renal interstitial fibrosis in rats [25][26][27]. Febuxostat-commonly used to treat gout-also modulates TGF-β1 signaling by activating the bone morphogenetic protein-7 (BMP-7) pathway. BMP-7 is a naturally existing inhibitor of the TGF-β1-dependent, pro-fibrotic pathway and protects the kidney against fibrosis, making it a promising drug target [28]. Finally, doxycycline-a tetracycline-class antibiotic-is known for its inhibitory action on MMP activity. Research has indicated that it has protective effects during renal ischemia-reperfusion injury and against fibrosis by attenuating MMP2 and 9 activity and TGF-β expression [29][30][31][32][33].
The implementation of renal normothermic machine perfusion (NMP) provides a unique avenue for treating isolated organs in an ex vivo manner. During NMP, kidneys are rewarmed to normothermic temperatures (35-37 • C) to restore metabolism and allow for functional evaluation of the graft [34][35][36]. Additionally, pharmacologically active substances, such as the ones described above, can be administrated during NMP to repair suboptimal donor kidneys [37]. The main advantage of treating an isolated organ is that systemic effects are avoided and, therefore, more potent drugs can be used.
Before administrating these drugs in a (experimental) perfusion setup, precision-cut kidney slices (PCKS) could be used as an intermediate experimental model to bridge the gap between in vitro and in vivo approaches [38]. PCKS is a novel technique to study multicellular processes as the organ architecture is maintained [39,40]. Thousands of PCKS can be obtained from just one kidney. This allows for testing of compounds using slices from the same biological source, thereby reducing biological variation. PCKS have been proven successful in investigating fibrotic processes and antifibrotic compounds in renal tissue [24,39,41]. By exposing kidneys to warm ischemia, and thus mimicking donation after circulatory death before slicing, it becomes possible to test compounds in a DCD-PCKS model [38].
Our aim was to test the antifibrotic potency of galunisertib, doxycycline, taurine, and febuxostat in a porcine DCD-PCKS model. PCKS were prepared from porcine kidneys that were exposed to 30 min of warm ischemia followed by 3 h of oxygenated hypothermic machine perfusion. We subsequently incubated PCKS for 48 h at 37 • C with the described compounds. To further elucidate the antifibrotic effects of galunisertib in a fibrotic environment, we cultured PCKS with TGF-β1.

Animal Model
All experiments were carried out with porcine kidneys. These kidneys were retrieved from a local slaughterhouse after a highly standardized slaughtering process. The pigs were anaesthetized using an electric shock followed by exsanguination. The kidney of choice was determined based on macroscopic morphology (e.g., undamaged condition, color, shape, arterial branching). All kidneys were flushed with 180 mL of ice cold 0.9% saline solution (Fresenius Kabi, Sèvres, France) after 30 min of warm ischemia time. This way each kidney underwent warm ischemic injury.

Hypothermic Machine Perfusion
As hypothermic machine perfusion (HMP) is standard clinical care for deceased donor kidneys in the Netherlands, kidneys were preserved and transported using HMP. Kidneys were surgically prepared and connected to the Kidney Assist Portable (XVIVO, Gothenburg, Sweden) perfusion machine. HMP was performed for 3 h at a set mean arterial pressure of 25 mmHg. A quantity of 230 mL of oxygenated (100 mL/min) University of Wisconsin (UW) perfusion solution (Bridge to Life Ltd., London, UK) was used at 3-5 • C.

Precision-Cut Kidney Slices
After HMP, the kidneys were immediately flushed with 120 mL of ice cold 0.9% saline solution. Using a 10 mm blade, the renal capsule was carefully removed. Cores were drilled from the cortex with a 6 mm biopsy punch and stored in ice cold UW cold storage solution.
The Krumdieck tissue slicer was attached to a water bath that cooled the Krebs-Henseleit buffer to 4 • C. Slices were collected and selected based on round and intact morphology. The slices had a weight between 4-6 mg with an estimated thickness of 300 µm. To ensure viability, the slices were collected in ice cold UW cold storage solution.
For the initial aim, part A, PCKS were individually incubated in 12-well plates under the conditions described in Table 1. The control group represents the standard conditions for incubating renal porcine tissue slices [38]. All slices were incubated at 37 • C in an incubator with 80% O 2 /5% CO 2 while shaking at a rate of 90 rpm for 48 h. The culture media with added compounds were refreshed after 24 h.

Inducing Fibrosis in PCKS
To further elucidate the antifibrotic effects of galunisertib, part B, we cultured the PCKS in a similar manner to methods described above. HMP was performed using UW cold storage solution instead of UW preservation solution. PKCS were incubated in 12-well plates supplemented with TGF-β1 to promote fibrosis and a treatment group with 10 µM galunisertib to target the induced fibrosis (Table 2).

Viability Analysis
Viability of the PCKS was evaluated by assessing the adenosine triphosphate (ATP) content as described by de Graaf et al., 2010 [40]. Three slices per timepoint and condition were individually stored in an Eppendorf tube with MiniBeads (Biospec products, Bartlesville, OK, USA) containing 1 mL of sonification solution (70% ethanol and 2 mM EDTA) and snap frozen in liquid nitrogen. A bioluminescence kit was used to measure ATP levels (Roche Diagnostics, Almere, The Netherlands). Luminescence was measured using a luminometer (LumiCountä, Packard, IL, USA). The obtained ATP value was normalized against the total protein content which was measured using a Pierce TM BCA protein assay kit. The final ATP content was expressed as pmol ATP/µg protein.

Gene Expression Analysis
Gene expression for multiple pro-fibrotic genes was assessed with quantitative realtime polymerase chain reaction (PCR). RNA was isolated from five pooled, snap-frozen, precision-cut tissue slices using TRIZOL reagent (Invitrogen, Waltham, MA, USA). Extracted RNA was reverse-transcribed at 70 • C for 10 min, at 37 • C for 50 min, and 15 min at 70 • C. Specific primers were used for PCR analysis ( Table 3). The PCR started with 20 s at 95 • C, followed by 40 cycles of 1 s at 95 • C, and 20 s at 60 • C. The plate was measured in the QuantStudio TM 7 Flex Real Time PCR System (Applied Biosystems, Waltham, MA, USA). Expression levels were corrected with GAPDH as a reference gene (∆Ct) and compared with a pooled control group (∆∆Ct). Next, the fold induction was calculated and the groups were normalized to the control (control set to 1).

Histological Analysis
Slices fixed in 4% formaldehyde and embedded in paraffin were sectioned at 4µm. An immunohistochemical staining for α-smooth muscle actin (α-SMA) was performed. In short, sections were blocked with 30% H 2 O 2 and incubated for 1 h with the α-SMA antibody (Sigma-Aldrich, Darmstadt, Germany), followed by an incubation with a second antibody (Rabbit anti-mouse immunoglobin HRP, #P0260, Dako Agilent, Amstelveen, The Netherlands), and a third antibody (Goat anti-rabbit immunoglobin HRP, # P0448, Dako Agilent), to amplify the signal. A quantity of 1% Normal Swine Serum was added (Dako Agilent). The antibody-antigen complex was visualized by adding diaminobenzidine (DAB) (Merck, Amsterdam, The Netherlands). Counterstaining with hematoxylin and dehydration completed the procedure.
The microscopic images were obtained using the NanoZoomer S360 (Hamamatsu, Hamamatsu, Japan) and the quantification, based on positive pixel count, of α-SMA staining was performed using Aperio ImageScope (Leica Biosystems, Richmond, IL, USA).

Statistical Analysis
GraphPad Prism 9.0 (GraphPad Software, Boston, MA, USA) was used for creating figures and performing statistical analyses. Values are shown as means with standard deviations. Significant differences were analyzed using one-way ANOVA with a Dunnett's multiple comparisons test. The cut-off for statistical significance was set at p < 0.05.

The Effect of Antifibrotic Drugs on DCD-PCKS
We first screened the effects of the different antifibrotic compounds in the absence of TGF-β1. ATP content of the PCKS significantly increased after 48 h incubation compared to baseline demonstrating that ATP production was restored after injury due to cold ischemia and slicing. Furthermore, ATP levels remained stable regardless of the treatment (Figure 1). This indicates that the viability of the PCKS was maintained during culture. Figure 2 shows that 48 h of incubation led to a significant increase in ACTA2, TGFB1, and FN2 expression (baseline compared to the control) indicating that fibrotic processes are already active within these 48 h. Supplementing the PCKS with galunisertib significantly lowered the gene expression of all measured fibrosis markers compared to the control with a p value of <0.0001 (Figure 2A-D). Doxycycline significantly lowered the gene expression of TGFB1, FN-2, and SERPINE1 with a p value of <0.05 ( Figure 2B-D). Febuxostat supplementation only significantly decreased mRNA expression of SERPINE1 (p < 0.0001) ( Figure 2D). Incubation with taurine did not attenuate any of the measured fibrosis markers. Together, these results show that galunisertib had the strongest antifibrotic effect during 48 h of PCKS incubation.   To observe whether antifibrotic effects were already visible at a protein level, α-SMA levels in tissue were analyzed ( Figure 3A). Α-SMA levels at baseline were significantly higher compared to the control after 48 h of incubation. The administration of galunisertib, To observe whether antifibrotic effects were already visible at a protein level, α-SMA levels in tissue were analyzed ( Figure 3A). A-SMA levels at baseline were significantly higher compared to the control after 48 h of incubation. The administration of galunisertib, doxycycline taurine or febuxostat did not significantly affect expression of α-SMA compared to the control after 48 h of incubation ( Figure 3B).

The Effect of Galunisertib on Fibrotic PCKS
As galunisertib showed the most significant antifibrotic effects, we then investigated the effects of galunisertib in PCKS that were cultured with TGF-β1 to induce fibrosis. ATP content significantly increased after 48 h compared to baseline (Figure 4). TGF-β1 and the combination of TGF-β1 and galunisertib did not affect ATP content of the PCKS indicating that the viability of the PCKS was maintained during the 48 h of incubation.

The Effect of Galunisertib on Fibrotic PCKS
As galunisertib showed the most significant antifibrotic effects, we then investigated the effects of galunisertib in PCKS that were cultured with TGF-β1 to induce fibrosis. ATP content significantly increased after 48 h compared to baseline (Figure 4). TGF-β1 and the combination of TGF-β1 and galunisertib did not affect ATP content of the PCKS indicating that the viability of the PCKS was maintained during the 48 h of incubation.

The Effect of Galunisertib on Fibrotic PCKS
As galunisertib showed the most significant antifibrotic effects, we then investigated the effects of galunisertib in PCKS that were cultured with TGF-β1 to induce fibrosis. ATP content significantly increased after 48 h compared to baseline (Figure 4). TGF-β1 and the combination of TGF-β1 and galunisertib did not affect ATP content of the PCKS indicating that the viability of the PCKS was maintained during the 48 h of incubation. We then observed the antibiotic effects by analyzing a larger panel of pro-fibrotic markers. Forty-eight hours of incubation led to a significant decrease in ACTA2 and SER-PINH1 expression (baseline compared to the control) ( Figure 5A,E). TGF-β1 promoted fibrosis in PCKS as shown by a significantly increased mRNA expression of TGFB1, FN1, SERPINE1, and SERPINH1 after 48 h of incubation compared to the control. (Figure 5B- We then observed the antibiotic effects by analyzing a larger panel of pro-fibrotic markers. Forty-eight hours of incubation led to a significant decrease in ACTA2 and SERPINH1 expression (baseline compared to the control) ( Figure 5A,E). TGF-β1 promoted fibrosis in PCKS as shown by a significantly increased mRNA expression of TGFB1, FN1, SERPINE1, and SERPINH1 after 48 h of incubation compared to the control. (Figure 5B-E). Galunisertib, however, clearly attenuated the expression of these pro-fibrotic markers after 48 h of incubation. mRNA expression levels for ACTA2, TGFB1, FN1, SERPINE1, SERPINH1, and COL1A2 were significantly lower for tissue slices incubated with galunisertib and TGF-β1 compared to just TGF-β1 ( Figure 5A-F). No differences were observed between the control and TGF-β + galunisertib group.

Discussion
Interstitial fibrosis formation in DCD renal allografts is a huge burden for the recipient as it leads to loss of renal function, the need for dialysis treatment and/or a re-transplant, and possibly premature death [42,43]. Unfortunately, research into transplant-related renal interstitial fibrosis is lacking due to the availability of proper translational models. Our goal was to assess the antifibrotic potencies of several compounds using porcine DCD-PCKS. This research shows that the combination of machine perfusion and PCKS provide a suitable way to assess antifibrotic pharmaceutical interventions in DCD kidney tissue. Furthermore, we convincingly demonstrated that galunisertib exhibits antifibrotic effects.
Incubation with doxycycline significantly downregulated three of the four measured fibrosis markers. In a previous study, we showed that the addition of doxycycline during HMP results in lower NGAL levels [32]. However, we did not see any significant differ-

Discussion
Interstitial fibrosis formation in DCD renal allografts is a huge burden for the recipient as it leads to loss of renal function, the need for dialysis treatment and/or a re-transplant, and possibly premature death [42,43]. Unfortunately, research into transplant-related renal interstitial fibrosis is lacking due to the availability of proper translational models. Our goal was to assess the antifibrotic potencies of several compounds using porcine DCD-PCKS. This research shows that the combination of machine perfusion and PCKS provide a suitable way to assess antifibrotic pharmaceutical interventions in DCD kidney tissue. Furthermore, we convincingly demonstrated that galunisertib exhibits antifibrotic effects.
Incubation with doxycycline significantly downregulated three of the four measured fibrosis markers. In a previous study, we showed that the addition of doxycycline dur-ing HMP results in lower NGAL levels [32]. However, we did not see any significant differences when looking at the mRNA expression of fibrosis markers. Moser et al. and Cortes et al. showed similar protective effects of doxycycline in rat kidneys exposed to ischemia-reperfusion injury [31,33].
Taurine did not seem to have any significant antifibrotic effects and febuxostat only on SERPINE1 mRNA expression, contrary to previous research [25,28]. This could be because of the administrated dosage. In previous studies, taurine was administered to whole animals instead of to single tissue slices. The administrated dosage is therefore difficult to translate, and a higher dosage might be necessary for PCKS. Furthermore, these previous studies were performed using rats. The progression of fibrosis in rodent models could have a different timeframe compared with the progression of fibrosis in larger species such as pigs or humans.
The variation in fibrosis induction could also lead to variation in fibrosis progression per model. In PCKS, the onset of pro-inflammatory and profibrotic responses can already be observed and targeted within 48 h because of the damage caused by warm ischemia, the mechanical stress of the tissue-slicing process, and cold ischemia. Here, we show that the addition of TGF-β1, a key mediator of fibrosis [13][14][15], greatly promoted these profibrotic responses by significantly increasing the expression of a broad range of profibrotic markers after only 48 h of incubation. Rodent fibrosis models do not fully mimic the clinical situation as it is often difficult to distinguish whether the given treatment targets the primary renal injury (e.g., ischemia-reperfusion injury, inflammation) causing attenuation of fibrogenesis [37], or actually targets the fibrosis [44]. Moreover, it usually takes weeks for rodents to develop fibrosis.
We selected several fibrosis-related genes corresponding to various processes involved in renal fibrosis. During the onset of fibrosis, TGF-β is one of the most important cytokines involved [13,14]. TGF-β has the ability to activate myofibroblasts that are characterized by de novo expression of α-SMA. These cells then aim to restore tissue integrity by producing and secreting ECM proteins, especially collagens and fibronectins [10]. HSP47-encoded by the gene SERPINH1-is a molecular chaperone that is required for collagen synthesis [45]. Additionally, PAI-1-encoded by the gene SERPINE1-promotes the accumulation of collagen and fibronectins and thus stimulates excessive tissue scaring [46].
We observed no significant differences in α-SMA formation between treatments. However, 48 h of incubation could be too short a timeframe to see differences on a protein level. We did observe significant differences in mRNA expression. Galunisertib caused a significant decrease in ACTA2 expression, the gene coding for α-SMA, after 48 h of incubation (Figures 2A and 5A). However, mRNA transcription is a much quicker process than protein translation.
As mentioned before, PCKS provide an elegant platform for testing multiple compounds using the same kidney, keeping biological differences minimal. Because of the size of the porcine kidney and the simplicity of the method, hundreds of slices can be produced and used to assess treatments under a variety of conditions (e.g., timepoints, concentrations, compounds) using various analytical techniques, such as qPCR, staining, and western blotting, allowing a close look at mechanisms of action. Porcine kidneys are anatomically similar to human kidneys and therefore provide a suitable model for translational research [47]. Slaughterhouse kidneys are commonly used in research on renal transplant-related questions and with promising results [38,[48][49][50][51][52][53]. Moreover, they spare us the need for using laboratory animals. Bigaeva et al. showed similar results on gene expression when incubating human (fibrotic) PCKS with 10 µM galunisertib [54], indicating that porcine PCKS provide translatable results.

Limitations
We observed differences between the baseline and control ATP and mRNA levels between our two aims. These differences could be due to the different glucose concentrations between UW machine perfusion solution used during HMP of the first aim and UW cold storage solution used during the second aim. Glucose is an important component for ATP production and different concentrations of glucose have significant impact on cell viability [55,56]. UW cold storage solution lacks glucose explaining the lower baseline ATP levels in the second aim. However, ATP levels for both aims follow the same trend and are still reliable, as each aim contains its own control.
Although PCKS provide a great way of testing multiple compounds while only needing one kidney, it does not provide insights into how compounds will react in full isolated organs. NMP with added interventions would therefore constitute a valuable follow-up experiment, providing a platform to assess the effects of promising treatments such as galunisertib on renal function in a transplant setting [57,58].

Conclusions
This study has provided insights into the antifibrotic potential of several drugs. We show that porcine DCD-PCKS provides an important translational model for answering transplant-and fibrosis-related questions. Furthermore, we show that galunisertib demonstrates strong antifibrotic effects, making it a promising antifibrotic compound for further research in a more preclinical model. Ultimately, galunisertib could be implemented during machine perfusion in a clinical setting as a treatment to prevent or to attenuate fibrosis in DCD kidneys. Institutional Review Board Statement: Not applicable, due to the use of slaughterhouse waste material.

Informed Consent Statement: Not applicable.
Data Availability Statement: All data is available upon request.