Discussion

Previous renal MPS models have targeted specific structures of the nephron, but regulation of xenobiotics and substrates in the blood is a concerted action of both filtration in the glomeruli and active reabsorption by the proximal tubule. The challenge in establishing an effective nephron MPS originates from fulfilling all required characteristics of kidney MPS: incorporation of renal cellular models (tubular structures or in vitro barrier models) (Jang et al., 2013; C. Sakolish et al., 2020; C. Sakolish et al., 2018; C. M. Sakolish et al., 2016; Tourovskaia et al., 2014; Weber et al., 2016), functional (filtration, absorption, and secretion) processes, and drug toxicity/ recovery testing. In this study, the glomerulus- and PCT- MPS addresses most of the features of a renal MPS, while also maintaining in vivo shear stress across the apical PCT (0.4-1.5 dyne-s/cm2) and recirculating 64.0% of flow into the bloodstream and 36.0% into the filtrate output. The tri-culture system of CIHP-1, HUVECs and HK-2 cells provides a novel, realistic approach by recapitulating the critical functions in both the glomerulus and PCT for long-term, overcoming the limitations of existing glomerulus-on-a-chip and PCT-on-a-chip designs, and assessing for drug toxicity.
Geometrically, many vital attributes contribute to the innovation of the glomerulus and PCT- MPS that have never been modeled previously in vitro . The MPS presents a dual in vitro barrier model in the glomerular filtration chamber and the PCT reabsorption device. Within the glomerular unit, the 30nm pore size PES membrane was selected based on the physiological size-exclusive filtration found on the human glomerular capillary walls (C. M. Sakolish & Mahler, 2017; C. M. Sakolish et al., 2019). The PCT barrier model highlights a multi-layered fluidic platform separated by a porous membrane, creating a dual channel system (apical and basolateral). Proximal tubule cells on the membrane selectively resorb solutes from the apical filtrate formation line to the basolateral bloodstream, reenacting the renal active and passive transport processes. This recirculation of flow was established through the T-Junction component, crucial in bloodstream circuit reservoir to simulate the removal of compounds that have been reabsorbed by the renal tubule. Additionally, the PCT device was constructed in a polycarbonate material instead of the traditional polydimethylsiloxane (PDMS), meaning that the devices were highly robust, autoclavable, and reused up to 10 times. Failure points, such as leakage, occurred in the presence of long-term exposure to humidity, high temperatures, and shear stress. Yet, the MPS design fosters low-cost manufacturing and minimizes wasted resources in single-use devices.
Throughout the MPS experiments, comparisons of the single and tri-culture systems determined the significance of modeling concerted structures of the nephron. The single culture is illustrative of the conventional PCT devices (Jang et al., 2013; Ng et al., 2012; Raghavan et al., 2014; C. M. Sakolish et al., 2019; Weber et al., 2016), but the unique attributes of the glomerulus and MPS system include three cell types, selected based on its key functions in filtration, reabsorption, and secretion. Filtration is operated through the glomerular filtration barrier mainly controlled by the glomerular and endothelial cells (Li et al., 2016). Co-culturing these cells have previously proved challenging through the technical difficulties of glomerular endothelial cells and podocytes. Since both cell types are highly differentiated, specialized, and interdependent cell types (Xinaris et al., 2016; Xinaris et al., 2012), selection of the cells in the model is vital in evaluating the MPS. CIHP-1 cells, representative of the podocytes, were selected based on its immortalized, homogenous characteristics. To ensure that the MPS establishes efficacy, homogeneity minimizes outliers and inconsistencies. Similarly, HUVECs were selected based on prevalent use in endothelial cell studies (Caulfield & Farquhar, 1974; Venturoli & Rippe, 2005) and derivation from humans. Reabsorption is mostly contributed by the proximal tubule cells as denoted by the HK-2 cells, which were selected based on extensive use previous work (Jang et al., 2013; C. M. Sakolish et al., 2016; C. M. Sakolish & Mahler, 2017; C. M. Sakolish et al., 2019). From the comparative results, the single culture and tri-culture were indicative of the significance of establishing a multi-cellular model in future kidney research. The differences in velocity flow rate and protein concentrations validate how the glomerulus and PCT MPS greatly impact the retention of functional cell-cell and cell-tissue interactions.
To achieve animal-byproduct free culture conditions, multiple media types were tested. Based on the results of the viability and proliferation testing (Figure S1 ), all three cell types were adapted to an optimal common cell culture medium (ESFM) prior to dynamic experimentation. For the glomerular functional assay, human serum albumin was used (FITC-HSA, 67.0 kDa). Cells within the glomerulus and PCT-MPS demonstrated an in vivo - like functional activity through the retention of HSA (~67 kDa) from circulating in the bloodstream pathway, while resorbing concentrations of glucose in filtrate output. The protein concentrations were depicted in both the plate reader fluorescence readings, but also in the confocal images, where there was an uptake of the green fluorescent protein in the podocytes (Figure S2 ). Previous works (Jang et al., 2013; Ng et al., 2012) found that dynamic conditions results in cytoskeletal reorganization and junctional reformation in PCT-derived cells. These outcomes were not clearly observed in the glomerulus and PCT MPS due to the low shear stress to exhibit healthy renal conditions and longer culture time. While human serum albumin results in this study do not distinguish between glomerular filtration and tubular reabsorption, a previous glomerulus and PCT MPS iteration (C. M. Sakolish & Mahler, 2017) featured transport with other proteins sizes. The MPS with only endothelial (HUVECs) and proximal tubule (HK-2) cells was subjected with 0.1 mg/mL of FITC-bovine serum albumin (FITC-BSA, 66.4 kDa) and FITC-ovalbumin (45.0 kDa). The HUVECs demonstrated more BSA localization in the junctions of the cell monolayer, whereas the HK-2 had an increase in BSA in vesicles, suggesting a presence of passive and active transport, respectively. Future works can incorporate different sized fluorescent compounds (i.e. dextran) to characterize the specificity of trans-glomerular filtration and tubular reabsorption (Caulfield & Farquhar, 1974; Ng et al., 2012; Venturoli & Rippe, 2005; Xinaris et al., 2016; Xinaris et al., 2012). For imaging in post-dynamic experiments, immunocytochemistry would typically be used to display functional biomarkers, such as VE-cadherin, clathrin, or nephrin (Musah et al., 2017; Petrosyan et al., 2019; Weber et al., 2016). In this study, alternate stains were used to avoid the use of animal products (Calcein AM, Hoescht 3342, and phallodin 568). Largely, the glomerulus- and PCT- MPS offers an effective tool for assessing preclinical drug adsorption, distribution, metabolism, elimination, and toxicity (ADMET) testing, modeling disease for drug treatment regime trials, and understanding cellular crosstalk.
Additional work is required to ensure confluent endothelial cell adhesion by day 7 on the PES membrane. Previous works have identified Collagen IV and laminin as a more physiological relevant extracellular matrix in renal studies (D. R. Abrahamson, 1991; D. R. Abrahamson, Hudson, Stroganova, Borza, & John, 2009; John & Abrahamson, 2001; Li et al., 2016; Mauer, 1994; Pavenstadt, Kriz, & Kretzler, 2003; C. Sakolish et al., 2018; Weber et al., 2016). Initially, glomerular endothelial and epithelial cells generate a glomerular basement membrane (GBM), a cytoskeletal template for renal cell adhesion. Following, fully differentiated podocytes further assemble matrix molecules into a highly intricate meshwork of collagen IV, laminin, heparin sulfate, entactin, agrin, and perlecan (D. Abrahamson, 1987; D. R. Abrahamson, 1985; Adler, 1992; John & Abrahamson, 2001; Miner, 1998; Sariola, 1984). Of those extracellular proteins, type IV collagen and laminin have ensured endothelial monolayers and better preservation of the podocyte phenotype based on the F-actin distribution, nephrin, and synaptopodin (Li et al., 2016). Recreating a kidney diseased model, such as chronic kidney disease and drug-induced kidney injuries (DIKI), on a molecular scale is essential in innovating safer therapeutic agents. Establishing predictive models of DIKI avoid unnecessary risk for patients and reduce financial burden of compound attrition in pharmaceutical development. Further investigation of renal drug candidates, such as cisplatin, gentamicin, perfluorooctanoic acid (PFOA), and angiotensin-converting enzyme inhibitors (ACEI) would permit a wider scope into the pharmacokinetics within the glomerular and proximal tubule MPS. In Sakolish et. al, 2019, a similar glomerulus and PCT MPS with only HK-2 cells was implemented to test altered glomerular filtration rate (GFR), hyperglycemia, nephrolithiasis, and drug-induced nephrotoxicity. Four conditions were subjected in the MPS with their respective disease model: (1) physiological flow of 0.8 dyne-s/cm2 and high shear flow of 5 dyne-s/cm2, (2) KSFM and glucose-treated SFM, (3) 0.3mg/ml of calcium oxalate monohydrate crystals (COM), and (4) 1.4mM of cisplatin and 0.27 mM of cyclosporine exposures for 72 hrs at 24 hr intervals. Based on the significant changes in cellular responses of these single culture MPS compared to the static conditions (Figure S3 ), this operates as a precursor for the current glomerulus and PCT tri-culture MPS’s potential application in assessing DIKI. More functional assays would build confidence in the physiologically credible characteristics. MPS under fluid-driven flow would have increased mechanical signaling in the primary cilia, microvilli, or glycocalyx, increased trafficking and expression of apical and basolateral transporters, and increased epithelial barrier function (Koepsell, 2013; Nieskens et al., 2016; Nieskens & Wilmer, 2016; Phillips et al., 2020).