Functional assessment of glomerular cells with hybrid bioink for coaxial cell printing
Tissue decellularized extracellular-matrix-based bioinks have been extensively applied in tissue engineering and 3D cell printing of tissues/organs to provide the tissue-specific native microenvironment to the resident cells 22-24. Recently, studies have reported the use of alginate blended with decellularized tissue bioinks to achieve the mechanical properties suitable for 3D coaxial cell printing 18, 21. To obtain printable bioinks for 3D coaxial cell printing of functional bGOAC, we first customized the hybrid bioink formulation by blending sodium alginate with KdECM bioink to promote cellular activity of GEs and podocytes.
We first assessed the cell viability and gene expressions of podocytes and GEs encapsulated independently in hybrid bioinks at different proportions of KdECM and sodium alginate (Supplementary Fig. 1A, B). After 7 days of culture, we observed significant increments in cell viability and gene expressions in glomerular cells (podocytes and GEs) encapsulated in the hybrid bioink at an equal ratio of 3% (w/v) KdECM mixed with 1% (w/v) sodium alginate (i.e., 3K1A). The values were superior to 3% (w/v) KdECM mixed with 2% (w/v) sodium alginate (i.e., 3K2A) and 3% (w/v) of pure sodium alginate (Supplementary Fig. 1A, B). Upon confirmation of the 3K1A bioink as our optimized hybrid bioink ratio, we compared the biological performance of the 3K1A with 3% (w/v) pure KdECM and 3% (w/v) collagen type I, an extensively used bioink for kidney tissue engineering. The customized hybrid bioink (3K1A) displayed superior cell viability for the GEs and podocytes compared with collagen type I (Fig. 1D, E, and Supplementary Fig. 2). Moreover, we also observed higher expressions of the vascular maturation marker for endothelial junctions (CD31) and podocyte-specific markers (podocin and nephrin) based on immunofluorescence staining of GEs and podocytes encapsulated in independent groups. The overall expressions of the cells encapsulated in hybrid bioink (3K1A) was higher than those in collagen type I bioink, but lower than those in pure KdECM bioink (Fig. 1F-H).
The customized hybrid bioink (3K1A) maintained high-cell viability, upregulated expression levels of vascular markers of glomerular endothelial cells, and podocyte markers.
Subsequently, we printed GEs and podocyte-laden bioinks using coaxial cell printing. This resulted in the formation of monolayered, hollow GE and podocyte tubes that are capable of perfusion in standard cell culture conditions (Fig. 2A, B). Both monolayered hollow tubes (GE and podocyte) exhibited a homogeneous distribution of cells in their respective tube walls. The live/dead images of the encapsulated cells in their respective monolayered hollow tubes showed prominent cell viability without distortions in tube structural fidelity after 7 days of culture (Fig. 2A, B).
We also confirmed based on immunofluorescence staining the sustained expressions of CD31 and zonula occludens (ZO-1) markers in the GE tube, and nephrin and podocin in the podocyte tube (Fig. 2C-H). These findings suggest that the markers CD31, ZO-1 in the GE tube, and nephrin and podocin in the podocyte tube were articulated along with their maturation process. The overall results of coaxial cell printed GE and podocyte hollow tubes indicate a promising applicability potential for glomerular printing.
Therefore, the 3K1A hybrid bioink was utilized to carry both GEs and podocytes for the printing of functional GFB.
In vitro modeling of glomerular filtration barrier
To date, several in vitro proximal tubule models have been introduced by microfluidics or 3D cell printing techniques 21, 25-27. In contrast, 3D cell printing of functional glomerular structure, has not yet been reported, and remains an ambitious goal. Using a coaxial cell printing method, we intended to replicate the specific glomerular barrier structure (inner layer: GEs, outer layer: podocytes and intervening GBM) by fabricating a bGOAC and evaluating whether this in vitro glomerular model could be utilized to examine the function of GFB, screen nephrotoxic drugs, and be used as the study of glomerular pathophysiology.
The functional bGOAC was prototyped using the customized bioink and coaxial cell printing process (Fig. 3 and Supplementary Movie 1). The fabrication process of the bGOAC involves the printing of a poly(ethylene/vinyl acetate) (PEVA)-based bGOAC chip body, coaxial cell printing of bilayer glomerular microvessel, and the fixation of the inlet–outlet of the printed bGOAC using agarose (Fig. 3A, i-iii, and 3B, i-iii). To visualize the discriminated inner layer (representing the monolayer of GEs) and outer layer (representing the monolayer of podocytes) of the glomerular model, we labeled inner and outer layers of the bilayer glomerular microvessel using red and green fluorescent beads, respectively. Both inner and outer layers of the printed, hollow, bilayer glomerular microvessel were successfully envisioned in its predefined place (Fig. 3C, D).
Figs. 3E and 3F represent the cross-sectional view of the printed, hollow, bilayer glomerular microvessel, reflecting distinctly defined inner and outer layers, each colored in either green or red. Despite the discriminated bilayers, the wall thickness of the inner and outer layers that respectively load GEs and podocytes, should be minimized to assist the formation of monolayer GEs and a podocyte layer. On average, the inner and outer diameters of the hollow bilayer glomerular microvessel were measured to be equal to 542 ± 41 and 684 ± 57 µm, respectively, whereas the inner and outer wall thicknesses (WTs) of the hollow bilayer glomerular microvessel were 71 ± 10 and 122 ± 24 µm, respectively (Fig. 3G, H).
Moreover, the inner/outer dimensions and layer thicknesses of the printed hollow bilayer glomerular microvessel were easily tuned 18, 20, 21. A perfusion test was performed with the optimized bGOAC (Fig. 3B, iv, and Supplementary Movie 2). During the entire perfusion process, the printed, bilayer glomerular microvessel lumen maintained good structural integrity without any leakage.
Upon the establishment of fabrication parameters, we cell printed the bGOAC using GEs and podocytes encapsulated individually in the customized hybrid bioinks (Fig. 4). The coaxial cell-printed, hollow glomerular microvessel possessed two distinct layers, with the inner layer containing the GEs, and the outer layer containing the podocytes. To allow cultures in the fabricated bGOAC, a peristaltic pump was connected to allow the circulation of GEs and culture media through the bilayer glomerular microvessel. This provided media to the GEs of the inner layer, while the podocyte culture media was supplied through the urinary compartment to support cellular growth in the outer layer of the bilayer glomerular microvessel. The viability of both cells in the wall of the printed bilayer glomerular microvessel was maintained on day 1 (~95%) and remained high until day 7 (~97%) (Fig. 4A, i, ii). Likewise, in a separate experiment, the bGOAC was stained for CD31 and nephrin on days 3 and 7 to observe the cellular phenotype within the printed bilayer glomerular microvessel (Fig. 4B-F). The mature and confluence expressions of CD31 and nephrin necessary for proper function of GFB was detected in the fabricated bGOAC, thus suggesting that the encapsulated GEs in the inner layer and podocytes in the outer layer of the bilayer glomerular microvessel generated the intact endothelium and epithelium, respectively (Fig. 4B-F). More importantly, mature expression of the critical tight junction marker ZO-1 was also verified on day 7 (Fig. 4G). To evaluate the long-term preservation of the cellular phenotype in the printed glomerular model, the marker expressions of CD31, tight junction ZO-1, and nephrin, were validated after 14 days. This confirmed the stable expressions and maintenance of these markers in the bGOAC (Supplementary Fig. 3A, B).
These results confirmed the successful fabrication of in vitro GFB on a chip via the coaxial cell printing method and our customized bioinks.
The recreation of correct functional GFB is a critical requirement for the development of a new therapeutic approach to chronic kidney disease as well as for the study of glomerular pathophysiology. Podocyte and GE layers in the developed GFB may not fully assure the correct function of glomerular capillary wall without the existence of GBM. During kidney development, podocytes and GEs are accountable for the synthesis of the GBM composition in the glomerular capillary wall 28, 29. In human GBM, COL IV, and LAM are its major structural components necessary for its function, as mutations in COL IV and LAM cause filtration defects and result in human kidney disease 30, 31.
We confirmed the deposition of COL IV and LAM in the printed podocyte tube, monolayer GE tube, and bGOAC on day 7. The immunofluorescence staining results verified the significant production and deposition of the COL IV and LAM proteins in the printed bGOAC, monolayer podocyte, and GE tubes (Fig. 5A, B). In addition, these results also revealed that production of COL IV and LAM were predominantly high in the podocyte tube compared with the GE tube, thus confirming that podocytes were mainly accountable for the assembly of these GBM proteins. In vivo GBM formation was directly associated with the interactions of the podocyte–GE cells. Thus, synthesis of neo-GBM in the bGOAC revealed the correct interaction and cellular crosstalk between the podocytes and GEs. This demonstrated that the coaxial, cell-printed bGOAC resembles the in vivo GFB.
Permselectivity function of the printed bGOAC
One important function of the GFB is to filter the molecules in the bloodstream based on their sizes 2. In vivo, the GFB is a highly specialized interface that retains large solutes (i.e., albumin) within the plasma while allowing higher conductance toward small to midsized solutes (i.e., inulin. Under the physiological condition, albumin is retained within the blood, whereas albumin leakage in the urine reflects the dysfunction of the GFB 10. We then conducted tests to assess whether printed bGOAC could also reconstitute the permselectivity function. For this, we perfused large molecules such as fluorescein isothiocyanate (FITC)–albumin and FITC–dextran (70 kDa) into the lumens of mature bGOAC, and in the monolayer GE, podocyte, and bare tubes (without cells).
The perfusable bGOAC prevented the large molecules (albumin and dextran) from leaking (Fig. 5 C and D) at 3 min, 20 min, and 30 min. More than 99% of albumin was retained within the bilayer glomerular microvessel after perfusion for 1 h (Supplementary Fig. S4). We also analyzed the albumin permeability within monolayer GE and podocyte tubes. Both monolayer GE and podocyte tubes exhibited moderately selective permeability toward albumin compared with bilayer glomerular microvessel (Supplementary Fig. S4). In addition, the podocyte tube retained albumin to selectively higher levels compared with the GE tube levels, thus demonstrating the podocytes to be the key player of albumin selective permeability in the bGOAC. To show the superiority of our glomerular model to other in vitro models, we have compared our bGOAC with the transwell, cocultured podocyte–glomerular endothelial barriers. The supplementary Fig. S4 revealed that the transwell coculture GFB displayed a pronounced leakage of albumin as compared with our cell-printed bGOAC, hence, indicating that the transwell, coculture podocyte–endothelial barrier is unable to execute as efficiently as the cell-printed bGOAC.
Furthermore, we evaluated the filtration capability of our fabricated bGOAC for inulin a small-sized molecule, as it is performed by in vivo GFB. Fig. 5E shows that inulin can freely pass the GFB-containing bilayer glomerular microvessel to the urinary compartment of bGOAC. These findings suggest that the cell printed bGOAC specifically replicates the normal function of the in vivo GFB that is able to perform the differential clearance of large and small molecules, as illustrated by the albumin and inulin outcomes.
In vitro modeling of drug-induced glomerular injury and proteinuria
In clinical therapeutics, drug-induced nephrotoxicity (DIN) is often inevitable and remains a major hurdle 32. Hence, an accurate prediction tool for DIN is of critical relevance as it can avoid renal damage to patients who are undergoing drug-based therapy. Additionally, the possibility to customize the design of drug types and doses based on the utilization of a functional screening glomerular system in vitro is envisaged to have tremendous value for the prediction of DIN and/or for new drug development.
To test the hypothesis that our cell-printed functional bGOAC represents a distinctive model for screening drug-induced glomerular injury state, we exposed the cell-printed glomerular model to various doses of Adriamycin for 3 days via perfusion through the bilayer glomerular microvessel (Fig. 6A). The cell viability in the bilayer glomerular microvessel was evaluated after treatment with different concentrations of Adriamycin (Fig. 6B, C). Cell viability in the bilayer glomerular microvessel was reduced after Adriamycin treatment. Notably, the podocyte functional genes (nephrin, podocin), endothelial cell-specific genes (CD31, eNOs), and tight junction marker (ZO-1) were also significantly decreased in the bilayer glomerular microvessel exposed to Adriamycin (Fig. 6D). In contrast, inflammatory marker VCAM-1 expression levels were drastically upregulated after Adriamycin exposure to the bilayer glomerular microvessel. Further, the immunofluorescence staining with ZO-1 and nephrin also confirmed the Adriamycin-induced injury state of the glomerular bilayer (Fig. 6E, F).
Proteinuria is routinely used in the clinic to assess renal injury (glomerular damage) 11. We have also assessed the albumin barrier function of bGOAC with and without treatment by perfusing the large molecular complex FITC–albumin. The glomerular chip treated with Adriamycin displayed remarkable albumin loss from the bilayer glomerular lumen and augmented the level of albumin leakage into the urinary part (Fig. 6G), as is detected in the course of Adriamycin-induced glomerular damage in vivo. Furthermore, the quantitative albumin clearance from the bGOAC was evaluated by perfusing FITC–albumin to normal and Adriamycin-injured glomerular microvessels (Fig. 6H). The printed glomerular model exposed with Adriamycin displayed substantial albumin leakage from the bilayer glomerular microvessel (40% loss of albumin) and increased albumin entry into the urinary part (Fig. 6H).
Diabetic nephropathy modeling within bGOAC
Diabetic nephropathy is a severe diabetic microvascular complication due to inadequate glycemic control, and is responsible for progressive kidney disease in diabetic patients. Hyperglycemia-associated injury of renal glomerular endothelial cells is acknowledged as the main instigation factor of diabetic nephropathy pathogenesis in diabetic conditions. Consequently, albumin leakage across the GFB is increased, which is an important pathological characteristic of diabetic nephropathy.
To demonstrate that our cell-printed bGOAC could mimic the hyperglycemia-induced glomerular injury, we perfused the media containing various concentrations of glucose (5, 15, 20 mM) in mature bGOAC for 3 days (Fig. 7A). Interestingly, we observed that exposure to high glucose concentration upregulated the expression level of the inflammatory marker intercellular adhesion molecule 1 (ICAM-1), and downregulated the expression of the glomerular functional markers (nephrin, eNOs) (Fig. 7B).
In addition, immunostained bGOACs with nephrin and ZO-1 showed that the expressions of nephrin and ZO-1 were drastically suppressed in diabetic bGOAC compared with normal bGOAC (Fig. 7C, D). The suppression of the tight junction protein ZO-1 expression in the injured glomerular leads to the enlargement of the filtration barrier. As a result, albumin clearance is promoted, which is the key clinical sign of diabetic nephropathy (Fig. 7E-G).
Altogether, results demonstrate that this cell-printed glomerular model signifies a distinctive platform to study glomerular pathophysiology.