micromachines Article
A Microfluidic Platform for Investigating Transmembrane Pressure-Induced Glomerular Leakage Ting-Hsuan Chen 1 , Jie-Sheng Chen 2 , Yi-Ching Ko 3 , Jyun-Wei Chen 2 , Hsueh-Yao Chu 2 , Chih-Shuan Lu 4 , Chiao-Wen Chu 2 , Hsiang-Hao Hsu 3, * and Fan-Gang Tseng 2,4, * 1 2
3 4
*
Department of Mechanical and Biomedical Engineering, City University of Hong Kong, Hong Kong, China;
[email protected] Department of Engineering and System Science, Frontier Research Center on Fundamental and Applied Sciences of Matters, National Tsing Hua University, Hsinchu 30013, Taiwan;
[email protected] (J.-S.C.);
[email protected] (J.-W.C.);
[email protected] (H.-Y.C.);
[email protected] (C.-W.C.) Department of Nephrology, Kidney Research Center, Chang Gung Memorial Hospital, Chang Gung University, College of Medicine, Taoyuan 33305, Taiwan;
[email protected] Institute of Nano Engineering and Microsystems, National Tsing Hua University, Hsinchu 30013, Taiwan;
[email protected] Correspondence:
[email protected] (H.-H.H.);
[email protected] (F.-G.T.); Tel.: +886-3-328-1200 (ext. 8181) (H.-H.H.); +886-3-571-5131 (ext. 34270) (F.-G.T.)
Received: 13 April 2018; Accepted: 3 May 2018; Published: 10 May 2018
Abstract: Transmembrane pressure across the glomerular filter barrier may underlie renal failure. However, studies of renal failure have been difficult owing to a lack of in vitro models to capture the transmembrane pressure in a controlled approach. Here we report a microfluidic platform of podocyte culture to investigate transmembrane pressure induced glomerular leakage. Podocytes, the glomerular epithelial cells essential for filtration function, were cultivated on a porous membrane supplied with transmembrane pressure ∆P. An anodic aluminum oxide membrane with collagen coating was used as the porous membrane, and the filtration function was evaluated using dextrans of different sizes. The results show that dextran in 20 kDa and 70 kDa can penetrate the podocyte membrane, whereas dextran in 500 kDa was blocked until ∆P ≥ 60 mmHg, which resembles the filtration function when ∆P was in the range of a healthy kidney (∆P < 60 mmHg) as well as the hypertension-induced glomerular leakage (∆P ≥ 60 mmHg). Additionally, analysis showed that synaptopodin and actin were also downregulated when ∆P > 30 mmHg, indicating that the dysfunction of renal filtration is correlated with the reduction of synaptopodin expression and disorganized actin cytoskeleton. Taking together, our microfluidic platform enables the investigation of transmembrane pressure in glomerular filter membrane, with potential implications for drug development in the future. Keywords: podocytes; glomerular leakage; transmembrane pressure; microfluidics
1. Introduction Transmembrane pressure across the basement membrane that separates epithelium and vascular endothelium may underlie the transportation of fluids, nutrients, or other metabolic factors across tissue interfaces and affect neighboring parenchymal tissues. In particular, it has an important role in glomeruli, the basic units of the kidney’s filtration function. Inside a glomerulus, the glomerular filter barrier consists of three cell layers [1]. Among them, the podocyte is a highly specialized cell type
Micromachines 2018, 9, 228; doi:10.3390/mi9050228
www.mdpi.com/journal/micromachines
Micromachines 2018, 9, 228
2 of 10
responsible for the glomerular permselectivity of proteins in the kidney [2]. Differentiated podocytes are mesenchymal-like cells with foot processes, a highly branched inter-digitating network. In between foot processes of opposing podocytes, the filtration slits establish the final barrier to urinary protein loss [3]. Thus, the factors that change the podocyte phenotype have attracted considerable attention as small changes of actin cytoskeleton may result in disappearance of actin-enriched foot processes and damage the filtration function [4–6], eventually causing leakage of blood cells or serum proteins into the urine [7,8]. Recent results suggest that net filtration pressure or transmembrane pressure may govern the fluidic movement across glomerular basement membrane essential for the filtration [9]. In normal conditions, the net filtration pressure is 10–30 mmHg [9]. However, it may exceed 60 mmHg in the condition of hypertension [10]. Thus, the pronounced increase of trans-glomerular pressure gradient that results in the enhanced mechanical stress may be an important cause of podocyte injury [11,12] and can eventually disrupt the renal filtration barrier [13]. There are many mechanical platforms for studying mechanical factors on cell activities. For example, microfluidics has been used to study shear stress in endothelial cell alignment and adhesion [14–16]. Mechanical compression or stretching was also applied to the entire culture [17–19] for studying the force dependence of cell orientation. In particular, because cells are pre-stressed with internal contractile forces, attachment to a more compliable substrate would cause a reduction in cell size as well as the internal forces [20–22], which later results in faster migration and optimized proliferation [22]. However, the existing platforms lack the ability to recapture the transmembrane pressure for cultured cells in a controllable manner [23], creating a difficulty when investigating the role of transmembrane pressure in physiology/pathology of glomerulus. Here we demonstrate a microfluidics-based in vitro model to investigate transmembrane pressure induced glomerular leakage. The schematic is shown in Figure 1. The glomerular filtration membrane was represented by a monolayer of podocytes with 100% confluence. The podocyte monolayer was cultured on an anodic aluminum oxide (AAO) membrane consisting of nanoscale transmembrane pores [11]. The membrane was sandwiched between two acrylic chambers to provide the transmembrane pressure (∆P), i.e., the net pressure difference between PH in the lower chamber to simulate glomerular capillaries hydrostatic pressure and PL in the upper chamber to simulate the combination of colloid osmotic pressure and Bowman’s hydrostatic pressure. A collagen-coated anodic aluminum oxide membrane was used to simulate the glomerular basement membrane. The filtration function was validated using dextran with different sizes. We found that dextran in 20 kDa and 70 kDa can penetrate the podocyte membrane whereas dextran in 500 kDa was blocked until ∆P ≥ 60 mmHg, which resembles the filtration function when ∆P was in the range of healthy kidney (∆P < 60 mmHg) as well as the hypertension-induced glomerular leakage (∆P ≥ 60 mmHg). Also, such damage was not self-repairable even if 10 mmHg of ∆P was subsequently applied. With the evidences showing downregulated expression of synaptopodin and disorganized actin cytoskeleton in high ∆P, our model suggests that high transmembrane pressure may result in a change of podocyte phenotype, eventually disrupting the filtration function. As an in vitro model recapturing transmembrane pressure in physiology/pathology of glomerulus, this approach paves the way for future drug development and potential therapy for pressure-induced renal failure.
Micromachines 2018, 9, 228 Micromachines 2018, 9, x FOR PEER REVIEW
3 of 10 3 of 10
Figure platform. (A) Figure 1. 1. Overview Overview of of the the microfluidic microfluidic platform. (A) Schematics Schematics of of the the device device where where podocytes podocytes were were cultured on porous membrane, which was sandwiched by an upper chamber supplying PLa and a cultured on porous membrane, which was sandwiched by an upper chamber supplying PL and lower H. The pressures in both chambers were controlled by peristaltic lower chamber supplying chamber supplying with PHwith . ThePpressures in both chambers were controlled by peristaltic pumps that pumps that infused the medium via inletand andmonitored outlet andbymonitored by pressure sensors. infused the medium via the inlet andthe outlet pressure sensors. (B) An image(B) of An the image of the real device; (C) a microscopic image of the differentiated podocytes monolayer. real device; (C) a microscopic image of the differentiated podocytes monolayer. Scale bar, 50 µm.Scale bar, 50 μm.
2. Materials and Methods 2. Materials and Methods 2.1. Podocyte Culture 2.1. Podocyte Culture We used conditionally immortalized cells with stable expression of nephrin and formation We used conditionally with stable expression of nephrin formation of of cell–cell contacts derivedimmortalized by Endlich’s cells group [24]. To cultivate them, in brief,and podocytes were cell–cell contacts derived by Endlich’s group [24]. To cultivate them, in brief, podocytes were transfected with a temperature-sensitive T antigen such that they become conditionally immortalized transfected temperature-sensitive antigen such that VA, they become conditionally at 33 ◦ C. The with growtha medium consists of RPMI T 1640 (Corning, Corning, USA) and 10% fetal bovine immortalized at 33 °C. The growth medium consists of RPMI 1640 (Corning, Corning, VA, USA) and serum (Corning, Corning, VA, USA) supplemented with 20–100 U/mL of mouse interferon gamma 10% fetal bovine serum (Corning, Corning, VA, USA) supplemented with 20–100 U/mL of mouse (INF-γ, Miltenyi Biotec, Bergisch Gladbach, Germany) to drive T-antigen expression. Later, to inactivate interferon gamma (INF-γ, Miltenyi Biotec, cycle Bergisch Gladbach,differentiation, Germany) to the drive T-antigen the T antigen that exits the cell proliferation and promotes cultures were expression. Later, to inactivate the T antigen that exits the cell proliferation cycle and promotes ◦ switched to medium lacking INF and transferred to an incubator set at 37 C [25]. differentiation, the cultures were switched to medium lacking INF and transferred to an incubator 2.2.atPodocyte Spreading and Proliferation on Porous Membrane set 37 °C [25]. To resemble the filtration barrier, porous membranes were used in the podocyte culture to mimic 2.2. Podocyte Spreading and Proliferation on Porous Membrane the glomerular basement membrane. The AAO membranes (Whatman, Inc., Little Chalfont, UK) resemble porous membranes were used the podocyte to withTo pore size 0.02the µmfiltration or 0.2 µmbarrier, were used to culture the podocytes. Forincomparison, weculture also used mimic the glomerular basement membrane. The AAO membranes (Whatman, Inc., Little Chalfont, polycarbonate (PC) membranes with 0.5 µm and 5 µm pore size (Whatman, Inc., Little Chalfont, UK) UK) size 0.02 μm or 0.2 μm were used to culture the podocytes. For comparison, we also withwith lowerpore porosity. used Before polycarbonate membranes with 0.5 was μmsterilized and 5 μm porealcohol size (Whatman, Inc., Little plating the(PC) podocytes, the membrane in 95% for 10 min and exposed Chalfont, lower porosity. to UV for UK) 24 h. with The membranes were either directly plated with cells or pre-coated with type I collagen Before plating the0.8membrane was sterilized in 95% forThe 10 podocytes min and (BD Biosciences, San the Jose,podocytes, CA, USA) in mg/mL solution for 24 h before cellalcohol seeding. 2 exposed to UV for 24 h. The membranes were either directly plated with cells or pre-coated with were plated on the membranes at a density of 10,000 cells/cm and cultured in a six-well culture ◦ type collagen (BDcell Biosciences, San the Jose, CA, USA) 0.8 mg/mL solution 24 hofbefore cell dish. I To study the proliferation, cultures were in conducted at 33 C forfor a total five days. 2 and cultured seeding. The podocytes were plated on the was membranes a ◦density 10,000growth cells/cm For studying the cell spreading, the culture started atat33 C with of regular medium for 48 h in a six-well culture the48cell proliferation, the cultures were conducted at 33 medium °C for a and then changed to dish. 37 ◦ CTo forstudy another h using culture medium lacking INF-γ. The culture total of five days. the of cell spreading, the was started at 33 °C regular was changed everyFor day.studying At the time measurement, the culture proliferation was obtained by with counting the growth medium for 48 h and then changed to 37 °C for another 48 h using culture medium lacking increase in cell numbers. For assessing the cell spreading, live cells were stained by 0.1% Calcein AM INF-γ. The cultureCarlsbad, medium CA, was USA) changed every the proliferation (Life Technology, in PBS at day. 37 ◦ CAt forthe 15time min of to measurement, observe the morphology in the was obtained counting the increase in cell numbers. For assessing the cell spreading, live cells following daysby using a fluorescence microscope. were stained by 0.1% Calcein AM (Life Technology, Carlsbad, CA, USA) in PBS at 37 °C for 15 min to observe the morphology in the following days using a fluorescence microscope.
Micromachines 2018, 9, 228
4 of 10
2.3. Pressure-Supplying Device The membrane was sandwiched between upper and lower acrylic chambers separately supplied with fresh medium (Figure 1). The pressure difference was achieved by peristaltic pumps and monitored by pressure sensors. Before attaching to the chambers, the membrane was coated with collagen for 30 min. Next, podocytes with 10,000 cells/cm2 were first cultured at 33 ◦ C for three days to reach 60% confluence, followed by eight days in 37 ◦ C for differentiation. Next, the membrane was sandwiched by the acrylic chambers and placed into an on-stage incubator to keep the culture under 37 ◦ C and 5% CO2 infusion. The upper chamber was supplied with lower pressure (PL ), which was adjusted for different ∆P, while the lower chamber was supplied with higher pressure (PH ) and kept as a constant. Note that podocytes were facing the chamber with lower pressure to simulate in vivo circumstances. 2.4. Dextran Filtration The ∆P was applied to the AAO membrane with confluent podocytes for 2 h. To evaluate the filtration function with different ∆P, the medium in both the upper and lower chambers was first removed, followed by infusing the medium with dextran into the lower chamber only. Dextrans with three representative sizes were used: 20 kDa, 70 kDa, and 500 kDa (Nanocs Inc., New York, NY, USA). One minute after infusion, the medium that penetrated from the lower chamber to the upper chamber was collected. To evaluate the self-repairing, 70 mmHg of ∆P was applied for 6 h, followed by supplying lower ∆P (10 mmHg) for another 6 h. Dextrans penetration samples were collected at 2 h, 6 h, 8 h and 12 h. The amount of dextran was measured using its fluorescent intensity by a plate reader (Victor3 , Perkin Elmer, Waltham, MA, USA). 2.5. Quantitative Real-Time Polymerase Chain Reaction (q-PCR) Total RNA was isolated using TRIzol reagent (Life Technology) according to the manufacturer’s instructions. q-PCR was performed on ViiA™ 7 Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). The gene expression was analyzed using TaqMan® Gene Expression Assay (Life Technology) containing FAM dye-labeled probe and primers (Assay ID: GAPDH, Mm99999915_g1; 18 S, Hs99999901_s1; αActin-4, Mm00502489_m1, and Synaptopodin, Mm03413333_m1) following manufacturer’s instructions. The comparative ∆CT method was used for relative quantification of mRNA expression. 18S were used as endogenous controls to obtain the ∆CT , and wild-type podocytes cultured in 37 ◦ C without ∆P were used as the calibrator to obtain the ∆∆CT . Finally, the relative mRNA expression was quantified by 2−∆∆CT as the final readout. 2.6. Fluorescence Staining of Actin After application of ∆P, the cells were fixed in 4% formaldehyde at 4 ◦ C for 15 min, followed by PBS rinsing three times. Next, 1% BSA in PBS was applied for 30 min for blocking, followed by incubation of phalloidin (1:100, Sigma-Aldrich, St. Louis, MO, USA) for 1 h and rinsing by PBS before image acquisition. 3. Results 3.1. Cell Proliferation on Porous Membranes We first studied the podocyte proliferation when cultured on porous membrane. Podocytes have been the center of focus because of the formation of foot processing and slit diaphragms, which are known as a final barrier to prevent urinary protein loss [26–28]. More importantly, unlike vascular endothelial cells, podocytes do not regenerate, suggesting a closer relationship to the progressive loss of renal filtration function observed in chronic kidney disease. Thus, while microvascular endothelial cells may also play a role in glomerular filtration and participate in the progression of glomerular disease [29], podocytes were the focus here in order to investigate the self-perpetuation of renal failure.
Micromachines Micromachines2018, 2018,9,9,228 x FOR PEER REVIEW
5 of of10 10
Podocytes cultured on an AAO membrane were compared with those on PC membrane with Podocytes on anpetri AAO membrane were compared with those on PCmicroscopy membrane(SEM) with lower porositycultured and regular dish with zero porosity. Scanning electronic lower porosity and regular petri dish with zero porosity. Scanning electronic microscopy (SEM) showed showed that the AAO membrane is compo and kept as a constant sed of nanoscale pores with a that the AAO membrane compo and kept as a closer constant pores with a density as high density as high as 50%, is providing a topology to sed thatofofnanoscale the glomerular basement membrane. as 50%, providing a topology closer to that of the glomerular basement membrane. Also, the actual Also, the actual pore size of AAO membranes 0.02 μm is in fact much larger than 0.02 µm (Figure pore size of AAO membranes 0.02 µm is in fact much larger than 0.02 µm (Figure 2A–D), which allows 2A–D), which allows for the penetration of a big dextran fraction of 500 kDa, which has a Stoke’s for the penetration of a big dextran fraction of 500 kDa, which has a Stoke’s radius of 10–15 nm. radius of 10–15 nm. The results of cell The results of cellproliferation proliferationshowed showedthat, that,on onan anAAO AAOmembrane membraneand andon onaapetri petridish dishwithout without ◦ any coating, the cell number doubled in about 48 h at 33 C, and they grew faster on 0.02 µm pores than any coating, the cell number doubled in about 48 h at 33 °C, and they grew faster on 0.02 μm pores on 0.2 µm pores or on a petri dish. By contrast, on a PC membrane, increased cell death was observed than on 0.2 μm pores or on a petri dish. By contrast, on a PC membrane, increased cell death was regardless the PC membrane’s size (Figure 2E). Thus, the results indicate podocytes prefer observed of regardless of the PC pore membrane’s pore size (Figure 2E). Thus, thethat results indicate that to grow on an AAO membrane or on a petri dish than on a PC membrane. Also, smaller pores on the podocytes prefer to grow on an AAO membrane or on a petri dish than on a PC membrane. Also, AAO membrane provide an additional positive effect for cell proliferation. smaller pores on the AAO membrane provide an additional positive effect for cell proliferation. Next, Next,we wecoated coatedcollagen collagenon onthe thesame samesets setsof ofmembranes membranesfor for24 24hhprior priortotoplating platingcells. cells.As Asthe the topology topology remained, remained, the the material material difference difference between between the the AAO, AAO,petri petridish, dish,and andPC PCmembrane membranewas was overridden overriddenby bythe thecoverage coverageof ofcollagen. collagen. The The results results showed showed that that the thecell cellnumber numberincreased increasedfor forall all cases (Figure 2F). Of note, on a PC membrane with a collagen coating, the growth rate rose to a level cases (Figure 2F). Of note, on a PC membrane with a collagen coating, the growth rate rose to a level similar onon AAO membranes, suggesting that the proliferation on a PCon membrane without similartotothat that AAO membranes, suggesting thatlower the lower proliferation a PC membrane coating is due to the material difference but not the topology. without coating is due to the material difference but not the topology.
Figure 2. SEM images for AAO membrane with (A) 0.2 μm and (B) 0.02 μm pores, and PC membrane Figure 2. SEM images for AAO membrane with (A) 0.2 µm and (B) 0.02 µm pores, and PC membrane with (C) 0.5 μm and (D) 5 μm pores. Scale bar, 1 μm (A,C), 100 nm (B), and 50 μm (D). (E) The with (C) 0.5 µm and (D) 5 µm pores. Scale bar, 1 µm (A,C), 100 nm (B), and 50 µm (D). (E) The growth growth curve of podocytes on AAO or PC membrane without collagen coating. (F) The growth curve curve of podocytes on AAO or PC membrane without collagen coating. (F) The growth curve of podocytes of AAO podocytes AAO orwith PC membrane with collagen (E,F), shown on or PC on membrane collagen coating. For (E,F),coating. data areFor shown asdata meanare ± SD (n = as 3). mean ± SD (n = 3).
3.2. Cell Spreading on Porous Membranes 3.2. Cell Spreading on Porous Membranes We further investigated the cell spreading in response to different membranes. Calcein staining We further investigated the cell spreading in response to different membranes. Calcein staining was used to reveal the cell shapes. The results showed that, when collagen coating is absent, cells was used to reveal the cell shapes. The results showed that, when collagen coating is absent, cells on on AAO membranes spread to a higher degree than on PC membranes and on a petri dish (Figure 3, AAO membranes spread to a higher degree than on PC membranes and on a petri dish (Figure 3, upper row). This suggests that AAO membranes can promote cell spreading, which may explain the upper row). This suggests that AAO membranes can promote cell spreading, which may explain the increased proliferation before applying a collagen coating. We then tested whether the collagen coating, increased proliferation before applying a collagen coating. We then tested whether the collagen which was shown to increase cell proliferation, can also have a positive effect on cell spreading. After coating, which was shown to increase cell proliferation, can also have a positive effect on cell coating with collagen on AAO and a PC membrane, podocytes showed higher spreading than that on spreading. After coating with collagen on AAO and a PC membrane, podocytes showed higher non-coated ones (Figure 3, lower row). Such an increase was also observed in cells on a petri dish. spreading than that on non-coated ones (Figure 3, lower row). Such an increase was also observed in cells on a petri dish.
Micromachines 2018, 9, 228 Micromachines 2018, 9, x FOR PEER REVIEW
6 of 10 6 of 10
Micromachines 2018, 9, x FOR PEER REVIEW
6 of 10
0.02 μm AAO
0.2 μm AAO
0.5 μm PC
5 μm PC
Culture Dish
0.02 μm AAO
0.2 μm AAO
0.5 μm PC
5 μm PC
Culture Dish
0.02 μm AAO + Collagen
0.2 μm AAO + Collagen
0.5 μm PC + Collagen
5 μm PC + Collagen
Culture Dish + Collagen
0.02 μm AAO + Collagen
0.2 μm AAO + Collagen
0.5 μm PC + Collagen
5 μm PC + Collagen
Culture Dish + Collagen
Figure 3. Podocytes incubated at 33 °C for 48 h. The images were taken by fluorescence microscopy. Figure 3. Podocytes incubated at 33 ◦ C for 48 h. The images were taken by fluorescence microscopy. Figure 3. Podocytes incubated at 33 °C for 48 h. The images were taken by fluorescence microscopy. Scale bar, 5 μm. Scale bar, 5 µm. Scale bar, 5 μm.
We also cultured podocytes at 37 °C and studied their morphology (Figure 4). The effect is ◦ C and studied their morphology (Figure 4). The effect is Wetoalso cultured podocytes at 37 We also cultured podocytes at 33 37°C. °CHowever, and studied their morphology (Figure 4).°C, The effectcell is similar the results from culture at compared to cells cultured at 33 where ◦ ◦ C, where cell similar to the results from culture at 33 C. However, compared to cells cultured at 33 similar to theseems resultspolygonal from culture 33 °C. However, to 37 cells 33 °C, cell morphology andatplane, podocytes compared cultured in °Ccultured spread at with an where arborized ◦ C spread with an arborized shape morphology seems polygonal and plane, podocytes cultured in 37 morphology seems polygonal and plane, podocytes cultured in 37 °C spread with an arborized shape and extended more foot processes (Figure 4, upper row). When the membranes were coated and extended more foot 4, upper row). When the were coated with shape and extended moreprocesses foot processes (Figure upper row). When the membranes were with collagen, this morphology was(Figure observed on4,both AAO and PCmembranes membranes (Figure 4, coated lower collagen, this morphology was observed on both AAO and PC membranes (Figure 4, lower row). with collagen, this morphology was observed AAOproliferation, and PC membranes (Figure was 4, lower row). Thus, consistent with the former results on of both podocyte cell spreading also Thus, consistent with the former results of podocyte proliferation, cell spreading was also increased by row). Thus, consistent with the former results of podocyte proliferation, cell spreading was also increased by collagen coating regardless of the underneath materials. collagen coating regardless of the underneath materials. increased by collagen coating regardless of the underneath materials.
0.02 μm AAO
0.2 μm AAO
0.5 μm PC
5 μm PC
Culture Dish
0.02 μm AAO
0.2 μm AAO
0.5 μm PC
5 μm PC
Culture Dish
0.02 μm AAO + Collagen
0.2 μm AAO + Collagen
0.5 μm PC + Collagen
5 μm PC + Collagen
Culture Dish + Collagen
0.02 μm AAO + Collagen
0.2 μm AAO + Collagen
0.5 μm PC + Collagen
5 μm PC + Collagen
Culture Dish + Collagen
Figure 4. Podocytes incubated at 37 °C for 48 h. The images were taken by fluorescence microscopy. Figure 4. Podocytes incubated Scale bar, 5 μm. ◦ C for Figure 4. Podocytes incubated at at 37 37 °C for 48 48 h. h. The The images images were were taken takenby byfluorescence fluorescencemicroscopy. microscopy. Scale bar, 5 μm. Scale bar, 5 µm.
3.3. Filtration Function in Response to ΔP 3.3. Filtration Function in Response to ΔP 3.3. Filtration in Response to ∆P We next Function implemented the transmembrane pressure to evaluate the filtration function. We We an next implemented the transmembrane pressure the of filtration function. We selected AAO membrane with a 0.02 μm pore size becausetoit evaluate is composed nanoscale pores with We next implemented the transmembrane pressure to evaluate the filtration function. We selected an AAO membrane with a 0.02 μm pore size because itthe is composed ofbasement nanoscalemembrane. pores with aselected density as high as 50%, providing a topology closer to that of glomerular an AAO membrane with a 0.02 µm pore size because it is composed of nanoscale pores with a density as a density as high as 50%, providing topology closer that of the glomerular basement Also, collagen coating was used asasuggested by thetostudies of cell proliferation and membrane. spreading. high as 50%, providing a topology closer to that of the glomerular basement membrane. Also, collagen Also, collagen coating was used as suggested podocytes by the studies cell proliferation and and spreading. Membranes with confluent and differentiated wereofattached to the upper lower coating was used as suggested by the studies of cell proliferation and spreading. Membranes with confluent Membranes with confluent and differentiated podocytes were attached to the upper and lower acrylic pressure chambers. Medium was pumped into the chambers and ΔP was monitored by and differentiated podocytes were attached to the upper and lower acrylic pressure chambers. Medium acrylic pressure chambers. Medium was1). pumped into the chambers ΔP was by pressure meters in each chamber (Figure After supplying ΔP for 2 h,and dextran wasmonitored added to the was pumped into the chambers and ∆P was monitored by pressure meters in each chamber (Figure 1). pressure meters each chamber (Figure with 1). After supplying ΔP for 2can h, dextran added to the medium to see in whether a membrane confluent podocytes preventwas transmembrane After supplying ∆P for 2 h, dextran was added to the medium to see whether a membrane with confluent medium to Three see whether a membrane with confluent podocytes can prevent penetration. representative sizes of dextran were used: 20 kDa represents smalltransmembrane molecules that podocytes can prevent transmembrane penetration. Three representative sizes of dextran were used: penetration. Threethe representative sizes of dextran used: 20 kDa represents small molecules that usually penetrate filtration membrane; 70 kDawere represents intermediate molecules that penetrate 20 kDa represents small molecules that usually penetrate the filtration membrane; 70 kDa represents usually penetrate filtration membrane; 70 kDa represents intermediate molecules that penetrate the membrane butthe are reabsorbed by tubule in vivo; and 500 kDa represent big molecules that intermediate molecules that penetrate the membrane but are reabsorbed by tubule in vivo; and 500 kDa the membrane reabsorbed by tubule intime vivo;ofand 500 kDa represent bigmedium molecules that cannot penetratebut theare filtration membrane. At the measurement, the regular in both represent big molecules that cannot penetrate the filtration membrane. At the time of measurement, cannotand penetrate the filtration At the time of measurement, medium both upper lower chamber wasmembrane. removed, followed by infusing medium the withregular dextran into theinlower the regular medium in both upper and lower chamber was removed, followed by infusing medium upper and lower chamber by infusing mediuminto withthe dextran the lower chamber only. One minutewas afterremoved, infusion,followed the medium that penetrated upper into chamber was with dextran into the lower chamber only. One minute after infusion, the medium that penetrated into chamber and only.quantified One minute after the medium thatThe penetrated into the upper chamber was collected based oninfusion, its fluorescent intensity. result showed that the penetration of the upper chamber was collected and quantified based on its fluorescent intensity. The result showed collectedwith andsize quantified its increased fluorescent intensity. The result showed that themmHg penetration of dextran 20 kDa based and 70on kDa when ΔP was increased from 10 to 60 (Figure dextran with size 20 kDa and 70 kDa increased when ΔP was increased from 10 to 60 mmHg (Figure 5). In addition, dextran with size 500 kDa was blocked when ΔP < 60 mmHg. This demonstrates the 5). In addition, dextran with size 500 kDa was blocked when ΔP < 60 mmHg. This demonstrates the
Micromachines 2018, 9, 228
7 of 10
that the penetration of dextran with size 20 kDa and 70 kDa increased when ∆P was increased from 10 to 60 mmHg (Figure 5). In addition, dextran with size 500 kDa was blocked when ∆P < 60 mmHg. Micromachines 2018, x FOR PEERREVIEW REVIEW 7 of7 10 2018, 9, 9, x FOR PEER of 10 This Micromachines demonstrates the filtration capability of the podocyte culture resembling the normal physiological conditions. More importantly, for ∆P ≥ 60 mmHg, dextran with size of 500 kDa startedMore toMore penetrate filtration capability thepodocyte podocyte culture resembling the physiological conditions. filtration capability ofofthe culture resembling thenormal normal physiological conditions. importantly, for ΔP ≥ 60 mmHg, dextran with size of 500 filtration kDa started to penetrate through the leakage. through the membrane, showing the malfunction of podocyte mimicking the glomerular importantly, for ΔP ≥ 60 mmHg, dextran with size of 500 kDa started to penetrate through the membrane, showing the malfunction of podocyte filtrationthe mimicking the glomerular leakage. Thus, by this Thus,membrane, when lowshowing ∆P was the supplied, the result demonstrates recapitulation of filtration function malfunction of podocyte filtration mimicking the glomerular leakage. Thus, when low ΔP was supplied, the result demonstrates the recapitulation of filtration function by this in in vitro model. When high ∆P was supplied, the simulated hypertension the filtration when low ΔP was supplied, the result demonstrates the recapitulation of damaged filtration function by thisfunction, in vitro model. When high ΔP was supplied, the simulated hypertension damaged the filtration vitro model. When high ΔP was supplied, the simulated hypertension damaged the filtration suggesting the in vivo-likethe response. function, suggesting in vivo-like response. function, suggesting the in vivo-like response.
Figure 5. Filtration capability of podocyte membrane using dextran with sizes of 20 kDa, 70 kDa, or
Figure 5. Filtration capability of podocyte membrane using dextran with sizes of 20 kDa, 70 kDa, or 500 500 kDa. Data are shown as mean ± SD (n = 3). 5. Filtration capability podocyte kDa.Figure Data are shown as mean ±ofSD (n = 3).membrane using dextran with sizes of 20 kDa, 70 kDa, or
500One kDa.ofData shown as mean SD (n = 3). malfunction is self-perpetuation, meaning that the the are crucial features of± glomerular condition will progress to terminal renal failure even if is theself-perpetuation, initial cause is removed. Thus, askcondition One One of the crucial features of glomerular malfunction meaning thatwethat the theloss crucial features of glomerular meaning whetherofthe of filtration function in this inmalfunction vitro modelisisself-perpetuation, self-healable. To address this, wethe will progress towill terminal renal failure even iffailure the initial cause isinitial removed. Thus, we ask whether the loss condition progress to terminal if theΔP is or removed. Thus, investigated the podocyte filtrationrenal function by even applying at 30cause mmHg 70 mmHg for we 6 h,ask of filtration function in this in vitro model is self-healable. To address this, we investigated the podocyte whether the loss of filtration function in this in vitro model is self-healable. To address this, we followed by applying 10 mmHg another 6 h. Under moderate ΔP, 30 mmHg, results showed that investigated the podocyte filtration ΔPfor at or 70 forΔP 6 h, filtration function by applying ∆P at 30function mmHgby or applying 70 mmHg 630h,mmHg followed bybymmHg applying 10 mmHg there was no loss of 500 kDa dextran observed during the entire course. In contrast, applying followed by applying 10 mmHg another 6 h. Under moderate ΔP, 30 mmHg, results showed that 6 h, which∆P, is the of hypertension, the dextran loss increased 0%500 to 77.5% anotherat670 h.mmHg Underfor moderate 30level mmHg, results showed that there was no from loss of kDa dextran there was no loss of 500 course. kDa observed theisentire course. In contrast, by applying (Figure 6). Importantly, thisdextran damaged filtration function notat fully self-repairable. applying observed during the entire In contrast, byduring applying ∆P 70 mmHg for 6 h,After which is theΔP level of ΔPmmHg at 10 mmHg another h, there still 67.9% of dextran loss for 500 kDa dextran, at 70 for 6 for h, which is6the level was of hypertension, the dextran loss increased fromsuggesting 0% to 77.5% hypertension, the dextran loss increased from 0% to 77.5% (Figure 6). Importantly, this damaged filtration that this in vitro modelthis is also self-perpetuated when high ΔP fully is applied, which is similar to the (Figure 6). Importantly, damaged filtration function is not self-repairable. After applying functionpathological is not fullybehavior self-repairable. After applying ∆P at 10 mmHg for another 6 h, there was still 67.9% of observed in vivo. ΔP at 10 mmHg for another 6 h, there was still 67.9% of dextran loss for 500 kDa dextran, suggesting
dextran loss for 500 kDa dextran, suggesting that this in vitro model is also self-perpetuated when high ∆P that this in vitro model is also self-perpetuated when high ΔP is applied, which is similar to the is applied, whichbehavior is similarobserved to the pathological behavior observed in vivo. pathological in vivo.
Figure 6. Podocyte filtration function. ΔP of 30 mmHg and 70 mmHg were applied for 6 h, followed by 10 mmHg for the following 6 h to investigate the recovery of filtration function. Data are shown as mean ± SD (n = 3).
Figure 6. Podocyte filtration function. ∆P ΔP of of 30 30 mmHg were applied for 6for h, followed Figure 6. Podocyte filtration function. mmHgand and7070mmHg mmHg were applied 6 h, followed by 10 mmHg for the following 6 h to investigate the recovery of filtration function. Data are shown as by 10 mmHg for the following 6 h to investigate the recovery of filtration function. Data are shown as mean ± SD (n = 3). mean ± SD (n = 3).
Micromachines 2018, 9, 228 Micromachines 2018, 9, x FOR PEER REVIEW
8 of 10 8 of 10
3.4. Synaptopodin Synaptopodin and and Actin Actin Expression Expression 3.4. We further further investigated investigated whether whether the the loss loss of of filtration filtration function function is is associated associated with with the the change change of of We podocyte phenotype. phenotype.Synaptopodin Synaptopodinisisanan important marker expressed in differentiated podocytes podocyte important marker expressed in differentiated podocytes [23]. [23]. The co-localized synaptopodin on actin filament is important for cytoskeletal rearrangement The co-localized synaptopodin on actin filament is important for cytoskeletal rearrangement and and processes foot processes [23,30]. Based onmRNA the mRNA expression of synaptopodin by q-PCR, we found foot [23,30]. Based on the expression of synaptopodin by q-PCR, we found that thatexpression the expression of synaptopodin decayed the increase ΔP (Figure 7A)became and became the levellevel of synaptopodin decayed with with the increase of ∆Pof(Figure 7A) and even even lower after∆P thewas ΔP switched was switched 10 mmHg for another 6 h (Figure Moreover, by staining lower after the to 10to mmHg for another 6 h (Figure 7B). 7B). Moreover, by staining the the actin, we observed a decrease of expression actin expression after applying = 70 for mmHg forfollowed 6 h and actin, we observed a decrease of actin after applying ∆P = 70ΔP mmHg 6 h and followed bymmHg ΔP = 10 for 6 7C). h (Figure 7C). Combined with the loss of filtration by ∆P = 10 formmHg 6 h (Figure Combined with the results of results the lossofofthe filtration function function (Figure 6), the results suggest that the non-self-healable renal failure due to high (Figure 6), the results suggest that the non-self-healable renal failure due to high transmembrane transmembrane could due to the loss of synaptopodin expression and disorganized pressure could bepressure due to the loss be of synaptopodin expression and disorganized actin cytoskeleton. actin cytoskeleton.
Figure 7. of of (A)(A) synaptopodin by by applying ΔP ∆P for for 6 h,6(B) Figure 7. q-PCR q-PCRanalysis analysisofofthe therelative relativeexpression expression synaptopodin applying h, synaptopodin by applying ΔP = 70 mmHg for 6 h or ΔP = 70 mmHg for 6 h followed by ΔP = 10 (B) synaptopodin by applying ∆P = 70 mmHg for 6 h or ∆P = 70 mmHg for 6 h followed by mmHg 6 h. for Data areData shown meanas± mean SD (n ± ≥ 2). Immunofluorescence staining of actin for ∆P = 10 for mmHg 6 h. are as shown SD (C) (n ≥ 2). (C) Immunofluorescence staining of culture applied with ΔP = 0 mmHg (left), ΔP = 70 mmHg for 6 h (middle) and ΔP = 70 mmHg for 6 h actin for culture applied with ∆P = 0 mmHg (left), ∆P = 70 mmHg for 6 h (middle) and ∆P = 70 mmHg followed by ΔP =by10∆P mmHg for 6 h for (right). for 6 h followed = 10 mmHg 6 h (right).
4. Conclusions 4. Conclusions In summary, we established a microfluidic-based model that recapitulates glomerular leakage In summary, we established a microfluidic-based model that recapitulates glomerular leakage by by transmembrane pressure. Transmembrane pressure is one of the major causes of renal failure. transmembrane pressure. Transmembrane pressure is one of the major causes of renal failure. However, However, owing to the lack of in vitro models, the understanding of the disease mechanism and owing to the lack of in vitro models, the understanding of the disease mechanism and development development of medical treatment remains limited. Based on our microfluidic platform with of medical treatment remains limited. Based on our microfluidic platform with optimized culture optimized culture condition, i.e., AAO membrane with 0.02 μm pores and collagen coating, we condition, i.e., AAO membrane with 0.02 µm pores and collagen coating, we found that 500 kDa found that 500 kDa dextran was blocked until ΔP ≥ 60 mmHg, which resembles the regular filtration dextran was blocked until ∆P ≥ 60 mmHg, which resembles the regular filtration function and function and hypertension-induced glomerular leakage. Importantly, once the filtration function hypertension-induced glomerular leakage. Importantly, once the filtration function was damaged, was damaged, this glomerular leakage was not self-repairable even after the ΔP was switched to 10 this glomerular leakage was not self-repairable even after the ∆P was switched to 10 mmHg for mmHg for a subsequent 6 h. Moreover, the synaptopodin and actin were down regulated when ΔP a subsequent 6 h. Moreover, the synaptopodin and actin were down regulated when ∆P increased, increased, suggesting that the transmembrane pressure induced glomerular leakage is closely associated with the loss of synaptopodin and actin. Together, this microfluidic device will provide a
Micromachines 2018, 9, 228
9 of 10
suggesting that the transmembrane pressure induced glomerular leakage is closely associated with the loss of synaptopodin and actin. Together, this microfluidic device will provide a platform technology to investigate the glomerular leakage, with implications for drug development and treatment of renal failure in the future. Author Contributions: T.-H.C. designed the experiment and analyzed the data; C.-S.L., C.-W.C., J.-S.C., Y.-C.K., J.-W.C. and H.-Y.C. performed the experiments; H.-H.H. and F.-G.T. supervised the project, giving technical and administrative support throughout the duration; T.-H.C. wrote the manuscript and it was edited by H.-H.H. and F.-G.T. Acknowledgments: This work was supported by the Research Grant of Linkou Chang-Gung Memorial Hospital, grant number CMRPG3B1353 & CMRPG3E1292, Chang Gung-Tsing Hua Research Grant (2013, 2015, 2016), MOST grant MOST 104-2221-E-007-072-MY3, and the General Research Fund (project Nos. 21214815, 11277516 and 11217217) from the Hong Kong Research Grant Council. Conflicts of Interest: The authors declare no conflict of interest.
References 1. 2. 3. 4. 5.
6.
7. 8.
9. 10. 11. 12. 13. 14. 15.
16.
Greka, A.; Mundel, P. Cell biology and pathology of podocytes. Annu. Rev. Physiol. 2012, 74, 299–323. [CrossRef] [PubMed] Pavenstadt, H.; Kriz, W.; Kretzler, M. Cell biology of the glomerular podocyte. Physiol. Rev. 2003, 83, 253–307. [CrossRef] [PubMed] Wennmann, D.O.; Hsu, H.H.; Pavenstadt, H. The renin-angiotensin-aldosterone system in podocytes. Semin. Nephrol. 2012, 32, 377–384. [CrossRef] [PubMed] Kerjaschki, D. Caught flat-footed: Podocyte damage and the molecular bases of focal glomerulosclerosis. J. Clin. Investig. 2001, 108, 1583–1587. [CrossRef] [PubMed] Hsu, H.H.; Hoffmann, S.; Di Marco, G.S.; Endlich, N.; Peter-Katalinic, J.; Weide, T.; Pavenstadt, H. Downregulation of the antioxidant protein peroxiredoxin 2 contributes to angiotensin ii-mediated podocyte apoptosis. Kidney Int. 2011, 80, 959–969. [CrossRef] [PubMed] Hsu, H.H.; Hoffmann, S.; Endlich, N.; Velic, A.; Schwab, A.; Weide, T.; Schlatter, E.; Pavenstadt, H. Mechanisms of angiotensin ii signaling on cytoskeleton of podocytes. J. Mol. Med. 2008, 86, 1379–1394. [CrossRef] [PubMed] Shankland, S.J. The podocyte’s response to injury: Role in proteinuria and glomerulosclerosis. Kidney Int. 2006, 69, 2131–2147. [CrossRef] [PubMed] Babelova, A.; Jansen, F.; Sander, K.; Lohn, M.; Schafer, L.; Fork, C.; Ruetten, H.; Plettenburg, O.; Stark, H.; Daniel, C.; et al. Activation of Rac-1 and RhoA contributes to podocyte injury in chronic kidney disease. PLoS ONE 2013, 8, e80328. [CrossRef] [PubMed] Brenner, B.M.; Troy, J.L.; Daugharty, T.M. The dynamics of glomerular ultrafiltration in the rat. J. Clin. Investig. 1971, 50, 1776–1780. [CrossRef] [PubMed] Brenner, B.M. Nephron adaptation to renal injury or ablation. Am. J. Physiol. 1985, 249, F324–F337. [CrossRef] [PubMed] Huang, C.F.; Cheng, H.C.; Lin, Y.; Wu, C.W.; Shen, Y.K. Study on cellar behaviors on different nanostructures by nanoporous alumina template. Int. J. Precis. Eng. Manuf. 2014, 15, 689–693. [CrossRef] Endlich, N.; Kress, K.R.; Reiser, J.; Uttenweiler, D.; Kriz, W.; Mundel, P.; Endlich, K. Podocytes respond to mechanical stress in vitro. J. Am. Soc. Nephrol. 2001, 12, 413–422. [PubMed] Ravera, M.; Re, M.; Deferrari, L.; Vettoretti, S.; Deferrari, G. Importance of blood pressure control in chronic kidney disease. J. Am. Soc. Nephrol. 2006, 17, S98–S103. [CrossRef] [PubMed] Wojciak-Stothard, B.; Ridley, A.J. Shear stress–induced endothelial cell polarization is mediated by Rho and Rac but not cdc42 or PI 3-kinases. J. Cell Biol. 2003, 161, 429–439. [CrossRef] [PubMed] Butcher, J.T.; Penrod, A.M.; García, A.J.; Nerem, R.M. Unique morphology and focal adhesion development of valvular endothelial cells in static and fluid flow environments. Arterioscler. Thromb. Vasc. Biol. 2004, 24, 1429–1434. [CrossRef] [PubMed] Hsiai, T.K.; Cho, S.K.; Reddy, S.; Hama, S.; Navab, M.; Demer, L.L.; Honda, H.M.; Ho, C.M. Pulsatile flow regulates monocyte adhesion to oxidized lipid-induced endothelial cells. Arterioscler. Thromb. Vasc. Biol. 2001, 21, 1770–1776. [CrossRef] [PubMed]
Micromachines 2018, 9, 228
17.
18.
19. 20.
21. 22. 23. 24.
25.
26. 27. 28. 29.
30.
10 of 10
Yoshigi, M.; Hoffman, L.M.; Jensen, C.C.; Yost, H.J.; Beckerle, M.C. Mechanical force mobilizes zyxin from focal adhesions to actin filaments and regulates cytoskeletal reinforcement. J. Cell Biol. 2005, 171, 209–215. [CrossRef] [PubMed] Tse, J.M.; Cheng, G.; Tyrrell, J.A.; Wilcox-Adelman, S.A.; Boucher, Y.; Jain, R.K.; Munn, L.L. Mechanical compression drives cancer cells toward invasive phenotype. Proc. Natl. Acad. Sci. USA 2012, 109, 911–916. [CrossRef] [PubMed] Menon, S.; Beningo, K.A. Cancer cell invasion is enhanced by applied mechanical stimulation. PLoS ONE 2011, 6, e17277. [CrossRef] [PubMed] Engler, A.J.; Griffin, M.A.; Sen, S.; Bonnetnann, C.G.; Sweeney, H.L.; Discher, D.E. Myotubes differentiate optimally on substrates with tissue-like stiffness: Pathological implications for soft or stiff microenvironments. J. Cell Biol. 2004, 166, 877–887. [CrossRef] [PubMed] Engler, A.J.; Sen, S.; Sweeney, H.L.; Discher, D.E. Matrix elasticity directs stem cell lineage specification. Cell 2006, 126, 677–689. [CrossRef] [PubMed] Lei, Y.G.; Gojgini, S.; Lam, J.; Segura, T. The spreading, migration and proliferation of mouse mesenchymal stem cells cultured inside hyaluronic acid hydrogels. Biomaterials 2011, 32, 39–47. [CrossRef] [PubMed] Shankland, S.J.; Pippin, J.W.; Reiser, J.; Mundel, P. Podocytes in culture: Past, present, and future. Kidney Int. 2007, 72, 26–36. [CrossRef] [PubMed] Schiwek, D.; Endlich, N.; Holzman, L.; Holthofer, H.; Kriz, W.; Endlich, K. Stable expression of nephrin and localization to cell-cell contacts in novel murine podocyte cell lines. Kidney Int. 2004, 66, 91–101. [CrossRef] [PubMed] Jat, P.S.; Noble, M.D.; Ataliotis, P.; Tanaka, Y.; Yannoutsos, N.; Larsen, L.; Kioussis, D. Direct derivation of conditionally immortal cell-lines from an h-2kb-tsa58 transgenic mouse. Proc. Natl. Acad. Sci. USA. 1991, 88, 5096–5100. [CrossRef] [PubMed] Haraldsson, B.; Nystrom, J. The glomerular endothelium: New insights on function and structure. Curr. Opin. Nephrol. Hypertens. 2012, 21, 258–263. [CrossRef] [PubMed] Jefferson, J.A.; Alpers, C.E.; Shankland, S.J. Podocyte biology for the bedside. Am. J. Kidney Dis. 2011, 58, 835–845. [CrossRef] [PubMed] Mathieson, P.W. The podocyte as a target for therapies-new and old. Nat. Rev. Nephrol. 2012, 8, 52–56. [CrossRef] [PubMed] Sun, Y.B.Y.; Qu, X.L.; Zhang, X.M.; Caruana, G.; Bertram, J.F.; Li, J.H. Glomerular endothelial cell injury and damage precedes that of podocytes in adriamycin-induced nephropathy. PLoS ONE 2013, 8, e55027. [CrossRef] [PubMed] Mundel, P.; Heid, H.W.; Mundel, T.M.; Kruger, M.; Reiser, J.; Kriz, W. Synaptopodin: An actin-associated protein in telencephalic dendrites and renal podocytes. J. Cell Biol. 1997, 139, 193–204. [CrossRef] [PubMed] © 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
