Hydrogel Tissue Construct-Based High-Content Compound Screening

Abstract

Current pharmaceutical compound screening systems rely on cell-based assays to identify therapeutic candidates and potential toxicities. However, cells grown on 2D substrata or in suspension do not exhibit the mechanical or physiological properties of cells in vivo. To address this limitation, the authors developed an in vitro, high-throughput, 3D hydrogel tissue construct (HTC)–based assay system to quantify cell and tissue mechanical properties and multiple parameters of physiology. HTC mechanics was quantified using an automated device, and physiological status was assessed using spectroscopy-based indicators that were read on microplate readers. To demonstrate the application of this system, the authors screened 4 test compounds—rotenone (ROT), cytochalasin D (CD), 2,4-dinitrophenol (DNP), and Rho kinase inhibitor (H-1152)—for their ability to modulate HTC contractility without affecting actin integrity, mitochondrial membrane potential (MMP), or viability. All 4 compounds dose-dependently reduced HTC contractility. However, ROT was toxic, DNP dissipated MMP, and CD reduced both intracellular F-actin and viability. H-1152 was found to be the best candidate compound since it reduced HTC contractility with minimal side effects. The authors propose that their HTC-based assay system can be used to screen for compounds that modulate HTC contractility and assess the underlying physiological mechanism(s) of compound activity and toxicity.

Introduction

The current “one-gene, one-protein, one-target” approach to drug discovery has encountered many unpredicted difficulties and has not yielded the expected increase in development of novel drugs. ¹ Despite enormous improvements in the capability to screen massive chemical libraries, the attrition rate of drug discovery programs due to safety concerns is high. ¹ Cell-based high-throughput (HTP) assay systems have been developed ² to predict biological responses to chemicals, identify therapeutic candidates, and eliminate potentially toxic compounds. Typically, these assays require analysis of HTP microscopic images to compute the distribution and magnitude of fluorescent indicators. ^3-5 However, cell-based assays use cultured cells on 2D substrata that do not recapitulate the 3D microenvironment for cells in vivo. ⁶ Furthermore, the 2D substrata prevent assessment of the mechanical properties of the cells. Cell mechanics is one of critically important physiological properties of cells in vivo since it dictates the structure and function of tissues and organs. Cellular traction and contractile force regulate, for instance, embryo development, cancer metastasis, homeostasis of blood pressure, and wound healing. Dysregulation of cell mechanics can trigger and exacerbate diseases, including cardiac fibrosis and atherosclerosis. ⁷ Despite the importance of cell and tissue mechanics, a simple technique for their analysis is lacking. Development of an in vitro tissue-based HTP assay system that can assess mechanical properties, as well as other physiological parameters of cells in a 3D microenvironment, will provide a powerful tool for drug discovery. ⁸

We have developed in vitro 3D hydrogel tissue constructs (HTCs) as a more natural environment for investigating cellular traction and contractile force. HTCs are formed by the extracellular-matrix remodeling activity of the cells, and different cell types (e.g., fibroblasts) can form HTCs that exhibit mechanical behaviors similar to in vivo tissues. ⁹ We demonstrated that the contractile force of the HTCs can be decomposed into active and passive components attributable to intracellular actomyosin activities and extracellular matrix stiffness, respectively. ¹⁰ To increase the throughput of HTC assays, we developed miniaturized HTCs and an automatic device, the Palpator™ ( Fig. 1a-d ), to quantify their mechanical properties. ¹¹ This system assesses HTC mechanics ( Fig. 1e ) without employing cell labeling, sophisticated microscopy, and image analysis. ^3-5 In addition, we demonstrate the assessment of cell physiological parameters using spectroscopic (both fluorescent and colorimetric) indicators and a microplate reader. Physiological parameters such as amount of intracellular F-actin, mitochondrial membrane potential (MMP), and cell viability were assessed in addition to HTC mechanics to generate high-content data for compound screening. In developing these assays, we discovered that HTCs improved the sensitivity and dynamic range of detecting the indicator signals due to increased cell density within the assay volume.

Fig. 1.

Mechanical quantitation of hydrogel tissue construct (HTC) contractile force. (a) HTC contractile force measurement device—the system includes (i) an isometric force transducer, (ii) probe baths, and an (iii) HTC stage on a (iv) temperature regulator plate. (b) The x-y-z motion robotic arm positions the force probe above the HTCs and lowers the probe onto the HTC for HTC stretching and force measurement. (c) Schematic of the robotic arm with attached force transducer and probe. (d) The probe is positioned 1 mm above the mid-plane of an HTC (i) and then lowered until it first (ii) touches and then (iii) stretches the HTC. (e) Representative forces recorded during HTC stretching. Arrows i, ii, and iii show force measurements associated with probe positions in (d).

To demonstrate the utility of this high-content screening system, we profiled HTC mechanical and physiological responses to 4 representative compounds that reduce cellular contractile force: Rho kinase inhibitor H-1152, actin filament disruptor cytochalasin D (CD), MMP uncoupler 2,4-dinitrophenol (DNP), and oxidative phosphorylation inhibitor rotenone (ROT). The effect of these compounds on HTC contractility was assessed using the Palpator™, whereas effects on the actin cytoskeleton, MMP, and HTC viability were assessed using spectroscopic indicator dyes. We envision that our HTC-based high-content screening system will enable the discovery of novel nontoxic compounds that can mediate cellular and tissue contractile forces for the treatment of fibrotic diseases.

Materials and Methods

Cell culture and HTC formation

Rat embryo fibroblasts (REF-52) were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Fisher Scientific, Pittsburgh, PA) supplemented with 10% fetal bovine serum (FBS; Atlanta Biologicals, Lawrenceville, GA). Cells were subcultured every 2 to 3 days and used between passages 40 and 70 to form HTCs that exhibit comparable force activity. To make HTCs, REF-52 cells were dissociated by incubating with 0.05% trypsin (Fisher Scientific) for 10 to 15 min, pelletized by centrifugation at 1000 g for 10 min, resuspended in 10% FBS DMEM, and diluted in HTC tissue solution to 8 × 10⁵ cells per mL. The HTC tissue solution consisted of 10% FBS DMEM, 1 mg/mL type 1 collagen (BD Biosciences, San Jose, CA) in 0.02 N acetic acid, sodium hydroxide (NaOH) to neutralize the acid, and 5× DMEM to compensate for the volume of collagen and NaOH. The HTC tissue solution was aliquoted (300 µL/well) into 8-well mini-construct chambers (MC-8™, InvivoSciences, McFarland, WI) ¹¹ and incubated at 37°C and 5% CO₂ for 30 min. Then, 350 µL of 10% FBS DMEM medium was added to each well for an additional 48 h of incubation.

HTC force measurement

The MC-8™s were placed on the stage of the Palpator™, which automatically inserted a probe into each well and stretched the individual HTC. The probe was connected to a force transducer that measured the resistance force induced in the HTCs. These forces were recorded, and a custom Matlab algorithm ¹¹ was used to analyze the data and quantify the active cell contractile force. To obtain stable measurements of contractile force, HTCs were preconditioned prior to force measurements. ¹¹ HTCs were preconditioned by being stretched 3 consecutive times on the Palpator™ device. Preconditioning was not necessary if subsequent force measurements were within 30 min of the previous stretch.

HTC TMRE labeling and MTT assay

HTCs were incubated in 100 nM tetramethylrhodamine (TMRE; Invitrogen, Carlsbad, CA) for 30 min and then in phenol red–free 10% FBS DMEM for 60 min. HTC TMRE signal was read on a Synergy HT plate reader (Biotek Instruments, Winooski, VT). A custom adapter plate was used to position the MC-8™s on the plate-holding rack, and the 543/590 filter set was used with a gain setting of 50.

3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium (MTT; Invitrogen) diluted in DMEM was added to each well to achieve 0.5 mg/mL and then incubated for 2 h. The formazan dye that forms in the cells was then extracted using 500 µL of isopropanol (Fisher Scientific) containing 0.1 N hydrochloric acid. Then, 200 µL of this solution was transferred to a 96-well plate and read on the Synergy HT plate reader. Absorbance at 570 and 650 nm was recorded, and the difference (A₅₇₀ – A₆₅₀) was used as the formazan signal.

HTC drug treatment

All chemicals were from Sigma-Aldrich (St. Louis, MO) unless otherwise noted. CD, ROT, DNP, and Rho kinase inhibitor (S)-(+)-2-methyl-1-[(4-methyl-5-isoquinolinyl)sulfonyl]homopiperazine (Calbiochem, Gibbstown, NJ) were suspended in DMSO. Stock solutions of CD, ROT, and H-1152 at 1 mM were serially diluted to 100, 10, and 1 µM in DMEM. DNP was diluted to 1 M in DMSO and then to 100, 10, and 1 mM in DMEM. Following HTC preconditioning and background force (i.e., predrug) measurement, 50 µL of the drug solutions was added to each well. The final DMSO concentrations for all drugs were at 1%, 0.1%, and 0.01% (respective to the drug concentrations), which did not exhibit significant toxicity effects on the HTCs (Suppl. Fig. 1). Then, 50 µL of DMEM was added to the control wells. Upon drug addition, medium in the wells was mixed by pipetting 4 times.

HTC fixing and F-actin quantification

Drug-treated HTCs were incubated in 0.5 mL of 4% paraformaldehyde (Sigma) solution (in PBS) for 30 min. Fixed HTCs were rinsed twice and then permeabilized by incubation in 0.5 mL of 0.1% Triton X-100 (Fisher Scientific) solution (in PBS) for 30 min. Permeabilized HTCs were rinsed twice with 0.5 mL Tris-buffered saline Tween-20 (TBST) and stained for 45 min with 0.5 mL of 1:200 diluted Alexa 568–conjugated phalloidin (A12380, Invitrogen) in TBST. HTCs were then rinsed thrice with TBST. Alexa-phalloidin fluorescence intensities were read on a Synergy HT plate reader using the TRITC filter set.

REF-52 cell treatment and fixing

REF-52 cells were plated on 35-mm culture dishes (50,000 cells) in 10% FBS DMEM (2 mL). Cells were incubated overnight and then treated with compounds. Concentrated CD (20 µM), H-1152 (10 µM), DNP (30 mM), and ROT (100 µM) were dissolved in DMEM and then diluted 10× into each plate of cells (220 µL per 2 mL medium). Cells were treated for 24 h, rinsed once with 2 mL of phosphate-buffered saline (PBS), and then fixed in 1 mL of 4% paraformaldehyde (Sigma) solution (in PBS) for 30 min. Fixed cells were rinsed twice and then stored in 2 mL of PBS.

REF-52 cell labeling and imaging

Fixed cells were permeabilized by incubation in 1 mL of 0.1% Triton X-100 (Fisher Scientific) in PBS for 15 min. Permeabilized cells were rinsed twice with 1 mL TBST buffer and then blocked with 1 mL of 5% goat serum in TBST for 1 h. Cells were stained with 1 mL of TBST containing 1:200 diluted Alexa 568–conjugated phalloidin (A12380, Invitrogen), 2% goat serum, and 400 nM of DAPI. Cells were stained for 30 min and then rinsed twice with TBST. Labeled cells were mounted with Vectashield (Vector Laboratories, Burlingame, CA), covered with a cover glass, and then sealed with nail polish. The plate was then inverted onto a SP5 confocal microscope (Leica Microsystems, Bannockburn, IL) and imaged with 63× water immersion objective. Alexa 568 was excited using the 543 laser line, and DAPI was excited using a MaiTai multiphoton laser.

Statistical analysis

Student’s t-test, 2-tailed, was used to assess the significance (p ≤ 0.05) of CD-treated versus control HTCs ( Fig. 2c ). The Z′ factor16 was used to evaluate the data quality of the Palpator™ ( Fig. 2a,b ), TMRE ( Fig. 3a,b ), and MTT ( Fig. 4a ) assays. In the MTT assay, DMSO was used as a positive control; results are presented in Supplemental Figure S1. Four replicate HTCs are tested for the majority (95%) of experimental conditions. Occasionally, n is reduced to 3 or 2 due to accidental physical damaging of the HTCs.

Fig. 2.

Drug induced dose-dependent reduction in hydrogel tissue construct (HTC) contractile force. Rho kinase inhibitor H-1152, cytochalasin D (CD), rotenone (ROT), and 2,4-dinitrophenol (DNP) dose-dependently reduced HTC contractile force. Forces were measured at (a) 3 h and (b) 24 h after being treated with varying concentrations of the drugs. Force measurements were expressed relative to control (medium-treated HTCs, F_c). Data show mean ± SEM, n = 3-4; Z factor > 0.44 (^#) and 0.59(*). (c) CD treatment reduced intracellular F-actin. HTCs treated with drugs for 24 h were fixed and labeled with Alexa 568–conjugated phalloidin. Fluorescence intensity was read on a plate reader and plotted relative to control, media-treated HTCs (I_c). Data show mean ± SEM, n = 3; ^@ p < 0.02 (Student’s t-test, 2-tailed vs. controls [Ctrl]). (d-r) REF monolayers were treated with (d, i, n) medium, (e, j, o) CD, (f, k, p) H-1152, (g, l, q) DNP, and (h, m, r) ROT. After (d-h) 3-h and (i-m, n-r) 24-h treatments, representative monolayers were fixed and stained with Alexa 568–conjugated phalloidin and DAPI. CD caused extensive actin depolymerization by 3 h (e). H-1152, DNP, and ROT had limited to no effects. CD- and ROT-induced cytotoxicity was evident by (j, m) cell shrinkage, (o) binucleation, and (r) nuclear disintegration. Note that scale bars in (j) and (o) are slightly larger. Images were captured on a Leica SP5 confocal microscope using a water-immersion 63× objective, bar = 50 µm.

Fig. 3.

Live fluorescence signal detection in hydrogel tissue construct (HTCs). (a, b) 2,4-Dinitrophenol (DNP) dose-dependently uncoupled HTC mitochondrial membrane potential (MMP). HTCs preloaded with tetramethylrhodamine (TMRE; 100 nM for 30 min) were treated with varying concentrations of cytochalasin D (CD), H-1152, DNP, and rotenone (ROT). TMRE fluorescence signal was measured using a plate reader at (a) 3 h and (b) 24 h posttreatment. Signal intensity was expressed relative to the control, medium-treated HTCs (I_c). Mean ± SEM shown, n = 2-4; Z factor > (a) 0.64 (^#) and (b) 0.46 (*). (c) DNP’s uncoupling effects were quantifiable in HTCs using a plate reader. Mean ± SEM shown, n = 4. (d) The same effects were not readily quantified in cell monolayers. Mean ± SEM shown, n = 11. (e) A representative tracing of microscopically quantified TMRE signal in the bottom cell layer of an HTC treated with DNP. Reductions in TMRE signal were higher in magnitude as compared to results from 2D cell monolayers (f). (f) A representative tracing of microscopically quantified TMRE signal in 2D rat embryo fibroblast (REF) cell monolayers treated with DNP.

Fig. 4.

Hydrogel tissue construct (HTC) contractile force and physiological phenotype-based screening of compounds. (a) Cytochalasin D (CD) and rotenone (ROT) exhibited dose-dependent cytotoxicity. Viability of drug treated HTCs was determined by MTT assay at 24 h. Absorbance of formazan (converted from MTT in viable cells) was read on a plate reader at 570 nm and expressed relative to control, medium-treated HTCs (A_c). Ten percent DMSO treatment was used as a positive control for this assay, which yielded A/A_C of 0.21 ± 0.045 (see Suppl. Fig. S1); Z factor for control vs. 10% DMSO was >0.49. Data show mean ± SEM, n = 3-4. (b-e) Phenotypic profiles were used to screen the compounds. Physiology data for each compound at 24-h treatment were normalized to the maximum (by 1 of the 4 compounds) and then fitted with linear or 4-parameter logistic functions (HTC contractile force [force, green lines], mitochondrial membrane potential [MMP, red lines], and MTT conversion [MTT, black lines]). A profile that decreases from 1 to 0 indicates that the compound had the maximal negative effect on that specific physiology. These phenotypic profiles simplify the compound selection process by making evident that H-1152 is the optimal candidate compound since it effectively reduced HTC contractile force (from 1 to 0) while exhibiting minimal uncoupling activity (mitochondrial toxicity) along with a reduction in MTT conversion (cytotoxicity). (f) Workflow of HTC screening. Duration indicates the amount of time needed to process each plate of HTCs. HTCs were synthesized and allowed to contract for 48 h prior to use. Shaded parallelograms indicate points of data acquisition. Repeated measurements are possible for in situ indicators (e.g., tetramethylrhodamine [TMRE]). Assays such as MTT require the sacrifice of the HTCs and are carried out at the end of the experiment. (g) Representative decision tree for HTC-based compound screening. Compounds from libraries are screened for acute (hours) contractility-reducing activity followed by their effects on cellular actin in the cytoskeleton (F-actin) and MMP. If a negative effect is observed for contractile force reduction, F-actin, or MMP, the compound is eliminated as nonactive or toxic. Longer term “chronic” treatments (for weeks) will enhance the sensitivity of viability assays to detect compound toxicity. Results from each assay will help eliminate and suggest potential drug hits or the need for further lead optimization.

Results and Discussion

Testing compounds for reduced HTC contractile force

We treated HTCs with CD, H-1152, DNP, and ROT for 3 and 24 h and profiled their effects on HTC mechanics. Three-hour treatment with H-1152 and CD dose-dependently reduced HTC contractile force with EC₅₀ ≈0.1 µM ( Fig. 2a ). DNP treatment also reduced force but required ~10,000-fold higher concentration, EC₅₀ ~1.3 mM. ROT was less effective at reducing force. At the maximal soluble concentration of 10 µM, force was reduced by only 30% from pretreated level as compared to the 70% observed in the other 3 compounds. Guided by these results, we further investigated how these compound treatments affect cell and tissue physiology.

Twenty-four-hour treatments enhanced the effects observed at 3 h ( Fig. 2b ). In particular, DNP’s EC₅₀ dropped from 1.3 mM to 20 µM. However, the highest concentrations of DNP, CD, and H-1152 did not further reduce force at 24 h. This suggests that these doses achieved maximal effect in 3 h. Taken together, these results demonstrate that we can quantify HTC contractile force and differentiate the compounds’ effects on HTC mechanics. Furthermore, the Z′ factors of the results at 3 and 24 h were 0.44 and 0.59, respectively. These values are above the National Institutes of Health’s (NIH’s) recommended value (Z′ factor ≥0.4) and confirm that this HTC force assay is reproducible and sufficiently robust for HTP screening.

Filamentous actin and HTC mechanics

Tissue force is maintained through the integrity of the cytoskeleton, especially through interactions between filamentous actin (F-actin) and myosin. ¹⁰ To quantify the amount of F-actin, compound-treated HTCs were stained with Alexa-phalloidin and measured on a plate reader. F-actin content in HTCs treated with CD (2 µM) was significantly reduced ( Fig. 2c ) as compared to controls. This is in agreement with our previous report, which correlated CD-mediated reduction in HTC contractile force with the loss of F-actin. ¹⁰ H-1152 and DNP, however, did not induce a loss in F-actin, which indicated different mechanisms of mediating HTC contractility.

To confirm the loss in F-actin, compound-treated monolayer cells were stained with Alexa-phalloidin and examined with a confocal microscope. Extensive loss in F-actin was observed by 3 h post CD treatment ( Fig. 2e ). In 24 h, CD treatment redistributed short F-actin segments within cells that are less spread ( Fig. 2j ) and binucleated ( Fig. 2o ). ¹² H-1152, DNP, and ROT did not dramatically affect F-actin morphology ( Fig. 2f-h ) even though H-1152 caused limited membrane ruffling and reduced phalloidin staining in the central region of the cells ( Fig. 2k ). Confluency of cell monolayers treated with DNP ( Fig. 2l ) and ROT ( Fig. 2m ) for 24 h was less than controls ( Fig. 2i ). Nuclear fragmentation in ROT-treated cells indicated the induction of apoptosis ( Fig. 2r ).

HTP assessment of mitochondrial membrane potential

To further identify the mechanisms by which HTC contractility can be mediated, we assessed changes in cellular MMP. A plate reader was used to measure the intensity of the fluorescent dye TMRE, which accumulates in the mitochondria depending on MMP. Three hours of DNP treatment uncoupled MMP and dose-dependently reduced TMRE signals with an EC₅₀ of ~340 µM ( Fig. 3a ). In 24 h, DNP treatment further lowered EC₅₀ to ~95 µM ( Fig. 3b ). This additional 21 h of DNP treatment reduced EC₅₀ of MMP 3.6-fold but dramatically dropped the EC₅₀ of HTC contractile force 65-fold ( Fig. 2a,b ). Even though DNP rapidly uncoupled MMP, its force-reducing effect was slower. The delayed force reduction was not due to continued diffusion of DNP into the HTCs. DNP diffusion within even noncontracted, large HTCs reached equilibrium within 90 min (Suppl. Fig. S2A). The rate of DNP diffusion in contracted HTC should be much higher. We do not expect that tetanic contraction contributes to the observed results since DNP treatment induced a gradual and continuous decline in HTC contractile force (see Suppl. Fig. 2B). CD, H-1152, and ROT treatments (at treatment concentrations that were efficacious at reducing HTC force) also reduced MMP but were limited to 10% to 20% ( Fig. 3a,b ). Similar to the force measurements, HTC-based TMRE measurements achieved good Z′ factors (0.46-0.64; Fig. 3a,b ) that support the use of HTC-based assays for HTP screening.

Results from these experiments also demonstrated superior sensitivity in detecting TMRE/MMP in the HTCs as compared to cell monolayers grown on 2D substrata. DNP-induced concentration-dependent reduction in TMRE signal was readily detectable in HTCs using a plate reader ( Fig. 3c ), but the same changes in 2D cell monolayers were not ( Fig. 3d ). Confocal microscopy was necessary to improve signal detection and confirmed DNP’s effect in the 2D cell monolayers ( Fig. 3e,f ). However, HTC results still showed superior sensitivity and improved dynamic range due to the integration of fluorescence signals from multiple cell layers. The entire HTC (4 × 4 × 0.8 mm³), which consists of ~500,000 cells, lies within the optical detection volume of the plate reader (5 mm diameter × 10 mm height), and thus the TMRE signal from a large number of cells was integrated to achieve a higher signal-to-noise ratio.

Quantification of HTC viability

The compound-induced reduction in HTC contractile force could have been due to toxicity. To investigate the compounds’ toxic effects, we performed HTC viability assays using MTT. Intracellular succinate dehydrogenase converts the yellow MTT to a purple-colored formazan. This change was quantified by differential absorbance spectroscopy (A₅₇₀-A₆₅₀) using a plate reader. Three-hour treatment with any compound did not result in a significant loss of viability (results not shown). In 24 h, CD and ROT treatments dose-dependently reduced HTC viability ( Fig. 4a ). The 0.2- and 2-µM CD treatments reduced MTT signal by 20%. This level of CD toxicity is in agreement with previously reported LD₅₀ of 5 to 30 µM in human epidermoid cell lines. ¹² ROT (10 µM) induced the greatest reduction of ~40% in MTT signal. This level of ROT toxicity is slightly higher than the previously reported 25% reduction in the MTT signal observed in HepG2 cells. ¹³ H-1152 and DNP treatments for 24 h did not reduce MTT signal. As positive toxicity control, HTCs were treated with 10% DMSO for 24 h, which resulted in an 89% reduction in the MTT signal. The corresponding Z factor for 10% DMSO was 0.85 (Suppl. Fig. S1).

HTC contractile force and physiological phenotype-based screening of compounds

To simplify compound screening decisions, experimental results were summarized as phenotypic profiles that reflect the impact of each compound on HTC physiology ( Fig. 4b-e ). The profiles showed that all 4 compounds reduced HTC force in a dose-dependent manner. This result is expected since the 4 compounds were initially selected for their ability to reduce HTC contractile force. However, DNP was also highly effective at reducing MMP ( Fig. 4b ), whereas CD moderately reduced MMP and HTC viability ( Fig. 4c ). ROT was the most toxic (i.e., reducing viability to the greatest extent) among the 4 compounds ( Fig. 4d ). We identified H-1152 as the best candidate compound that reduces HTC contractile force (with a low EC₅₀ of 0.1 µM) with limited side effects on MMP and HTC viability ( Fig. 4e ). This result was not surprising since Rho kinase inhibition, in general, is known to reduce cellular and tissue contractility and is thus therapeutic for many cardiovascular diseases. ¹⁴ Examples include the Rho kinase inhibitors Y27632, which has been shown to decrease end-diastolic pressure, ¹⁵ and fasudil, which has revealed no human safety concerns in phase II clinical studies for atherosclerosis and hypercholesterolemia treatment. In Japan, fasudil (Eril®) is used clinically as a vasodilator for treating ischemia, vasospasm, hypertension, and angina.

Classify chemical compounds through physiological parameters obtained using HTC-based assays

As demonstrated, in addition to the HTC contractile force, MMP and viability were incorporated into the assay workflow to concurrently monitor multiple physiological parameters. Additional assays, including intracellular calcium levels, oxygen consumption, and free radical production, can be inserted at the end point of the workflow ( Fig. 4f ). HTCs can also be harvested for metabolomic, proteomic, or genomic analyses after the real-time live-HTC assays. Since each real-time live-HTC assay takes only a few minutes, dynamic changes in HTCs’ physiological parameters can be monitored by repeating the assays over time. To increase the throughput of HTC applications, we can readily manufacture and implement 96-well mini-construct chambers since the size of the HTCs conforms to 96-well dimensions. ¹¹ Therefore, HTC-based screening can thus be integrated into existing HTP compound screening workflows to enable rapid classification of compounds according to their effects on tissue mechanics and underlying biological activities (e.g., cytoskeletal integrity, metabolism, or gene expression).

A decision tree illustrates how HTC-based assays can be applied to identify drug candidates that exhibit similar characteristics to the Rho kinase inhibitor (H-1152) to progressively and systematically reduce the number of candidate compounds ( Fig. 4g ). Through the acute-response assays, compounds are tested for their ability to rapidly reduce HTC contractile force (within minutes to hours). Those that are not active will be eliminated from the candidate compound list. The remaining active compounds will then be screened for their effects on F-actin and MMP. Compounds that reduce F-actin and MMP are considered toxic or at least undesirable and eliminated from the list. Through chronic (weeks) treatments, the compounds’ toxicity will be further tested using HTC viability assays such as MTT. The compounds that do not exhibit long-term toxicity will be identified as lead compounds for further testing or chemical optimization.

There are currently no other means to easily simultaneously assess the mechanics and physiology of HTCs and not in high-throughput format. One potential alternative is the micro-post-based technique proposed by Vandenburgh et al. ⁴ that quantifies tissue contractile force by the extent of micro-post deflection. However, this approach is not as flexible in that the sensitivity range of force measurement is limited to ~1.2 mN, which is determined by the stiffness, width, and the maximum amount of micro-post deflection. The sensitivity range of the transducer used in our technique is approximately 4 times larger, from 0 to 49 mN. Furthermore, only our technology offers researchers the ability to stretch the tissues and investigate strain-induced changes in cell physiology, including morphology, growth, differentiation, and responses. In particular, the ability to reconstitute and measure the Frank-Starling relationship in engineered myocardium ⁹ is of great importance to cardiovascular research.

Summary

In this report, we demonstrated how HTCs can be used to phenotypically screen for compounds that reduce tissue contractile force yet have minimal effects on the actin cytoskeleton, mitochondrial membrane potential, and cell viability. The HTCs provided cells a more in vivo–like 3D microenvironment to recapitulate the morphology and physiology observed in native tissues. In addition, optical assays using HTCs demonstrated superior sensitivity for detecting fluorescent indicators since emission signals were integrated from multiple cell layers. Most Z′ factors of the HTC contractility, MMP, and MTT assays satisfied NIH’s criteria (≥0.4) for use in HTP screening applications. We envision that the HTC-based assay system can serve to identify drug candidates that mediate tissue mechanics and facilitate the identification of their efficacy, toxicity, and mechanism of action in the early stages of drug discovery.