In this study, a biological actuator and a self-stabilizing, swimming biorobot with functionalized elastomeric cantilever arms are seeded with cardiomyocytes, cultured, and characterized for their biochemical and biomechanical properties over time.
In recent years, hybrid devices that consist of a living cell or tissue component integrated with a synthetic mechanical backbone have been developed. These devices, called biorobots, are powered solely by the force generated from the contractile activity of the living component and, due to their many inherent advantages, could be an alternative to conventional fully artificial robots. Here, we describe the methods to seed and characterize a biological actuator and a biorobot that was designed, fabricated, and functionalized in the first part of this two-part article. Fabricated biological actuator and biorobot devices composed of a polydimethylsiloxane (PDMS) base and a thin film cantilever were functionalized for cell attachment with fibronectin. Following functionalization, neonatal rat cardiomyocytes were seeded onto the PDMS cantilever arm at a high density, resulting in a confluent cell sheet. The devices were imaged every day and the movement of the cantilever arms was analyzed. On the second day after seeding, we observed the bending of the cantilever arms due to the forces exerted by the cells during spontaneous contractions. Upon quantitative analysis of the cantilever bending, a gradual increase in the surface stress exerted by the cells as they matured over time was observed. Likewise, we observed movement of the biorobot due to the actuation of the PDMS cantilever arm, which acted as a fin. Upon quantification of the swimming profiles of the devices, various propulsion modes were observed, which were influenced by the resting angle of the fin. The direction of motion and the beating frequency were also determined by the resting angle of the fin, and a maximum swim velocity of 142 µm/s was observed. In this manuscript, we describe the procedure for populating the fabricated devices with cardiomyocytes, as well as for the assessment of the biological actuator and biorobot activity.
Biorobots are devices based on living cells that are incorporated within a mechanical backbone that is usually composed of soft, elastic materials, such as PDMS or hydrogels1. The cells undergo rhythmic contractions, either spontaneously or in response to stimuli, and thus function as an actuator. The power generated from cell contraction drives various biorobots. Mammalian heart cells (cardiomyocytes) and skeletal muscle cells are often used for biorobot actuation due to their contractile properties. Aside from cardiomyocyte and skeletal muscle cells, other cell types, such as insect muscle tissues2 and explanted muscle tissues3, have been used. Insect muscle tissues enable the operation of biological actuators at room temperature.
The function and performance of a biorobot are chiefly determined by the strength and consistency of the biological actuator (i.e. muscle cells), while the mechanical backbone structure primarily determines the mechanisms of locomotion, stability, and power. Since these devices are solely driven by forces generated by cells, there are no chemical pollutants or operating noises. Therefore, they form an energy-efficient alternative to other conventional robots. Various literature sources have discussed the different methods to integrate living cells and tissues into biorobots1,4,5. Advances in microfabrication and tissue engineering techniques have enabled the development of biorobots that can walk, grip, swim, or pump5,6. In general, cells are cultured directly onto the mechanical (polymeric) backbone as a confluent cell sheet or they are molded into 3-dimensional actuating structures within scaffolds such as rings and strips. Most often, biorobots are fabricated using cardiomyocyte sheets6,7, as these cells have an innate ability to exhibit spontaneous contraction without external stimuli. On the other hand, reports on skeletal muscle cell sheets are limited due to their need for stimuli to initiate contractions in vitro in order to initiate membrane depolarization8.
This protocol first describes how to seed cardiomyocytes on a functionalized biological actuator made of a thin PDMS cantilever. It then describes in detail the seeding and analysis of the swimming profiles. The cantilever is functionalized with a cell adhesive protein such as fibronectin and is seeded confluently with cardiomyocytes. As the cells seeded on the device contract, they cause the cantilever to bend and thus to act as an actuator. Over time, as the cells mature, we trace the changes in surface stress on the device by analyzing videos of the cantilever bending. The biological actuator developed here can be used to determine the contractile properties of any cell type, such as the fibroblasts or induced pluripotent stems cells, as they undergo differentiation.
Much of the earlier research on biorobots has been focused on developing biological actuators, while optimization of the biorobot architecture and functional capabilities were largely neglected. Recently, a few studies have demonstrated the implementation of swimming modes in biorobots that are inspired by nature. For example, swimming biorobots with flagella-based motion6, jellyfish propulsion9, and bio-hybrid rays4 have been engineered. Unlike other works in literature, here we focus on varying the properties of the mechanical backbone to create a self-stabilizing structure. The biorobot developed in this study is capable of maintaining a constant pitch, roll, and immersion depth as it swims. These parameters can be modified by varying the thickness of each base composite. The fabrication steps involved in developing the PDMS actuator, the submergible biorobot, and the functionalization of the device are described in detail in Part 1 of this two-part article, as well as in our recent work7.The technique developed here can pave the way for the development of novel, highly efficient biorobots for various applications, such as cargo delivery.
The isolation process followed in this study is similar to the process described in an earlier work10, as well as in recently published work7. The microfabrication methods used for fabricating the PDMS actuators and biorobot devices are described in detail in Part 1 of this two-part manuscript. The protocol section of this manuscript describes the steps involved in seeding cardiomyocytes onto the fabricated PDMS actuator and the biorobot following their functionalization with cell adhesive proteins.
All procedures described here have been carried out using an approved protocol and in accordance with the regulations of the Institutional Animal Care and Use Committee of the University of Notre Dame.
1. Cell Seeding and Culture
2. Biochemical Characterization
3. Imaging
4. Image Analysis of the Biological Actuators on a Stationary Base
5. Analysis of Swimming Biorobots
6. Analysis of Protein Expression
Note: The mounted samples prepared in steps 2.2.4 and 2.2.5 were imaged using a confocal microscope. Images were acquired at 20X, 40X, and 60X magnification sequentially in three channels simultaneously: 460 nm, 488 nm, and 594 nm. A set of 5 images were captured at 40X magnification, from different positions for each sample, and each channel was saved as an individual .TIFF file. The exposure setting was determined by the magnification of the objective used and was set constant for all the captures at that magnification.
The biological actuator made of a thin PDMS cantilever (25 µm in thickness) and cardiomyocytes constitutes the core of the swimming biorobot, as shown in the schematic and screenshot of the devices in Figure 1. The cells start to exhibit contractions after 24 h in culture, and bending of the cantilever arms was observed by day 2. The side profile of the device was recorded every day, and the surface stress was quantified from the bending of the cantilever arms using a customized image analysis script7. The static and dynamic stresses were extracted from the surface stress on each day (Figure 2a). Note that static stress (cell traction force) is the contractile stress the cells exhibit on the surface at their resting state and dynamic stress (cell contraction force) is the stress generated by the cells at maximum contraction.
The biological actuator continued to exhibit spontaneous contractions for up to 10 d, and data was collected for the first 6 days. As the cells matured in culture, a gradual increase in static and dynamic stress was observed over time (Figure 2a). There was a large standard deviation in the measurement of forces due to differences in cell maturation between different samples. The cells exhibited a maximum cell traction force of 50 mN/m and a maximum contraction force of about 165 mN/m on day 6. Analysis of all individual data for multiple samples showed a strong positive correlation between the static and dynamic stresses, as both showed an increase over time.
In order to quantify the maturation of the cardiomyocytes over time, the expression levels of some of the structural and functional cardiomyocyte proteins, such as cardiac troponin I, gap junction protein connexin-43, and actin filaments, were calculated. As seen in Figure 2b, a steady increase in intensity measurements over time were observed, corresponding to the expression of the respective proteins. This increase in protein expression confirms the maturation of the cells (cardiac troponin I)15, cell growth and spread (i.e. increase in actin expression), and an increase in cell interconnectivity (i.e. increase in the number of gap junctions).
The static and dynamic stresses plotted in Figure 2a were quantified using a customized computer script, as described in the earlier section. Once the recorded images of the biological actuator were loaded into the system, the script allowed for the tracking of the movement of the cantilever arm with the help of manually assigned markers, as shown in Figure 3. A gradual increase in the bending angle of the cantilever arms was observed over time in culture, which corresponded to the increase in dynamic contraction and static cell traction force exhibited by the cells7. The results of the calcium flux assay and the related video are provided in the previously published work7.
The biorobots exhibited spontaneous contractions on day 2 after cell seeding and were able to actively swim horizontally for up to 10 d. Due to the force balance between the weight of the biorobot and the buoyancy, the device maintained a stable position at the air/medium interface. Displacement or movement of the biorobots was driven by the synchronous contractions of the cardiomyocytes, which caused the bending of the thin cantilever arms and functioned as an actuator. It was observed that the velocity of the biorobots and the distance moved with each stroke decreased after 6 d in culture. Since the muscle cells used in this study were primary cardiomyocytes isolated from neonatal rats, the seeded cell population also contained other cell types, such as cardiac fibroblasts, which are highly proliferative16. As the cardiac fibroblasts proliferate and spread over time in culture, they can suppress the contractility of cardiomyocytes. The result is consistent with other studies that have shown that the activity of primary cells inherently declines after the first week in culture16. In future research, the cell culture could be treated with anti-mitogen agents to block non-myocyte proliferation to increase the lifetime of the biorobots.
We observed that the resting angle of the cantilever arm after the fabrication determined the swimming profile of the biorobots, providing either a horizontal or vertical mode to the swimming profiles of the biorobots. Here, "horizontal" and "vertical" refer to the resting angle of the cantilever with respect to a horizontal axis and do not refer to the direction of swimming motion7. The vertical biorobots had a resting angle of about 110° and contracted at an angle of 90°, while the horizontal-mode biorobots had a resting angle of about 45° and contracted about the horizontal axis (0°). Also, we observed a wide range of beating frequencies across all of the samples and broadly classified them as either a high-frequency (HF) or low-frequency (LF) mode. Figure 4 compares the velocity, beating frequency, and distance traveled for a single stroke for three main types of biorobots: horizontal HF, horizontal LF, and vertical mode. The horizontal HF biorobots exhibited an average beating frequency of 1.6 ± 0.417 Hz, while the horizontal LF biorobots maintained a steady 1.09 ± 0.134 Hz. Although the vertical biorobots exhibited only 0.86 ± 0.07 Hz, they exhibited a higher swim velocity of 142 µm/s and exhibited the greatest distance traveled at 160 ± 642 µm. On the other hand, the horizontal LF traveled about 48 ± 21.2 µm with each stroke at a speed of 67.3 µm/s, while the horizontal HF biorobot had a velocity of 84.4 µm/s and covered a distance of 61.5 ± 17.7 µm per stroke.
Figure 1: Biological Actuator and Biorobot. (A) Schematic of the biological actuator seeded with a confluent cell sheet in relaxed and contracted state (top panel) and a screenshot of the device in culture (bottom panel). As shown in the figure, the biological actuator is composed of a thin, functionalized PDMS cantilever (25 µm thick) seeded with a sheet of cardiomyocytes. This actuator forms the core of the swimming biorobot, also as shown. (B) Schematic of a single-arm biorobot seeded with a confluent cell sheet (top panel) and a screenshot of the device in culture (bottom panel). Please click here to view a larger version of this figure.
Figure 2: Biomechanical Analysis of the Cardiomyocytes. (A) The dynamic contraction force and static cell traction force increased as the cardiomyocytes matured; sample size = 6 for each parameter. The dynamic and static stress expressed by the cells increased over time as the cells matured and developed in culture. A maximum static force of 50 mN/m and a dynamic contraction force of 165 mN/m were observed on day 6. (B) Quantification of the fluorescence intensity for the protein markers cardiac troponin-I, connexins-43, and actin; sample size = 4 for each parameter. The fluorescence signal for all three markers increased throughout the culture, suggesting an increase in the expression of these functional proteins over time in culture. The error bars in (A) and (B) represent the standard deviation for each parameter quantified. Please click here to view a larger version of this figure.
Figure 3: Quantification of ROC: Radius of Curvature (ROC) Calculations using a Customized Image Analysis Script. The ROC found during a contraction is illustrated in the figure. Multiple points are manually picked along the cantilever, shown as a small green "X." Once entered for calculation, a best-fit circle is drawn for the points provided, as shown by the green circle going through the cantilever. Please click here to view a larger version of this figure.
Figure 4: Comparison of the Average Swim Velocity, Beating Frequency, and Distance Moved/Stroke for 3 Biorobots with Different Swimming Profiles: Horizontal Low Frequency (Horizontal LF), Horizontal High Frequency (Horizontal HF), and Vertical. The measurements are normalized to the value of horizontal LF. The error bars represent the relative standard deviation for each parameter quantified. The horizontal LF and HF each had a cantilever arm with a resting angle along the horizontal axis and displayed beating frequencies of 1.09 ± 0.134 Hz and 1.59 ± 0.417 Hz and swim velocities of 67.3 µm/s and 84.4 µm/s, respectively. The vertical biorobot exhibited a beating frequency of 0.862 ± 0.075 Hz and an average swim velocity of 142 µm/s. Please click here to view a larger version of this figure.
The procedure outlined here describes a successful seeding method for PDMS-based actuators and biorobots, which facilitates the attachment of cardiomyocytes. Furthermore, the process of image acquisition and the subsequent analysis that characterizes the behavior of the cells and the performance of the devices was described.
We observed spontaneous contraction of cells on the cantilever arms after 24 h; the intensity of contractions continued to increase steadily over time and reached a maximum on day 6, after which the intensity decreased slowly. Although the cantilever arms of the biological actuator were only 4 mm long, large deflections up to 2.5 mm were observed, especially after 6 days in culture. The low Young's modulus (750 kPa) and ultra-low thickness of the cantilever (25 µm) allowed for such large deflections, which resulted in the strong propulsion of the biorobots. In order to assess the mechanical characteristics of these cells, we quantified the cell contraction and traction forces generated on each day. Thin film cantilevers have been used for measuring contractility and mechanical stress induced by cardiomyocytes or other muscle cells13,17.
The measured maximum dynamic contraction force of 165 mN/m in this study is comparable to the literature13, where cells seeded on a hydrogel cantilever exhibited a systolic stress of about 82.8 ± 22.4 mN/m. The measured forces and the increase in stress over time were relatable to the maturation of the cells seeded on the device, as seen from the corresponding increase in contractile and cytoskeletal protein expression. The electromechanical coupling also increased over time, as seen from an increase in expression of the gap junction protein connexins-43.
The biorobot device developed here falls in the category of ostraciiform swimmers18, where the propulsion is provided by the wagging of a tail and deflection is limited to the caudal fin. The maximum swim velocity of 142 µm/s, exhibited by the vertical mode biorobot, is between the measured values of other popular swimming biorobots in the literature, which are flagella-based biorobots with speeds of about 9.7 µm/s6 and jet-propulsion mode devices with speeds of 6 – 10 mm/s9.
In recent years, many biological machines or biorobots have been developed using soft, elastic materials, such as PDMS and hydrogels, and are seeded with contractile muscle cells. Walking and swimming biorobots have garnered increased focus as an alternative to traditional robots due to their potential to function as energy-efficient, agile robots with self-repair potential1,5. Unlike other biorobots in the literature, the biorobot developed in this study can maintain its own pitch, roll, and immersion depth7. However, the longevity of the cardiac cells used here is one of the biggest limitations, as the life span is restricted to 10 d. To overcome this limitation, contractile cell types from other species may be used to drive the device, rather than mammalian-derived cells. For example, various works by Akiyama et al. have explored the use of contractile insect wing cells for actuation, as they have longer life spans and can survive at atmospheric temperature2.
Currently, the cells are cultured isotropically on the cantilever surface, which limits the net contractile force generated. In future studies, one of the possible modifications that can be included will be to incorporate micropatterns on the cantilevers to attain anisotropic alignment of the seeded cells9. Also, electrodes could be inserted for electrical stimulation to enhance the force generated by the cells4. With continued advancement in technology and manufacturing techniques, these devices can pave the way for the development of novel biological machines with diverse applications, such as cargo delivery. For instance, the base of the biorobot developed here can be easily modified to carry small packages (i.e. cargo)7. Therefore, in this work, we have provided an alternate approach to develop a biorobot, focusing on varying the properties of the mechanical backbone to create a self-stabilizing structure. Moreover, the biological actuator developed here can be used to determine the contractile stress of other cell types, such as fibroblasts and induced pluripotent stem cells.
The authors have nothing to disclose.
M. T. Holley is supported by the Graduate Fellows program of the Louisiana Board of Regents, and C. Danielson is supported by the Howard Hughes Medical Institute Professors Program. This study is supported by NSF Grant No: 1530884.
Chemicals and reagents | |||
Cardiomyocytes (primary cardiac cells) | Charles River | NA | Isolated from 2-day old neonatal Sprague Dawley rats |
Dulbecco’s modified eagle’s media (DMEM) | Hyclone Laboratories | 16750-074 | with 4500 mg/L glucose, 4.0 mM L-glutamine, and 110 mg/L sodium pyruvate |
Fetalclone III serum | Hyclone industries, GE | 16777-240 | Fetal bovin serum (FBS) |
Dulbecco’s phosphate buffer (PBS) | Sigma-Aldrich | D1408-100ML | |
Penicillin-G sodium salt | Sigma-Aldrich | P3032 | |
Goat serum | Sigma-Aldrich | G9023 | |
4,6-diamidino-2-phenylindole dihydrocholride powder (DAPI) | Sigma-Aldrich | D9542 | |
Fibronectin from bovine plasma | Sigma-Aldrich | F1141 | Solution (1 mg/ml) |
Calcein-AM and ethidium homodimer-1 kit (Live/Dead Assay) | Molecular Probes | L3224 | |
Calcium Fluo-4, AM | Molecular Probes | F14217 | calcium indicator dye |
Tyrodes salt solution | Sigma-Aldrich | T2397 | buffer solution |
Pluronic F-127 | Molecular Probes | P3000MP | nonionic surfactant-20 % solution in Dimethylsiloxane (DMSO) |
16% Parafomaldehyde | Electron microscopy | 15710 | Caution: Irritant and combustible |
Triton x-100 | Sigma-Aldrich | X-100 100 mL | cell lyses detergent, (4-(1,1,3,3-Tetramethylbutyl)phenyl-polyethylene glycol, t-Octylphenoxypolyethoxyethanol, Polyethylene glycol tert-octylphenyl ether) |
ProLong gold antifade reagent | Molecular Probes | P10144 | Mounting agent |
Alexa Fluor 594 Phalloidin | Molecular Probes | A12381 | Actin filament marker |
Goat anti-rabbit IgG (H+L) secondary antibody, Alexa Fluor 594 conjugate | Molecular Probes | A-11012 | |
pha | Molecular Probes | A-11001 | |
Anti-connexin 43 antibody | Abcam | ab11370 | Gap junction marker |
Anti-cardiac troponin I antibody | Abcam | ab10231 | Contractile protein |
16% EM grade paraformaldehyde solution | Electron microscopy | 100503-916 | |
Polydimethylsiloxane (PDMS) | Elsevier | Sylgard 184 | |
Materials and Equipment | |||
Camera | Thor Labs | DCC1545M | |
LED light strip | NA | NA | Any white LED without spectrum emission |
Confocal microscope | Nikkon C2 | NA | Confocal microscope with three filter set. |
Zooming lens | Infinity | Model# 252120 | |
Software | |||
Matlab | Mathworks | NA | Used in Section 4) for biological actuator analysis. |
Image J | National Institute of Health | NA | Java-based image processing software. Used in Section 5) for biorobot analysis. Free Image Processing and Analysis software in java. (https://imagej.nih.gov/ij/) |
Thor Cam | Thor Labs | NA | Camera operating software |