Development of a functional sarcopenia model utilizing a microcantilever microphysiological system as a phenotypic disease model

Abstract

Purpose

Sarcopenia is a skeletal muscle (SKM) condition marked by reduced muscle mass and function and is often observed as a comorbidity with inflammatory conditions.

Methods

We have developed a defined in vitro functional model to study sarcopenia by incubating human-induced pluripotent stem cell–derived SKM cells with 3 nM TNF-alpha (TNFα) for 96 h. The long-term effects of this treatment were studied for up to 32 days in culture. Muscle function was evaluated by measuring myotube width as well as microcantilever contraction amplitude, force, and fatigue index.

Results

This model revealed the maintenance of a significant reduction in myotube width for 32 days after TNF-α treatment; moreover, a decrease in myotube contraction amplitude and an increase in fatigue indices were recorded for 32 days. In addition, disease progression and tissue degradation of the sarcopenic conditions were monitored for higher reactive oxygen species production for both acute (day 4) and chronic (day 40) timelines.

Conclusion

This is the first microphysiological system for sarcopenia evaluation that can be used to evaluate therapeutics. This device can now be combined in a multi-organ microphysiological system to study sarcopenia induced by an inflammatory response from other organs or as a comorbidity in the platform.

Graphical Abstract

Development of the first defined in vitro functional model of sarcopenia on a microcantilever platform. This study recapitulates physiological signatures of disease progression characterized by loss of muscle mass and function and an increase in reactive oxygen species

Introduction

Sarcopenia is characterized by loss of muscle mass and function. It is mostly considered either an inflammatory or an age-related condition affecting the skeletal muscle (SKM). In inflammation-inducing diseases such as diabetes mellitus and cancer, the inflammatory molecules affect SKM both acutely and chronically [1–7]. Specifically, one of the prominent cytokines, TNF-alpha (TNFα), has been shown to affect the contractile machinery and increase reactive oxygen species (ROS) production in SKM, contributing to sarcopenia [8–12]. As a condition, the elaborate mechanisms leading to sarcopenia development and progression are yet to be fully understood, but it is frequently observed as a comorbidity along with cardiac hypertrophy, diabetes, cancer, and other diseases associated with aging [13–15].

A promising in vitro approach to study sarcopenia-related comorbidity is to utilize microphysiological systems housed in an interactive network of multiple organs. Interconnected multi-organ systems provide real-time measurements of how different organs’ physiological processes contribute during the manifestation of disease for both acute and chronic conditions [16]. As the first step towards developing a comorbidity model of sarcopenia, a defined model would need to be established for sarcopenic SKM myotubes, which could subsequently be used in multi-organ studies [17–19]. While earlier studies have focused on understanding aging in skeletal muscle by comparing cells from donors of different age groups, a defined model that focuses on the development of disease in vitro has not been proposed [20–22].

In the present study, we have utilized our long-term defined model of SKM development to induce sarcopenia using TNFα. The major hurdle faced in long-term in vitro studies of functionally active SKM myotubes is to prevent them from detaching or peeling from the surfaces after multiple contractions over a period of time. The key feature of our model system that enables maintenance of long-term SKM myotube culture is the defined surface modification and extracellular matrix combination employed. This has been achieved by covalently linking the silane N-(3-Trimethoxy-silyl-propyl-diethylene-triamine), DETA, to glass coverslips and silicon microcantilevers as it is a positively charged molecule to enable charge-charge interactions with extracellular matrix (ECM) materials for better adsorption and adherence of the proteins [23]. This sarcopenia model has been evaluated through various morphological, functional, metabolic, and physiological parameters.

Myotube width was measured to assess the changes in muscle mass upon sarcopenia induction during disease progression. To observe the change in muscle strength, the amplitude of myotube contraction and the rate at which the myotubes fatigued were evaluated. Further, the production of ROS and mitochondrial potential to estimate the physiological state of cells in culture upon sarcopenia induction was measured. The myotubes were then integrated with a microcantilever-based microphysiological system to test the force output as a clinically relevant functional readout of muscle strength. The functional system was then evaluated for ROS production compared to healthy SKM. This defined, multifaceted microphysiological system features various stages of sarcopenia, serving as a model for pre-clinical studies and disease monitoring. It can also be integrated with other organ systems in multi-organ platforms to understand organ-organ interactions to mimic behavior in the aged population [17–19, 24, 25].

Results

X-ray photoelectron spectroscopy confirmed incorporation of the protein coating on the DETA-silane surfaces

An X-ray photoelectron spectroscopy (XPS) spectrum of DETA-silane and DETA-Elastin-Collagen (DETA-silane with extracellular matrix proteins coating) is shown in Fig. 1a and b, respectively, indicated the presence of silicon, carbon, nitrogen, and oxygen. The decrease in intensity of silicon (Si2p), as shown in Fig. 1c, suggests that the DETA peaks were being covered by a protein layer, in this case, the DETA-Elastin-Collagen surface modification. The amount of carbon (C1s) detected on the surface increased from the DETA only coverslips compared to DETA-Elastin-Collagen, as well as a substantial increase in the carbonyl carbon peak was observed at 288 eV, confirming protein adsorption on the DETA modified surfaces (Fig. 1d).

Fig. 1.

X-ray photoelectron spectroscopy (XPS) analysis showing incorporation of the extracellular matrix proteins elastin and collagen on the DETA surfaces. a XPS spectrum of coverslips modified with the DETA-silane. b XPS spectrum of the protein coating (elastin and collagen) on DETA-silane modified coverslips. c Si2p high-resolution XPS spectrum. d C1s high-resolution XPS spectrum indicated the growth of a carbonyl carbon peak at 288 eV as well

Development of a sarcopenia model by induction with TNF-alpha

The overall procedure for sarcopenia induction is shown in Fig. 2a. The control cultures are shown in Fig. 2b–d, while sarcopenia cultures are shown in Fig. 2e–g. iPSC-derived myoblasts were proliferated in myocult medium to attain 60–70% confluency (Fig. 2b and e). The cultures were then primed by performing a half medium switch to DKHI medium (day − 2) [23]. After 48 h, the cultures were switched to a differentiation medium (noted as day 0) by performing a half medium switch to NbActiv4 medium. To induce sarcopenic conditions, 3 nM TNFα was used at this stage of growth. A full medium change was performed with NbActiv4 after 96 h (day 4). Figure 2d indicates the alignment of myoblasts for myotube formation in control conditions, whereas Fig. 2g indicates myoblasts similar to those in Fig. 2f, depicting the arrest of myotube formation and impaired myoblast migration upon incubation with TNFα for 96 h.

Fig. 2.

Sarcopenia induction using TNFα: a timeline indicating the culture conditions employed for the control and sarcopenic phenotype. For both conditions, myoblasts were allowed to proliferate in myocult medium, then primed in DKHI medium, followed by switching to differentiation medium; TNF-alpha (3 nM) was added only to the sarcopenic system. b Phase image for control condition at day − 2. c Phase image for control conditions at day 0. d Phase image for control condition at day 4. e Phase image for sarcopenic condition day − 2. f Phase image for sarcopenic condition day 0. g Phase image for sarcopenic condition day 4

Myotube width reduction after sarcopenia induction

Cultures were maintained for 32 days after differentiation. Myotube width (values indicated in Table 1) was measured on days 10, 14, 17, 21, 28, and 32 in control (Fig. 3a–f) and sarcopenic conditions (Fig. 3g–l). Statistical quantification of the myotube width indicated that the p-values calculated using Student’s T-test were less than 0.00001 for all days (Fig. 3, lower-most panel).

Table 1.

Average myotube width measurement to quantify reduced muscle mass upon sarcopenia induction over 32 days in culture. Average myotube width is expressed as average ± standard error, N = number of myotubes used for analysis. This data was sampled from three independent experiments

Day	Condition	Myotube width (average ± standard error)	N
10	Control	10.76 µm ± 0.23	99
	Sarcopenic	9.11 µm ± 0.28	98
14	Control	10.36 µm ± 0.29	122
	Sarcopenic	8.10 µm ± 0.20	132
17	Control	10.35 µm ± 0.31	80
	Sarcopenic	8.07 µm ± 0.27	74
21	Control	11.88 µm ± 0.28	119
	Sarcopenic	8.32 µm ± 0.31	88
28	Control	12.33 µm ± 0.44	123
	Sarcopenic	9.41 µm ± 0.32	120
32	Control	11.30 µm ± 0.56	104
	Sarcopenic	7.08 µm ± 0.38	84

Fig. 3.

Decrease in myotube width upon sarcopenia induction. Representative phase images in the control condition, a day 10, b day 14, c day 17, d day 21, e day 28, and f day 32. Representative phase images in the sarcopenic condition, g day 10, h day 14, i day 17, j day 21, k day 28, and l day 32. Scale bar equals 50 µm. Bottom graph showing quantification of myotube width over 32 days in control and sarcopenic conditions. This data was sampled from three independent experiments. Error bars represent standard error

Myosin heavy chain expression in SKM myotubes

Representative immunocytochemical analysis of control myotubes on day 32 is shown in Fig. 4a–c, and for sarcopenic myotubes in Fig. 4d–f. Thirty-day-old multi-nucleated myotubes in both control and sarcopenic conditions expressed myosin heavy chain, as shown in Fig. 4a and d, respectively. Corresponding phase images in Fig. 4b and e confirmed the prominent presence of thinner myotubes in sarcopenic cultures. It is noteworthy that in both conditions, in long-term culture, striated myotubes were observed. Even though sarcopenic myotubes express myosin heavy chain, as confirmed by immunocytochemical studies, their amplitude of contraction was lower as compared to healthy myotubes, indicating that the force generation was reduced upon treatment with TNF-alpha to establish the sarcopenia model.

Fig. 4.

Immunocytochemical analysis of SKM myotubes with myosin heavy chain protein (red) with DAPI labeled nuclei (blue). Representative images from control group. a Image showing myosin heavy chain and nuclear stain. b Phase image. c Nuclear stain using DAPI: representative images from sarcopenic condition. d Merged image showing myosin heavy chain and nuclear stain. e Phase image indicating myotubes. f DAPI stain. Arrows indicate striated myotubes. Scale bar = 50 µm

Amplitude of myotube contractile activity was decreased upon sarcopenia induction

Myotubes in cultures were electrically stimulated at frequencies of 0.33 Hz and 0.5 Hz for 15 s each, and the amplitudes of their contractile activity were measured for control and sarcopenic conditions (Fig. 5a–f). Figure 5a–d shows traces for control and sarcopenic conditions at 0.33 and 0.5 Hz on days 14 and 32, with amplitudes normalized to corresponding controls. Figure 5a and b show comparative amplitudes on day 14 at 0.33 Hz and 0.5 Hz. Figure 5c and d indicate representative normalized amplitudes on day 32 at 0.33 Hz and 0.5 Hz for control and sarcopenic conditions. Figure 5e and f represent quantification of comparative contractile myotube amplitudes on days 14, 17, 21, 28, and 32 for 0.33 Hz and 0.5 Hz. The normalized amplitudes of sarcopenic conditions at 0.33 Hz stimulation frequency on days 14, 17, 21, 28, and 32 were 0.50 ± 0.06 (n = 24, p < 0.005), 0.52 ± 0.06 (n = 24, p < 0.00001), 0.66 ± 0.08 (n = 24, p < 0.005), 0.87 ± 0.08 (n = 23, p = 0.258), and 0.71 ± 0.09 (n = 16, p < 0.05). The normalized amplitudes for sarcopenic conditions at stimulation of 0.5 Hz on days 14, 17, 21, 28, and 32 were 0.47 ± 0.06 (n = 24, p < 0.001), 0.46 ± 0.05 (n = 24, p < 0.000001), 0.61 ± 0.07 (n = 24, p < 0.0005), 0.81 ± 0.07 (n = 23, p < 0.1), and 0.71 ± 0.10 (n = 16, p < 0.05) These results indicate that 96 h of sarcopenia induction with TNF-α leads to decreased amplitude of myotube contraction over 32 days.

Fig. 5.

Quantification of contractile amplitude for control (blue) and sarcopenia (orange) induced SKM myotubes. Representative myotube contractions on day 14 for cultures stimulated at two frequencies: a 0.33 Hz and b 0.5 Hz. Representative myotube contraction profile on day 32 at stimulation frequencies of c 0.33 Hz and d 0.5 Hz. Amplitudes plotted for control and sarcopenic conditions on days 14, 17, 21, 28, and 32, at stimulation frequencies of e 0.33 Hz and f 0.5 Hz. The amplitudes for the sarcopenic condition were normalized to the control group on corresponding days

Myotube fatigue index was increased with sarcopenia induction

Myotubes were stimulated at frequencies of 1 Hz, 2 Hz, and 4 Hz for 15 s, and their performance was evaluated utilizing the clinically relevant measurement of fatigue index. Representative stimulation curves for both conditions on days 14, 17, 21, 28, and 32 and corresponding quantified values of fatigue index are listed in Table 2 and graphically represented in Fig. 6.

Table 2.

Fatigue index values (shown as average ± standard error, n) for contracting myotubes upon repeat stimulation over 32 days were obtained using the pixel subtraction method. These datasets were sampled from three independent experiments and indicate that, over 32 days, sarcopenic myotubes fatigued faster than control myotubes. Corresponding p-values have been shown for Student’s T-test

Day	Condition	Fatigue index (1 Hz)	p	Fatigue index (2 Hz)	p	Fatigue index (4 Hz)	p
14	Control	0.34 ± 0.02, 35	0.063	0.23 ± 0.02, 35	0.05	0.18 ± 0.02, 35	< 0.001
	Sarcopenia	0.40 ± 0.02, 41		0.29 ± 0.02, 41		0.32 ± 0.03, 41
17	Control	0.33 ± 0.02, 35	< 0.001	0.26 ± 0.03, 35	< 0.005	0.22 ± 0.03, 35	< 0.05
	Sarcopenia	0.47 ± 0.03, 34		0.38 ± 0.03, 34		0.32 ± 0.04, 34
21	Control	0.29 ± 0.02, 37	< 0.005	0.20 ± 0.02, 37	< 0.001	0.26 ± 0.03, 37	< 0.01
	Sarcopenia	0.41 ± 0.02, 36		0.32 ± 0.02, 36		0.38 ± 0.03, 36
28	Control	0.29 ± 0.03, 24	0.07	0.22 ± 0.03, 24	0.151	0.26 ± 0.03, 24	0.284
	Sarcopenia	0.38 ± 0.04, 24		0.29 ± 0.04, 24		0.31 ± 0.04, 24
32	Control	0.37 ± 0.04, 16	0.03	0.27 ± 0.04, 16	0.03	0.25 ± 0.03, 16	0.01
	Sarcopenia	0.49 ± 0.04, 16		0.41 ± 0.05, 16		0.38 ± 0.04, 16

Fig. 6.

Measurement of fatigue index by stimulation of cultures at 1 Hz, 2 Hz, and 4 Hz. Representative myotube contraction traces on days 14, 17, 21, 28, and 32 at stimulation frequencies of 1 Hz, 2 Hz, and 4 Hz for control (blue) and sarcopenic (orange) conditions. Corresponding quantification of fatigue indices at three different frequencies is shown in the lower panel

Force measurement of myotubes on microcantilever systems upon sarcopenia induction

Microcantilever systems can be used for force measurements of myotubes and is another clinically relevant functional readout as it is indicative of muscle activity. Force readouts are represented as the voltage amplitude of the photodetector. Representative force characteristics of myotubes in control vs sarcopenic conditions on days 14, 17, 21, and 28 are shown in Fig. 7a–d. Figure 7e indicates SKM myotubes on microcantilever chips for the control condition on day 14 and Fig. 7f for the sarcopenic condition on day 14. The decrease in force output and the maintenance of sarcopenic phenotype can be observed over 28 days with repeat testing from average amplitude of photodetector voltage (in µV) for control and sarcopenic conditions (Table 3, Fig. 7g).

Fig. 7.

SKM force measurements with the microcantilever system. Representative force characteristics for control and sarcopenic conditions on a day 14, b day 17, c day 21, and d day 28, upon repeat testing. e Phase images under bright field showing SKM myotubes on the microcantilever chip on day 14 in control conditions. f Phase image under bright field showing the presence of myotubes on day 14 in sarcopenic conditions. g Average amplitude of photodetector voltage representing force produced by SKM myotubes in control and sarcopenic conditions on days 14, 17, 21, and 28

Table 3.

Average SKM force amplitudes obtained from control and sarcopenic conditions with repeated stimulation of the same systems to observe the development of the sarcopenic phenotype. The average amplitude values (shown as average amplitude (µV) ± standard error, n) indicate the voltage of the photodetector (µV) with standard error, and n represents the number of individual cantilevers tested per condition. Corresponding p-values obtained from Student’s T-test have been indicated

	Control	Sarcopenic	p
Day 14	8.67 ± 0.87, 134	3.33 ± 0.63, 84	p < 0.00001
Day 17	5.50 ± 0.41, 136	3.31 ± 0.53, 95	p < 0.005
Day 21	6.64 ± 0.49, 152	4.43 ± 0.73, 58	p < 0.05
Day 28	5.95 ± 0.67, 109	3.86 ± 0.65, 86	p < 0.05

Sarcopenic cells exhibit increased reactive oxygen species and active mitochondria

TNFα-induced sarcopenic myotubes also exhibited an increase in reactive oxygen species (ROS) in addition to the significant functional deficits. The accumulation of ROS has been suggested to contribute to the loss of SKM mass and strength, and in the overall process of organismal aging [26–28]. As a consequence, ROS production can alter the structure of nucleic acids, proteins, and lipids, leading to the deterioration of cellular structure and tissues [29]. Myoblasts were cultured in normal versus sarcopenic conditions as previously described, followed by the labeling of cells with DCFDA, a fluorogenic dye that is deacetylated by cellular esterases to a non-fluorogenic species within the cell, in which the presence of ROS further oxidizes it to highly fluorescent 2′,7′–dichlorofluorescein (DCF). Flow cytometry analysis was performed for the DCF-labeled cells on day 4, which was considered as “short-term” sarcopenic conditions. The data is summarized in Fig. 8a–f. Gating of Live, singlet cells followed by FSC-H and SSC-H revealed that control cells exhibited a 90.98% DCF signal, detected in the FITC channel, whereas sarcopenic cells exhibited a 96.55% DCF signal (Fig. 8a–c). The mean fluorescence intensity for sarcopenic cells was significantly higher compared with control cells (Fig. 8e). In addition, DCF labeling of cells was analyzed at day 40, which was considered as chronic sarcopenic conditions. Results indicated that sarcopenic cells exhibited approximately 60% DCF labeling compared with 40% for control cells (Fig. 8d). Moreover, confocal microscopy confirmed similar DCF labeling for cells at day 40 as observed via flow cytometry, in which sarcopenic cells exhibited higher ROS compared with control cells (Fig. 8f).

Fig. 8.

Detection of ROS in sarcopenic versus control cells. a Flow cytometry analysis revealed that sarcopenic cells at day 4 exhibited higher ROS than control cells. Gating strategy for single cells, followed by FSC-H vs SSC-H to isolate the myoblast population. Viability was further confirmed by DRAQ5 staining of gated singlets (data not shown). b DCF labeling of cells at day 4 was detected in the FITC channel. c Quantification of DCF-positive myoblasts as a readout for ROS based on day 4 flow cytometry. d DCF labeling of cells at day 40. e Quantification of mean fluorescence intensity (MFI) based on gated DCF-positive populations. f Confocal images showed DCF labeling of control and sarcopenic cells, further processed using ImageJ

All graphs and statistical analyses were generated using GraphPad Prism. A two-tailed paired t-test revealed p** < 0.01 for analyzed populations. Since sarcopenic cells indicated increased ROS induced by TNF-alpha-mediated inflammatory signaling, mitochondrial membrane potential was investigated to determine if it was higher in sarcopenic cells. Increasing concentrations of TNF-alpha were titrated into the culture, and a quantitative TMRE assay was performed. At day 4, results showed that increasing the concentration of TNFα resulted in increases in the mitochondrial membrane potential of sarcopenic cells (Fig. 9a) with no changes in cell viability (Fig. 9b). At day 8, 10 nM TNF-alpha-induced sarcopenic cells indicated decreased mitochondrial potential compared with control cells (Fig. 9c). These results demonstrate evidence of increased mitochondrial activity in sarcopenic cells at day 4, which is a time when myoblasts exhibit increased demand of mitochondrial energy in the process of myotube fusion. This was further confirmed by the higher DCF signal observed at day 4 via flow cytometry and confocal microscopy (Fig. 8). At day 8, sarcopenic cells induced with 10 nM TNFα exhibited lower membrane potential with approximately 150% TMRE values compared to control cells at 180% TMRE.

Fig. 9.

Viability and mitochondrial potential measurements in sarcopenic versus control cells. a Myoblasts cultured under control and increasing concentrations of TNFα for 4 days were labeled with TMRE, followed by fluorescence measurements at emission (488 nm)/excitation (575 nm) using a microplate reader. b In parallel, viability was acquired by incubating the same batch of cells with Cell Titer Glo reagent for 10 min, followed by luminescence recordings using a microplate reader. c Control and sarcopenic cells were labeled with TMRE and fluorescence measurements were acquired on day 8 as previously described

These results indicate that normal myoblasts fuse into functional multi-nucleated myotubes while TNFα-induced sarcopenic cells fuse into thinner myotubes, dampening their functional contractile force. The damaging action of reactive oxygen species might compromise the mitochondrial activity and thus lowers the energy output required for proper myotube formation of sarcopenic cells (Fig. 10). Our model revealed several hallmarks that recapitulate different stages of sarcopenia, providing a clinically relevant functional model for future human-centric clinical studies and therapeutic development.

Fig. 10.

Schematic indicating the development of sarcopenic conditions. Muscle precursor cells differentiate into myoblasts, in which healthy cells undergo primary and secondary fusion forming thick and contractile multi-nucleated myotubes. In contrast, TNFα acts as a mitogen on myoblasts, inducing their proliferation and causing a decrease in their fusion rate into myotubes. The aging process and the accumulation of ROS lower the functional contractile force of sarcopenic cells

Discussion

Sarcopenia is symptomatically defined by the loss of muscle mass and muscle function through a combinatorial effect of multiple players, including lifestyle (activity and nutrition), chronic illness, inflammatory load and aging. Studies have shown that we lose 50% of our skeletal muscle mass linearly by the age of 80, with the decline starting as early as 40 years of age [30]. At the molecular level, it is a collective effect of a reduction in growth-promoting factors like testosterone, insulin-like growth factor-1 (IGF-1), estrogen, growth hormone and an increase in inflammatory molecules such as TNFα, TGF-beta, IFN-gamma and IL6, with increasing age [31]. TNFα has been utilized to study its effects and the downstream pathways of inflammation in both in vitro and animal models [26, 30].

Interestingly, in a disease such as sarcopenia, the effect on force generation does not just hamper basic physical function but the overall health of a human being [32–34]. Thus, there is a pressing need for a functional in vitro developmental model where clinically relevant force dynamics can be quantified and assessed for therapeutic development to address sarcopenia progression before advancing to clinical trials.

Given the developing understanding and complex interplay of the sarcopenia condition, here we have developed a defined model to understand the effects of inflammatory load on skeletal muscle (aging adults, using TNFα) and compared to the control condition (healthy adults). TNFα was used to induce sarcopenia in human iPSC-derived SKM cells in vitro in a defined system, using a 96-h treatment at the stage when the myoblasts start to differentiate into myotubes. During the 96 h of treatment, it was observed that the cells did not initiate the process of fusion into myotubes in the TNFα treated sarcopenic condition, whereas fusion into myotubes started in the control condition. This can be because TNFα is a mitogen which keeps myocytes in cell cycle during the incubation period and hence, prevents them from fusing into myotubes [35]. Once TNFα was removed, on day 4 after differentiation, the sarcopenic cells started to differentiate, but the sarcopenic conditions showed decreased fusion and the myotubes formed were thinner than those in the control condition. This indicates that TNFα has irreversibly influenced myocyte physiology. As mentioned above, catabolic factors like TNFα increase with age and hence compromise the ability of satellite cells (skeletal muscle stem cells) to form healthy myotubes in older adults and adults with comorbidities. This model is highly promising as the addition of TNFα treatment at the starting phase helps to stimulate the regeneration and repair process of skeletal muscle myotubes in older adults with higher inflammation in comparison to healthy adults. Essentially, it exhibits how the regeneration of skeletal muscle itself is compromised in the presence of high oxidant and inflammatory niches in aging adults.

To investigate this in a long-term model system, these cultures were observed up to 32 days after differentiation. Myotube width comparisons between control and sarcopenic conditions up to day 32 indicated a statistically significant decrease in myotube width upon sarcopenia induction, indicating an overall decrease in muscle mass, which is a phenotype of sarcopenia. A decline in function of muscle upon induction of sarcopenia was observed, and these myotubes had a decreased strength of contraction as indicated by the amplitude, and they also fatigued faster as shown by fatigue index measurements, a clinically relevant functional readout. It was noted that even though sarcopenic myotubes do not contract with as much force as the healthier myotubes, they fatigue faster. These measurements were conducted up to day 32 by consecutively testing the same set of cultures on days 14, 17, 21, 28, and 32 to observe how the phenotype developed over time. For force measurements, a microcantilever system was used, where a decline in force in the sarcopenic condition was observed consistently up to 28 days by repeat testing of the same systems. The two key observations which emerge from this study were that sarcopenia induction using TNF-alpha reproduces both phenotypes of sarcopenia: decrease in muscle mass and reduction in muscle strength.

We investigated acute and chronic reactive oxygen species (ROS) production upon sarcopenic induction. It was observed acutely, on day 4, and chronically, on day 40, that ROS production was higher in the sarcopenic cultures as compared to control cultures. Higher ROS could lead to a reduction in mitochondrial potential, which was observed with a TMRE assay conducted on day 8. Earlier studies have indicated that higher ROS production could adversely affect the force dynamics of SKM [36]. Myotube formation is a multi-step process, where first individual myocytes fuse into myotubes, which then combine to form myofibers. In sarcopenic cultures, the ROS overload induced by TNFα appears to prevent the myoblasts from fusing efficiently to enable the formation of higher order myotubes (Fig. 10). TNFα has been shown to have a mitogenic action that stimulates the proliferation of SKM [35]. Since sarcopenic cells exhibit thinner myotubes, we speculate that TNFα induction is causing cells to proliferate while impeding their ability to fuse into myotubes. This observation, however, remains to be further investigated.

The microcantilever-based model offers a platform to understand the inflammatory response of exercise, which is an extensively debated topic. The present consensus is that, depending on the physiological status of an individual, exercise exhibits differential benefits. Since our system offers the ability to exercise the muscle through electrical stimulation of the cantilevers [37], it can be employed to quantify the endogenous secretion of inflammatory molecules during exercise and their long-term impact on SKM physiology [36, 38–40]. Our future goal is to integrate sarcopenic microcantilever systems with multi-organ microphysiological devices and observe how inflammatory molecules from other organs affect muscle physiology to study comorbidities in a defined condition and screen putative pharmacological interventions [41].

In summary, as our understanding of the exact temporal interplay of all involved factors and their downstream signaling pathways contributing to sarcopenia is still growing, we have modeled the clinically relevant physiological features of sarcopenia by developing a defined model system utilizing one key player (TNFα). Our system can be used to evaluate the loss of muscle mass (measurement of skeletal muscle myotube width), the loss of muscle function (muscle fatigue index, muscle contractile force), and inflammatory state (biochemical assays for ROS detection and mitochondrial activity). This system lays the foundation stone to model the complexity of the disease in a serial and defined manner by utilizing the complex array of molecules contributing to the disease development and progression.

Conclusion

Here, we report the first defined in vitro model of sarcopenia, where its major physiological characteristics are mimicked, i.e., loss of muscle mass and function. We have reported myotube thinning, reduction in contractile amplitude and increase in fatigue as a sarcopenic phenotype. We have utilized a microcantilever-based system to measure muscle force and monitor anatomical and physiological alterations as the sarcopenic phenotype develops and further be used for chronic drug efficacy and toxicity studies. This model could further be integrated into a multi-organ system to observe the off-target toxicity effects of such drugs.

Experimental section

Surface preparation

Glass coverslips and silicon microcantilevers were used for our experiments.

N1-(3-Trimethoxysilylpropyl)diethylenetriamine (DETA)-silane modification on glass coverslips

Glass coverslips of size 22 mm × 22 mm (VWR) were acid cleaned with hydrochloric acid: methanol (1:1), concentrated sulfuric acid and boiled in water, each for 1 h, and rinsed with deionized water between acid washes. The coverslips were arranged in ceramic racks and dried after rinsing with ethanol and then plasma treated (Harrick Plasma cleaner) using an oxygen plasma at 750 mTorr for 20 min. Simultaneously, 0.1% DETA-silane (United Chemical Laboratories, Inc., T2910) was prepared in freshly distilled toluene in a glovebox (MBraun, Stratham, NH) under inert conditions. The stacked coverslips were placed in a reaction vessel, and 0.1% DETA-silane was added and allowed to react for 30 min while Linearly increasing the temperature to and maintaining it at 85 °C for the last 10 min. The reaction was allowed to cool and coverslips were rinsed thrice in toluene and then were heated again to 85 °C in toluene. The surfaces were cured overnight at 100 °C [23, 42–48].

Surface modification and patterning of cantilevers

Silicon cantilevers were acid cleaned as described above (“X-ray photoelectron spectroscopy confirmed incorporation of the protein coating on the DETA-silane surfaces” section) and plasma treated for 20 min, 750 mTorr, and allowed to react in 0.15% PEG-silane with 0.03% hydrochloric acid in distilled toluene at room temperature for 1 h. The coverslips were rinsed serially twice in toluene, twice in ethanol and finally in deionized water and allowed to dry overnight. The PEG-silane modified cantilevers were patterned using a laser at 2 W for 45 s. Within 2 h of ablation, the surfaces are coated with DETA-silane to backfill the ablated region using the reaction described in the “X-ray photoelectron spectroscopy confirmed incorporation of the protein coating on the DETA-silane surfaces” section [23, 49, 50]. 

Protein deposition, cell culture, and sarcopenia induction

Protein/DETA coated glass coverslips and silicon cantilevers were sterilized with 100% ethanol in a biosafety cabinet. Upon drying, surfaces were transferred to 6 well plates (for coverslips) and 12 well plates (for cantilevers). The surfaces were coated with 0.056 µg/mm² of elastin (E7402, ThermoFisher Scientific) and 0.031 µg/mm² of collagen (A1048301, ThermoFisher Scientific) and stored at 4 °C overnight; stock solutions of proteins were prepared in PBS. The suspension was aspirated and human iPSC-derived SKM cells (differentiated from a human iPSC cell line as described previously [51–53]) were plated in myocult medium at a cell density of 100 cells/mm² on coverslips and 300 cells/mm² on cantilevers. Once the cultures were 60–70% confluent, a half medium change with DKHI medium was performed. After 48 h, a half medium change was performed with NbActiv4 medium (BrainBits, LLC), and this was considered day 0. For sarcopenia induction, 3 nM of TNFα was introduced at this step and maintained for 96 h. Since we present a long-term model, the dose was optimized to 3 nM to present the features of the disease state without compromising the life of tissue in culture.

A full medium change was performed after 96 h, and cells were maintained in NbActiv4 with a half medium change every 3 days. Myocult medium was prepared using Dulbecco’s modified Eagle’s medium (DMEM) with 1000 mg/L D-glucose (36,253, STEMCELL technologies) and human MyoCult™-SF Expansion 10 × Supplement (05982, STEMCELL Technologies). DKHI (previously noted as HI) medium was prepared using the protocol mentioned in previous publications [23, 51–53]. The key rationale for using human iPSC-derived skeletal muscle cells was to observe the effect of TNFα on these regenerating myotubes from satellite cells and the eventual reflection of this into their functional activity.

XPS

To prepare samples for XPS, DETA-silane coated coverslips were sterilized in ethanol and dried in a biosafety cabinet and analyzed. For the DETA-Elastin-Collagen sample, DETA-silane-coated coverslips were coated with elastin-collagen as in the “Development of a sarcopenia model by induction with TNFα” section. The PBS was aspirated before analysis, and samples were allowed to dry. Two samples were analyzed per group, and three random points of 400 mm spot size were chosen on each sample for analysis. A Thermo Scientific K-alpha XPS spectrophotometer was used to acquire survey spectrum at 200 eV pass energy, high-resolution silicon (Si2p) XPS spectrum at 50 eV pass energy and high-resolution carbon (C1s) XPS spectrum at 50 eV pass energy.

Phase imaging and myotube width measurement

Cultures for control and sarcopenia conditions were imaged using an inverted phase-contrast microscope (Zeiss, Axiovert 200) and myotube width measurements were performed in ImageJ using phase images taken at 50 × magnification and N denotes the number of myotubes used for analysis.

Functional measurements using phase microscopy

Cultures on coverslips were assembled each in single-organ acrylate housings, and a printed bridge circuit (PCB) board was attached. For testing, the cultures were switched to carbon dioxide–independent medium prepared using Leibovitz 15 with L-glutamine (11415064, GibcoTM, ThermoFisher Scientific), B27 supplement (17504044, GibcoTM, ThermoFisher Scientific), and antibiotic–antimycotic (15240062, GibcoTM, ThermoFisher Scientific). After testing, a full medium change was performed with Nbactiv4 + medium. Chlorinated silver electrodes connected to the PCB board were submerged in a medium to facilitate the stimulation of cultures. The single-organ housing was placed under an inverted microscope, and functional testing was performed by simultaneously stimulating the cultures and recording their electrical activity at frequencies of 0.33 Hz, 0.5 Hz, 1 Hz, 2 Hz, and 4 Hz for 15 s. This was performed using a program in LabVIEW (National Instruments, Austin, TX, USA) and analyzed using a script written in Python. For comparison between control and sarcopenic conditions, quantified amplitudes of electrical activity at 0.33 Hz and 0.5 Hz were measured, and fatigue index at 1 Hz, 2 Hz, and 4 Hz. The same cultures were repeat tested on days 14, 17, 21, 28, and 32 [23, 53].

Functional measurements using silicon microcantilevers

Silicon microcantilever chips were assembled in single-organ housings and connected with a PCB board and chlorinated silver wires. Testing was performed in a carbon dioxide–independent medium as described in the “Myosin heavy chain expression in SKM myotubes” section. A laser was focused at the tip of each cantilever, which would deflect upon stimulation of the cultured myotubes. This deflection was measured as a voltage utilizing a photodetector, which is directly related to the force generated by the contracting myotubes on the cantilevers. The cultures were stimulated through a pulse stimulator (model 2100, A-M system) at a frequency of 0.5 Hz and recorded for 10 s using a program written in LabVIEW and analyzed using a Python script [23, 54].

Immunocytochemistry and confocal imaging

Cells were fixed using 4% paraformaldehyde for 20 min, and the membrane was permeabilized with 0.1% triton for 15 min. The cells were kept at 4 °C in goat blocking buffer for 1 h. Myosin heavy chain (A1.4101, Developmental Studies Hybridoma Bank) primary antibody (anti-mouse) in 0.1% triton and goat blocking buffer was used to stain overnight. The primary antibody was removed and coverslips were washed thrice with PBS. The cells were incubated in a secondary antibody (AlexaFluor 568 goat anti-mouse, Invitrogen) for 2 h in the dark. Three rinses of PBS were performed with the addition of DAPI in the second rinse. The coverslips were serially washed for 2 min each in 50%, 70%, 90%, 100% ethanol, and allowed to dry. The coverslips were mounted on glass slides using a mounting solution and sealed with enamel, and the slides were stored at 4 °C until imaging. A Zeiss confocal microscope with VoloCity software was used to image the myotubes [42, 54].

DCFDA and flow cytometry assays

Myoblasts were plated at 100 cells/mm² on 1 mg Matrigel-coated white opaque, clear flat-bottom, 96-well plates. Sarcopenia was induced with 3 nM or increasing concentrations of TNFα for 96 h. The medium was removed and cells were washed with 1 × PBS (metal-free) twice, followed by the addition of the final 20 μM DCFDA (113851, Abcam) as per manufacturer’s instructions. After 45 min of labeling, cells were washed twice with 1 × PBS and fluorescence values were obtained at emission (485 nm)/excitation (535 nm) with a microplate reader. Data were graphed and statistical analysis (one-way ANOVA and unpaired t-test) was performed using GraphPad Prism. For confocal microscopy (Zeiss confocal microscope), myoblasts were plated on DETA-Elastin-Collagen coated coverslips, in which sarcopenia was induced as previously described. Cells were incubated with DCFDA for 45 min and the medium was removed followed by two washes with 1 × PBS. Cells were fixed with 4% paraformaldehyde for 20 min, followed by three washes with 1 × PBS. Cells were mounted onto glass slides with ProLong Gold Antifade Reagent containing DAPI (8961, Cell Signaling). Images were acquired using a regular inverted microscope and further analyzed using ImageJ software. For flow cytometry, cells were plated on DETA-Elastin-Collagen coated coverslips and lifted with Gibco TrypLE after DCFDA incubation, followed by centrifugation and pelleting. Cells were stained with DRAQ5 (65–0880-92, Thermo Fisher) for viability, and further analyzed using Beckman Coulter CytoFlex Flow Cytometer. Final data were processed using CytExpert software.

Viability and TMRE assays

Myoblasts were plated at 100 cells/mm² on 1 mg Matrigel-coated white opaque, clear flat-bottom, 96-well plates as described above, and sarcopenia was induced. Cells were subsequently washed in 1 × PBS and then incubated with either TMRE (113852, Abcam) in DMEM/F12 medium or Cell Titer Glo (G7571, Promega) for mitochondrial potential or viability, respectively, as per manufacturer’s instructions. The medium was removed and cells were washed with 1 × PBS. For TMRE, fluorescence was measured at emission (488 nm)/excitation (575 nm) with a microplate reader. For viability, luminescence was recorded as the final readout. Final data and statistical analysis were performed using GraphPad Prism.

Statistical analysis

Statistical analysis was conducted using a two-tailed Student’s T-test considering samples with unequal variances, and the p-values are indicated. The values have been shown as average ± standard error.

Abbreviations

SKM

Skeletal muscle

Author contributions

HJ designed and conducted the experiments, analysed the data, interpreted the results, and wrote the manuscript. LHG and RE helped in conducting flow cytometry experiment and analysis. JJH conceived the idea, designed the experiments, interpreted the results, and wrote the manuscript.

Funding

Funding for this research was provided by the National Institutes of Health grant number R44AG059511.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Conflicts of interest

The authors confirm existing competing financial interest but there has been no financial support for this research that could have influenced its outcome. The only author with competing interest is J.J.H. who has ownership interest and is Chief Scientist and member of the Board of Directors in a company that may benefit financially as a result of the outcomes of the research work reported in this publication.