Changes in Cellular Crosstalk between Skeletal Muscle Myoblasts and Bone Osteoblasts with Aging

Musculoskeletal function declines with aging, resulting in an increased incidence of trips and falls. Both bone and muscle experience age-related losses in tissue mass that alter their mechanical interactions in a well characterized manner, but changes in the biochemical interactions between bone and muscle with aging are not well understood. Of note, insulin-like growth factor 1 (IGF-1), a potent growth factor for bone and muscle, can be negatively altered with aging and may help explain losses in these tissues. We recently developed a co-culture system for simultaneous growth of bone mesenchymal stem cells (MSCs) and muscle satellite cells (SCs) to investigate the biochemical crosstalk between the two cell types. Here, we utilized an aging rat model to study cellular changes between young and old rat MSCs and SCs, in particular whether 1) young MSCs and SCs have increased proliferation and differentiation compared to old MSCs and SCs; 2) young cells have increased IGF-1 and collagen expression as a measure of crosstalk compared to old cells; and 3) young cells can mitigate the aging phenotype of old cells in co-culture. Rat MSCs and SCs were either mono- or co-cultured in Transwell® plates, grown to confluence, and allowed to differentiate for 14 days. Across the 14 days, cell proliferation was measured, with differentiation and crosstalk measurements evaluated at 14 days. The results suggest that in both young and old, proliferation is greater in mono-cultures compared to co-cultures, yet age and cell type did not have a significant effect. Differentiation did not differ between young and old cells, yet MSCs and SCs demonstrated the greatest amount of differentiation in co-culture. Finally, age, cell type, and culture type did not have a significant effect on collagen or IGF-1 expression. These results suggest co-culture may have a controlling effect, with the two cell types acting together to promote differentiation more than in mono-cultures, yet this response was not altered by age. In general, results for old cells had higher variability, suggesting a wider variety in the aging phenotypes demonstrated in these animals. This study was the first to use this rat aging model to investigate changes between bone and skeletal muscle cells, however further investigations are required to determine what signaling changes occur in response to age. Determining these signaling changes could lead to new targets for mitigating the progression of aging.


Introduction 29
Musculoskeletal performance declines with aging, resulting in an increased incidence of trips 30 and falls in the elderly. 1 These changes are compounded by concomitant losses in bone mass 31 (osteopenia, osteoporosis) and skeletal muscle mass (sarcopenia), leading to high rates of fracture 32 and muscle injury in the elderly population. 3,4 Under the mechanostat paradigm, these detrimental 33 muscle changes, which manifest in decreased muscle loading on the skeleton, will induce negative 34 bone adaptation and bone loss. 2 Aged individuals with decreased bone mass (osteopenia) often 35 also present with reduced muscle mass and function (sarcopenia). 3,4 However, although elderly 36 populations often experience both osteopenia and sarcopenia, age-related declines in either bone 37 from overuse, and signaling molecules such as inflammatory and growth factors are reduced. 7,8 Similarly, bone loss associated with aging is the result of reduced functional ability of bone 48 osteoblasts to produce new bone, as well as a loss of signaling factors. 9 49 While these signaling changes have been investigated in either muscle or bone alone, the 50 investigation of the signaling changes in tandem have been under characterized. The concurrent 51 age-related changes in bone and muscle signaling are not fully understood, although the 52 importance of this crosstalk has been established in several key studies. In elegant parabiosis 53 studies, in which the vascular systems for young and old rat pairs were surgically merged, old rats 54 experienced some muscle rejuvenation, including increased muscle mass and improved muscle 55 fiber structure, although specific factors from the young rats stimulating this change remained 56 unidentified. 17,18 Further, in vivo mouse studies demonstrated faster healing of bone fractures in 57 the presence of additional xenografted muscle tissue, 19,20 but again the specific crosstalk factors 58 stimulating tissue repair were ill-defined, where specific signaling factors were not discussed. 59 Therefore, a more specific method is required to investigate the specific crosstalk changes. Our 60 gastrocnemius and tibialis anterior (TA) muscles and the tibia and femur from one leg. Rat bone 93

Cell differentiation 139
On differentiation day 14, cells were rinsed with phosphate buffered saline (PBS), fixed in 140 10% zinc-buffered formalin (VWR, Randor, PA) for 30 min, and then rinsed again with PBS. To 141 assess mineral deposition for MSC differentiation, a 2% Alizarin Red S (Fisher) solution was 142 applied to each sample for 5 min, rinsed with deionized water, and imaged using an iPhone 6S 143 camera (Apple, Cupertino, CA) and EVOS ® XL light microscope (Life Technologies, Carlsbad, 144 CA). Mineral deposition was quantified by solubilizing the Alizarin Red S stain with a 0.5 N HCl 145 + 5% sodium dodecyl sulfate (SDS) solution for 30 min and then taking absorbance readings at 146 405 nm. 147 To assess SC differentiation, immunocytochemistry (ICC) procedures for skeletal myosin 148 expression were performed. On differentiation day 14, cells were rinsed with DMEM, fixed in 149 10% zinc-buffered formalin for 10 min, and then rinsed with tris-buffered saline (TBS). A blocking 150 solution of TBS with 2% normal goat serum (NGS) was applied to the cells and allowed to incubate 151 at 4°C overnight. Cells were then allowed to warm to room temperature for 10 min. A rat-specific 152 primary antibody for skeletal muscle myosin heavy chain (MA1-35718, ThermoFisher) was 153 diluted 1:5 in TBS-NGS. Wells were rinsed with a TBS-0.05% Tween 20 (TW20, ThermoFisher) 154 solution, and primary antibodies were applied. Cells were incubated for 1 h at room temperature, 155 followed by an overnight incubation at 4°C in a humidified chamber with gentle shaking. Cells 156 were again allowed to return to room temperature while the secondary antibody (AlexaFluor 584, 157 A-21044, ThermoFisher) was diluted 1:10 in TBS-NGS. Cells were rinsed in TBS-TW20, and the 158 secondary antibody was applied for 2 h at room temperature. Wells were aspirated and rinsed with 159 TBS-TW20, and a diluted DAPI (ThermoFisher) solution was prepared (1 ug/mL diluted in TGS- Since IGF-1 aids growth in both bone and muscle, its expression was assessed in the 170 conditioned media of both mono-and co-cultures at differentiation day 14 using an ELISA kit 171 (ERIGF1, ThermoFisher), according to the manufacturer's instructions. A standard curve was 172 produced, and samples expressing IGF-1 levels outside of the standards' range were omitted. 173 To assess the effect of aging on collagen production, ICC was performed to evaluate the 174 amount of collagen I deposited by the cells. The ICC protocol was the same as for the cellular 175 differentiation analysis, except with a rat-specific collagen I primary antibody (PA1-36145, 176 ThermoFisher) and a secondary antibody (A-11034, ThermoFisher) used at a 1:10 dilution in TBS-177

Statistical analyses 180
Statistical analyses were performed with Prism (version 6.07; GraphPad Software, La Jolla, 181 CA) using a significance level of 0.05. All results were averaged across the groups and expressed 182 in the form of mean ± standard error of the mean due to unbalanced samples. The effects of age 183 (young, old) and cell type (Mono-SC, Mono-MSC, Co-SC, Co-MSC) were examined using two-way ANOVAs. In cases where differentiation day was considered, a two-way ANOVA was 185 performed with repeated measures. Post-hoc pairwise comparisons within each factor were 186 evaluated with Tukey tests if interactions were previously found to be significant. 187 in mono-culture (p < 0.05, Fig. 3B), with no difference in the amount of myosin expressed between 208 young and old cells, either in mono-or co-culture (p > 0.05). In mono-culture, myosin expression 209 did not differ between muscle and bone cells regardless of age (p > 0.05). However, myosin 210 expression in muscle cells seemed to trend higher than in bone cells. In co-culture, bone cells 211 generally had little to no expression of myosin relative to their muscle cell counter-parts (p < 0.05). 212 Age had no effect on the amount of myosin expressed in these co-cultures. Similarly, when 213 controlling for intensity or number of cells within the field, these differences remained the same 214 (data not shown). Representative images of these condition are shown in Figures 3 C, D, and E. 215

Mineralization results 217
Alizarin red results in mono-culture showed that bone cells had or tended to have 218 significantly more amounts of mineralization relative to muscle cells (p < 0.05), yet age did not 219 have a significant effect (p > 0.05, Fig. 4A). In co-culture, however, all cell types and ages had 220 very little mineralization (p > 0.05). Between mono-and co-cultures, mineralization was higher in 221 bone cells in mono-culture than in co-culture. These data suggests that osteoblasts more 222 successfully differentiated into osteoblasts in mono-culture than in co-culture. 223

IGF-1 Content 225
ELISA results for IGF-1 expression in conditioned media demonstrated no difference 226 between cell or culture type (Fig. 5). In mono-culture, bone cells tended to have lower expression 227 of IGF-1 compared to muscle cells, but these differences were not significant; young and old cells 228 had similar IGF-1 expression in mono-culture. In co-culture, IGF-1 expression did not differ 229 between cell types or between young and old cells. Furthermore, no differences were found in the amount of expression between mono-and co-cultures. These results suggest that IGF-1 expression 231 is not altered by communication between these two cell types in aging. 232 233 Discussion 234 Understanding the bone-muscle interactions has become increasingly important to better 235 understand musculoskeletal diseases. Understanding these interactions with regards to aging could 236 lead to novel therapeutic options to improve aging outcomes. Here, we utilized a previously 237 optimized co-culture system to investigate the bone-muscle interactions in both young and old 238 cells. Our hypothesis was that young cells would generally demonstrate greater proliferation and 239 differentiation compared to old cells. Furthermore, when placed in co-culture, young cells paired 240 together would experience increased crosstalk between the cell types; young cells paired with old 241 cells would be able to improve the old cells' proliferation, differentiation, and crosstalk; and old 242 cells paired together would perform the worst in regard to proliferation, differentiation, and 243 crosstalk. The results of these experiments demonstrated that our hypotheses were partially 244 supported. 245 246

Cellular proliferation 247
Our primary hypotheses for cellular proliferation predicted that young cells would proliferate 248 more than old cells, and cells in co-culture would synergistically undergo greater proliferation 249 compared to the mono-cultured cells. Neither of these hypotheses were supported by the 250 alamarBlue ® results, which showed no difference in proliferation between young and old cells and 251 greater proliferative ability at later days of differentiation for mono-cultures compared to co-252 cultures. Other studies investigating the proliferative potential between two cell types have typically demonstrated a greater proliferation response for co-cultures compared to mono-254 cultures. 23,24 The differences between these finding and our own, however, lie in the selection of 255 animal model, cell type, and type of co-culture. In our case, we are the first to use this animal 256 model of aging in a co-culture study between SCs and MSCs. The decreased proliferation found 257 in our co-cultures may be the result of a regulatory effect between the two cell types in co-culture. 258 This response may mimic those found in the in vivo state, where bone-muscle interactions can 259 restrict the growth of both tissue types. 25 More interestingly, however, is that this restrict effect 260 was conserved across age in our data. Aging can hinder many aspects of the cellular lifecycle, such 261 as a characteristically reduced proliferative capacity. 26-28 Our old cell data did not demonstrate this 262 reduced proliferation compared to young cells and may indicate a less severe aging phenotype in 263 our cells than present in other studies. This less severe phenotype was observed not only in 264 proliferation but many of the cellular activities in our data set. compared to co-cultures, with little effect of age. We hypothesized that MSCs would have more 290 mineralization in co-culture, with greater effects in young cells, neither of which were supported 291 by our data. We posit two explanations for these observations: 1) SCs inhibited MSC 292 differentiation or 2) media mixing between the two wells diluted differentiation media in MSC 293 wells or a combination of the two. In the first case, experiments exploring bone-muscle crosstalk 294 using isolated monocultures and conditioned media have demonstrated that signaling factors from 295 muscle, such as myostatin and ciliary neurotrophic factor 1 (CNF-1), inhibit osteoblast 296 differentiation. 32,33 These factors are secreted by muscle after sufficient growth occurs to inhibit 297 further muscle growth. In our preliminary studies and those of others, SCs have been observed to 298 reach myoblast fusion as early as 10 days, 34-36 whereas MSCs differentiation can take as long as 21 days in some studies 37,38 . We suggest that SCs were reaching a more differentiated state more 300 quickly than MSCs and may have secreted these inhibitory factors before MSCs could completely 301 differentiate. We also found that growth of myoblasts occurred more quickly than MSCs and 302 would result in monolayers lifting off the plate after 14 days, precluding us from continuing to 21 303 days to see if MSCs could reach a more differentiated state. In the second case, the Transwell ® 304 setup required that each cell type receive its own differentiation media, yet the pores within the 305 Transwell ® did not inhibit mixing between the two wells. While another study reported that this 306 mixing can take several days 39 , and may not have been a factor in this study, we cannot exclude 307 this possibility. In either case, the presence of significant mineralization in mono-cultures suggest 308 that that the differences have arisen due to the co-culture setup itself. Furthermore, the lack of 309 difference between young and old cultures continues to reinforce that the cells developed a less 310 severe aging phenotype. 311 312

Cellular crosstalk 313
To investigate crosstalk, we investigated the role of IGF-1 in signaling between the two cell 314 types, as well as collagen I expression in the two cell types. With age, muscle commonly loses its 315 ability to repair and replace tissue effectively and instead deposits collagen. Additionally, bone 316 experiences collagen loss, resulting in an increased incidence of fractures and microcracks with 317 age. 9 IGF-1 is a potent growth factor of both muscle and bone and can enhance collagen 318 production. We hypothesized that, with age, IGF-1 expression would be decreased, leading to 319 decreases in collagen I deposition in both SCs and MSCs. In co-culture experiments, we 320 hypothesized that old cells plated with young cells would behave more similarly to young cells 321 and have similar IGF-1 and collagen I expression, but our results did not support these hypotheses.
Instead, our results show that IGF-1 expression (Fig. 5) and collagen expression (Fig. 3A) did not 323 differ across the culture conditions, cell types, or age. Many studies have highlighted the role of 324 crosstalk between muscle and bone and the importance of this crosstalk for the growth and 325 development of both. 5 In this study, we focused on IGF-1, as it is an important mediator of growth 326 and is often reduced with age. 40 However, IGF-1 has only been minimally investigated during 327 bone-muscle crosstalk. 328 Our results with no age-related changes in IGF-1 expression were surprising, considering the 329 differences in IGF-1 expression observed in several aging studies, but this discrepancy may result 330 from the differences in study design such as using in vitro experiments instead of in vivo 331 experiments. IGF-1 expression is frequently measured in serum samples 41 or in whole tissue 332 samples. 42 While both bone and muscle can secrete and react to their own IGF-1 signals in an 333 autocrine fashion, perhaps the IGF-1 expression in bone and muscle in an in vivo environment are 334 enhanced by additional IGF-1 from the bloodstream and elsewhere. In the isolated in vitro 335 environment used here, we found that IGF-1 is clearly secreted by bone and muscle but may only 336 promote initial cell growth and not the continued growth and differentiation. Therefore, while we 337 did not observe differences in IGF-1 expression associated with age or co-culture conditions in 338 this isolated system, it cannot capture the system-wide IGF-1 signaling that occurs in vivo. 339 Furthermore, IGF-1 is merely one bone-muscle crosstalk factor that can be detrimentally affected 340 by aging; studies of additional signaling factors could provide a more complete picture of age-341 related crosstalk changes in vitro. 43-45 342 IGF-1 was also investigated due to its significant role in downstream collagen production. 343 We anticipated diminished amounts of IGF-1 expression with aging would lead to decreased 344 collagen deposition in old cells compared with young cells. In this context, the lack of differences in collagen I expression between cell types, ages, or culture types may not be so surprising. In 346 typical in vivo environments, bone has approximately 36% of its composition as collagen I, and 347 this content decreases with age. 46,47 Conversely, muscle has 2-6% collagen I, and content increases 348 with age. 48,49 However, collagen I deposition in vitro was not different between bone and muscle 349 cells. Since collagen is deposited by mature cells, and given the early maturation stage of our cells, 350 perhaps the differentiation time was not sufficient to allow significant collagen deposition. 351 Furthermore, in co-culture, the lack of difference in collagen deposition may be indicative of a 352 lack of influence of the crosstalk between the two cell types. While differences in IGF-1 expression 353 or collagen deposition were not observed with co-culture, we cannot exclude the possibility that 354 other signaling pathways or protein expression were not altered by crosstalk. Further studies are 355 required to probe this possibility. 356 357

Conclusions 358
In summary, we demonstrated that co-culture of bone and muscle cells regulates proliferation 359 more so than for cells grown in mono-culture, yet age or cell type did not affect proliferation. 360 Furthermore, differences in cellular differentiation or crosstalk were minimal, with little to no 361 effect of age, cell type, or culture condition on myosin expression, mineralization, IGF-1 362 expression, or collagen deposition. This study was the first to use the F344 x BN F1 hybrid rat in 363 a study for cellular aging while also investigating the crosstalk between SCs and MSCs. The lack 364 of differences between young and old cellular proliferation, differentiation, and crosstalk 365 demonstrate that this particular model may not be ideal for age-related cellular studies. However, 366 the use of this model to successfully grow SCs and MSCs should provide evidence to continue 367 using this system for in vitro aging crosstalk studies. The use of a Transwell ® co-culture system provides a framework with which to understand the individual contributions of each cell type, 369 which cannot be isolated in direct co-culture models. Many potential bone-muscle crosstalk factors 370 besides IGF-1 could be contributing to the concomitant age-related changes observed in bone and 371 muscle in vivo, and future studies should focus either on developing in vitro systems that better 372 mimic the in vivo aging environment and/or examining contributions of different signaling 373 pathways to tissue aging. Overall, this study demonstrated that the intricate interactions between 374 bone and muscle depends on a variety of cues, and advancing our understanding of these cues is 375 essential for developing better models for mimicking aging and better treatments to mitigate 376 functional musculoskeletal deficits with aging.  Figure 1. Study design. SCs and MSCs were isolated from rat tissue and cultivated in monoculture and co-culture and allowed to differentiate for 14 days before outcome metrics were assessed. Figure 2. AlamarBlue ® reduction across differentiation. Proliferative ability differed between some mono-and co-cultures during differentiation days 7-14 for A) young muscle, B) young bone, C) old muscle, and D) old bone. Data are mean ± SEM. *p < 0.05 for mono-culture (blue) vs. coculture (red). **p < 0.05 for mono-culture (blue) vs. co-culture (green).

Figure 3. Immunocytochemistry analysis of cultured cells. Collagen I expression (A) and
skeletal myosin expression (B) after 14 days of differentiation. No differences in collagen expression were found across cell type or age, or in mono-vs. co-culture (p < 0.05). Myosin expression was typically increased in SCs in co-culture (*,p < 0.05) and tended to be increased compared to MSCs in mono-culture. Age had no effect between these conditions. Representative images from the various conditions are shown in (C), (D), and (E). Data are mean ± SEM.

Figure 4. Alizarin red solubilization data from fixed cells. Solubilization results (A) and
representative images (B). In mono-culture, MSCs tended to have or had a significantly greater amount of mineralization compared to SCs (p < 0.05). In co-culture, MSCs paired with old SCs had significantly less mineralization compared to their mono-culture equivalent (p < 0.05). MSCs paired with young SCs had similar amounts of mineralization compared to mono-culture (p < 0.05). Data are mean ± SD. Bars with the same letter above them are significantly different.