A study protocol was approved, and permission granted for this community-based study by the University of Missouri-Kansas City Institutional Review Board (UMKC IRB #15-454).

Participants were recruited via flyers and word of mouth at the two different sites where Stay Strong Stay Healthy had been regularly offered free of charge to older adults. The first site was an Independent Senior Living facility in Kansas City, MO (“Independent Living Facility”); while the second one was a community center maintained by the Clay County Senior Services in Smithville, MO (“Community Center”) that provides different activities to older adults living in the community. After being told about the nature of the study, those willing to participate were consented by the co-author responsible for the exercise sessions (AV). The work was conducted first at the Independent Living Facility and afterwards at the Community Center and run, along with the physical and balance evaluations by one of the co-authors (AV), an experienced SSSH trainer who was at the time with University of Missouri (MU) Extension. All participants were asked to answer the “Physical Activity Readiness Questionnaire” (PAR-Q), a short “Find your Start Point” questionnaire (optional), and to present a release form signed by a physician before enrolling in the study.

Stay Strong Stay Healthy (SSSH) was chosen as the exercise intervention for this study since no difference was seen in an initial study between the two exercise programs used at the time, PEPPI and SSSH (Abreu et al., 2014). However, this time we investigated the response to exercise in two different populations according to dwelling, older adults in an independent living or still living at their homes. The 8-week, 1-h, twice-a-week SSSH program consisted of warm-ups and upper and lower body strengthening exercises with and without free weights. More detailed information about SSSH, including a graphical description of the exercise, can be found in the literature (Ball et al., 2013).

Five functional performance tests were used to evaluate changes in physical fitness before and after the 8-week SSSH exercise program: 30-s chair stand, chair sit-and-reach, back scratch, 8-foot up-and-go and grip strength. These tests have been successfully applied to evaluate fitness in older adults and their use have been widely reported in the literature. They are described in detail on previous papers (Ball et al., 2013; Abreu et al., 2014). The tests along with a balance evaluation were performed at the end of the first and last day of the SSSH program.

Balance evaluation was done differently at both study locations. First, at the Independent Senior Living Facility, the evaluation consisted of an adapted Berg Balance Scale, six selected tasks with scores varying between 0 and four for all tasks, which were sitting to standing, standing with eyes closed, reaching forward with outstretched arm, retrieving object from floor, turning to look behind, standing with one foot in front and standing on one foot. The Berg Balance Scale has been successfully used to evaluate the effects of exercise in older adults (Gleeson et al., 2014; Aartolahti et al., 2020; Sadjapong et al., 2020). Later, at the Community Center, three unassisted standing positions (mountain pose, tandem stand and one-legged stand), which have been previously described in Ball et al. (Ball et al., 2013) were recorded in seconds.

The biomarkers were quantified in serum using ELISA assays following the respective manufacturers’ instructions. The following sandwich ELISA kits were used: Fast and slow sTnT, CTX-1 and Tnmd from MyBio Source; COMP, HA and SOST from R&D System; OCN and alpha-actinin-2 from Invitrogen. At the end of the first and last day of SSSH, after the physical fitness and balance evaluations, participants had their blood drawn into serum separator tubes by a skilled nurse and kept in cold until back in the lab. The length of time for blood drawing was consistent among participants, at both locations and at both times, within 15-35 min after exercise. Serum was then separated by manufacturer’s instructions, and aliquoted into eight tubes to be used on the different ELISA assays. Aliquots were stored at -80°C until used and serum samples were thawed only once. For each biomarker, in addition to the measured serum concentrations, we also calculated each subject’s percent change between the two measurements, before and after SSSH.

Non-parametric statistical tests were used throughout the study since results were not normally distributed. Differences between measurements before and after SSSH for the biomarkers (serum concentration), functional tests and balance scores were evaluated using the Wilcoxon Signed Ranks Matched Paired test. When comparing the biomarkers results by type of dwelling, the Mann-Whitney U tests were applied to the percent changes between before and after SSSH. Statistical values were reported as mean and standard deviation for pre- and post-SSSH measurements; median and interquartile range for the differences between post- and pre-SSSH. For all statistical tests the type 1 error rate was set at p < 0.05.

From the initial 21 participants recruited, two dropped from the study for unknown reasons, both from the Independent Living Facility, resulting in a final sample size of 19 participants, ten (10) at the Independent Living Facility and nine (9) at the Community Center, corresponding to an attrition of approximately10%. Attendance for individual subjects varied between 87.5 and 100% for the 16 SSSH classes at the two places. Except for one male subject, all others were females, and ages varied between 65 and 90 years with mean (SD) age of 72.9 (5.6) years.

Participants showed improvements in flexibility and upper/lower body strength based on the statistically significant changes seen after the SSSH exercise program for all functional performance tests. Table 1 shows for each test mean and SD (pre- and post-SSSH), median and interquartile range (IQR), mean percent change (in relation to initial measurement), and p-value (Wilcoxon Signed Ranks Matched Paired test).

As mentioned in the previous section, two different evaluations for balance were used at the Independent Living facility (adapted Berg Balance Scale) and at the Community Center (standing poses). The reason for these two evaluations is explained in the Discussion section. The results from the Berg test were [mean (SD)] before and after the SSSH classes, respectively, 18.9 (2.8) and 21.9 (1.2), representing a significant (p = 0.02) improvement of 15.9%. In relation to the Community Center, results were 24.7 (3.5) and 26.1 (4.9) before and after SSSH, respectively, a non-significant result (p = 0.20).

Comparisons between potential serum biomarkers for muscle, tendon/ligament and cartilage, and bone before and after an 8-week exercise program (Stay Strong Stay Healthy) are shown respectively in Figures 1–3. Table 2 shows for each biomarker mean and SD (pre- and post-SSSH), median and interquartile range (IQR), mean percent change (in relation to initial measurement), and p-value (Wilcoxon Signed Ranks Matched Paired test).

For all participants, regardless of dwelling, the difference in mean serum levels between before and after SSSH was statistically significant only for SOST, p = 0.003), in this case, a decrease in serum SOST after SSSH in relation to before SSSH. For all other comparisons, p-values varied between 0.117 and 0.952.

When reviewing the data, more specifically percent changes, it was noticed that for some biomarkers there was a clear distinction in pattern of percent increase versus decrease (for the paired values) depending on the type of dwelling. Percent changes were then tested for all biomarkers utilizing the Mann-Whitney U test, two-tailed. A significant difference for sTnT fast (p = 0.014) was found, with mean percent increase of 20.8% for those individuals in the independent living facility [median (IQR) = 19.0 (53.7)] versus a decrease of 16% for those still living in the community [median (IQR) = -19.9 (26.8)]. Furthermore, HA and OCN had close to significance p-values (respectively p = 0.050 and p = 0.060), which warrants further investigation on larger studies. Figure 4 shows percent changes (between pre- and post-SSSH) for sTnT fast for the Independent Living Facility and Community Center groups.

All measures of flexibility and upper/lower body strength significantly improved after 8 weeks of exercise intervention (SSSH). Currents results were not only in line with a previous study (Abreu et al., 2014), but also consistent to other works that used functional tests to evaluate on older adults the effects of different exercise programs (Sousa et al., 2014; Zhuang et al., 2014; Furtado et al., 2015). Handgrip strength measurements, when considering age and gender, were compatible with reference values for residents in the US (Wang et al., 2018). They were also in the range but slightly higher in absolute values (and post-exercise percent increase) than in that previous work, despite the current shorter time in this case of the SSSH Program (eight versus 10 weeks). Similar to functional tests, exercise programs in general lead to improvements in grip strength (Labott et al., 2019). Although functional tests and handgrip strength are valid indicators of the benefits of exercise to older adults, they may be less useful to make a distinction between different exercise regimens because in the great of majority of studies there is an improvement in those parameters.

In relation to balance, unfortunately how balance was evaluated needed to be changed. At the Independent Facility, the initial place of this study, the evaluation was done using the adapted Berg Balance Scale. However, during the post-SSSH evaluation one participant fell. The co-author running that part of the study decided to change how balance was evaluated at the Community Center. Although the adapted Berg Balance Scale was able to show improvements in balance following SSSH for the Independent Facility participants, the simpler evaluation at the Community Center did not show any difference. indicating the need of a better way of balance assessment.

When analyzing the data, the large variability in the serum biomarkers measurements, with standard deviations (SD) around 50% of the means, was clearly noticeable. Interestingly, this large variability was also seen in similar works in the literature (mentioned later in the Discussion). Possible explanations are the diverse biological processes that can influence the serum levels of those proteins (and polysaccharide) and how much there is still to know about them.

When participants were not divided by type of dwelling, only SOST serum levels were significantly different between pre- and post-SSSH. SOST has been described a “bone formation brake”, being one of the two major inhibitors of the pathway related to bone formation by osteoblasts (Rauch and Adachi, 2016). Therefore, a decrease in mean serum SOST after SSSH suggests less inhibition of bone deposition as part of bone homeostasis. Furthermore, when considering the effect of resistance exercise on muscles, the decline seen in circulating SOST supports further investigation on the role of muscle-bone crosstalk on the beneficial overall impact of exercise on musculoskeletal health.

When the data was analyzed considering the type of dwelling (Independent Living versus Community Center) other significant (or close to significance) differences between pre- and post-SSSH emerged in the Community Center group for TnT fast, HA and OCN. Which introduces an interesting point. Does the daily life level of activity affect the response of older adults to exercise? One possible way of answering this question would be the monitoring of the level of activity of each participant outside the exercise classes by wearable activity trackers. It would also be interesting to consider other types of dwelling, for example, assisted living and nursing homes. It is opportune to mention that studies investigating the response of older adults to exercise have focused predominantly on community-dwelling populations.

This study results were comparable to previously reported findings for COMP (Wakitani et al., 2007; Das Gupta et al., 2017; Jayabalan et al., 2021); HA (Montazeri et al., 2005; Singh et al., 2014); OCN (Khwanchuea and Punsawad, 2021; Nobrega da Silva et al., 2021); and SOST (Bhattoa et al., 2013).

In relation to the other proteins (sTnT, alpha-actinin and Tnmd), only one study has investigated previously serum levels sTnT (Abreu et al., 2014), while there are only genetic works on the physiological or pathological roles in humans of alpha-actinin and Tnmd.

The current work built upon previous work that found a decrease in serum skeletal muscle troponin T (sTnT), along with improvements in physical performance and grip strength, in older adults after two similar but distinct 10-week exercise programs, “Peer Exercise Program Promotes Independence (PEPPI)” and “Stay Strong Stay Healthy (SSSH)”, both designed having in mind the needs and limitations of older adults (Abreu et al., 2014). No difference in outcomes was found between PEPPI and SSSH, despite their slight differences (Abreu et al., 2014) so for this study the only exercise intervention was SSSH, this time at a shorter time (8 weeks) according to MU Extension current directives. Also, that initial study focused more on community-dwelling older adults and no special attention was paid to other types of dwelling and their intrinsic differences in terms of everyday activity level. Approximately half participants in this study were community-dwelling older adults, while the other half lived in an independent-living facility. Interestingly, sTnT serum concentration results from that previous work are more in line with sTnT slow in the current study.

In Figure 5, the most important results are summarized graphically, questioning the role of a crosstalk between muscle-bone in the overall response of the musculoskeletal system of older adults to exercise. Interesting points to note are that slow twitch sTnT, more associated with postural muscles, did not seem to be affected by SSSH, despite improvements in balance; the potential role of HA as biomarker for the effects of exercise in older adults, despite its widespread importance in the body; and that SOST and OCN moved in opposite directions of change, consistent with their respective anti-anabolic anabolic roles in bone homeostasis. Also, both studies with similar values for OCN intriguingly investigated overweight adolescents, which may be explained by the role of OCN, on its uncarboxylated form, in energy metabolism (Nobrega da Silva et al., 2021).

The main limitations of this work are its sample size and lack of diversity in the population studied, which limit the generalization of the results. Also, balance evaluation was limited and incomplete. On the other hand, this current work results, despite the relatively small sample size and large standard deviations, are encouraging towards a larger study on a more diverse population and include a more rigorous balance evaluation, as previously done (King et al., 2016; King et al., 2019).