Use of adult mesenchymal stromal cells in tissue repair: impact of physical exercise

Abstract

Physical exercise (PE) has unquestionable beneficial effects on health, which likely extend into several organ-to-cell physiological processes. At the cell scale, endogenous mesenchymal stromal cells (MSCs) contribute to tissue repair, although their repair capacities may be insufficient in paucicellular or severely damaged tissues. For this reason, MSC transplantation holds great promise for tissue repair. With the goals of understanding if PE has beneficial effects on MSC biology and if PE potentiates their role in tissue repair, we reviewed literature reports regarding the effects of PE on MSC properties (specifically, proliferation, differentiation, and homing) and of a combination of PE and MSC transplantation on tissue repair (specifically neural, cartilage, and muscular tissues). Contradictory results have been reported; interpretation is complicated because various and different species, cell sources, and experimental protocols, specifically exercise programs, have been used. On the basis of these data, the effects of exercise on MSC proliferation and differentiation depend on exercise characteristics (type, intensity, duration, etc.) and on the characteristics of the tissue from which the MSCs were collected. For the in vitro studies, the level of strain (and other details of the mechanical stimulus), the time elapsed between the end of exposure to strain and MSC collection, the age of the donors, as well as the passage number at which the MSCs are evaluated also play a role. The combination of PE and MSC engraftment improves neural, cartilage, and muscular tissue recovery, but it is not clear whether the effects of MSCs and exercise are additive or synergistic.

INTRODUCTION

There is evidence that regular physical exercise (PE) has unquestionable beneficial effects on health: for instance, regular aerobic fitness exercise reduces by 33% the risk of developing cardiovascular diseases (54); exercise-induced mobilization of circulating endothelial progenitor cells, involved in vascular repair and angiogenesis, may explain this positive effect (43). There is also evidence that these benefits depend on the PE nature: for instance, brisk walking has superior effects to slower pace to reduce cardiovascular-related mortality (47).

Focusing on the cell scale, literature provided evidence that PE promotes hematopoietic stromal cell and progenitor cell (HSPC) mobilization through adaptations in the medullary-niche architecture and increases in skeletal-muscle cytokine production (3, 10). In addition, the HSPC numbers in peripheral blood transiently increase following heavy exercise (10). In training of runners, HSPC numbers increase in both the bone marrow (BM) and extramedullary organs but decrease following heavy exercise (10).

Mesenchymal stromal cells (MSCs) behave as plastic-adherent, self-renewing cells with colony-forming capability that can differentiate into various cell types including osteocytes, chondrocytes, and adipocytes in vitro (39). MSCs are identified by the expression of a subset of cell markers (such as CD73, CD90, and CD 105⁺) and by the absence of hematopoietic markers (CD45, CD34, CD14, CD11b, CD79, or CD19 and HLA-DR) (14). MSCs, which form a reservoir of progenitors for the replacement of damaged or aging cells (8), have been isolated from various tissues (20, 50). Although direct replacement of lost or damaged cells by MSC differentiation has been observed for bone (16), cartilage (13), and liver (49), in vivo, MSCs have now been demonstrated to synthetize and release paracrine factors, highlighting their primary role as a stromal support for tissue regeneration (42). Such replacement capacities, however, may either be at a very low rate (7) or sometimes overcome, for instance, in paucicellular (32) and severely damaged tissues (56). In such cases, MSC transplantation holds substantial promise for tissue repair (2, 6, 12, 31, 34), specifically for the treatment of cardiovascular (34), neurodegenerative (2), and osteo-articular lesions (6, 12, 31). We, therefore, hypothesized that PE may improve endogenous- and transplanted-MSC properties.

One investigator (C. Bourzac) searched PubMed, using the following strategy: (mesenchymal stem cells or mesenchymal stromal cells) and (exercise or physical exercise) and (transplantation or implantation) and (tissue repair or tissue healing) with English and French language restriction. Inclusion criteria were thorough description of the exercise protocols, data on MSC properties, and concomitant (<48 h between MSC transplantation and start of the exercise) use of MSC and exercise for tissue repair. In vivo studies pertaining to low-magnitude vibrations, believed to mimic high-impact exercises (18, 44), were also included. In vitro studies (stretching…), hematopoietic stem cells studies or limb unloading studies were excluded. The present review outlines the effects of PE on MSC properties (specifically, proliferation, differentiation, and homing), as well as the combined effects of PE and MSC transplantation in the repair of neural, cartilage, and muscular tissues in humans and rodents.

EFFECTS OF PHYSICAL EXERCISE ON MESENCHYMAL STROMAL CELL PROPERTIES

Details regarding exercise regimes and MSC characterization are provided in Table 1.

Table 1.

Effects of physical exercise on mesenchymal stromal cell properties

Properties/Exercise	Exercise Regime	Effects	Method of Assessment	MSC Origin	MSC Characterization	Species	Ref.
Proliferation
MCR	14 m/min (1st wk) to 24 m/min (10th wk), 45 min/day, 3 days/wk, 10 wk	= CFU-f (29% as many as in sedentary controls but P > 0.05)	CFU count	BM (tibia, femur)	Adherence to plastic, Sca-1+, c-kit+	Mice	(3)
MCR	14 m/min (1st wk) to 28 m/min (5th wk), 45 min/day, 5 days/wk, 5 wk	↑ CFU-f (21 ± 2 vs. 1.8 ± 2 × 106 Cells in sedentary controls, P < 0.05)	CFU count	BM (tibia, femur)	Mononucleated, Sca-1+ Lin− CD45−CD31− CD51+cells	Mice	(32)
	19.3 m/min, 5° slope, 60 min/day, 8 wk	↑ CFU-f of 3rd passage BM and AT-MSC vs. sedentary controls ↑ CFU-f of 6th passage AT-MSC vs. sedentary controls = CFU-f of 6th passage BM-MSCs vs. sedentary controls	CFU count	BM (femur) AT(inguinal fat pad)	CD11b− CD45− CD79+ CD90+ Differentiation in osteocyte and adipocyte	Rats	(27)
	13 m/min, 30 min/day, 5 days/wk, 5 wk	↑ CFU (and size of colonies) of TSC	CFU count + morphology	Tendon (patellar)	Oct-4, Nanog, Nucleostemin	Rats	(58)
	12 m/min, 30 min/day, 5 days/wk, 1 wk	= MSC numbers in humerus (vs. sedentary controls, P = 0,39) ↑ MSC numbers in femur (vs. sedentary controls, P = 0,044)	Flow cytometry	BM (humerus, femur)	Sca-1+, c-Kit+, CD44+, CD90.2+, CD105+	Mice	(50)
	14 m/min (1st wk) to 28 m/min (5th wk), 45 min/day, 5 days/wk, 5 wk	↑ MSC numbers (4 ± 2 × 106 vs. 16 ± 2 in sedentary controls, P < 0.05) after 21 days in culture	Flow cytometry	BM (tibia, femur)	Sca-1+ Lin− CD45− CD31− CD51+	Mice	(32)
		↑ % of CD105+ MSC	Immunochemistry	BM (tibia, femur)	Sca-1+ Lin− CD45− CD31− CD51+	Mice	(32)
	Single bout, warming 8 m/min, 2 m/min, increase every 10 min until 16 m/min, 16 m/min 30 min, 18 m/min 20 min, 0% slope	↑ PDGFRα+ MSC numbers (68.43 ± 4.43% vs. 57.02 ± 8.00% in sedentary controls, P = 0.01)	BrdU incorporation	BM (tibia, femur)	Sca-1+, c-Kit+, Lin−, CD45−, PDGFRα+	Mice	(15)
	13 m/min, 30 min/day, 5 days/wk, 5 wk	↓ PDT of TSC by 35% vs. sedentary controls (P < 0.05)	Cell count	Tendon (patellar)	Oct-4, Nanog, Nucleostemin	Rats	(58)
	50 min/day, 5 days/wk, 3 wk; speed not reported	↓ PDT of TSC vs. sedentary controls (P < 0.05) ↑ TSC numbers by 88% (vs. sedentary controls, P < 0.05)	Cell count and morphology	Tendon (patellar)	Oct-4, NS, Sca-1, SSEA-1	Rats	(57)
LMHFV	15 min/day, 5 day/wk, 90 Hz, 0.2 g peak acceleration	↑ MSC numbers by 46,1% (vs. sedentary controls, P = 0.022)	Flow cytometry	BM (femur, tibia)	Sca-1+, Pref-1+	Mice	(29)
DRE	17 m/min, 30 min, −20° slope, acute single bout	↑ Muscle-derived MSC	Flow cytometry	Muscle (gastrocnemius, soleus)	Sca-1+ CD45−	Mice	(51)
CRT	Knee flexions, 33 min/day, 3 days/wk, 12 wk (75–85° range of motion, 2–3 s tempo, eccentric loading = 120% of concentric loading)	↑ MSC numbers vs. before training (P < 0.01)	Immunohistochemistry Ki-67	Muscle (vastus lateralis)	Mononucleated, PDGFRα+ CD90+	Humans	(17)
MCR	15 m/min, 30 min/day, 5 days/wk, 3 mo	↓ MSC apoptosis (vs. sedentary controls, P < 0.05)	Caspase-3 expression RT-PCR	BM (femur, tibia)	Adherence to plastic	Rats	(21)
	19.3 m/min, 5° slope, 60 min/day, 8 wk	= MSC apoptosis	Caspase-3 activity	BM (femur) AT (inguinal fat pad)	CD11b− CD45− CD79+ CD90+, osteocyte and adipocyte differentiation	Rats	(27)
VCE	66–73 m/day for 12–13 min/day, 4 wk	= CFU-f vs. sedentary controls	CFU count	BM (tibia)	Adherence to plastic	Mice	(34)
Differentiation
Osteogenic
ICR	95% maximum oxygen uptake; 28 m/min, 50 min/day, 6 days/wk for 5 wk, 5° slope	= ALP-positive CFU-f vs. sedentary controls	ALP staining + cell count	BM (femur)	Adherence to plastic	Mice	(59)
MCR	85% maximum oxygen uptake; 18 m/min, 50 min/day, 6 days/wk for 5 wk, 5° slope	↑ ALP-positive CFU-f vs. sedentary controls (P < 0.05)	ALP staining + cell count	BM (femur)	Adherence to plastic	Mice	(59)
	14 m/min (1st wk) to 28 m/min (5th wk), 45 min/day, 5 days/wk, 5 wk	↑ ALP activity (19 ± 1 vs. 16 μU/mL ± 1 in sedentary controls, P < 0.05)	Colorimetric assay	BM (tibia, femur)	Sca-1+ Lin− CD45− CD31− CD51+	Mice	(32)
	15 m/min, 30 min/day, 5 days/wk, 3 mo, 0° slope	↑ ALP activity in osteopenic rats (vs. sedentary osteopenic controls, P < 0.05) at 7 but not 14 and 21 days of culture	Colorimetric assay	BM (femur)	CD45− CD54 CD73 CD90	Rats	(36)
	19.3 m/min, 5° slope, 60 min/day, 8 wk	↑ ALP activity in 3rd and 6th passage BM-MSCs (vs. sedentary controls, P < 0.05)	pNPP	BM (femur) AT (inguinal fat pad)	CD11b− CD45− CD79+ CD90+ Differentiation in osteocyte and adipocyte	Rats	(27)
	15 m/min, 30 min/day, 5 days/wk, 3 mo	↑ ALP expression in adult rats (vs. sedentary controls, P < 0.05) = ALP expression (vs. young rats, P > 0.05) at 7, 14, 21 days of culture	qRT-PCR	BM (femur, tibia)	Adherence to plastic	Rats	(21)
	19.3 m/min, 5° slope, 60 min/day, 8 wk	↑ ALP-mRNA expression with 3rd and 6th passage MSC (vs. sedentary controls, P < 0.05) = ALP-mRNA expression with 3rd and 6th passage MSC (vs. sedentary controls, P < 0.05)	RT-PCR	BM (femur) AT (inguinal fat pad)	CD11b− CD45−CD79+ CD90+ Differentiation in osteocyte and adipocyte	Rats	(27)
	15 m/min, 30 min/day, 5 days/wk, 3 mo	↑ OCL expression in adult rats (vs. sedentary controls, P < 0.05) = OCL expression (vs. young rats, P > 0.05), at 21 days of culture = Sialoprotein expression in adult rats (vs. sedentary controls, P > 0.05) = Col I expression in adult rat (vs. young rats, P > 0.05)	qRT-PCR	BM (femur, tibia)	Adherence to plastic	Rats	(21)
	14 m/min (1st wk) to 28 m/min (5th wk), 45 min/day, 5 days/wk, 5 wk	↑ OCL level (170 ± 1 vs. 130 ng/mL ± 1 in sedentary controls, P < 0.05) ↑ OPN level (48 ± 2 vs. 32 ng/mL ± 2 in sedentary controls, P < 0.05)	ELISA	BM (tibia, femur)	Sca-1+ Lin− CD45− CD31− CD51+	Mice	(32)
Differentiation
Osteogenic
MCR	19.3 m/min, 5° slope, 60 min/day, 8 wk	↑ OCL mRNA expression for 3rd passage BM-MSCs (vs. sedentary controls, P < 0.05) = OCL mRNA expression for 6th passage BM-MSCs (P > 0.05) ↑ Col I mRNA expression for 3rd passage BM-MSCs (vs. sedentary controls, P < 0.05), = Col I mRNA expression for 6th passage AT-MSC (P > 0.05) = OCL mRNA expression for 3rd and 6th passage AT-MSC (vs. sedentary controls, P > 0.05) = Col I mRNA expression for 3rd and 6th passage AT-MSC (vs. sedentary controls, P > 0.05)	RT-PCR	BM (femur) AT (inguinal fat pad)	CD11b− CD45− CD79+ CD90+ Differentiation in osteocyte and adipocyte	Rats	(27)
	14 m/min (1st wk) to 28 m/min (5th wk), 45 min/day, 5 days/wk, 5 wk	↑ % of absorbed alizarin red dye (81 ± 9 vs. 56 ± 3 in sedentary controls, P < 0.05)	Alizarin red staining + DO	BM (tibia, femur)	Sca-1+ Lin− CD45− CD31− CD51+	Mice	(32)
	19.3 m/min, 5° slope, 60 min/day, 8 wk	↑ %Absorbed alizarin red dye for 3rd passage BM-MSCs (vs. sedentary controls, P < 0.05) = %Absorbed alizarin red dye for 6th passage BM-MSCs (vs. sedentary controls, P < 0.05) = %Absorbed alizarin red dye for 3rd and 6th passage AT-MSC (vs. sedentary controls, P < 0.05)	Alizarin red staining + DO	BM (femur) AT (inguinal fat pad)	CD11b− CD45− CD79+ CD90+ Differentiation in osteocyte and adipocyte	Rats	(27)
	15 m/min, 30 min/day, 5 days/wk, 3 mo	↑ Numbers of mineralized nodules (vs. sedentary controls, P < 0.05)	Silver stain + count	BM (femur, tibia)	Adherence to plastic	Rats	(21)
	15 m/min, 30 min/day, 5 days/wk, 3 mo, 0° slope	↑ Numbers of mineralized nodules in osteopenic rats (8.85 ± 0.9, 8.27 ± 2.7, 6.47 ± 0.3 vs. 0.51 ± 0, 1.70 ± 0.1, 0.02 ± 0.03 in sedentary osteopenic controls, at 7, 14, and 21 days respectively, P < 0.05, but < that in nonosteopenic rats, P < 0.05)	Silver stain + count	BM (femur)	CD45 CD54 CD73 CD90	Rats	(36)
	19.3 m/min, 5° slope, 60 min/day, 8 wk	= Numbers of mineralized nodules for 3rd and 6th passage MSC (vs. sedentary controls, P < 0.05)	Alizarin red staining + DO	AT	CD11b− CD45− CD79+ CD90+ Differentiation in osteocyte and adipocyte	Rats	(27)
	14 m/min (1st wk) to 28 m/min (5th wk), 45 min/day, 5 days/wk, 5 wk	↓ %Absorbed Oil red O dye (17 ± 3 vs. 46 ± 3 in sedentary controls, P < 0.05)	Oil red O staining + DO	BM (tibia, femur)	Sca-1+ Lin− CD45− CD31− CD51+	Mice	(32)
Adipogenic
MCR	19.3 m/min, 5° slope, 60 min/day, 8 wk	↓ %Absorbed Oil red O dye for 3rd passage MSC (vs. sedentary controls, P < 0.05) = %Absorbed Oil red O dye for 6th passage MSC (vs. sedentary controls) ↓ %Absorbed Oil red O dye 3rd and 6th passage MSC (vs. sedentary controls, P < 0.05)	Oil red O staining + DO	BM AT	CD11b− CD45− CD79+ CD90+ Differentiation in osteocyte and adipocyte	Rats	(27)
	19.3 m/min, 5° slope, 60 min/day, 8 wk	↓ PPARγ2 mRNA for 3rd passage MSC (vs. sedentary controls, P < 0.05) ↓ C/EBP-α mRNA for 6th passage MSC (vs. sedentary controls, P < 0.05)	RT-PCR	BM, AT	CD11b− CD45− CD79+ CD90+ Differentiation in osteocyte and adipocyte	Rats	(27)
Tenogenic
MCR	13 m/min, 30 min/day, 5 days/wk, 5 wk	↑ Col I (5-fold; vs. sedentary controls, P < 0.05) ↑ Col III (3-fold; vs. sedentary controls, P < 0.05) ↑ Tenomodulin (2-fold; vs. sedentary controls, P < 0.05) ↓ >50% PPARγ2, Runx2 and Col II (vs. sedentary controls, P < 0.05)	RT-PCR	Tendon (patellar)	Oct-4, Nanog, Nucleostemin	Rats	(58)
Osteogenic
LCR	65% maximum oxygen uptake; 8 m/min, 50 min/day, 6 days/wk for 5 wk, 5° slope	= ALP-positive CFU-f vs. sedentary controls ↑ ALP-positive CFU-f (vs. sedentary controls, P < 0.05) at 2 wk ↑ Size of mineralized nodules (vs. sedentary controls, P < 0.05) at 2 wk	ALP staining + count	BM (femur)	Adherence to plastic	Mice	(59)
VCE (light)	66–73 m/day for 12–13 min/day, 4 wk		ALP staining + count Alizarin red staining area	BM (tibia)	Adherence to plastic	Mice	(34)
Cycling ergometer (intensive)	Exercise to exhaustion (max 8 min)	= Proliferation rate of 3rd passage MSC cultured with serum collected before warm-up (7%) vs. immediately after exhaustion (5%) vs. 1 h after exhaustion (9%) = Apoptosis for 3rd passage MSC before warm up, immediately after exhaustion and 1 h after exhaustion	immunohistochemistry Ki-67 Immunochemistry activated caspase-3	BM (femur)	Gradient density, adherence to plastic	Humans	(45)
		↑ Migration of 3rd passage MSC by 120% (immediately after exhaustive exercise, vs. before exercise, P < 0.05) ↑ Migration of 3rd passage MSC by 132% (1 h after exhaustive exercise, vs. before exercise, P < 0.05)	Boyden chamber	BM (femur)	Gradient density, adherence to plastic	Humans	(45)
CTE	Warming 8 m/min, increasing 2 m/m every 10 min until 16 m/min, 16 m/min for 30 min, then 18 m/min 20 min, 0% slope	Secretion of G-CSF, SCF, IL-2, IL-17, STNFRI, IL-3, thrombopoietin, IL-5, and IFN-γ was decreased	Cytokine assay	BM (femur, tibia)	Lin−, CD45−	Mice	(15)

ALP, alkaline phosphatase; AT, adipose tissue; BM, bone marrow; BrdU, 5-bromo-2-deoxyuridine; CFU, colony-forming unit; Col I, type I collagen; Col II, type II collagen; Col III, type III collagen; CRT, contraction resistance training, DRE, downhill running exercise, Fabp4, fatty acid binding protein 4 adipocyte; HU, hind limb unloading; ICR, intensive continuous running; LCR, low-intensity continuous running; MCR, moderate continuous running; MSC, mesenchymal stem cells; OCL, osteocalcin; OPN, osteopontin; Osx, Osterix; PDT, population doubling time; pNPP, p-nitrophenyl phosphate assay; PPARγ2, peroxisome proliferator-activated receptor γ2 (a master transcription factor during adipogenesis); TSC, tendon stem cells; VCE, voluntary climbing exercise.

Physical Exercise Affects Mesenchymal Stromal Cell Proliferation and Apoptosis

In some diseases, for instance osteoporosis, the number of MSCs is insufficient for bone regeneration (19). PE is recommended in osteoporotic people to improve bone health. Is PE altering MSC proliferation and apoptosis? The effects of various exercise regimes on MSC proliferation and apoptosis were investigated in rodent models and human athletes (Table 1 and Fig. 1).

Fig. 1.

Effects of physical exercise on mesenchymal stromal cell (MSC) properties. The effects of physical exercise on MSC proliferation, apoptosis, and homing (A) and differentiation (osteogenic, adipogenic and tenogenic) (B) capacities are presented. Green arrows indicate a positive effect (increased capacity), red arrows indicate a negative effect (decreased capacity), and blue arrows indicate an absence of effect (similar capacity) of physical exercise on MSC properties.

In rodents, most reports pertaining to MSC proliferation have been focusing on the effects of moderate continuous treadmill exercise (CTE) on BM-MSC, and results from these studies are conflicting. For instance, when compared with results obtained from sedentary controls, a higher [21 ± 2 vs. 16 ± 2, P < 0.05 (33)] or a similar [P > 0.05 (3)] number of colony-forming units (CFU)-f was obtained with BM-MSCs collected from femurs or tibiae of mice submitted to moderate CTE for several weeks. Parameters in these studies that may have contributed to the observed discrepancies include duration of the exercise protocols [5 (33) vs. 10 (3) wk; see Table 1 for further details], age at the time of BM-MSC collection [10 (33) vs. 15 (3) wk], and cell culture duration before CFU-f count [21 (33) vs. 14 (3) days]. When considering an acute single bout exercise, the proliferation of BM-MSCs collected from 4-wk-old C57B1/6 mice (assessed by the number of 5-bromo-2-deoxyuridine-positive MSC) submitted to CTE was increased (P = 0.01) in runners, compared with respective results in sedentary controls (15). Similarly, an acute single bout of downhill running (eccentric exercise) resulted in increased numbers (2.2-fold, P < 0.001) of muscle-derived (m) MSC in skeletal muscles collected from wild-type mice (age, sex not reported) compared with respective results obtained from sedentary mice (51). The appearance of these mMSC in skeletal muscles was shown to be regulated by α₇-integrin (51).

Regarding tendon tissue, population-doubling time of patellar tendon-derived stem cells (TSC) collected from 20-wk-old Sprague-Dawley rats (sex not reported) exposed to moderate CTE was 35% shorter (P < 0.05) compared with respective results from sedentary controls (60). Moreover, the colony numbers of TSC and the number of TSC per colony, were higher (P < 0.05) in the exercising, compared with sedentary rats (60). Similarly, the numbers of TSC collected from the patella of aging (9- and 24-mo-old) mice submitted to moderate CTE (50 min/day, 5 days/wk for 3 wk; speed not reported) were higher (P < 0.05) compared with controls (59).

In the studies above, the CFU-f numbers may have been affected by cell passage and tissue origin. Supporting evidence was provided by the results that the CFU-f numbers were higher (P < 0.05) for passage 3 but not passage 6 MSCs when collected from bone marrow (from 13- to 14-wk-old male Wistar rats exposed to moderate CTE) compared with sedentary controls. In contrast, the CFU-f numbers were higher (P < 0.05) for either passage 3 or passage 6 MSC when collected from adipose tissue (AT) (28). Within a given tissue, for instance bone marrow, MSC proliferation may also be affected by the specific site of cell collection. Supporting evidence was provided by the results that BM-MSC numbers collected from the humerus of 6-wk-old, female Hsd:ICR mice were similar (P = 0.39) between mice submitted to moderate CTE and sedentary controls, whereas in the femur CTE lowered (P = 0.044) the number of BM-MSCs (52). Since the baseline density in BM-MSCs was similar (P = 0.61) between the two bones in the control group, the lowered number in the running group was interpreted as MSC having been signaled away from their progenitor state in the femur (52).

Voluntary-climbing exercise (VCE) on a 100 cm mesh-wire tower (66–73 m/day for 12–13 min/day for 4 wk) did not affect the CFU-f numbers for BM-MSCs collected from the femurs of 8-wk-old mice at 2 and 4 wk of the experiment (35). Unfortunately, these results are based on a single study. In contrast, low-magnitude high-frequency vibrations (LMHFV; provided by a vertically oscillating platform at 90 Hz, 0.2 g, 15 min/day, 5 days/wk for 6 wk) applied to 7-wk-old male mice, increased, although not significantly (P = 0.23), MSC numbers in bone marrow compared with respective results obtained from control mice with no vibrations (30).

In humans, the effects of a single bout of intense exercise (46) or contraction resistance training (17) were evaluated on MSC proliferation. Five athletes (age and sex not reported) were submitted to a single bout of cycling ergometer exercise (warm-up for 12 min followed by exhausting exercise), and sera were collected before warm-up, immediately after exhaustion, and 1 h after exercise, and used as supplement in in vitro culture of human BM-MSCs (donor age and sex not reported) (46). The percentages of proliferative, passage 3-BM-MSCs were similar whether these cells were cultured with serum collected before warm-up (7%), at exhaustion (5%), and 1 h after exercise (9%) (46). In contrast, the numbers of mMSC, collected from the vastus lateralis of young (mean age: 23.4 yr) men were increased (P < 0.01) following concentric (one leg) or eccentric (the contralateral leg) contraction resistance training, compared with results obtained before exercise initiation (17).

Considering MSC apoptosis, the results reported in the literature are inconclusive. In one in vivo study, moderate CTE (15 m/min, 30 min/day, 5 days/wk for 3 mo) increased (P < 0.05) cleaved caspase-3 expression in passage 4 BM-MSCs from 6-mo-old, female, Wistar rats compared with sedentary rats (22). The percentage of BM-MSCs per field in both groups, however, was similar, suggesting that PE is associated with both a higher rate of apoptosis and an increased rate of cell proliferation (22). In another study, moderate CTE (19.3 m/min, on a 5° incline, 60 min/day for 8 wk) did not affect apoptosis of BM-MSCs from 13- to 14-wk-old, male Wistar rats (28). Similar results were obtained in vitro: the number of apoptotic human BM-MSCs was similar after the BM-MSCs were cultured with athletes’ sera obtained either immediately after exhaustive exercise or 1 h later (46).

In summary, literature reports provided evidence that the effects of exercise on MSC proliferation and apoptosis depend on exercise characteristics, specifically type (such as treadmill running, climbing, vibration, etc.), intensity (low, moderate, intense, etc.), duration of exposure to exercise conditions (acute single bout or chronic), age of the subjects, and origin of the tissue from which the MSC were collected. The time elapsed between the end of exposure to exercise and MSC collection, as well as the passage number at which the MSC are evaluated, may play a role.

Physical Exercise Affects the Differentiation Capacity of Mesenchymal Stromal Cells

Little literature data is available to determine which specific cell-lineage differentiation may be driven by PE (Table 1).

Considering running exercise, moderate CTE may increase the osteogenic potential of BM-MSCs. Supporting evidence was provided by the results that alkaline phosphatase (ALP)-positive CFU-f was increased (P < 0.01) for BM-MSCs obtained from 2-mo-old, male mice subjected to CTE, compared with respective results obtained from sedentary mice (61). Similarly, ALP activity for BM-MSCs obtained from 5-wk-old mice was enhanced (P < 0.05) when these mice were subjected to moderate CTE; the adipogenic potential of BM-MSCs, collected from exercising mice, was however decreased (P < 0.05) (33). Moreover, moderate CTE decreases osteogenic differentiation associated with age or osteoporosis. Supporting evidence is provided by the results that the osteocalcin (at 21 days of cell culture) and ALP (at 7, 14, and 21 days of culture) expressions for BM-MSCs obtained from 6-mo-old, female Wistar rats, participating in a moderate CTE, were comparable to those obtained from 1-mo-old, sedentary rats at the same culture time points (22). Additionally, moderate CTE [using the same exercise protocol (22)] resulted in increased (P < 0.05) ALP activity in BM-MSCs collected from 5-mo-old, osteopenic, female, Wistar rats, only after 7 days of cell culture, compared with results obtained from sedentary control rats (37). Increased capacity to produce mineralized nodules was also observed with BM-MSCs collected from these osteopenic (after 7, 14, and 21 days of cell culture in osteogenic medium) (37) or 6-mo-old (22) rats subjected to moderate CTE compared with their respective sedentary controls. The number of mineralized nodules remained, however, lower (P < 0.05) than that obtained with BM-MSCs collected from nonosteopenic (37) or 1-mo-old (22) rats not exposed to exercise. Evidence that nitric oxide-synthase inhibition decreased ALP activity in nonosteopenic and in active osteopenic (but not in sedentary) rats (37) led to the conclusion that moderate CTE enhanced the osteogenic capacities of BM-MSCs through the nitric oxide-mediated pathway.

In contrast, high- or low-intensity CTE resulted in similar osteogenic capacities (as assessed by the numbers of ALP-positive CFU-f) between runners and control mice (61). LMHFV (applied at 0.2 g, which is lower than earth’s gravitational field) is low-intensity exercise, even relative to normal weight bearing. In 7-wk-old, male, C57BL/6J mice, LMHFV (applied at 0.2 g, 90 Hz, 15 min/day, 5 days/wk, for 6 wk) resulted in increased BM-MSCs osteogenic capacities (showed by upregulated transcription of Runx2) compared with control mice not subjected to vibration (30). Downregulation of peroxisome proliferator-activated receptor-γ was also observed in mice subjected to LMHFV. It was therefore suggested that these changes in osteogenesis and adipogenesis obtained with LMHFV were not related to exercise-induced increased metabolic activity (considering the low loading and low signal magnitude) (45). It was also demonstrated that MSC differentiation was driven to musculoskeletal lineage, away from an adipocytic fate (45).

Similar to moderate CTE, VCE (exercise of moderate intensity) enhanced (P < 0.05) the osteogenic potential following 2 wk of the exercise program in 7- to 8-wk-old, male mice. The results, however, were not different after 4 wk of the exercise program (35).

In the case of TSC, exposure to moderate CTE resulted in increased (P < 0.05) expression of the tenocyte genes compared with sedentary controls (60). Exposure to CTE also resulted in decreased, by >50% (P < 0.05), non-tenocyte-related gene expression, compared with results from sedentary controls (60).

Along with MSC proliferation, MSC differentiation capacities may have been affected by cell passage and tissue origin. In a study that compared mRNA-levels from bone marrow- and AT-derived MSC collected from rats submitted to moderate CTE (28), higher (P < 0.05) ALP-mRNA expression was observed in passage 3 and passage 6 BM-MSCs, compared with respective results in MSC obtained from sedentary controls. Upregulation of osteocalcin and type I collagen mRNA expression was observed only for passage 3 BM-MSCs (28). ALP staining of for passage 3 and passage 6 cells and mineral content of passage 3 cells were increased only for the BM-MSCs under osteogenic culture conditions after exercise training. The results (mRNA levels, ALP staining, and mineral content) obtained with AT-MSC from exercising rats were, however, similar to those obtained with either for passage 3 and passage 6 AT-MSC from sedentary rats (28). Exposure to exercise resulted in lower (P < 0.05) adipogenic capacity only for passage 3 BM-MSCs and for passage 3 and passage 6 AT-MSC, compared with sedentary controls (28).

In summary, literature reports provided evidence that similar exercise programs result in different patterns of gene expression depending on the tissue from which the MSC were harvested. In addition, various populations of MSC exhibit different patterns of gene expression depending on the type of exercise to which the animals tested are exposed.

Physical Exercise Affects Mesenchymal Stromal Cell Homing

MSC homing is the migration and arrest of MSC within the vasculature of a tissue followed by transmigration across the endothelium (24) and engraftment in the target tissue. MSC homing to injured tissues depends on the systemic and local inflammatory state (40). Does PE affect MSC homing (Table 1)?

The effects of PE on MSC migration capacities were evaluated using sera collected from five athletes (age and sex not reported) submitted to exhaustive exercise on an ergometer as a chemoattractant in a Boyden chamber (46). Sera were collected before warm-up, immediately after exhaustive exercise, and 1 h later. After 8 h of incubation, migration of passage 3-human BM-MSCs was increased by 120% (P = 0.021) and 132% (P = 0.015) using sera collected immediately after exhaustive exercise and 1 h later, respectively, compared with results obtained with sera collected before exercise (46). The observed migratory function of BM-MSCs could be in response to chemical compounds in the serum following PE rather than to a direct effect of the physical stimulus on the cells. Supporting evidence is provided by a markedly (P < 0.05) increased concentration of interleukin-6 (known to promote increased motility of MSC) in the serum after exhaustive exercise (46).

Physical Exercise Affects Mesenchymal Stromal Cell Secretome

MSCs have been demonstrated to synthetize and release paracrine factors, providing support for tissue repair (42). Literature pertaining to the effects of exercise on MSC secretome is limited to one study. Following an acute single bout of CTE, secretion of G-CSF, SCF, IL-2, IL-17, STNFRI, IL-3, and thrombopoietin from BM-MSCs collected from 4-wk-old C57B1/6 mice was increased compared with respective results in sedentary controls. Conversely, secretion of IL-5 and IFN-γ was decreased (15). Some of these factors are involved in HSPC proliferation and mobilization into peripheral circulation.

MSCs are considered ideal therapeutic targets for regenerative medicine. Considering the effects of PE on MSC properties, PE could be used as an adjuvant for MSC transplantation for tissue repair.

EFFECTS OF PHYSICAL EXERCISE COMBINED WITH MESENCHYMAL STROMAL CELL TRANSPLANTATION ON NEURAL, CARTILAGE, AND MUSCULAR TISSUE REPAIR

MSC therapy has gained great interest in neurological, osteo-articular, and cardiovascular diseases. Since PE is recommended for the prevention and management of such conditions, literature reports have mainly focused on the combined effects of PE and MSC for neural, cartilage, and muscular tissue repair (Table 2 and Fig. 2).

Table 2.

Detailed effects of the combination of exercise and mesenchymal stem cells on tissue repair

Tissue/Type of exercise	MSC Origin	Model of Injury	MSC Route of Administration	Effects	Ref.
Neural
Swimming	Heterologous BM	Sciatic nerve transection	At injury site	= Sciatic function index, vertical locomotor activity, ankle angle, compound-muscle action potential characteristics 4 wk after sciatic nerve transection (vs. swimming alone, water 30°C, P > 0.05) ↑ Sciatic function index, vertical locomotor activity, ankle angle, compound-muscle action potential characteristics 4 wk after sciatic nerve transection (vs. BM-MSCs alone, P < 0.05)	(53)
	Heterologous BM	Sciatic nerve crushing	At injury site	↑ Sciatic function index, vertical locomotor activity, ankle angle, compound-muscle action potential peak amplitudes, and onset latencies 4 wk after sciatic nerve crushing (vs. either BM-MSCs alone or swimming alone, water 0°C, P < 0.05) = Vertical locomotor activity and ankle angle (vs. BM-MSCs alone, P < 0.05)	(57)
MCR	Heterologous BM	Cerebral ischemia	Intravenous	↓ Neurological severity score (5.33 ± 0.82) vs. BM-MSCs alone (7.33 ± 0.82) or exercise alone (7.50 ± 1.37), P < 0.01	(62)
Cartilage
MCR	Heterologous BM	Osteochondral defect	At injury site	↓ Total Wakitani score (vs. exercise alone, BM-MSCs alone, P < 0.05) at 4 wk postinjury = Total Wakitani score (vs. exercise alone, BM-MSCs alone, P > 0.05) at 2 wk postinjury = Total Wakitani score (vs. exercise alone, BM-MSCs alone, P > 0.05) at 8 wk postinjury ↑ % area Col II (vs. exercise alone, P < 0.05) at 4 wk postinjury = % area Col II (vs. exercise alone, BM-MSCs alone, P < 0.05) at 4 wk postinjury	(55)
Muscle
MCR	Heterologous BM	Myocardial infarction	Intravenous	= Myocardial infarction extension (vs. exercise alone, BM-MSCs alone, P > 0.05) = LVID-D, LVID-S, LVFW-D, IVS-D, IVS-S, ejection fraction, fractional shortening (vs. exercise alone, BM-MSCs alone, P > 0.05 ↓ Time to peak extension (vs. BM-MSCs alone, P < 0.05)	(26)
Downhill running	Heterologous		At site	↑ DiI-negative fibers (indirect new fiber synthesis) in close proximity to DiI-positive MSC (vs. no exercise, P < 0.05)	(51, 63)

BM, bone marrow; Col II, type II collagen; LVID-D, left ventricular internal dimension in diastole; LVID-S, left ventricular internal dimension in systole; LVFW-D, left ventricular free wall dimension in diastole; IVS-D, thickness of the interventricular septum in diastole; IVS-S, thickness of the interventricular septum in systole; MCR, moderate continuous running; MSCs, mesenchymal stromal cells.

Fig. 2.

The effects of a combination of physical exercise and mesenchymal stromal cell (MSC) transplantation on tissue repair. The effects of a combination of physical exercise and MSC transplantation on cartilage repair and on the functional recovery of cardiac, cerebral, and neural injuries are presented. Green arrows indicate a positive effect of this combination on tissue repair, and blue arrows indicate the absence of effect.

Effects of Physical Exercise and MSC Transplantation on Neural Tissue Repair

Exercise training (walking on a 20°-upslope treadmill for 1 h/day, 5 days/wk, for 2 wk) elicited enhanced axon regeneration but not functional recovery following peripheral nerve injury in female Lewis rats (age not reported) (11). The synergetic effects of exercise and BM-MSCs transplantation were evaluated in adult, male, Sprague-Dawley rats (age not reported) for their recovery from sciatic nerve injury (53, 57). The combination of 30 min of swimming in 30°C water (every day for 7 days, starting 12 h after nerve transection) and heterologous BM-MSCs transplantation (at the time of surgery) at the injury site improved nerve functional recovery 4 wk postinjury compared with BM-MSCs alone but not compared with swimming alone (53). The same protocol (swimming and BM-MSCs), except for an average water temperature of 16.5°C, was applied to a crush model of sciatic nerve injury in rats (57). Compared with animals exposed to either BM-MSCs alone or swimming alone, the sciatic function index, vertical locomotor activity, ankle angle, compound-muscle action potential peak amplitudes, and onset latencies were improved (P < 0.05) in the animals exposed to combined swimming and BM-MSC transplantation 4 wk after injury. The vertical locomotor activity and ankle angle, however, were not different when the animals were exposed to either combined swimming and BM-MSC transplantation or BM-MSC transplantation alone 10 days after injury (57). The differences observed in functional recovery between the two aforementioned studies may be explained by the two different models of injury used, and thus the severity of the lesion was induced. Alternatively, cold-induced decrease in inflammation may have played a role.

In a model of cerebral ischemic injuries in adult, male, Sprague-Dawley rats (age not reported), moderate CTE (12 m/min, 30 min/day for 14 days, starting 48 h after injury) associated with intravenous administration of BM-MSCs 24 h after injury resulted in greater improvement in neurological function compared with results obtained from animals exposed to either BM-MSCs alone (P < 0.01) or exercise alone (P < 0.01) (62). The neurological function improvement was attributed to inhibition of apoptosis in the implanted BM-MSCs induced by exercise and thus to subsequent better engraftment of these cells (44).

In summary, the literature reports provided evidence that combination of exercise and BM-MSC engraftment improves neurological tissue recovery compared with either exercise or BM-MSC transplantation alone. These effects may depend on exercise characteristics, specifically type (such as treadmill running, swimming), exercise conditions, but also injury characteristics (such as transection, crush, ischemia…). It is not clear from these studies, however, whether the effects of MSC and exercise are additive or synergistic.

Effects of Physical Exercise and MSC Transplantation on Cartilage Tissue Repair

When injured, the hyaline cartilage in a joint is replaced by a fibrocartilage with inferior mechanical properties. Cartilage repair of femoral osteochondral defects was compared following injection of either heterologous, passage 3-BM-MSCs, or phosphate-buffered saline solution into the femoro-tibial joint of 8-wk-old, male, Wistar rats, associated or not with moderate CTE (12 m/min, 30 min/day, 5 days/wk starting 48 h after injection, and 4 wk after osteochondral defect creation) for periods of 2, 4, or 8 wk (55). Two weeks postinjury, the results were similar for the rats that had been exposed to both BM-MSCs and exercise and for those that had received only BM-MSCs. Four weeks after injection, there was improvement (P < 0.05) in cartilage repair in the rat group that had received both BM-MSCs and exercise compared with the results obtained from the rats that had been exposed only to exercise, only to BM-MSCs, or to sham group. At 4 wk postinjury, the percentage area of type II collagen (differentiation biomarker) was higher (P < 0.05) in the rat group that had received both BM-MSCs and exercise compared with the animals exposed only to exercise and to the animals not exposed to either exercise or BM-MSCs. At 8 wk, although not significant, improvement in cartilage repair was, however, lower in the rat group that received both BM-MSCs and exercise, compared with animals exposed to either only exercise or only BM-MSCs (55). This result suggests that long exposition to moderate exercise may be detrimental to the tissue repaired by the combination of BM-MSC injection and exercise. This study did not determine whether the beneficial effect of the experimental conditions on cartilage repair was a time-dependent, additive, or synergistic effect (55). The effects of the time elapsed between osteochondral defect creation and exercise initiation was not evaluated.

Effects of Physical Exercise and MSC Transplantation on Cardiac Tissue Repair

Therapeutic strategies following myocardial infarction include, but are not limited to, MSC therapy and aerobic exercise to prevent adverse cardiac remodeling and ventricular dysfunction. However, little is known about the combined effects of MSC transplantation and exercise. The effects of moderate CTE (60% of the maximal running speed, 60 min/day, 5 days/wk, for 12 wk, starting 24 h after injury) associated with immediate BM-MSC therapy on recovery from myocardial infarction were evaluated in 40-day-old male Wistar rats. Although exposure to either BM-MSC therapy or exercise separately had beneficial effects on the infarcted heart remodeling and function, the combination of the two treatments had no synergistic effects (26). Explanations regarding these results are currently unknown, and further investigation is warranted.

DISCUSSION

Taken together, the results reported in the literature provide evidence that the effects of exercise on MSC functions depend on exercise characteristics, specifically type (treadmill running, climbing, etc.), intensity (moderate, intense, etc.), duration of exposure, etc. Specifically, moderate exercise increases BM-MSC proliferation and osteogenic differentiation (3, 33, 61). In contrast, either high- or low-intensity running exercises have either none or detrimental effects on the proliferation and osteogenic differentiation of BM-MSCs (3, 55). The type of exercise may also play a role in MSC proliferation and differentiation since VCE did not result in any changes in MSC proliferation but resulted in increased osteogenic and adipogenic differentiation cell capacities. Moreover, LMHFV, a low-intensity exercise that generates mechanical signals in bone tissue several orders of magnitude below strains generated during strenuous exercise, resulted in increased BM-MSC osteogenic differentiation (30, 52). It is difficult to determine whether the aforementioned differences are actually due to the various types of exercise tested or to intrinsic differences in exercise intensity (35).

Furthermore, literature reports provided evidence that the tissue characteristics (specifically, origin) from which MSC are collected affect MSC proliferation and differentiation after exposure to exercise (28). This outcome is not surprising since differences in the cell proliferation and differentiation capacities of various MSC niches in several species in the absence of exercise stimulation have been reported (1, 21, 23, 25, 27, 29, 41, 58).

In addition, the effects of exercise on MSC proliferation and differentiation may be affected by the age of the donors or the passage number of MSCs. Supporting evidence was provided by the results that the CFU-f numbers were higher (P < 0.05) when passage 3 BM-MSCs (from 13- to 14-wk-old male Wistar rats) exposed to moderate CTE were used compared with cells from sedentary controls but not when passage 6-BM-MSCs were used (28). This result highlights the importance of MSC passage number since the cell differentiation process is affected as this number increases.

An important limitation in most of the studies reported in this review is that the effects of exercise on MSC properties, specifically, proliferation and differentiation, were evaluated by in vitro tests, sometimes after a period of cell culture (Table 1). Differences between groups in MSC properties in vivo may have been attenuated by several weeks of cell culture in vitro in a similar culture medium. In contrast, some differences may have been enhanced by optimal culture conditions, in which MSC interactions with their environment were not replicated. Whether the data in vitro reflect events in vivo is therefore difficult to confirm. Direct evidence of the effects of exercise (through LMHFV application) on MSC adipogenic differentiation in vivo, however, was provided by tracking MSC and their differentiated states using green fluorescent protein in irradiated mice (45). The effects of LMHFV on MSC differentiation in vivo were also indirectly demonstrated by analyzing pertinent gene expression directly on “fresh bone marrow” (i.e., without any cell culture) (30).

Another limitation pertains to the various methods of MSC isolation and identification used in the reported studies (Tables 1 and 2). These methods usually included adherence to plastic dish and multilineage differentiation or automatic cell sorting based on specific surface markers. These surface markers, however, varied widely between studies and so did methods to assess differentiation potential. Consequently, the MSC populations evaluated were probably not equivalent, which likely contributed to some of the discrepancies between studies. Unfortunately, exclusive, specific surface markers or features that categorically identify MSC have not been reported, yet MSC surface markers and differential potential depend on, for instance, species from which they were isolated (38).

Regarding the effects of combinations of exercise and MSC transplantation on tissue repair, literature reports provided evidence of, at least, additive effects on neural tissue functional recovery and cartilage repair (55, 57, 62). Treadmill running may be more effective for neural functional recovery than swimming, although the small number of reported studies prevents from drawing definitive conclusions. In contrast, the beneficial effect of the combination of exercise and MSC transplantation on structural and functional cardiac remodeling in infarcted rats is not definitively resolved yet (26). Differences in methodological procedures (such as exercise protocols, time elapsed between injury exercise initiation or BM-MSC transplantation, MSC dose…), however, make comparisons between studies difficult. Moreover, young animals, i.e., with inherent better regeneration capacities, were used in most of the studies. Results must, therefore, be interpreted with caution.

Tissue repair improvement observed with a combination of MSC transplantation and physical exercise may be explained in different ways. First, moderate PE induces the systemic release of growth factors (such as vascular endothelial growth factor, platelet-derived growth factor…) and cytokines that may improve the environment for MSC survival at the site of injury. PE may also induce local inflammation, which in turn activates both transplanted and resident MSC to exert paracrine, neurotrophic, or immunomodulatory effects, improving tissue healing. In such a model, the extent of inflammation would parallel the extent of exercise-induced tissue damage. This hypothesis is, however, unlikely in the studies reported here, since MSC were implanted in an already acutely inflamed environment (24–48 h after injury induction) (26, 53, 55, 57, 62). Moreover, literature rather provides evidence that PE promotes an anti-inflammatory environment (4), and it was demonstrated, at least in muscles, that inflammation was not the primary regulator of MSC accumulation (51). Second, PE may directly stimulate transplanted MSC to differentiate into resident cell population (such as osteocytes, chondrocytes, tenocytes. . . ). Although a possibility, supporting evidence is lacking since MSC were not labeled and actual engraftment and differentiation not confirmed in all but one (45) studies reported in this review. Third, MSC themselves may directly sense mechanical signals [such as the osteocyte in bone (9)] and adapt their paracrine or immunomodulatory effects to alter tissue healing. Supporting evidence was provided in vitro by the results that MSC submitted to high-strain magnitude or low-intensity vibration loading regime incorporating refractory periods exhibited decreased adipogenesis compared with controls without refractory periods (48). Finally, PE may deter MSC from engaging undesirable differentiation into adipocytes or fibrotic cells in the tissue repair process. Supporting evidence was provided by the results that LMHFV deterred MSC away from adipogenesis (through downregulation of peroxisome proliferator-activated receptor-γ) toward osteogenesis (45). These hypotheses are not mutually exclusive and a combination of different mechanisms is likely.

Overall, the results of the literature are difficult to interpret and compare because of major differences in the experimental protocols used. For instance, the various exercise regimes performed were not fully described (regarding intensity, maximum oxygen uptake, exercise duration, etc.) and the species, sex and age of the experimental subjects were different. The population and the differentiation stage of MSC were not always defined, and cells of various passages were used. Nonetheless, there is strong evidence that MSC respond to mechanical strain (5) and thus probably modulate tissue repair.

CONCLUSIONS

Moderate, continuous, running exercise has beneficial effects on MSC proliferation and differentiation, specifically, improved osteogenic and tenogenic properties. The effects of a combination of exercise and MSC transplantation, however, remain inconclusive; cellular and molecular changes induced by physical activity are not completely elucidated but have great research interest (36).

Future studies should aim at comparing the effects of various types of exercise on MSC properties using standardized protocols and include detailed information regarding all aspects of the exercise regime applied.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

C.B. prepared figures; C.B. drafted manuscript; M.B., S.P., and H.P. edited and revised manuscript; C.B., M.B., S.P., and H.P. approved final version of manuscript.