Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Introduction

Friedenstein and coworkers were the first to identify a cell population in bone marrow, distinct from hematopoietic stem cells (HSCs), with the ability to generate in vivo a heterotopic osseous tissue capable of supporting hematopoiesis ( Chap. 1). This population, now referred to as mesenchymal stromal cells (MSCs), was subsequently exploited to establish long-term bone marrow cultures in vitro, which provided a unique opportunity to dissect the cell-type specific interactions and soluble factors that regulate various aspects of hematopoiesis. These studies revealed much about the phenotype and function of MSCs and as such revealed for the first time insight into their unique paracrine functions [1]. Therefore, we will begin by examining the role of MSCs in hematopoiesis, which continues to evolve.

Role of MSCs in Hematopoiesis

A number of reviews have been published describing in detail the important role played by MSCs in regulating hematopoiesis [2, 3]. Therefore, the topic is discussed only briefly here. MSCs secrete several classes of proteins including cytokines, chemokines, growth factors, neuropeptides, and extracellular matrix proteins that modulate hematopoiesis. For example, the matrix proteins fibronectin, laminin, vitronectin, thrombospondin, haemonectin, thrombopoietin, tenascin, and collagens function as organ and lineage-specific binding proteins for hematopoietic cells [4–12]. These proteins also directly bind cytokines and growth factors and present them in biologically active forms to hematopoietic cells, which then stimulate growth and maturation [13–15]. Many growth factors including stem cell factor, granulocyte-macrophage colony-stimulating factor (GM-CSF), granulocyte colony-stimulating factor (G-CSF), macrophage colony-stimulating factor, kit ligand, leukemia inhibitory factor, interleukin 1 beta (IL-1B), interleukin 3 (IL-3), interleukin 6 (IL-6), interleukin 7 (IL-7), interleukin 8 (IL-8), interleukin 11 (IL-11), insulin-like growth factor 1 (IGF-1), and transforming growth factor beta 1 (TGFβ1) are also secreted by MSCs [16, 17]. In addition, their expression may be altered in response to external stimuli, thereby providing a mechanism to modulate hematopoiesis in response to stress, infection, and injury. For example, treatment of MSCs with IL-1β, IL-6, and lipopolysaccharides (LPS) stimulates, whereas treatment with interferon α suppresses expression of GM-CSF and G-CSF [18]. Activin A, a potent stimulator of erythroid differentiation and negative regulator of B cell lymphopoiesis, is also strongly upregulated in MSCs by inflammatory cytokines and suppressed by glucocorticoids [19]. Moreover, platelet-derived growth factor (PDGF) enhances the ability of MSCs to support growth of colony forming unit-granulocytes macrophages (CFU-GM) by inducing secretion of GM-CSF, IL-3, and IL-6 [20]. Other factors secreted by MSCs that have been shown to play a role in hematopoiesis include Flt-3 ligand [21], hepatocyte growth factor (HGF) [22], jagged1 [23], substance P [24], and calcitonin gene-related protein [25]. Additionally, secreted frizzled-related protein-1 (sFRP1) inhibits osteoclast formation [26] as well as maintains homeostasis of HSCs in marrow via extrinsic regulation of beta-catenin [27].

More recent studies have identified MSC subpopulations in marrow, discriminated based on secretion of specific cytokines, that bind to functionally distinct T and B cell lineages [28]. For example, MSCs expressing CXCL12 (chemokine, CXC motif, ligand 12) have been shown to interact specifically with pre-pro-B cells and memory plasma cells, while MSCs that express IL-7 but lack expression of CXCL12 interact specifically with memory CD4+ T cells. These MSC subpopulations also express vascular cell adhesion molecule 1 but lack expression of endothelial cell adhesion molecule 1. These studies indicate that the bone marrow reticular system may be inordinately complex and contain distinct stromal subtypes that specifically interact with different hematopoietic lineages to sustain hematopoiesis [1]. Various groups have also reported that MSCs and or MSC-derived osteoprogenitors physically interact with HSCs in bone marrow and as such contribute to the HSC niche (see Chap. 3).

It is important to note that human CD34  +  HSCs as well as CFU-GM, CFU-megakaryocytes, and blast-forming unit erythrocytes also secrete various growth factors, cytokines, and chemokines that function via autocrine and paracrine mechanisms to regulate hematopoiesis [29, 30]. For example, human CD34+ HSCs stimulate secretion of G-CSF and IL-6 from MSCs [31]. Therefore, hematopoiesis is orchestrated by the concerted action of many secreted proteins, expression of which is controlled both by changes in the external environment and cross-talk between stromal and hematopoietic cell lineages within marrow.

Regeneration Versus Replacement: Shifting Paradigms of MSC Function

Early studies demonstrating the multi-potency of MSCs led to the assumption that the cells function in tissue homeostasis by serving as a reservoir of connective tissue progenitors and as such were first employed clinically to treat osteogenesis imperfecta [32]. The established role of MSCs in supporting hematopoiesis was also exploited in clinical trials to speed hematopoietic recovery following bone marrow transplantation [33]. Although results from these trials were encouraging, reports published in the early 2000s indicating that MSCs possessed unexpected plasticity as evidenced by their transdifferentiation into cardiomyocytes [34], astrocytes [35], hepatocytes [36], lung epithelium [37, 38], and other lineages sparked renewed interest in the therapeutic potential of the cells. While further research revealed that “transdifferentiation” occurred at a low frequency in vivo [39], MSCs continued to demonstrate measurable therapeutic effects in experimental animal models of disease. Consequently, this prompted a major paradigm shift with respect to the anticipated function of MSCs. Rather than contribute directly to tissue replacement via directed differentiation, the ability of MSCs to alter the tissue microenvironment via paracrine signaling rapidly established itself as the principle mechanism by which the cells affected tissue repair and regeneration. This new paradigm has rapidly gained acceptance following the identification of various proteins secreted by MSCs that have demonstrated angiogenic, anti-apoptotic, anti-inflammatory, and trophic effects in various disease models.

The MSC Transcriptome

The diverse array of paracrine-acting factors secreted by MSCs was predicted in large part by analysis of their transcriptome using genomics-based approaches. For example, our laboratory was the first to analyze the transcriptome of human [40, 41] and primary rodent MSCs [42] via serial analysis of gene expression (SAGE). Interrogation of these databases revealed expressed transcripts encoding proteins that regulate a variety of functions necessary to maintain homeostasis of bone and bone marrow including angiogenesis, hematopoiesis, cell migration and communication, neural activities, immunity, and defense. These findings were confirmed by other groups [43–45]. We also validated the biological activity in vitro and in vivo of a subset of secreted proteins identified by SAGE. For example, we demonstrated that primary mouse MSCs ameliorated bleomycin-induced lung injury in mice and that secretion by cells of interleukin receptor 1 antagonist (IL-1RN), a protein with potent anti-inflammatory activity, contributed significantly to this effect [46]. We also showed that neural regulatory factors secreted by MSCs promoted survival and neurite outgrowth from neuroblastoma cells and spinal nerves from the dorsal root [47]. Finally, we identified a large number of angiogenic factors secreted by MSCs that induce growth and branching morphogenesis of human vascular endothelial cells [48]. Results from genomics-based studies have been codified by proteomics-based analyses showing that MSCs secrete chemokines [49], chemoattractants [50], and angiogenins [51] that play roles in tissue injury and repair.

SAGE analysis also revealed that a single cell-derived colony of human MSCs simultaneously expressed a diverse array of lineage-specific mRNAs characteristic of skeletal and muscle tissue [40]. These data suggested that MSCs exist in a ground state with respect to mRNA expression, similar to that proposed for HSCs [52], and as such are poised for rapid differentiation in response to external stimuli. A similar ground state may exist for MSCs with respect to their paracrine function (Fig. 9.1). For example, MSCs constitutively secrete a number of mitogens including fibroblasts growth factor 2 (FGF-2), brain-derived neurotrophic factor (BDNF), IGF1, and HGF, angiogenins including vascular endothelial growth factor-A (VEGF-A), angiopoietin 1 (ANG1), and CRY61, as well as various cytokines and chemokines, such as GM-CSF, G-CSF, IL-6, CXCL12, and TGFβ1 that are important for bone homeostasis and hematopoiesis. However, a growing number of studies have shown that secreted levels of these proteins are altered and/or induced in MSCs following exposure to external stimuli, such as infection, inflammation, and changes in oxygen concentrations. Importantly, these conditions typify the microenvironments encountered by MSCs when transplanted ectopically to sites of tissue injury or disease. For example, MSCs express several toll-like receptors (TLRs) that allow cells to sense and respond to infectious agents [53]. TLR activation leads to secretion by MSCs of pro-inflammatory cytokines and chemokines [54], growth factors [55], and soluble mediators that regulate immune cell function [56]. MSCs also express receptors for tumor necrosis factor-alpha (TNF-α) and IL-1, which upon ligand binding induce expression of secondary mediators of inflammation, such as IL-6, and other proteins including HGF [57], monocyte chemotactic protein-1 (MCP1), cathepsin L, and several matrix metalloproteases [58]. Importantly, engagement of these receptors also leads to secretion of several potent anti-inflammatory proteins including IL-1RN [46] and TNF-α-induced protein 6 (TNF-αIP6) [59], which promote healing by limiting the extent of tissue inflammation. MSCs also induce mature dendritic cells to secrete interleukin 10 (IL-10), which has anti-inflammatory properties, and T helper 2 cells to secrete interleukin 4 (IL-4), which induces IL-10 secretion from macrophages [60]. Exposure of MSCs to interferon-gamma (IFN-γ) has also been shown to augment secretion of HGF, IL-10, TGFβ1, and indoleamine 2,3-dioxygenase, thereby enhancing their immunomodulatory effects [61] (Chap. 6).

Fig. 9.1
figure 00091

MSCs are poised for rapid lineage specification and activation of paracrine signaling in response to external stimuli. (a) Genomics-based studies have shown that a single cell-derived colony of human MSCs expresses mRNAs characteristic of skeletal and muscle tissue [40], ­indicating that the cells are poised for rapid lineage specification in response to specific external stimuli as illustrated by the red (adipogenic), blue (chondrogenic) and yellow (osteogenic) arrows. (b) MSCs have also been shown to constitutively secrete a variety of cytokines/chemokines, ­angiogenins, and mitogens that play important roles in tissue homeostasis. Moreover, secreted levels of these proteins can be significantly altered in response to a variety of external stimuli, such as infection, inflammation, and hypoxia. Importantly, these conditions typify the microenvironments encountered by MSCs when transplanted ectopically to sites of tissue injury or disease. Therefore, these data suggest that the paracrine functions of MSCs also exist in a ground state and are poised for rapid activation in response to tissue injury and disease. Protein expression in MSCs is constitutive (black arrows), induced by infectious and inflammatory agents (red arrows) or by hypoxic conditions (blue arrows)

Full size image

Similarly, hypoxia has been shown to elicit a pro-angiogenic program in human MSCs by stimulating expression of VEGF, IL-6, and IL-8 and suppressing expression of TNF-α, interleukin 12 (IL-12), and tissue inhibitor of metallo-protease 1 [62]. Others have confirmed and extended these studies by demonstrating that hypoxia also induces expression of FGF2, HGF, IGF1 [55] and IL-1β, TNFα, and IL-10 [63]. Consistent with these studies, MSCs derived from heme oxygenase-1 (HO1) knockout mice secreted lower levels of CXCL12, VEGF-A, and HGF and exhibited a lower angiogenic potential in vitro [64]. Therefore, HO1 also contributes to the paracrine response of MSCs following exposure to environmental stressors.

The ability of MSCs to respond to and modulate the inflammatory response has broad implications with respect to their use in clinical medicine. For example, although inflammation plays a central role in the elimination of infectious agents and reparation of tissues following injury, unremitting inflammation also is characteristic of many disease states, such as nonhealing wounds, interstitial lung disease, arthritis, psoriasis, inflammatory bowel disease, and others. Owing to their potent anti-inflammatory effects, MSCs are being evaluated in phase I/II clinical trials for the treatment of Crohn’s disease, osteoarthritis, muscle and skeletal trauma, and diabetes (www.clinicaltrials.gov). Moreover, inflammation also contributes prominently to other common maladies such as obesity and drug addiction and plays a role in allograft rejection. In the later case, MSCs may modulate regulatory T cell-dependent allograft acceptance by limiting the extent of tissue inflammation. Therefore, similar to their capacity to undergo multi-lineage differentiation, MSCs may also be primed to respond to a broad array of aberrant tissue microenvironments and restore homeostasis to these microenvironments via the secretion of paracrine-acting factors (Fig. 9.1). This capacity is exemplified by the broad therapeutic effect of MSCs demonstrated in the following disease models.

Paracrine Effects in Ischemic Diseases

Myocardial Infarction

From their onset, clinical studies have demonstrated the safety of intracoronary infusion of MSCs for the treatment of myocardial infarction and shown this yields a measurable improvement in overall left ventricular function [65] (Chap. 6). Factors implicated in contributing to the therapeutic effect of MSCs in myocardial infarction include IGF, HGF, VEGF, and FGF, and in some cases, MSCs have been genetically engineered to overexpress these factors to augment their therapeutic effect [66–69]. Importantly, the mechanism of MSC action in myocardial infarction is complex as the cells exhibit beneficial effects at various stages of disease progression. For example, MSCs exhibit anti-apoptotic activity and protect cardiomyocytes from hypoxia-induced death by downregulating expression of the pro-apoptotic protein Bax and augmenting expression of FGF, VEGF, and CXCL12 in heart tissue [70]. In addition, MSCs engineered to overexpress the murine thymoma viral oncogene homolog 1 (AKT1) were found to be superior to wild-type MSCs for cell therapy of acute myocardial infarction in a rat model [71]. Herein, genetic modification altered the repertoire of secreted paracrine factors in MSCs based on the finding that conditioned media from AKT1-modified but not wild-type MSCs exerted cardioprotective effects in vivo. A subsequent study identified ­soluble frizzled-related protein 2 as a key AKT1-regulated paracrine factor secreted by MSCs responsible for reparative effects and myocardial survival [72]. Similarly, overexpression in MSCs of GATA-binding protein 4 (GATA-4) has also been shown to augment secretion of IGF-1 and VEGF and enhance the cells’ cardioprotective effects in vivo [73].

MSCs also promote neovascularization in infracted myocardium, which is necessary to prevent cell death, promote tissue remodeling and improve overall cardiac function [74]. For example, autologous MSC administration in a rat model of myocardial ischemia significantly increased capillary density within the ischemic heart tissue [70]. Enhanced vasculogenesis appears to be a common outcome seen following MSC administration in other ischemic diseases. For example, in animal models of limb ischemia, local delivery of MSCs augments collateral perfusion. This effect is mediated, in part, via paracrine mechanisms since antibodies against VEGF and FGF-2 partially inhibit the capacity of MSC-conditioned media to promote proliferation of endothelial and smooth muscle cells [75]. Other factors secreted by MSCs, such as HGF and IGF-1, augment aortic endothelial cell growth and survival, a response not observed with fibroblast conditioned media [76]. MSCs modified to overexpress GATA-4 or preconditioned by exposure to hypoxic conditions in vitro also exhibit enhanced anti-apoptotic and angiogenic effects on endothelial cells [73, 77]. Hypoxic preconditioning enhances expression of VEGF, IL-6, MCP1 and CXCL12 in MSCs as well as other unidentified factor(s) that activate the phosphatidylinositol 3-kinase (PI3K)-AKT pathway in endothelial cells. The latter is consistent with the fact that PI3K signaling mediates angiogenesis in vascular endothelial cells [78].

The paracrine action of MSCs in myocardial infarction is exemplified by the observation that the cells exhibit cardioprotective effects when administered not only locally (transcardial and/or intraventricular) but also intravenously. In the latter case, most MSCs accumulate rapidly in lung tissue in the first few hours after administration and then are slowly released over a few days into the circulation in low numbers. Lee et al. [59] reported that MSCs trapped in emboli within lung tissue secrete high levels of TNF-αIP6, which antagonizes the function of TNF-α and ameliorates tissue inflammation that contributes to the pathogenesis of myocardial ischemia. These data are consistent with other studies showing that MSC administration attenuates increases in CD68-postiive inflammatory cells and expression levels of MCP-1in heart tissue in a rat model of acute myocarditis [79]. Nguyen et al. [80] further showed that intracoronary injection of MSC-conditioned media reduced cardiac troponin-T levels and improved cardiac output, stroke volume, and wall motion score index in a swine model of acute myocardial infarction.

Ventricular remodeling in response to ischemic injury is typically characterized by hypertrophy and apoptosis of cardiomyocytes and tissue fibrosis. Depending upon the size of the infarction, aberrant remodeling can lead to decreased cardiac output and increased susceptibility to a second heart attack. Paracrine factors produced by MSCs limit the extent of aberrant remodeling by supporting regeneration of cardiomyocytes [81]. However, the identity of factors responsible for this effect is indeterminate. MSCs may promote regeneration of myocardium by stimulating growth and survival of cardiac progenitor cells (CPC). For example, human growth hormone and IGF-1 are part of an autocrine loop that maintains muscle tissue integrity, but their expression declines rapidly with aging. Similarly, HGF is necessary for CPCs to migrate to areas of tissue damage and promote repair [82]. Therefore, IGF-1 and HGF secreted by MSC may promote myocardial regeneration by stimulating ­proliferation of CPCs resident in heart tissue.

Whether MSCs also prevent tissue fibrosis remains unclear based on the available data. We previously showed that MSCs prevent fibrosis in a mouse model of acute lung injury, but this effect was a consequence of their potent anti-inflammatory activity [46]. MSC administration has been shown to reduce fibrous tissue deposition in heart in rat models of ischemia/reperfusion injury [83] and dilated cardiac myopathy [84]. However, whether this outcome is secondary to the anti-apoptotic and angiogenic activity of the cells remains uncertain. It is anticipated that HGF secreted from MSCs has anti-fibrotic activity based on its ability to suppress expression of TGFβ-1 [85, 86]. Moreover, intramyocardial injection of IGF-1/HGF affinity bound alginate biomaterial has been shown to reduce fibrosis, attenuate infract expansion, and increase vessel formation at the site of infarct in a rat model of acute myocardial infarction [87]. Therefore, MSCs may limit the extent of tissue fibrosis via secretion of factors that antagonize TGFβ-1 activity. However, it is important to note that MSCs express an array of collagens, matrix proteins, and metalloproteases and under certain conditions can adopt a myofibrocyte phenotype [88]. Therefore, understanding how expression of these proteins is altered when MSCs encounter an ischemic or fibrotic milieu is necessary to better clarify their anti-fibrotic potential. Contradictory reports exist regarding the anti-fibrotic effects of MSCs in models of liver injury [89, 90], which may be related to differences in the timing of cell administration and extent of liver damage. MSCs have been reported to prevent renal fibrosis although this outcome may also be secondary to the anti-inflammatory effects of the cells in this model [91, 92].

In summary, evidence supporting a paracrine hypothesis of MSC action in myocardial infarction includes poor engraftment and retention of MSCs in heart tissue following transplantation, minimal capacity to transdifferentiate into cardiomyocytes in vivo, ectopic overexpression of genes that augment expression levels of secreted proteins enhances the therapeutic efficacy of MSCs in vivo, MSC-conditioned media has cardioprotective effects in vitro and in vivo, and neutralizing antibodies against secreted proteins diminishes the therapeutic effect of MSCs. Nevertheless, it is likely that the list of cardioprotective factors secreted by MSCs is incomplete. Other cardioprotective factors produced by MSCs may include insulin-like growth factor binding protein 7 (IGFBP7) [93], which stimulates prostacyclin production in cultured bovine endothelial cells [94]. Prostacyclin is both antithrombotic and a vasodilator. Therefore, IGFBP7 may play a beneficial role in myocardial infarction and peripheral vascular disease by inhibiting thrombosis and vasoconstriction. Mining of genomic databases is likely to facilitate discovery of additional paracrine-acting factors secreted by MSCs that contribute to their cardioprotective effects in vivo.

Stroke

Similar to effects in myocardial infarction, MSCs have also been shown to positively impact various stages of disease progression in stroke. Although studies defining the role of individual proteins are lacking, the prevailing data indicate that paracrine factors secreted by MSCs reduce ischemic damage [95] and apoptosis [96, 97], induce neurogenesis [98], angiogenesis, synaptogenesis [99], neurite outgrowth [100, 101], enhance neuroplasticity [102], and restore cognitive functions [103]. Proteins secreted by MSCs that have therapeutic effects in myocardial ischemia are also implicated in providing a therapeutic benefit in stroke. For example, the ability of MSCs to promote neuronal cell survival and ameliorate neurological deficits in a rat model of middle cerebral artery occlusion (MCAO) was partially attenuated when cells were transduced prior to injection with a VEGF-RNAi lentivirus [104]. Other studies have reported that IGF1 expression is upregulated in MSCs engrafted within the infarct border in the brains of rats subjected to MCAO, and endogenous levels of VEGF, epidermal growth factor (EGF), and FGF-2 are also upregulated in brain [105]. Lastly, exposure of human MSCs to extracts from ischemic brain tissue augments expression of BDNF, nerve growth factor (NGF), VEGF, and HGF [106]. Therefore, the ischemic brain microenvironment is capable of altering the paracrine activity of MSCs, similar to that seen by exposing cells to hypoxic conditions in vitro. Additionally, trophic factors produced by MSCs engrafted within the ischemic brain are anticipated to modulate the production and expression levels of autocrine and paracrine factors produced by the brain parenchyma [97, 107–110]. This feed forward affect may account for the potent therapeutic effects of MSCs in vivo.

Genetic modification of MSCs has also been used to enhance their therapeutic effects in experimental stroke models. For example, MCAO rats administered human MSCs engineered to overexpress ANG1 and VEGF showed enhanced structural and functional recovery as compared to untreated rats and those administered wild-type MSCs [111]. Placental growth factor gene-modified MSCs also elicited greater angiogenesis and a larger reduction in lesion volume compared to native MSCs [112]. Moreover, BDNF and glial cell line-derived neurotrophic factor (GDNF) but not ciliary neurotrophic factor (CNTF) or neurotrophin 3 (NT-3) gene-modified MSCs are reported to exhibit enhanced capacities to promote functional recovery and reduce infarct size in MCAO rats [95, 113, 114]. In addition to trophic factors, MSCs also secrete extracellular matrix proteins that support the growth of neurons, astrocytes, and oligodendrocytes in vitro by increasing their metabolic rate and protecting cells from nutrient and growth factor deprivation [115]. However, whether these proteins are secreted from cells engrafted within brain tissue has yet to be examined.

MSCs may also improve functional recovery after stroke by modulating cytokine expression in the brain. For example, MSC administration increases brain expression levels of IL10, which has anti-inflammatory and neuro-protective activity [97, 116] and decreases levels of the pro-inflammatory cytokine TNFα [116]. MSCs also increase bone morphogenetic protein 2/4 expression in ischemic astrocytes, which enhances subventricular progenitor cell gliogenesis by activating relevant signaling pathways [107]. The cells also increase tissue plasminogen activator activity and downregulate plasminogen activator inhibitor 1 activity within the ischemic boundary of MCAO mice and in astrocytes cultured in vitro. These changes resulted in enhanced neurite outgrowth form cortical neurons [101]. Importantly, behavioral recovery and neurogenesis in a rat stroke model was more pronounced when animals were administered early versus late passage MSCs. Moreover, endogenous levels of trophic factors, such as GDNF, NGF, VEGF, and HGF, were higher in early passage MSC-treated brains [117]. These findings illustrate that culture expansion and/or methods of isolation may significantly impact the paracrine functions of MSCs.

Paracrine Effects in Lung Disease

Pulmonary Fibrosis

Our lab was the first to show that MSC administration ameliorated inflammation and fibrosis in a mouse model of bleomycin-induced lung injury [37]. A critical result from these studies was the demonstration that MSC administration at the time of bleomycin exposure but not 1 week later significantly reduced the extent of neutrophil infiltration into lung tissue and upregulation of pro-inflammatory cytokine expression. Based on this result, it was apparent that the principal effect of MSCs was anti-inflammatory in nature and this was subsequently linked to secretion of high levels of IL-1RN [46]. Other studies have shown that MSC administration suppresses bleomycin-induced increases in TGFβ1, IGF-1, and PDGF in lung tissue and laminin and hyaluronan expression in bronco-alveolar lavage (BAL) in rats [118]. Similarly, MSCs derived from umbilical cord blood attenuate expression of TGFβ1, IL-10, IFN-γ, and macrophage migration inhibitory factor in mice exposed to bleomycin [119]. In related studies, coculture of polarized human alveolar epithelial type II cells after inflammatory insult with MSCs was shown to preserve their protein permeability. ANG1 secretion was responsible for this beneficial effect in part by preventing actin stress fiber formation and claudin 18 disorganization via suppression of nuclear factor kappa-B activity [120].

In lung as in heart, it remains unclear whether MSCs actually exhibit anti-fibrotic effects or if their capacity to block fibrosis is merely a consequence of their potent anti-inflammatory properties. Salazar et al. [121] recently showed that MSCs from mouse bone marrow and human umbilical cord blood secrete high levels of PDGF-AA and TGFβ1 and stimulated growth and collagen production of lung fibroblasts. Interestingly, antagonism of TGFβ1 reduced collagen expression in lung fibroblasts, but their growth was inhibited by the Wnt antagonist sFRP1. These data suggest that MSCs also secrete Wnt ligands that stimulate fibroblast proliferation. Lee et al. [88] also recently reported that treatment of MSCs with connective tissue growth factor upregulates collagen type 1 and tenacin-C expression as well as collagen type III, fibronectin, and matrix metallo-protease type I. Further exposure to TGFβ1 induced the cells to adopt a myofibroblast phenotype and undergo fibrogenesis instead of ectopic mineralization in vivo. These studies suggest that MSCs may be pro-fibrotic under specific conditions.

Endotoxin-Induced Lung Injury

MSCs from mouse bone marrow can also significantly reduce lung inflammation, edema and decrease levels of IFN-γ, IL-1B, IL-6, macrophage inflammatory protein 1-alpha, and IL-8 in peripheral blood in a mouse model of endotoxin-induced acute lung injury [122]. Human MSCs produce a similar effect that is paracrine in nature based on their capacity to downregulate expression of TNF-α by macrophages stimulated with LPS in vitro [123]. In related studies, Lee et al. [124] showed that human MSCs also restored alveolar epithelial fluid transport and lung fluid balance in an ex vivo perfused human lung preparation injured by E. coli endotoxin. Importantly, conditioned media from MSCs yielded a similar outcome and knockdown studies demonstrated that keratinocyte growth factor (KGF) secreted from MSCs contributed to their therapeutic effect in this model.

Asthma

In a ragweed-induced mouse asthma model, Nemeth et al. [125] demonstrated that MSCs administered i.v. at the time of antigen challenge significantly reduced the extent of eosinophil infiltration and mucus production in the lung; decreased expressed levels of IL-4, IL-5, and interleukin 13 (IL-13) in BAL; and also lowered serum levels of IgG1 and IgE. In this study, allogeneic and autologous MSCs exhibited similar therapeutic effects, while skin fibroblasts reduced the total number of cells in BAL but not the number of eosinophils compared to untreated mice. Skin fibroblasts also significantly reduced circulating levels of IL-13 but not IL-4. Treatment of mice with MSCs from TGFβ1 knockout mice failed to have a therapeutic effect, consistent with in vitro studies showing that IL4 within BAL from ragweed-challenged mice induced TGFB1 expression in MSCs. Similar results were obtained by Bonefield et al. [126] in an ovalbumin model of asthma. In the latter study, MSCs also reduced systemic levels of IL-1β, suppressed inducible nitric oxide synthase expression from lung infiltrating monocytes, and enhanced IFN-γ levels in BAL fluid.

Paracrine Effects in Wound Healing and Diabetes

Wound Healing

MSCs constitutively express a variety of growth factors important for wound healing including PDGF, EGF, TGFβ1, VEGF, KGF, FGF2, and HGF [127]. Moreover, expression of PDGF, EGF, KGF, and HGF is significantly elevated when MSCs are exposed to LPS or IL-1B. Similarly, treatment of MSCs with superfusates from wounded abdominal tissue upregulated expression of TGFB1, EGF, and VEGF. Human MSCs and adipose-derived cells also secrete appreciable levels of IGF1, VEGF, and HGF, the latter two of which were upregulated by exposure to TNF-α [128]. Therefore, exposure to the wound microenvironment appears to induce in MSCs expression of various angiogenins and mitogens that normally participate in the wound healing process. MSCs have also been reported to secrete significantly higher levels of ANG1, KGF, IGF-1, PDGF, and erythropoietin as compared to dermal fibroblasts [129, 130], which may explain their superior healing capacity [130]. They also secrete high levels of cysteine-rich angiogenic inducer 61 (CYR61), and depletion of this protein from MSC-conditioned media abolished their angiogenic activity [51]. This is consistent with studies showing that MSC-conditioned media also stimulates wound healing [129]. Therefore, paracrine signaling via release of angiogenins (VEGF, EPO, CYR61), growth factors (EGF, FGF2, IGF-1, KGF, PDGF, TGFβ1), and other soluble factors promote angiogenesis, keratinocyte proliferation, and migration and may also modulate the activity of inflammatory cells. MSCs also secrete a large array of extracellular matrix molecules including collagens, fibronectin, and various matrix metalloproteinases and as such may directly contribute to tissue repair by functioning akin to fibroblasts.

Diabetes

MSCs have been shown to normalize blood glucose levels when administered to streptozotocin-induced hyperglycemic mice [131, 132]. In these studies, MSCs were shown to increase the number of pancreatic islets and beta cells producing mouse insulin and also prevent renal damage. Consistent with these results, other studies have shown that islets from MSC-treated animals expressed high levels of pancreas/duodenum homeobox protein 1 and insulin and that peripheral T cells from these animals exhibited a shift toward IL10/IL13 production [133]. Coculture of human pancreatic islets with MSCs also improves their adenosine-5′-triphosphate/adenosine-5′-diphosphate ratio and insulin secretory function in vitro. MSC-conditioned media was shown to contain high levels of IL-6, VEGF-A, HGF, and TGFβ1, factors known to improve the survival, function, and angiogenesis/revascularization of islets [134]. Consistent with this result, Xu et al. [135] reported that exposure of MSCs to rat pancreatic extracts significantly upregulated secretion of IGF-1, VEGF, and FGF2 and conditioned media from pancreatic extract-treated MSCs was able to lower blood glucose levels when administered to diabetic rats. Finally, MSCs also reportedly restore normoglycemia in diabetic rats by promoting vascularization and enhancing survival of islet grafts via secretion of VEGF [136].

Paracrine Effects in Neurodegenerative Diseases

The anti-inflammatory and immunomodulatory affects of MSCs may be advantageous in the treatment of various neurodegenerative diseases since inflammation is thought to contribute significantly to their pathogenesis. For example, elevated levels of pro-inflammatory cytokines in brain tissue are detected in mouse models of lysosomal storage diseases, and the degree of inflammation has been shown to coincide with the onset of clinical symptoms in these models [137–139]. In most cases, microglia activation occurs in response to aberrant neural cell function or as part of a wider stress response in the brain and typically precedes neuronal cell loss. Inflammation is also a prominent feature in other neurodegenerative disorders such as Alzheimer’s and Parkinson’s disease [140].

With respect to animal models of storage disease, injection of unmodified MSCs into the cerebellum markedly reduces the extent of microglial and astrocyte activation and reduces levels of macrophage colony-stimulating factor, a microglial activator, in a mouse model of Niemann-Pick type C disease [141]. Similar results were also obtained with adipose tissue-derived stem cells [142]. In the latter studies, cell transplantation directly to the cerebellum resulted in rescue of Purkinje neurons as evidenced by their enhanced electrical activity and suppression of neuro-inflammation based on decreased glial cell activation and decreased expression levels of IL-1B, IL-6, and TNF-α protein in the cerebellum.

MSCs administration also delays disease progression, improves motor performance, and decreases microglial activation and astrogliosis in the spinal cords of mice carrying a glycine 93 to alanine (G93A) mutation in the superoxide dismutase 1 gene (SOD1), a model of amyotrophic lateral sclerosis (ALS) [143, 144]. The therapeutic effect of MSCs in this model is also likely paracrine in nature based on studies showing that exposure of MSCs to extracts from the brains or spinal cords of SOD1 mutant transgenic rats significantly upregulated expression of VEGF-A, HGF, and NGF and suppressed expression of FGF2 and IGF1. Moreover, spinal cord but not brain extracts induced expression of BDNF and GDNF [145]. Other studies have shown that exposure to G5 supplement, FGF2, and CNTF induces mRNA and protein expression of the glutamate transporter 1 (GLT-1) in MSCs, which results in an enhanced ability of the cells to uptake aspartate. This result suggests that MSCs may be neuro-protective by restoring glutamate homeostasis in response to disease [146]. Importantly, MSCs from SOD1(G93A) mutant rats showed reduced aspartate uptake despite expressing higher level of GLT1 mRNA and protein. These results indicated that MSCs from mutant rats expressed a nonfunctional GLT1 receptor and therefore were unable to protect neurons from ­glutamate toxicity. This finding questions the suitability of autologous stem cell grafts for treatment of familial forms of ALS. Similarly, Cho et al. [147] reported that secreted levels of several trophic factors including FGF2, HGF, IGF-1, CXCL12, and VEGF-A are decreased in MSCs derived from the bone marrow of human ALS patients. Nevertheless, direct administration of MSCs from human ALS patients into the cistern magna of mice engineered to overexpress the human mutant SOD1(G93A) gene resulted in a dose-dependent increase in life span and survival of motor neurons in the ventral horn of the spinal cord [148]. Therefore, patient-specific MSCs may still provide some degree of therapeutic benefit when administered in vivo, despite their reduced paracrine activity.

Direct administration of MSCs into the cerebellum of newborn Lurcher mutant mice, a model characterized by the selective early postnatal death of Purkinje cells, resulted in significant improvement in motor function as evidenced by improvement in the rotarod test. At 2-month posttransplant, histological analysis demonstrated a significant increase in Purkinje cell numbers in treated versus control mice and revealed that many of the surviving MSCs in brain were juxtaposed to the Purkinje cell layer in the cerebellum. This outcome was attributed to secretion by MSCs of BDNF, NT-3, and GDNF, neurotrophins important for Purkinje cell survival [149]. MSCs also decreased the extent of glial activation, oxidative stress, and apoptosis within the hippocampus and improved memory function and learning in a mouse model of acute Aβ-induced Alzheimer’s disease [150].

Collectively, these studies indicate that paracrine mechanisms also contribute significantly to the therapeutic effect of MSCs in a variety of neurodegenerative diseases. This is consistent with genomics-based studies showing that MSCs secrete various neurotrophins as well as other factors that promote neural cell survival and neurite outgrowth under stressful conditions and following injury in vivo [47, 151, 152]. This capacity of MSCs is likely related to the fact that bone and marrow are innervated by nervous tissue, providing a means by which sympathetic efferent input can modulate hematopoiesis. Despite the benefits afforded by MSCs in the aforementioned disease models, it should be noted that one study has shown that MSC-conditioned medium promotes glial cell activation and upregulates expression of TNF-α and IL-6 in organotypic cultures of the hippocampus, leading to neuronal cell death [153]. These results should be carefully assessed as they suggest that MSC-based therapies may yield adverse or unforeseen outcomes that may exacerbate the disease state. The latter necessitates incorporating dose response studies into translational and clinical trials and metrics to measure potential adverse side effects.

Conclusions

Only a decade ago, MSCs were thought to function specifically as a reservoir of progenitor cells to maintain homeostasis and facilitate repair of connective tissues following injury. While exploring their potential use in tissue engineering remains an avid area of research, the field in general has witnessed a major paradigm shift with respect to the function and potential clinical applications of MSCs. Specifically, it is now believed that the principal mechanism by which MSCs affect tissue repair is by paracrine signaling. This paradigm shift has gained support from genomics and proteomics-based studies, which revealed that MSCs secrete an array of ­proteins that exhibit angiogenic, anti-inflammatory, immunomodulatory, and trophic activity, and studies demonstrating that MSCs or MSC-conditioned media ameliorate disease and promote tissue repair in a broad array of experimental animal models of disease and human clinical trials. Moreover, evidence is mounting that MSCs express a large repertoire of surface receptors that coordinately regulate paracrine function in response to changing environmental conditions, such as hypoxia, inflammation, and infection. Therefore, it appears the pendulum has swung full circle. In the 1970s, cytokine secretion by MSCs was exploited to establish long-term bone marrow cultures, which led to the identification of the HSC. Most MSC-based therapies currently being pursued today also exploit the paracrine activity of MSCs. Despite these advances, aspects of MSC biology remain indeterminate including their origin during development, the molecular mechanisms that regulate self-renewal and lineage specification, and how paracrine functions are specified within populations. Efforts aimed at addressing these questions will likely lead to the procurement of even more potent cell-based vectors for clinical therapies. Therefore, MSC research likely will remain an exciting and rewarding enterprise for some time to come.