Mesenchymal stem/stromal cell-based therapy: mechanism, systemic safety and biodistribution for precision clinical applications

MSC-based mechanisms of action

Several possible mechanisms by which MSCs exert their beneficial effects have been proposed. Early studies reported that MSCs could migrate to sites of injury and then differentiate into functional cells [40], or that they could fuse with compromised cells to regenerate damaged tissues [41, 42]. More recent studies have demonstrated that paracrine factors [43, 44], mitochondrial transfer [45], and extracellular vesicle secretion [46] have important roles in mediating the effects of MSCs.

Paracrine effects

MSCs secrete paracrine factors, including cytokines, chemokines, growth factors, and miRNAs. MSC transplantation or administration of isolated secreted factors enables MSC paracrine factors to get to injured tissues, to help restore a healthy microenvironment to promote tissue repair [47] (Table 2). MSC paracrine factors play important roles in immunomodulation [48, 49], tissue regeneration and healing [50, 51], anti-fibrosis [52, 53], anti-apoptosis [54], and angiogenesis [55]. As such, many studies have focused on altering culture conditions in order to steer the secretome of MSCs towards therapeutic agents. Alterations have included using MSCs from different types of tissue [56, 57], oxygen concentration [58], growth factor incubation or cytokine pretreatment [59], passage number [60–62], three-dimensional spheroid culturing [63], and mechanical strain [64].

The capability of MSCs for immunomodulation has made them a useful treatment approach for inflammatory disorders such as multiple sclerosis [65], Crohn’s disease [66], GVHD [67], systemic lupus erythematosus [67], and type I diabetes [68]. Immunomodulation is dependent on crosstalk between MSCs and the immune microenvironment of the target tissue. In an inflammatory microenvironment, proinflammatory cytokines, including IL-1β, IL-6, IL-23, IFN-γ, and TNF-α, can stimulate MSCs to secrete anti-inflammatory factors such as TNFα stimulated gene (TSG)-6 [69], nitric oxide (NO) [70], IL-10 [71], galectins [72], prostaglandin E2 (PGE2) [73], and transforming growth factor (TGF)-β [3, 71]. Upon exposure to these MSC-secreted anti-inflammatory signals, nuclear factor (NF)-κB activity and consequent inflammatory cytokine expression in macrophages, dendritic cells, and T cells are inhibited, and immune cells will express higher levels of anti-inflammatory cytokine IL-10 as a result [3, 74]. The MSC paracrine factors also interact with other immune cells and have been reported to skew macrophage polarization towards the M2 phenotype, which downregulates both innate and adaptive immune responses [75]. Regulatory T cells (Treg) were also reported to stimulate MSCs to secrete indoleamine 2,3-dioxygenase (IDO), thereby augmenting the Treg response and attenuating acute liver injury [3, 76].

In addition to their immunomodulation ability, MSCs are able to secrete factors that can promote cell proliferation, increase angiogenesis, and reduce cell apoptosis. For example, MSCs can secrete growth and angiogenesis-promoting factors such as basic fibroblast growth factor (bFGF) [77], insulin-like growth factor (IGF) [78], TGF-β [3, 55], stromal cell-derived factor (SDF)-1α [79], secreted frizzled-related protein-1/2 (SFRP1/2) [80, 81], angiopoietins, and vascular endothelial growth factor (VEGF) [82, 83].

It has been demonstrated that MSCs can inhibit fibrosis via paracrine factors [84]. Chronic inflammation is a major factor that drives the fibrosis process, which can alter the normal architectural structure of tissues and lead to deteriorated functioning. Because MSCs can be used to reduce inflammation, they have become an attractive therapeutic strategy for suppressing fibrosis. MSC-derived conditioned medium (CM) was shown to attenuate liver fibrosis by reducing Th17 cells in a IDO-dependent manner [85]. MSC-secreted interleukin 1 receptor antagonist (IL-1Ra) was also shown to inhibit stellate cell activation and decrease type I collagen expression, a key component of liver fibrosis [86]. Administration of MSC-CM also reduced fibrotic score and collagen deposition in both bleomycin- and silica-induced lung injury models [87, 88]. In MSC-treated cells, levels of HGF, KGF, and BMP-7 increased while levels of TGF-β1 and TNF-α decreased. These results suggest that the anti-fibrotic effect of MSCs may be mediated via paracrine mechanisms [88]. In support of this, a bleomycin-induced lung injury model showed that the stanniocalcin-1 (STC-1) secreted by MSCs in response to TGF-β1 exerted antifibrotic effects by reducing oxidative stress, endoplasmic reticulum (ER) stress, and TGF-β1 production in alveolar epithelial cells [89]. Likewise, MSCs were able to decrease the expression of fibrosis-associated tissue inhibitor of matrix metalloproteinase 1 (TIMP)-1, to improve cardiac function in a myocardial infarction model [90].

Mitochondrial transfer

Mitochondrial dysfunction is a hallmark of the aging process, and has been implicated in the pathogenesis of numerous diseases [91]. MSC-based mitochondrial transfer has therefore been a promising therapeutic strategy, by either replenishing or replacing the damaged mitochondria in targeted diseased cells [92]. Studies have observed increased tunneling nanotube (TNT) and gap junction formation with mitochondrial transfer between MSCs and injured epithelial/endothelial cells under inflammatory or hypoxic conditions, and MSC-derived mitochondria transfers could prevent apoptosis of recipient cells [93–95]. In addition, it was found that iPSC-derived MSCs could attenuate alveolar damage and fibrosis via mitochondrial transfer by TNT [96]. The tissue origin of MSCs may affect mitochondrial transfer ability. For example, iPSC-derived MSCs were shown to be more effective at mitochondria transfer compared with MSCs derived from bone marrow [96]. Mechanistically, mitochondrial transfer was found to alleviate epithelial injury through mitochondrial Rho-GTPase Miro1 regulation in an asthma model [97].

Despite these beneficial findings of MSC-mediated mitochondrial transfer, there are also potential risks, as mitochondrial transfer can increase the risk of tumor promotion. In acute myeloid leukemia (AML), NOX2 stimulated mitochondrial transfer from BMMSCs to cancer cells, and this promoted the survival of the cancer cells [98]. Mitochondrial transfer also increased the resistance of leukemic cells to chemotherapeutic agents, and transfer occurred bidirectionally [99, 100]. In an in vitro co-culture of BMMSCs and T cell acute lymphoblastic leukemia (T-ALL) cells, upon induction of oxidative stress by the addition of chemotherapeutic agents, T-ALL cells transferred their mitochondria to BMMSCs, but received few mitochondria from the BMMSCs, raising the chemoresistance of the T-ALL cells [99]. Neutralizing the cell adhesion molecule ICAM-1 and disrupting intercellular mitochondrial transfer restored the sensitivity of the T-ALL cells to the chemotherapeutic agent [99].

Extracellular vesicle (EV) transfer

MSC-derived extracellular vesicles (EVs) have raised increasing interest as a non-cellular alternative to MSC-based therapy, as this approach eliminates concerns of unintended lineage differentiation [101]. EVs refer to exosomes, microvesicles, and apoptotic bodies, and are membrane-enclosed entities secreted by a cell in response to stimulation or apoptosis. The size and contents of these vesicles are highly variable and heterogeneous, involving proteins, mRNAs, and miRNAs [101]. Their role in MSC-mediated cellular therapy remains elusive due to their heterogeneous nature, but it is currently believed that they play an important role in many biological processes and intercellular communication [101].

Exosomes from MSCs have shown beneficial effects in disease models of autoimmune uveitis [102], retinal detachment [103], myocardial infarction [104], type 1 diabetes [105], wound healing [106], bone repair [107], burn injury [46], traumatic brain injury [108], spinal cord injury [109], and several other conditions [110]. The most commonly suggested mechanism responsible for the effects of exosomes is via their capability to regulate immune cells and immune microenvironments. MSC-derived exosomes can suppress the expression of pro-inflammatory cytokines TNF-α, IL-1β, IL-6, IL-17, IFN-γ, and MIP-1α in immune cells [103, 105, 109, 111]. Additionally, MSC-derived exosomes significantly increased the levels of anti-inflammatory cytokines IL-4, IL-10, and TGF-β in a type 1 diabetes animal model [105]. In a drug-induced liver injury model, MSC-derived exosomes enhanced the local expression of cytokines TGF-β and HGF, both of which are key factors in liver regeneration [112]. The underlying mechanism involved changes in the immune cell population, including increased M2 polarization [106, 108, 109], increased Th2 and regulatory T cell differentiation [105, 112], decreased Th17 differentiation [111], and decreased local immune cell infiltration [102].

In addition to promoting immunomodulation, MSC-derived exosomes participate in other biological processes. MSC-derived exosomes were found to promote neoangiogenesis in diabetic and burn wounds via increased VEGF-A expression, the Wnt4/β-catenin pathway, and increased tube formation and proliferation of endothelial cells [106, 113]. MSC-derived exosomes also activate Akt, ERK, and STAT3 pathways and induce expression of HGF, IGF1, NGF, SDF1, and TGF-β, which critically regulate wound healing and tissue repair [114]. In addition, MSC-derived exosomes can aid in tissue repair by enhancing autophagy and inhibiting apoptosis [103].

In contrast to microvesicles and exosomes from MSCs, apoptotic bodies are entities specifically generated by cells during apoptosis. Apoptotic bodies containing ubiquitin ligase RNF146 and miR328-3p were shown to help maintain MSC multipotency via the Wnt/β-catenin pathway [115]. In support, it was recently shown that apoptotic bodies released from donor MSCs improved myocardial infarction via autophagy regulation in recipient cells [116].

The lack of consistent or standardized methods to isolate and identify EVs presents a challenge for current therapeutics. A recent study has shown that compared to EVs, MSC-CM resulted in more effective immunomodulation [117]. Further studies are necessary to decipher the optimal MSC culture conditions and the specific subpopulations of secreted components that contribute to the most effective therapeutic benefit.

Clinical applications of MSC-derived EVs have gained increasing interest, as many of the safety concerns of MSC-based therapy might be avoided, including undesired differentiation of implanted cells in tumor formation/promotion risks, and the cell-derived secondary ischemic damage by vessel clotting. As MSC-derived EVs are still in their clinical infancy, there is currently little information on clinical safety. To monitor biodistribution, most of the in vivo studies utilize lipophilic dyes to label the EVs [118, 119]. While the injected MSC-derived EVs migrated and accumulated at the injured tissue, they also aggregated in the lung, liver, and spleen [118, 119].

MSC safety consideration: Tumor initiation, promotion, and suppression

MSC-related cell therapy is a promising therapeutic strategy because of the high immune modulation ability and the absence of tumor initiation risk of MSCs. However, there is still concern that MSCs can pose a risk for promoting tumor cell growth [120, 121]. MSCs share some characteristics with fibroblast cells, which are able to transform into cancer-associated fibroblasts (CAFs) in tumor niches. The tumor niche involves local fibroblasts, endothelial cells, immune cells, and cancer associated MSCs. Increasing evidence shows that the tumor niche is not only trophic to cancer cells, but also highly associated with tumor initiation and growth, and is able to increase cancer stemness-related properties, including the capacity for cell migration, invasion, and chemotherapy resistance. Therefore, cancer treatment strategies have expanded from solely targeting the tumor cells, to altering the tumor milieu.

Since MSCs have an excellent ability for homing to tumor sites, the possibility for therapeutic MSCs to transform into cancer-associated MSCs exists. Several studies have examined the effect of MSCs on different types of tumor cells. Not surprisingly, conclusions among these studies are unclear (Fig. 2). Studies using MSCs from different tissue origins, different cultivation processes, and different cancers can lead to diverse results and interpretations.

On the other hand, taking advantage of the ability of MSCs to home to tumor sites enables MSCs to serve as therapeutic carriers that deliver anticancer agents to appropriate sites [122]. As highly progressive and late stage malignancies constitute a major health burden, for which current treatments are unsatisfactory and curative therapies are unavailable, MSC-related drug carriers may provide new hope for cancer treatments, particularly for late stage cancers.

MSC Promotion effects on tumor cell growth and metastasis

The underlying mechanisms responsible for MSC tumor promotion are complicated and diverse (Table 3). They are classified below according to MSC type and signaling pathway, and are listed systematically in Table 4 and summarized in Fig. 3.

Cell type

Signaling pathways

MSC suppression effects on tumor growth

While MSCs utilize diverse mechanisms for tumor promotion, they suppress tumor growth mainly by inducing apoptosis of tumor cells. MSCs have been shown to suppress the growth of breast [167–169], brain [148, 170–174], lung [170, 175], liver [175, 176], ovarian [167, 177, 178], bone [167, 179], esophageal [168], bladder [180], colorectal [170] and hematological malignancies [181–183]. The underlying mechanisms responsible for MSC tumor suppression are classified below as well in Table 5, and are summarized in Fig. 4.

Apoptosis, autophagy and senescence

The majority of MSC tumor suppressing mechanisms involve increasing tumor cell apoptosis and impeding cell cycle progression. Upregulation of caspase-3, caspase-9, p16, p21, p53, TRAIL, pro-apoptotic BAX, ATG5, ATG7, BECLIN1 and cellular H₂O₂ levels [148, 167–170, 173, 180, 182]; activation of Smac/DIABLO [173]; and downregulation of survivin, XIAP, cyclin D1, Cdk4, Cdk6, cyclin A2, cyclin E1, AKT/pAKT, Bcl-2, β-catenin, c-Myc, pro-caspase-7, PCNA, Bcl-xL and MMPs have been demonstrated to be involved in the MSC-dependent tumor cell apoptosis seen with MSC-CM, MSC cell lysate (CL) and with direct cell–cell interaction [167, 168, 170–176, 178, 180].

Boosting immunity

The immunomodulation ability of MSCs is also correlated with tumor suppression. Lin et al. demonstrated that 3 kDa MWCO-WJMSC-CM concentrate can induce immunogenic cell death in lymphoma cells, which showed decreased viability and increased apoptosis, as well as increased levels of the ER stress markers eLF2a and XBP-1. Increased levels of surface damage-associated molecular pattern markers ecto-CRT, ecto-Hsp70 and ecto-Hsp90, as well as extracellular ATP and high mobility group box 1 were also observed. When cocultured with WJMSC-CM-treated lymphoma cells, dendritic cells had enhanced CD80 and CD86 expression. Yet lymphoma cells treated with WJMSC-CM concentrate had decreased CD47 and PD-L1 expression [181].

Anti-angiogenesis

In addition to directly inhibiting tumor cell growth, coculturing with hBMMSCs or hUCBMSCs decreased angiogenesis in glioblastoma. The underlying mechanism may involve the down-regulation of PDGF-BB and IL1β secretion or decreases in FAK, VEGF or Akt [184, 185]. Another attractive source of MSCs, human endometrial mesenchymal stem cells (EnSCs), also show an ability to decrease tumor growth and to increase angiogenesis in ovarian cancer by inhibiting AKT phosphorylation and decreasing expression of VEGFA and HIF-1α, possibly via nuclear translocation of FoxO3a [186].

Anti-migration and invasion

hUCBMSCs are also reported to decrease glioblastoma cell invasion and migration by increasing PTEN or Mad1 expression and downregulating PI3K/AKT, c-Myc/ERK or EGFR/c-Met activities [187, 188]. Inhibition of Wnt signaling has been shown to decrease tumor growth and migration after treatment with hUCMSC- or hADMSC- CM in bile duct cancer and breast cancer, respectively [189, 190].

Other mechanisms

Human BMMSC-secreted oncostatin M (OSM) has been reported to inhibit tumorigenicity and EMT by activating the OSM/STAT1 signaling pathway in lung adenocarcinoma cells [191]. Decreased cancer cell proliferation was also correlated with suppressed NF-κB expression and activity in HCCs and BCCs by MSCs derived from fetal bone marrow or fetal dermal tissue [192]. Vascular wall-resident MSCs as well as hBMMSCs displayed a capacity for decreasing the risk of lung metastasis after radiation-induced injury in breast cancer and melanoma by downregulating endothelial MMP2 and SASP factors CCL2 and Plau/uPA, which were induced by radiation injury [193]. In addition to suppressing tumor progression, hUCMSCs promote granulocytic differentiation of immature myeloid cancer cells in acute promyelocytic leukemia (APL), which drives the disease into remission by activating MEK/ERK pathways [194].

Engineered MSCs

Another promising strategy to treat progressive malignancy is the use of engineered MSCs, which show a remarkable ability to suppress tumor progression [195]. UCBMSCs with exogenous overexpression of CXCR1 and CXCR4 displayed enhanced tropism towards gliomas [196]. In addition, irradiation of glioma cells enhanced IL-8 expression, which promoted the tropism of hUCBMSCs equipped with TRAIL migration to tumors, and further induced tumor cell apoptosis [197]. hBMMSCs overexpressing TRAIL can also induce apoptosis in CD133-positive primary glioma cells in vitro [198]. Modified interleukin-12 (IL-12p40N220Q; IL-12 M), which enhances expression of the IL-12p70 heterodimer that is necessary for induction of Th1 and CTL immunity, was overexpressed in hUCBMSCs and found to significantly decrease tumor growth and angiogenesis, as well as to increase the survival of glioma-bearing mice and to confer tumor-specific long-term T-cell immunity [199].

In human glioma studies, IL-24-hUCMSCs promoted tumor cell apoptosis, and IFN-beta-hBMMSCs were shown to prolong animal survival [200, 201]. Meanwhile, IFN-beta-WJMSCs and IFN-beta-hBMMSCs exhibited the ability to suppress tumor growth in bronchioloalveolar carcinomas [202] and HCCs, respectively, the latter exerting its effect by increasing expression of p21, p27 and FOXO3a, as well as decreasing protein levels of cyclin D1, pRb and AKT [203]. In addition, engineered BMP4-secreting hADMSCs could suppress tumor cell migratory ability and increase survival in glioblastoma [204]. As for hematological cancers, treatment with hUCMSCs equipped with Tandab (a tetravalent bispecific tandem diabody with two binding sites for CD3 and two for CD19) combined with IDO pathway inhibitor showed significantly decreased B cell lymphoma growth by way of decreasing CD98 and Jumonji, and by restoring the proliferation of T cells [205]. Another study demonstrated that UC-MSCs overexpressing IDO can inhibit proliferation of leukemia cells [206]. hWJMSCs engineered with scFvCD20-sTRAIL fusion protein, which targets CD20-positive cells and induces apoptosis through sTRAIL, inhibited proliferation in B cell lymphoma [207]. Another study showed that hWJMSCs transfected with vector coding sTRAIL driven by AFP promoter had significant antitumor activity in HCC [208]. Decreased tumor growth was also observed in gastric cancer and in epithelial ovarian cancer using hUCBMSCs delivering TNFSF14 or IL-21, respectively [209, 210]. In a syngeneic pancreatic tumor mouse model, IL15-hUCBMSCs inhibited tumor growth and increased survival of tumor-bearing mice. The IL15-hUCBMSCs induced NK- and T-cell accumulation at the tumor site and established tumor-specific T-cell memory immunity [211]. Cytosine deaminase-expressing hADMSCs serving as a prodrug converting vehicle, showed significant decreases in colorectal cancer growth in the presence of prodrug 5-fluorocytosine [212].

Summary of promotion and suppression effects of MSCs in cancer

MSCs can contribute to tumor promotion as well as to tumor suppression. Although it may appear that these effects occur randomly, closer examination provides a more promising picture. Summarizing a total of 110 reports, (excluding engineered MSCs) reveals that in 58.6% of the studies, BMMSCs promoted tumor growth, while 9.8% of studies found that BMMSCs suppressed growth. Although the tendency of ADMSCs is not as obvious as that of BMMSCs, they also exhibit a preference for tumor promotion (Fig. 2). In general, MSCs derived from reproduction-related sources, including placenta, umbilical cord, Wharton’s jelly, and umbilical cord blood, show a higher likelihood for tumor suppression (Fig. 2). In regards to tumor type, we found that BMMSCs show an overwhelming promoting effect on cancers of the bone (100%, 6/6), breast (100%, 7/7) and GI tract, (liver, bile duct, colorectal, gastric and esophageal; 93.75%, 15/16) (Fig. 2a).

MSCs demonstrate an impressive suppressive ability in hematological cancers. In all 7 studies, MSCs from different tissue types showed tumor suppression. Similarly, in a total of 8 studies of MSCs and ovarian cancer, only one study reported that MSCs promoted tumor growth (Fig. 2b). To date, there is no report showing a tumor promoting effect for MSCs from placental tissue.

MSCs can exert their effects directly by contacting tumor cells, or indirectly by secreting soluble factors and microRNAs that the affect the tumor cells. The mechanisms by which different types of MSCs promote or suppress the growth of different tumor types are complicated (Tables 2, 3, 5, Fig. 4). Factors that may affect the properties of MSCs and cause different outcomes, include (1) the origin of the MSCs; (2) different processes of isolation, purification, and expansion of MSCs; and (3) different culture conditions and passages of the MSCs. Most of the results described herein were derived from direct or indirect in vitro co-culture systems or from in vivo co-injection experiments, but the underlying mechanisms were not always examined. It will be necessary to elucidate these underlying mechanisms, as well as to find potential biomarkers of MSC-tumor interactions for future clinical applications of MSCs.

Biodistribution of therapeutic cells in a preclinical evaluation

In light of the tremendous potential of MSCs for treating various diseases, it is necessary to define the systemic distribution and to quantify the administered cells in order to facilitate our understanding of the safety and efficacy of MSC-based cell therapy. This information is critical in clinical trials since it is vitally important to know whether the transplanted cell products home to the target diseased sites to deliver their intended effects. Indeed, several factors can affect the pharmacokinetics (PK) of the administered MSCs, including cell size, cell source, immunological features and labeling, detection methods, route of administration, and size of the animal model.

Factors that affect the biodistribution of MSCs

The typical diameter of a MSC is between 15–30 μm; in comparison, lymphocytes have a diameter of only 4–12 μm [213]. Furthermore, MSCs become larger after serial ex vivo cell passaging [214]. The relatively large size of MSCs explains their initial mechanical entrapment at lung capillary systems after intravenous administration, a phenomenon referred to as the pulmonary first-pass effect [26, 215]. Redistribution to liver, spleen, and other inflamed tissues subsequently takes place in the following hours to days, with gradual clearance from the lungs [26]. In some studies, MSCs were still detected in the lungs up to 150 days after transplantation in vivo [216]. MSCs retained at the lungs potentially decrease the number of cells available for therapeutic effects [217]. To decrease the mechanical entrapment of MSCs at the lungs, several strategies may be implemented, including pretreatment with the vasodilator sodium nitroprusside in order to increase the effective diameter of the pulmonary capillary system; delivery via an extravascular route; or delivery via multiple smaller doses [215, 217, 218]. Although administering MSCs intra-arterially may decrease the extent of mechanical entrapment at the lungs [219], the effect of cell size still has important implications, as larger MSCs may be associated with vascular occlusions that could cause subsequent ischemia and infarcts of unintended tissues and organs [220, 221]. Engineering of MSCs might potentially alter this adverse effect. For example, by overexpressing integrin α4 (ITGA4), which mediates leukocyte trafficking of MSCs, Cui et al. observed that cell aggregation of MSCs were significantly decreased, and MSC-associated cerebral embolism was ameliorated in rat model of stroke [222]. Furthermore, the risk of embolism has been found to be positively associated with cell dose of infusion and low infusion velocity [223].

In addition, aging of either donor or recipient could affect the biodistribution of inoculated MSCs, with decreased transplantation efficiency observed with aged donor MSCs and recipients [224]. Furthermore, when MSCs were extracted from older donors, they exhibited lower proliferative and differentiation capabilities [225, 226]. The culture condition also plays a role in the kinetics of administered MSCs. For example, hypoxic preconditioning increased MSC migration to injured tissue via enhanced HGF/cMET signaling and MSC recruitment, thus affecting biodistribution of the administered cells [227].

Immunogenic reactions also affect clearance and biodistribution of injected cells, as the allogeneic MSCs are not completely immune-privileged [228]. When MSCs are transplanted in an allogeneic host, the transplanted MSCs have decreased survival compared with their survival in a syngeneic host [229]. Formation of antibodies against injected MSCs could explain the reduced effectiveness and increased adverse effects that were observed with repeated inoculations in some studies [230].

Furthermore, the injected cells can also trigger an instant blood-mediated inflammatory reaction (IBMIR), which causes reduced graft survival and thromboembolism [231]. A portion of injected MSCs do not reach their intended destination due to the host’s immune reaction, embolization, and micro-ischemia [232]. Previous literature has demonstrated that the extent of IBMIR is related to the level of tissue factor (TF) expressed by MSCs; expression levels vary among different tissue origins of MSCs, and with culture conditions [233]. Compared with ADMSCs and UCMSCs, BMMSCs express lower levels of TF [233]. Thus, selecting TF-deficient BMMSCs may reduce the risk of IBMIR and improve the chances for clinical success. Otherwise, co-treatment with an anticoagulant may be an important consideration for clinical applications [234].

Methods of tracking MSCs in vivo

A critical step in generating pharmacokinetic models of cell products is tracking the fate of cells following transplantation. An ideal quantification technique should have the following features: high sensitivity and specificity; long-term detection and monitoring; and spatiotemporal resolution. The advantages and disadvantages of currently available methods for quantitative MSC detection are summarized in Table 6. Polymerase chain reaction (PCR) has been used to track human MSCs in murine xenogeneic models by detecting human DNA [19, 235–237]. The low limit of detection of quantitative PCR enables detection of 100 MSCs per gram of organ tissue, making it feasible to detect MSCs in patient biopsies. Both flow cytometry and optical imaging require labeling MSCs with fluorescent dyes or proteins. Flow cytometry enables estimation of the number of live MSCs per weight unit of tissue, and optical imaging uses a variety of dyes, such as 4′,6-diamidino-2-phenlindole (DAPI), that can bind reversibly or irreversibly to the MSCs [238–241]. The use of red fluorescent protein (RFP) or green fluorescent protein (GFP) expressing MSCs has the advantage of providing viability information of transplanted cells [242]. However, the transfection efficiency is not consistent, and the transfected cells could have altered potency and expression and cannot be accurately tracked over time [243]. Therefore, the biodistribution and quantitative data produced by fluorescent protein labeling methods may be incomplete. Bioluminescence imaging (BLI) which utilizes luciferase reactions also has the advantage of providing viability information of transplanted cells, but this method suffers from poor tissue penetration and low spatial resolution. MSCs can also be labeled with gold nanoparticle and tracked by computed tomography (CT) image in vivo [244, 245]. These gold nanoparticles have advantage of exerting negligible influence on viability, proliferation, and differentiation ability of labeled MSCs, and offer good spatial resolution and long-term tracking when used in conjunction with CT modality [244]. However, sensitivity is relatively poor, and there is still difficulty deriving quantitative information from CT images [246].

Magnetic resonance imaging (MRI) can be used to track MSCs in vivo by labeling MSCs with superparamagnetic iron oxide nanoparticles (SPIONs) or fluorine-19 (¹⁹F). Direct labeling of MSCs with SPIONs is possible as these agents are readily taken up by MSCs and show up as hypointense signals on MRI [247]. However, some studies have shown that proliferative and differentiation capabilities of MSCs could be affected when labeled at higher concentrations [247]. The downside of SPION labeling is that the specificity of SPION-labeled cells could be low and the signals could be hard to differentiate from acutely injured tissues containing hemorrhages. In contrast, ¹⁹F-labeling offers better specificity as endogenous fluorine level is low, minimizing background interference and is a better labeling agent when the targeting lesion involves hemorrhage [248]. In general, MRI offers good spatial resolution but suffers from poor temporal resolution. Positron emission tomography (PET), single-photon emission computed tomography (SPECT) [26, 249–251] and radioisotope labeling [26, 252, 253] have been used to image and track the migration dynamics, and inter-patient variability of MSCs in clinical patients, but quantifying cell numbers with these methods is difficult and only semi-quantitative information on the biodistribution of the transplanted cells can be obtained. Photoacoustic imaging, which combines ultrasonography with optical imaging, is another attractive approach, as ultrasonography has the unique advantage of providing real-time information while still maintaining good spatial resolution. By using gold nanorods coated with reactive oxygen species (ROS) sensitive dye as probe, Dhada et al. were able to also detect viability of implanted cells [254]. However, photoacoustic imaging suffers from operator dependent variability [255]. More recently, multimodal imaging probes that combine the advantage of different imaging modalities have been developed, including PET/MRI imaging agent [256], SPECT/MRI/fluorescent imaging agent [257], and SPECT/MRI/BLI imaging agent [258].

An ideal cell tracking method should be biocompatible and nontoxic, require no genetic modification, have single-cell detection sensitivity, and permit quantification of cell numbers at any anatomic location. Optical imaging utilizing nanoparticles as exogenous contrast agents is suitable for this purpose, although the technique is mainly used for animal models in preclinical experimentation due to the limited penetration depth of visible photons into tissue. Among various exogenous contrast agents, fluorescent nanodiamond (FND) has emerged as an attractive option because it is chemically inert and inherently biocompatible [259, 260]. A viable application of FNDs for background-free imaging and quantitative tracking of MSCs in animal models beyond rodents has been demonstrated using magnetic modulation [261–263]. The magnetic modulation fluorescence (MMF) method uses magnets to modulate the fluorescence intensity of FNDs. This technique, which allows background-free imaging, together with the inertness of FNDs and the large quantity of the nanoparticles taken up by the cells, has permitted studies of the biodistribution and pharmacokinetics of FND-labeled MSCs in preclinical settings. This strategy can also be applied to the characterization of cell-based products in order to accelerate their progression towards commercialization to meet the needs of patients. The technique has excellent compatibility with time-gated fluorescence imaging, which has been shown to be a powerful means of acquiring high-contrast fluorescence imaging of FND-labeled cells in tissues. The ability to find single cells is particularly valuable for ex vivo histological detection of MSCs in clinical trials. This combined approach represents an appealing alternative to hazardous radioisotope labeling techniques in cell tracking applications. The technique can be used with immune cells, stem cells, and other cell types used for cell therapy. Here, we put these technologies together, and describe how they could be used to contribute to the development of pharmacokinetic modeling of MSC-based cell products.

An FND-based platform to track therapeutic cells in vivo

The ability to monitor the behavior of transplanted cells in vivo is required for cell therapy. When cellular products are submitted for investigational new drug (IND) status, pharmaceutical studies must provide evidence of not only the safety of the cell product, but also information regarding cell location, cell migration, PK and pharmacodynamics (PD), and cell biodistribution after transplantation in animal models. There are three critical issues that must be addressed for cell therapy: (1) whether therapeutic cells maintain their potency after transplantation, (2) the appropriate dosage for curing diseases and (3) a route of administration and a formulation that permits successful drug delivery. Over the past decade, the traditional concepts, confined to low molecular weight organic compounds and large biomolecules, have been challenged with the advent of new drugs based upon cells, which we refer to here as cell therapy. As for all drugs, understanding the pharmacology of cell-therapy products is critical for their effective application in the clinical setting. For example, tissue section and PCR does not provide sufficient information of cell behavior in vivo, because these procedures select a sample from a population, making it difficult to provide PK and PD information for the whole animal. In contrast, the FND-labelled tracking technique provides a new method to achieve high throughput whole organ treatment and analysis, providing accurate pharmacology information, such as PK, PD and biodistribution of the cellular therapy (Fig. 5a). This method not only provides immediate and highly specific cell localization data after gathering histological sections from the animal, but also provides a one-step, one-tube analysis for any kind of animal tissue. Compared to the qPCR sampling method, this protocol can provide more accurate data for whole organ/tissue analysis and takes less time for validation and analysis.

We use a healthy mouse model to demonstrate that the FND-labelled platform can provide evidence of cell biodistribution. Figure 5b shows the biodistribution analysis of FND-labelled placenta choriodecidual membrane-derived MSCs (pcMSCs) for one week in a mouse model using the FND-based labelling platform. Our results show that the majority (up to 70%) of FND-labelled pcMSCs localized to the lungs after intravenous administration, which is consistent with the pulmonary first-pass effect [217, 264]. The trapping of MSCs in the lungs is due to space restriction [265], as pcMSCs are more than ~ 20 μm in diameter and much larger than the width of the micro-capillaries of the lung. After intravenous infusion, FND-labelled pcMSCs disappeared from the lungs as time passed, and migrated to other tissues/organs such as the liver and spleen, or to injured sites. Nevertheless, the number of FND-labelled pcMSCs decreased in the heart and kidneys (Fig. 5b).

As it has been reported that MSCs will migrate to injured sites [266], we induced an ischemia–reperfusion injury to the left kidney in our animal model (Fig. 6a) and examined whether FND-labelled pcMSCs injected into the portal vein would appear in the injured kidney, to test the concept that MSCs will migrate to sites of injury. In our mouse model with healthy kidneys, the number of pcMSCs in the kidneys decreased over time (Fig. 6b, upper panel) and the decrease was evident in both the left and right kidneys. (Fig. 6b, lower panel). In contrast, in the mouse model with the injury the number of FND-labelled pcMSCs in the injured kidney was highest on day 5 (3%; Fig. 6c). As seen in the lower panel of Fig. 6c, the injured kidney (L kidney) had significantly more FND-labelled pcMSCs than the healthy kidney (R kidney). The percent of FND-labelled pcMSCs remained consistent over time (~ 0.25%) in the healthy right kidney (R kidney) (*P < 0.5, **P < 0.01, ***P < 0.001, ****P < 0.0001.) (Fig. 6c, lower panel). Given these data, it appears that the percentage of MSCs that migrate to kidneys is limited to about 4%, and it appears that the kidneys have the ability to redistribute MSCs in vivo. In addition to providing fast and accurate results, this technique is completely safe to the cell tissue. The FND-labelling technique does not alter any properties of the cell, including cell viability, proliferation, differentiation and immunomodulation, making this method very biocompatible.

Clinical applications of MSCs in cell therapy: safety and potency

The potential and promise of MSC therapy is highly anticipated in recent and coming decades. As with all emerging new medical technologies, patient safety is always the first priority. As we have discussed, although the ability to modulate immune environment and promote tissue regeneration have been well reported in preclinical studies, the aspect regarding tumor induction or promotion is still one of the many concerns. The MSCs derived from different tissue origins or expanded under different culture conditions present different immune profiles which may result in tumor promotion [126]. Additionally, as the double sided blades of the MSCs’ strong immune modulation ability [262], evaluation of both the specific MSC properties as well as the patient’s immune conditions is strongly needed. The patient’s immune condition both before, during, and after treatment should be closely monitored.

Some reports showed that artificial engineering process may decrease the tumor induction and increase tumor-suppressing function of MSCs [263]. However, genetically engineered MSCs also raise other safety concerns. Although several clinical trials claimed the safety of MSC-treated patients, however, most of the trials only showed short-term safety and are without the examination of tumor-associated biomarkers [267, 268].

A recent systematic review and meta-analysis reappraised 55 randomized controlled trials and over 2000 patients to investigate the safety of systemically inoculated MSCs [39]. The risk of fever was significantly greater in the group of patients receiving MSCs. There was no significant increase in the risk of infection, thrombo-embolic events, malignancy or ectopic tissue formation, while the risk of death was significantly lower in the MSC-treated patients. Among the included studies, severe adverse events, including treatment related fever, in-stent thrombosis with death, acute coronary artery occlusions after intra-coronary delivery, grade 1 anaphlyactoid reaction, gastric ulcer perforation, hypersensitivity reaction, and anal cancer, have been reported to be possibly related to MSC treatment. Although the conclusion of the meta-analysis ends on a promising note, it was also emphasized that an a priori plan to monitor safety should be outlined in every clinical study design, including immediate allergic reactions, local complications (hematoma formation, local infection), vascular obstructions (dyspnea, oliguria, myocardial infarction, venous thromboembolic events), systemic complications (systemic infection, abnormal liver or renal function), malignancy or ectopic manifestation of implanted MSCs, and other disease-specific safety considerations [39].

Additionally, patients with medical history of ischemic diseases, cardiovascular diseases, lung fibrosis, concurrent neoplasm, and family history of hereditary cancer should be carefully reviewed during MSC treatment. The cell dose, infusion route and rate should be documented. The product profiles of the MSCs from different tissues and different generation processes, such as transcriptome, epigenome, proteomic data, cell populations, potential potency biomarkers, preclinical data from cell and animal studies, should be provided.

The therapeutic efficacy of MSCs in different disease indication is still under evaluation, as most of the studies to date have been limited to phase 1 and phase 2 studies (Fig. 1b and 1c, Additional file 2). As we have discussed in this review, the differences in MSC tissue origins and the variety of cell culture conditions would be some of the important factors determining MSC potency in vivo [269]. Thus, the development of surrogate potency assays using preclinical animal model is needed [270]. Recently the International Society for Cellular Therapy (ISCT) have announced some strategies to identify the potential effective factors of MSC action mechanism, including the combined the matrix assay and multiple techniques, such as quantitative RNA analysis for the specific genes, flow cytometry analysis for cell surface markers, and the protein-based assay of secretome [271]. Potency assessments in evaluating cell pharmacology, cell delivery route, as well as the cell-drug interaction are still under development to improve the MSC precision therapy [272–275]. Although the matrix assays were reported to serve as a platform to identify the biomarkers for MSC potency in vitro [276, 277], whether this in vitro assays are able to identify the MSC potency are still under discussion. For example, the use of allogeneic human peripheral blood mononuclear cells for mixed lymphocyte reaction (MLR) assays is a popular assay to demonstrate the MSC immunomodulation capacity. However, the lack of robustness, accuracy, and reproducibility is of concern [278–280]. Additionally, the correlation between the in vitro assays and in vivo pre-clinical/clinical data requires further evaluation.

Cryopreservation could be another factor affecting MSC potency. It has been documented that the MSC cryostorge, the so-called “cryo stun effect”, may decrease MSC therapeutic efficacy, leading to failures in MSC clinical trials [278]. Recently, a systematic review regarding the impact of cryopreservation on BMMSCs showed that the cryopreservation appears to affect the cell viability, apoptosis, cellular attachment, immunomodulation, and metabolism of BMMSCs [279]. Furthermore, these impaired viability or functions of the MSCs can be restored, partially or totally, by following an acclimation period [279–281], or by IFNγ licensing before cryopreservation [282].

In summary, the use of standardized potency assays should be incorporated into future MSC product release criteria. Thus, development of surrogate potency assays for different disease indications should be highlighted. The optimal process of cryopreservation and thawing may be another important factor requiring further attention.

Ethics approval and consent to participate

The kidney ischemia/reperfusion animal model and its study was approved by the Institutional Animal Care and Use Committee of National Taiwan University (with IACUC Approval No. 20180123) and conducted with compliance of the standards established in the Guide for the Care and Use of Laboratory Animals.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.