The potential of mesenchymal stem‐cell secretome for regeneration of intervertebral disc: A review article

Low back pain is a crucial public health problem that is commonly associated with intervertebral disc de‐ generation and has vast socio‐economic impact worldwide. Current treatments for disc degeneration are conservative, non‐surgical, or surgical interventions, and there is no current clinical therapy aimed at directly reversing the degeneration. Given the limited capacity of intervertebral disc (IVD) cells to self‐repair, treatment aiming to regenerate IVDs is a topic of interest and mesenchymal stem cells (MSCs) have been identified as having potential in this regeneration. Recent studies have revealed that the benefits of MSC therapy could result from the molecules the cells secrete and that play principal roles in regulating essential biologic processes, rather than from the implanted cells themselves. Therefore, the objective of this study is to review the potential use of the MSC secretome to regenerate IVDs. Current evidence shows that the secretome may regenerate IVDs by modulating the gene expressions of nucleus pulposus cells (upregulation of keratin 19 and downregulation of matrix metalloproteinase 12 and matrix Gla protein) and stimulating IVD progenitor cells to repair the degenerated disc.


Introduction
Low back pain (LBP) is a crucial public health problem (Driscoll et al. 2014). It has an enormous socioeconomic impact worldwide and reduces the patient's quality of life remarkably (Katz 2006; Husky et al. 2018. About 577 million people are affected by LBP, and its cost exceeds 100 billion dollars per year in the United States alone (Katz 2006; Wu et al. 2020. Globally, LBP is the leading cause of disability (Wu et al. 2020).
Low back pain is commonly associated with interver tebral disc degeneration (IDD) (Clouet et al. 2019). Inter vertebral disc degeneration may be present in more than 90% of people, but many of them have no signs of the dis ease (Cheung et al. 2009). Degeneration of the disc often begins in the second decade of life, earlier than other con nective tissues in the human body, and is viewed as one of the inevitable consequences of aging (Siemionow et al. 2011; Kepler et al. 2013. As degeneration occurs, this disturbed the disc's ability to perform its mechanical func tions (Roberts et al. 2006). At the cellular level, degen eration of the disc is characterized by increased degrada tive enzyme production, increased apoptosis, increased in flammatory cytokines expression, decreased extracellular matrix production, and neurovascular ingrowth (Kepler et al. 2013).
Intervertebral disc (IVD) has a limited ability to re pair itself following injury and degeneration (Vadalà et al. 2016). Current treatments for disc degeneration are con servative treatments, nonsurgical, or surgical interven tions aimed for muscular stabilization and symptomatic re lief with no clinical therapy targeting to reverse the degen erated disc itself (Wei et al. 2014). However, the clinical results of surgical interventions such as spinal fusion and total disc replacement remain suboptimal (Van Den Eeren beemt et al. 2010).
Considering the limited capability of the IVD to re pair itself, clinical therapy aimed at the regeneration of in tervertebral disc has become an appealing research topic. Disc degeneration results in changes in the biochemical microenvironment in IVD that challenge the successful ap plication of some potential biological therapies, such as decreased nutrient and oxygen supplies, declined pH, and increased cellular apoptosis (Kepler et al. 2013; Loibl et al. 2019. One of those potential treatments is using mes enchymal stem cells (MSCs) for IVD regeneration. Some studies conducted in animal models have shown its poten tial regenerative effect in IVD degeneration (Bach et al. 2014; Freeman et al. 2016; Vadalà et al. 2016). How ever, recent studies have revealed that MSCs therapy ben efit could be due to their vast secreted molecules, which play a principal role in regulating many essential biologic processes, rather than from the implanted cells themselves (Vizoso et al. 2017). Therefore, this study aims to review the potential use of MSCs secretome to regenerate the in tervertebral disc.

Pathophysiology of IDD
The intervertebral disc is a cartilaginous structure located between the vertebral body of the spinal column (Urban and Roberts 2003). They provide flexibility for motion and act as a shock absorber. The IVD is composed of three parts; the nucleus pulposus (NP) located centrally, the annulus fibrosus (AF) that surrounded the NP, and the cartilaginous end plates (EP) that separates IVD from the adjacent vertebral bodies (Kepler et al. 2013).
The NP is an avascular and immuneprivileged struc ture that consisted of three different cell types: NP stem/progenitor cells (NPPCs), notochordal cells (NCs), and chondrocytelike cells (NP) (Hunter et al. 2003; Erwin 2010; Erwin and Hood 2014. The NP's extracellular ma trix (ECM) is mainly composed of type II collagen and has a considerably higher proteoglycan concentration than AF (Kepler et al. 2013). Matrix component of the NP consists of proteoglycan aggrecan and type II collagen in a ratio of approximately 27:1 (Mwale et al. 2004). Nucleus pulpo sus contains a large amount of proteoglycan, up to its 50% dry weight (Walker and Anderson 2004). Since proteogly cans are negatively charged and highly polar, they attract water into the ECM, thus maintain the highwater content in the IVD (Kepler et al. 2013).
The AF consists of concentrically arranged type I col lagen fibers that serve as a border containing the inner NP (Kepler et al. 2013). The AF can be divided into the outer annulus and inner annulus (Walker and Anderson 2004). The collagen fibers in the outer annulus are not oriented uniformly instead, they are aligned at approximately 30°to the longitudinal axis of the spine and alternate their direc tion with each lamella (Walker and Anderson 2004). This characteristic gives optimal tensile strength to maintain the NP in place during spine movement (Freemont 2009). Ini tially, there is a transition zone between the NP and the inner annulus, but this distinction disappears as the disc degeneration begins (Roberts et al. 2006).
The cartilaginous end plate separates NP from the ver tebral bone and gives resilience to prevent the load trans mitted through the IVD from fracturing the vertebral bone (Freemont 2009). The EP is an avascular organ, but there are capillary networks near the central portion of the EP that are directly connected with the vertebral body vascu lature (Erwin and Hood 2014).
The IVD is an avascular organ because they lose their blood supply in the first decade of life (Rodriguez et al. 2011). The IVD cells have adapted to function in this con dition by relying on diffusion and convection for nutrient and metabolite exchange (Mokhbi Soukane et al. 2007). Nutrients and metabolites are transported to and from the IVD by diffusion from the blood vessels at the outer NP peripherally and the cartilaginous EP centrally (Holm et al. 1981). Hence, NP cells congregate near the cartilaginous EP where specialized capillary layers between the bony and cartilaginous EP provide nutritional supply (Urban et al. 1978).
In the second decade of life, the blood supply to the cartilaginous EP is diminished, consequently, their dif fusional capacity is decreased (Boos et al. 2002). This condition leads to a change in the microenvironment of the IVD which becomes acidic because of lactic acid's buildup (Kepler et al. 2013). Exposure to this acidic condi tion has a profound effect on the cellmatrix turnover be cause it decreases the IVD cells ability to produce ECM (i.e., sulfated glycosaminoglycan and tissue inhibitors of metalloproteinases1). However, it does not inhibit degradative enzyme production, such as matrix metallo proteinases (MMPs) (Razaq et al. 2003). These microen vironment changes will lead to ECM breakdown and there fore the IVD degeneration. Other conditions that could compromise the cartilaginous EP's vascular supply are vasoconstriction (e.g., from nicotine or vibration expo sure) (Deyo andBass 1989; Wilder andPope 1996), vaso occlusive process (e.g., atherosclerosis and arterial steno sis) (Kauppila 2009), and end plate sclerosis (Roberts et al. 1996).
Usually, innervation of the IVD is limited to the outer AF, but during the degeneration process, the nociceptive nerve endings grow deeper into the disc and play a role in pain transmission from the IVD (Bogduk et al. 1988; Freemont et al. 2002. This neuronal ingrowth is induced by nerve growth factor (NGF) secreted by vascular tissue accompanying them (Freemont et al. 2002). Some studies also found that brainderived neurotrophic factor (BDNF), a substance secreted by the IVD cells, especially during the degenerative process also appears to encourage neuronal ingrowth (Gruber et al. 2008; Kepler et al. 2013). An other possible source of pain in patients with IVD degen eration is the upregulation of proinflammatory cytokines, especially tumor necrosis factor alpha (TNFα) (Takahashi et al. 1996; Bachmeier et al. 2007. Another molecular basis of IVD degeneration is cel lular senescence (Kepler et al. 2013). Cellular senescence is an irreversible and progressive loss of replicative capa bility of the cells (Hayflick and Moorhead 1961). Based on the underlying cause, cell senescence is divided into two groups, stressinduced premature senescence (SIPS) and replicative senescence (Kepler et al. 2013). Replica tive senescence is caused by the loss of telomeres, the tip of the chromosome which serves as protection from ge nomic instability (Victorelli and Passos 2017). The telom eres shorten in each replication, when they run out, it will lead to permanent cell cycle arrest (Victorelli and Passos 2017). The cellular mechanism underlying the pathophys iology of SIPS is the accumulation of unrepairable DNA damage caused by reactive oxygen species (ROS) from mechanical injury or inflammatory cytokines release (Tou ssaint et al. 2000). Although senescence is a physiological process, studies found that this process is accelerated dur ing disc degeneration (Le Maitre et al. 2007; Kim et al. 2009).

Mesenchymal Stem Cells (MSCs)
Stem cells are cells with the ability to renew themselves and differentiate into various specialized cell types (Wei et al. 2013). Stem cells can be categorized into embryonic stem cells (ESCs), adult stem cells, and induced pluripo tent stem cells (iPSCs) (Ullah et al. 2015). ESCs are pluripotent stem cells that have distinctive selfrenewal ability, genomic stability, can differentiate to most lin eages, and hold promise for regenerative medicine (Ullah et al. 2015). However, their use is restricted for ethical reasons and tissue rejection problems following transplan tation in patients (Takahashi and Yamanaka 2006). iPSCs are made from adult cells by introducing four transcrip tion factors, cMyc (avian myelocytomatosis virus onco gene cellular homologue), Oct3/4 (octamerbinding tran scription factor 3/4), Klf4 (kruppellike factor 4), and Sox2 (sexdetermining region Y) Yamanaka 2006; Ullah et al. 2015). iPSCs share many properties with ESCs, but their genomic stability is still questionable (Ul lah et al. 2015). Due to the limitation of ESCs and iPSCs, great attention has come to MSCs, which are free from both ethical reasons and genomic stability problems (Wei et al. 2013).
MSCs are adult stem cell which possesses the ability to differentiate into connective tissue cells' lineages in cluding bone, IVD, ligament, muscle, and fat (Richard son et al. 2010). In accordance with the International So ciety for Cellular Therapy, MSCs can be identified us ing three criteria: they must adhere to the plastic sur face; express CD73, CD 90, and CD105 and does not ex press CD14, CD34, CD45, or CD11b, CD79α or CD19 and HLA Class II; and finally, they must be able to dif ferentiate into osteoblasts, chondroblasts, and adipocytes (Dominici et al. 2006). However, these criteria still pro duce a relatively heterogeneous progenitor cell population (Loibl et al. 2019). Therefore, many researchers are try ing to solve this problem by preselecting particular MSCs populations; for example, CD271-MSCs have a higher potential to differentiate into nucleus pulposus than their CD271+ counterparts (JezierskaWozniak et al. 2017). Furthermore, another study found that MSCs subpopula tions expressing CD146 or CD271 markers performed bet ter in repairing cartilage (PérezSilos et al. 2016).
Mesenchymal stem cells therapy for tissue repair de pends not only on the ability of MSCs to differentiate into specific cell types but also on their immunomodulatory and trophic effects (Wei et al. 2014). MSCs' therapeu tic effect in IVD degeneration can occur in various ways. First, interactions between nucleus pulposus cells (NPCs) and implanted MSCs induce differentiation of MSCs to wards a more chondrogenic change (Vadalà et al. 2008). Second, MSCs pose a trophic effect by secreting vari ous growth factors and cytokines that promote angiogen esis, stimulate differentiation and proliferation of progen itor host cells, and inhibit fibrosis formation (Caplan and Dennis 2006). This trophic effect also has been reported by another study showing direct celltocell contact be tween MSCs and NPCs increases NPCs viability in a co culture system (Yamamoto et al. 2004). Third, MSCs have immunomodulatory effects, they can exhibit pro inflammatory or antiinflammatory phenotype depending on the balance between the cytokines released into the sur rounding microenvironment (Keating 2012). MSCs have been shown to be indirectly having stimulatory and in hibitory effects on Bcell differentiation, proliferation, and antibody production (Fierabracci et al. 2015). These ef fects seem to be mediated by other cell types and depend on the inflammatory environment (Comoli et al. 2008).
MSCs based therapy is generally safe (Comella et al. 2017) and also appears to be able to avoid allogeneic rejec tion due to their lack of MHCII, CD 40, CD86, and CD80 expression on their cell surface, thus they can escape from Tcell recognition (Ryan et al. 2005). However, the com plication arising from the implantation procedure may also occur such as osteophyte formation due to MSCs migra tion (Vadalà et al. 2012). Another challenge is determin ing which patients will benefit from the disc regeneration, as patients seek medical help for pain, not for degeneration (Bendtsen et al. 2016).
For MSCs to effectively exert their proposed regener ation effect, they must withstand the IVD microenviron ment. MSCs depend on glycolysis for their energy source, and if glucose is removed, they will die rapidly (Moya et al. 2017). Furthermore, an acidic condition also has a detrimental effect on cellular activity and viability of the disc cells and MSCs (Wuertz et al. 2009). The IVD has a harsh microenvironment even in its healthy state due to low oxygen and nutritional supply, hyperosmolarity, and high mechanical loading which pose a challenge in clinical trials (Wuertz et al. 2009; Loibl et al. 2019). These condi tions become heavier in the degeneration process because of inflammation and increased acidity (Kepler et al. 2013).
Due to the poor nutrient supply within the IVD, ECM proteins production by MSCs or the stimulation of ECM synthesis by the host cells is likely restricted and may not be sufficient to promote full IVD regeneration (Loibl et al. 2019). It has been reported that MSCs could survive and able to differentiate after administration to the IVD (Sakai et al. 2005; Henriksson et al. 2009), but the problem lies in the nutrient supply (Loibl et al. 2019). Because of the IVD's avascular nature, they can only support limited cell numbers (Smith et al. 2011). An additional number of cells from the MSCs will interfere with the nutritional balance in the IVD because increased cell number results in in creased nutrient demand (Loibl et al. 2019). Furthermore, because of the diminished nutrient supply, the implanted MSCs will eventually die (Loibl et al. 2019). Another problem arises from the slow capacity of the MSCs to pro duce ECM. A study found that MSCs' ability to produce glycosaminoglycans (GAG), a component responsible for maintaining disc hydration and height, ranges from 0.017 to 0.086 mg GAG/million cells/month (Allon et al. 2010; McCorry et al. 2016. As GAG's concentration in the nor mal disc nucleus is about 70 mg/mL, it would take decades to restore 25% of disc tissues through implanting stimu lated MSCs (Bendtsen et al. 2016; Loibl et al. 2019.
The problem with nutritional supply that resulted in cell death suggests that the principal effects of MSCs may be mediated by paracrine mechanisms (Maguire 2013). In contrast to the original paradigm that the MSCs' mech anism of action was based on their capability to replace cells, recent studies have shown that MSCs' secreted molecules are responsible for their therapeutic effects (Madrigal et al. 2014). Therefore, MSCs secretome has gained much interest for its potential use in regeneration and tissue repair (Baglio et al. 2012; Maguire 2013.

The Secretome of MSCs
The secretome is a set of molecules released by the stem cells, including cytokines, chemokines, antiinflammatory factors, growth factors, and even proteins delivered by extracellular vesicles (EVs) (Maguire 2013; Eleuteri andFierabracci 2019). The latter can be classified according to their origin, size, density, and surface marker into exo somes, microparticles, and apoptotic bodies (Beer et al. 2017). The composition of the secretome can vary de pending on the change in its microenvironment (Vizoso et al. 2017). MSCs secretome has multiple mechanisms of action such as immunomodulation and antiinflammatory activity (Kyurkchiev 2014), neurotrophic and neuropro tective effects (Caseiro et al. 2016), antiapoptotic activ ity (Li et al. 2015a) , angiogenesis regulation (Kagiwada et al. 2008), and regenerative capacity (Osugi et al. 2012; Di et al. 2017. Table 1 lists some of the molecules se creted by the MSCs and their functions. MSCs secretome provides some advantages over cell based therapy in the regenerative medicine field: (1) se cretome resolves safety and risk problem related to MSCs implantation such as immune compatibility (Herberts et al. 2011), migration of the cells outside the implantation site which can result in osteophyte (Vadalà et al. 2012) or emboli formation (Tatsumi et al. 2013), and tumori genicity (Herberts et al. 2011); (2) secretome can be eval uated for dosage, safety and potency like other conven tional pharmaceutical agents (Vizoso et al. 2017); (3) The use of MSCscultured conditioned media (CM) can reduce several problems that are encountered in clinical applica tions of stem cells, such as safety, time, and expense (Os ugi et al. 2012), therefore increase the possibility of mass production; (4) and lastly, the biologic products in the se cretome can be modified to desired specific effects (Vizoso et al. 2017).

Immunomodulation and anti-inflammatory activity
MSCs have been reported to have an immunomodulatory property and antiinflammatory effect on both innate and adaptive immune systems through various mechanisms, notably via cytokine and chemokine secretions (Abuma ree et al. 2012). Transforming growth factor beta (TGFβ) is a cytokine produced and constitutively secreted by the MSCs (Kyurkchiev 2014). TGFβ has an essential role for the immunomodulatory property of the MSCs, includ ing inhibition of effector Tcell function and proliferation; attenuation cytokine production and cytolytic activity of natural killer (NK) cells; conversion of naive T cells into T reg; and suppression of dendritic cells (DCs), B cells, and macrophages (Yoshimura and Muto 2011). Besides TGF β, MSCs also secrete Galectin1 (Gal1) constitutively and Galectin9 (Gal9) when induced by proinflammatory stimuli (e.g. IL1β and IFNγ) (Gieseke et al. 2013). Both Gal1 and Gal9 share immunomodulatory property via inhibition of Th1 and Th17 cells proliferation, but Gal 9 is more potent to induce T cells death (Gieseke et al. 2010(Gieseke et al. , 2013. Prostaglandin E2 (PGE2) is another main effector for the antiinflammatory effect of MSCs, with its cellular target mainly are monocytes, macrophages, pe ripheral blood mononuclear cells (PBMCs), NK cells, and transitional processes of monocyte differentiation into ma ture DCs (Van Elssen et al. 2011). PGE2 exerts its anti inflammatory effect by reducing IL6, TNFα, and vascu lar permeability in mice models of sepsis (Németh et al. 2009).

Neuroprotective and neurotrophic effects
Numerous studies have reported neuroprotective and neurotrophic effects of MSCs secretome (Eleuteri and Fierabracci 2019). A group of growth factors such as brainderived neurotrophic factor (BDNF), nerve growth

Anti-apoptotic activity
MSCs prevent cell death by secreting molecules such as monocyte chemoattractant protein1 (CCL2/MCP1) and hepatocyte growth factor1 (HGF1) which have antiapoptotic activity (Kyurkchiev 2014; Madrigal et al. 2014). However, CCL2/MCP1 is also reported to have proapoptotic activity, and the balance between these two effects depends on the microenvironment and cytokine profile (Rafei et al. 2009). A study also reported that MSCs increase the antiapoptotic Bcl2 levels and de crease the proapoptotic Bax levels in rat models of my ocardial infarction treated with MSCs (Yao et al. 2005). Furthermore, another study found that MSCs secretome from the normal human uterine cervix (hUCESCs) pro moted apoptosis in cancer cells both in vitro and in vivo (Eiró et al. 2014).

Angiogenesis regulation
Angiogenesis is the formation of new blood vessels from the existing ones, which is necessary for the wound heal ing process (Vizoso et al. 2017). Recent studies reported that the beneficial angiogenic effects of MSCs are me diated by their secretome (Maacha et al. 2020). This angiogenic property is attributable to the secretion of growth factors such as vascular endothelial growth fac tor (VEGF) and HGF1 (Madrigal et al. 2014). However, a study reported that MSCs also secrete tissue inhibitor of metalloproteinase1 (TIMP1) that carries an anti angiogenic effect (Zanotti et al. 2016). The data indicate that the balance of these pro and antiangiogenic factors may be modified by hypoxic conditions and chemokine (Vizoso et al. 2017). Moreover, a study indicated that these bioactive molecules are carried by the EVs to per form their angiogenic modulation (Maacha et al. 2020).

Regenerative capacity
Previously, the regenerative properties of MSCs are at tributed to their ability to differentiate into specialized cell types, but recent evidence showed that it is due to their secreted byproducts acting at a distance that mediates re generative outcomes (Basu and Ludlow 2016). A study re ported that MSCs secrete tumor necrosis factorinducible gene 6 protein (TSG6) that promotes corneal epithelial wound healing in diabetic mice models by stimulating mi togenic activity of endogenous corneal progenitor cells and enhancing the colonyforming efficiency (Di et al. 2017). This regenerative property was also demonstrated in a study conducted by Osugi et al. that used conditioned media to accumulate paracrine factors of the MSCs to re generate bone defect in rat models (Osugi et al. 2012). They showed that MSCsconditioned media (MSCsCM) could stimulate the migration of endogenous MSCs and accelerate new bone formation (Osugi et al. 2012). This is caused by a cooperative effect between VEGF and insulin like growth factor1 (IGF1) that promotes angiogenesis and osteogenesis (Osugi et al. 2012). Hingert et al. (2020) have reported that conditioned media from human MSCs (hMSCs) has the potential to regenerate the IVD by in creasing disc cell viability and ECM production. These effects are probably caused by the presence of growth fac tors in the CM (Hingert et al. 2020). Apart from being an immunomodulator, TGFβ may also have a role in the regenerative capacity of MSCs. In a study conducted by Matta et al. (2017), they reported that TGFβ1 was able to promote cell proliferation, increase healthy ECM protein synthesis, and decrease cell apoptosis in NP cells from de generated human disc, therefore promote its regeneration.

Extracellular Vesicles (EVs) of MSCs
Besides growth factors and cytokines, most of the MSCs also secrete a large amount of micro and nanovesicles as components of their secretome, either constitutively or upon activation signals (Baglio et al. 2012). The biological properties of these vesicles have not been understood com pletely, but their potential as mediators for cell communi cation has gained much attention, especially for exosomes (Vizoso et al. 2017). The exosomes are a subclass of ex tracellular nanovesicles with a diameter of 40150 nm and a density of 1.09-1.18 g/mL that are derived from special ized intracellular compartments known as multivesicular bodies (MVBs) or late endosomes (Baglio et al. 2012; Vi zoso et al. 2017 (Figure 1).
They are able to transfer proteins, lipid molecules, and functional genetic material such as microRNAs (miRNAs) and messenger RNA (mRNAs) to the other cells (Valadi et al. 2007). The presence of RNA in the exosomes opens up its potential use as drug and gene carrier in the field of regenerative medicine and tissue engineering (Lamich hane et al. 2015). A study showed that there is a control mechanism for the sorting of miRNAs into the vesicles (Collino et al. 2010). However, this loading mechanism of miRNAs into exosomes may be modified selectively by engineering an extraseed sequence (hEXO motif) (San tangelo et al. 2016). Therefore, by selectively modifying exosomes cargo, miRNAs may be specifically transported into the target cells which lack these miRNAs for special ized function (Rader and Parmacek 2012).
Recent evidence found that regenerative properties previously credited for stem cells are actually mediated by their secreted exosomes (Basu and Ludlow 2016). Some preclinical studies have been done to demonstrate these effects, such as a study conducted by Nakamura et al., which demonstrated that MSCsderived exosomes, pri marily due to their miRNA content promote muscle regen eration by enhancing angiogenesis and myogenesis pro cess in a cardiotoxin muscle injury model (Nakamura et al. 2015). Another study conducted by Zhang et al. also demonstrates that MSCs exosomes improve functional re covery of rats with traumatic brain injury (TBI) model by promoting neurogenesis, angiogenesis, and reducing neu roinflammation (Zhang et al. 2015). MSCsderived EVs administration showed similar effects as MSCs in the treat ment of focal brain ischemia in C57BL6 mice, improving neurological impairment, neuroregeneration, and inducing longterm neuroprotection (Doeppner et al. 2015).
Even with all these beneficial and protective effects of MSCsEVs, it is mandatory to be cautious when using engineered MSCsEVs in clinical therapy. A study con ducted by Zhu et al. showed that exosome from bone mar row MSCs promoted angiogenesis and tumor growth in a mouse xenograft model of gastric carcinoma, this effect may be mediated by increasing tumor cells VEGF expres sion through activation of extracellular signalregulated kinase1/2 (ERK1/2) pathway (Zhu et al. 2012). This find ing is not entirely unexpected because MSCs have been re ported to contribute to tumor growth (Roorda et al. 2009).
Other challenges that emerge are how to avoid artifacts and ensure the reproducibility of studies since there are no available methods to ensure absolute purification and char acterization of the EVs (Eleuteri and Fierabracci 2019). Minimal Information for Studies of Extracellular Vesicles (MISEV) guidelines has summarized recommendation of how to characterize the EVs properly that depends on the presence of at least one protein of these three categories (Théry et al. 2018): (1) GPIanchored or transmembrane proteins associated with the endosome and/or plasma membrane, e.g., MHC class I, tetraspanins, or integrins; (2) cytosolic proteins, e.g., caveolins (CAV*), flotillins1 and 2 (FLOT1/2), or heat shock proteins HSC70 (HSPA8); and (3) nonEVs coisolated proteins, e.g., lipoprotein, al bumin (ALB), or TammHorsfall protein.
Furthermore, other two categories recommended to be analyzed for studies that focused on one or more EVs sub type: (1) proteins associated with intracellular compart ments other than endosomes and plasma membrane, e.g., histones (HIST1H**), or cytochrome C (CYC1); and (2) proteins that can bind to specific EVs surface receptors, e.g., collagen (COL**), or fibronectin (FN1).

Modification of MSCs secretome
The MSCs' secretome contents appear to match the IVD tissue requirements (Wangler et al. 2021). Wangler et al. (2021) have demonstrated that by exposing MSCs with healthy IVD, MSCs respond with releasing secretome that induces immunomodulation; and when exposed to trau matic and degenerative IVD, MSCs respond with releasing secretome that stabilizes ECM turnover. Recent evidence suggests that preconditioning MSCs could affect their se cretory profile, hence improve the therapeutic effects of their secretome (Vizoso et al. 2017). These in vitro pre conditioning included hypoxia (Ejtehadifar et al. 2015), proinflammatory stimuli (CroitoruLamoury et al. 2011), tridimensional culture (Bartosh et al. 2010), and pharma cological compounds .

Hypoxia
In the context of cell culture, hypoxia refers to oxygen ten sion of less than 10% (Das et al. 2010). Generally, hypoxic preconditioning of MSCs increases the cytoprotective and regenerative effects of MSCs . Ef fects of hypoxia are mediated by hypoxiainducible fac tors (HIF1α), which induce the expression of angiogenic factors such as VEGF and interleukin6 (IL6) (Ejtehadi far et al. 2015). Since neovascularization is the first step in the regenerative process of damaged tissues, this may re sult in a better therapeutic effect of preconditionedMSCs with hypoxia . Furthermore, MSCs' proliferation rate and viability are increased under hypoxic conditions (Ejtehadifar et al. 2015).

Pro-inflammatory stimuli
Exposure to the proinflammatory stimuli, particularly IFNγ induces MSCs to release indoleamine 2,3 dioxy genase (IDO) enzyme, which has an immunosuppressive effect (CroitoruLamoury et al. 2011; Kyurkchiev 2014. IDO exerts this effect by decreasing tryptophan and/or the accumulation of kynurenine, which then decreases cytotoxic T cells activity (Soliman et al. 2010). Other cytokines (e.g. IL1, TNFα, IFNβ, and IFNγ), and lipopolysaccharides are also able to induce production of IDO enzyme, although to a lesser degree (Croitoru Lamoury et al. 2011). Another study also reported that preconditioning of MSCs with TNFα enhances prolifera tion, osteogenic differentiation, and mobilization of MSCs through upregulation of bone morphogenetic protein2 (BMP2) (Lu et al. 2013). Moreover, tolllike receptors (TLR) 2/6, receptors of the innate immune response, are also reported to be able to stimulate MSCs' angiogenic ac tivity (Grote et al. 2013).

Tri-dimensional (3D) culture configuration
Typically, MSCs are cultured in vitro in monolayered sys tems however a new approach via tridimensional config uration such as spheroid culture has been reported to stim ulate a higher level of trophic factors secretion than with monolayer culture (Madrigal et al. 2014). It is worth not ing that cells located at the center of spheroid configuration will be exposed to a hypoxic condition hence increasing their viability and proliferation rate as mentioned above. Conditioned media from MSCs spheroids inhibit the pro duction of IL6, IL23, IL12p40, CXCL2, and TNFα from LPSstimulated macrophages and encourage higher pro duction of prostaglandin E2 (PGE2) (Vizoso et al. 2017). A study conducted by Bartosh et al. showed that hM SCs cultured in a spheroid culture expressed and secreted higher levels of antiinflammatory molecule TSG6 com pared with hMSCs cultured in a monolayered structure (Bartosh et al. 2010). Moreover, when they administered hMSCs to the zymosaninduced peritonitis mouse model, they showed that spheroid hMSCs culture have more ef fective antiinflammatory effects than monolayered hM SCs culture (Bartosh et al. 2010).

Pharmacological compounds
In some specific cases, preconditioning MSCs with phar macological compounds may be considered as an alter native approach . A study reported that MSCs preconditioned with atorvastatin seemed to in crease migration of MSCs and improves cardiac perfor mance due to upregulation of CXCR4 expression in rats with acute myocardial infarction models (Li et al. 2015b) . In concordance with this finding, another study found that preconditionedMSCs with oxytocin improved car diac function in ischemia/reperfusion injury rat models (Kim et al. 2012). Moreover, Liu et al. demonstrated that preconditioning MSCs with curcumin resulted in bet ter heart function, smaller infarct size, higher cells reten tion, decreased myocardial apoptosis, promoted neovas cularization, and enhanced VEGF secretion in myocardial ischemiareperfusion injury (IRI) rat models (Liu et al. 2015).

FIGURE 2
Possible mechanisms of MSCs secretome to tackle the IVD degeneration. The degenerated disc is characterized by increased ECM breakdown, neurovascular ingrowth, cellular senescence, and upregulation of proinflammatory cytokines (Freemont et al. 2002;Kepler et al. 2013). The MSCs could tackle these changes via enhancing disc cell viability, decrease disc cell apoptosis, modulate NPCs gene expression, stimulate the IVD progenitor cells differentiation, and give an anti-inflammatory effect (Brisby et al. 2013;Kyurkchiev 2014;Lv et al. 2014;Matta et al. 2017).

MSCs Secretome: Pre-clinical Evidence of Potential Use in the IVD Regeneration
In the animal models, MSCs have been demonstrated to be able to regenerate the disc (Bach et al. 2014; Free man et al. 2016; Vadalà et al. 2016. MSCs and degener ated NP cells mainly communicate through extensive bidi rectional membrane components exchange and microvesi cles (Strassburg et al. 2012). An experimental study con ducted in the bovine proinflammatory/degenerative disc models showed that MSCs have an immunomodulatory paracrine effect via their secretome products . Another study conducted by Lv et al. showed that NPlike cells that were treated with MSCs condi tioned media (MSCsCM) showed upregulation of ker atine 19 (KRT19) and downregulation of matrix metal loproteinase 12 (MMP12) and matrix gla protein (MGP) (Lv et al. 2014). Since KRT19, MMP12, and MGP have been associated with IVD degeneration, it is suggested that MSCsCM could regenerate healthy NP cells . It was further proposed that the IVD progen itor cell populations present in the degenerated IVD may be stimulated by MSCs secretome and take part in repair attempts (Brisby et al. 2013) (Figure 2).

Conclusions
The MSCs secretome is better than MSCs in some aspects such as safety, production, storage, product shelf life, and their potential as a readily available biological therapeu tic agent. The potential use of MSCs secretome for IVD regeneration is mediated by their ability to modulate NP cells' gene expressions (upregulation of KRT19 and down regulation of MMP12 and MGP) and stimulate the IVD progenitor cells to repair the degenerated disc. This re view still lacks evidence from clinical trials regarding the use of MSCs secretome to regenerate the degenerated disc.

Authors' contributions
R devised the study. R, CRSP, DT, DNU, HS, F collected the main literature related to the study. R, DNU, HS, F, FAR, HBN analyzed and interpreted the data. R, CRSP, DT, DNU, HS, F wrote the manuscript. FAR, HBN made the illustration for the manuscript. CRSP, DT provided critical revision of the article. All authors read and ap proved the final version of the manuscript.

Competing interests
The author declare that they have no competing interest.