Role of iron and iron‐related proteins in mesenchymal stem cells: Cellular and clinical aspects

Mesenchymal stem cells (MSCs) are located in various tissues where these cells show niche‐dependent multilineage differentiation and secrete immunomodulatory molecules to support numerous physiological processes. Due to their regenerative and reparative properties, MSCs are extremely valuable for cell‐based therapy in tackling several pathological conditions including COVID‐19. Iron is essential for MSC processes but iron‐loading, which is common in several chronic conditions, hinders normal MSC functionality. This not only aggravates disease pathology but can also affect allogeneic and autologous MSC therapy. Thus, understanding MSCs from an iron perspective is of clinical significance. Accordingly, this review highlights the roles of iron and iron‐related proteins in MSC physiology. It describes the contribution of iron and endogenous iron‐related effectors like hepcidin, ferroportin, transferrin receptor, lactoferrin, lipocalin‐2, bone morphogenetic proteins and hypoxia inducible factors in MSC biology. It summarises the excess‐iron‐induced alterations in MSC components, processes and discusses signalling pathways involving ROS, PI3K/AKT, MAPK, p53, AMPK/MFF/DRP1 and Wnt. Additionally, it evaluates the endogenous and exogenous saviours of MSCs against iron‐toxicity. Lastly, it elaborates on the involvement of MSCs in the pathology of clinical conditions of iron‐excess, namely, hereditary hemochromatosis, diabetes, β‐thalassaemia and myelodysplastic syndromes. This unique review integrates the distinct fields of iron regulation and MSC physiology. Through an iron‐perspective, it describes both mechanistic and clinical aspects of MSCs and proposes an iron‐linked MSC‐contribution to physiology, pathology and therapeutics. It advances the understanding of MSC biology and may aid in identifying signalling pathways, molecular targets and compounds for formulating adjunctive iron‐based therapies for excess‐iron conditions, and thereby inform regenerative medicine.


| Overview of iron uptake and utilisation
Iron is crucial for numerous cellular and physiological processes.
Following dietary iron uptake by the duodenal enterocytes (1-2 mg/ day), iron is absorbed into the circulation, where it binds to the ironcarrier protein transferrin that transports iron to all cells of the body.
Transferrin-bound-iron binds to transferrin receptors on cellsurfaces and iron is taken up by the cells. The imported iron is incorporated into functional enzymes/proteins such as haemoglobin (oxygen transport), myoglobin (oxygen transport), cytochromes (electron transport chain and drug metabolism), nitric oxide synthases (signal transduction) and ribonucleotide reductase (de novo synthesis of deoxyribonucleotides). Iron is exported out of the cells via the transmembrane protein ferroportin, which is present on the surfaces of hepatocytes and macrophages, the iron-storing and ironrecycling cells, respectively, as well as the enterocytes and placental cells. Excess iron is stored in the iron-storage protein ferritin, which exists both intracellularly and in circulation. As mammals do not possess the mechanisms for removal of excess iron, iron homoeostasis is tightly regulated at cellular and systemic levels. Natural means of iron exit from the body is through menstruation and through regular sloughing of intestinal cells (Ganz, 2011;Sharp & Srai, 2007).
In iron-loaded conditions, iron levels in the plasma exceed the ironbinding capacity of transferrin, which is manifested as increased transferrin saturation. This leads to elevation in non-transferrin-bound iron (NTBI) in the circulation and the entry of NTBI into the cells via membrane metal channels causing excessive intracellular iron accrual.
This can no longer be accommodated by intracellular ferritin, leading to increment in "free iron", which is believed to collect as labile iron pool (LIP) (Tanaka et al., 2019). Excess "free iron" can accelerate the Fenton reaction and increase the generation of reactive oxygen species (ROS) to levels that are beyond the quenching capacity of cellular antioxidants, thereby dysregulating intracellular redox homoeostasis. Unregulated ROS is toxic as it can directly damage proteins, lipids, membranes and DNA, and/or stimulate cell signalling pathways that alter gene/protein expression . Also, excess iron can directly simulate signalling pathways that may exacerbate disease pathology (Mehta et al., 2018). Excess iron deposits in the heart, liver, pancreas, bone marrow (BM), joints, endocrine glands and kidneys, and damages these organs causing heart diseases, hepatic fibrosis/cirrhosis, glucose intolerance/diabetes, arthropathy, impotence, impairment of hematopoiesis and kidney dysfunction . Whartenby, 2014;Kholodenko et al., 2019). There is a controversy regarding the exact origin of these cells. These nonhematopoietic, fibroblastic, immunologically naive cells are pleotropic in nature and perform two main functions. First, a subpopulation of this group plays a direct role in regeneration and tissue repair due to their ability to differentiate into various cell types. In vitro, MSCs have been shown to differentiate into cells with specialised functions such as adipocytes (fat cells), chondrocytes (cartilage cells), osteoblasts (which mature into osteocytes (bone cells)), myocytes (muscle cells) (Atashi et al., 2015;Iansante et al., 2018), functional hepatic-like cells (Lee et al., 2004) and neuron-like cells (Pittenger et al., 2019), while its differentiation into cardiomyocytes, lung cells and gut epithelial cells is debated (Glenn & Whartenby, 2014). In vivo, MSCs can selfrenew and undergo microenvironment-dependent differentiation into various cell types of mesodermal lineage. For example, cell tracing experiments in mice have shown the ability of perivascular MSCs to differentiate into adipocytes, myoblasts, profibrotic fibroblasts and follicular dendritic cells (Pittenger et al., 2019). The second function of MSCs is to provide physical support in the perivascular spaces and exert paracrine effects by secreting growth factors, regulatory compounds, and extracellular vesicles/exosomes containing protein, DNA, messenger RNAs (mRNAs) and microRNAs (some of which can have immunosuppressive effects) (Crivelli et al., 2017;Glenn & Whartenby, 2014;Park et al., 2018). Through these functions, MSCs form specific niches to support several physiological processes including BM homoeostasis, bone turnover, haematopoiesis, wound healing and angiogenesis (Glenn & Whartenby, 2014;B. Zhang, Wu, et al., 2015) and demonstrate antiapoptotic, antifibrotic, anti-inflammatory, proangiogenic and immunomodulatory properties (Pittenger et al., 2019). Due to their reparative characteristics and easy in vitro expansion and differentiation, MSCs make an excellent option for cell-based therapy, thereby helping to evade the challenges of whole organ transplantation in treating pathological conditions. With MSCs being one of the most sorted cell types for cell therapy, there are more than one thousand MSC-utilising clinical trials listed on the EU and NIH (US national library of Medicine) clinical trial registers.

| Rationale and significance of understanding the role of iron in MSCs
Iron is essential for all cell types including the MSCs. However, excess iron has detrimental effects on all cell types and MSCs are no exception. As the MSCs have physiological and therapeutic significance, understanding MSCs from an iron perspective can provide better insights into MSC processes and functionality. For instance, the bone marrow mesenchymal stem cells (BM-MSCs) (in conjunction with hematopoietic stem cells [HSCs]) play a vital role in creating the exclusive BM niche for haematopoiesis and are located in a low oxygenic BM environment (Pittenger et al., 2019). BM-MSC's support towards haematopoiesis is dependent on iron regulation in these cells as iron-damaged MSCs are unable to support haematopoiesis effectively. This is bound to aggravate the pathology of haematological conditions that show iron overload; for example, βthalassaemia and myelodysplastic syndromes (MDS). Application of iron-related knowledge of BM-MSC physiology could aid in formulating adjunctive therapies to restore BM-MSC functionality in such cases. Also, it may help in developing safer therapeutic strategies that not only reduce the chemotherapeutic agent-induced damage to the BM niche but also enhance the ability of iron chelators to penetrate deeper into the BM environment and enhance BM transplantation outcomes (Crippa et al., 2019).
In addition to the aforementioned haematological diseases, ironloading is frequently observed in chronic pathologies of the liver, brain and metabolic syndromes (Mehta, Ahmed, et al., 2019;, where pathogenesis could be partly attributed to iron-loading. Since the MSCs are located in various tissues, it is very likely that iron-damaged MSCs contribute to the pathogenesis. For example, BM-MSC dysfunction has been described in β-thalassaemia (Crippa et al., 2019), diabetes (Fijany et al., 2019) and MDS (Poon et al., 2019), and this dysfunction could be due to ironloading of BM-MSCs. Another example is aging, which increases the possibility of iron accumulation (Ashraf et al., 2018;Jung et al., 2008;Xu et al., 2008) while decreasing the fraction of MSCs obtainable from the BM (Pittenger et al., 2019). It is possible that these two aspects are correlated and the reduced ability to repair injured tissue in older people could be (partly or entirely) attributed to the effect of excess iron on MSC's reparative properties. Thus, knowledge of MSC physiology from an iron perspective could help in formulating supplementary therapeutic approaches to restore lost MSC functionality in such cases. This reiterates the significance of understanding the effect of iron-loading on the MSCs, and envisages the application of this knowledge to a wide range of diseases that show iron-loading.
Reviewing MSC functionality and signalling mechanisms under iron-loaded conditions can help identify cellular pathways and molecular targets of therapeutic significance. Based on the studies so far, some links between iron-related pathologies and MSC's therapeutic potential can be established. For example, hepatic ironloading increases the predisposition to hepatocellular carcinoma (Kowdley, 2004), and MSCs have been found to hinder tumour cells in hepatocellular carcinoma via specific signalling pathways (Ai et al., 2019). Similarly, neuronal iron accumulation is common in Alzheimer's disease (Duce et al., 2010) and MSCs have shown to increase neurogenesis and neuronal differentiation via Wnt signalling in an Alzheimer's disease model , probably to replace the damaged neurons.
Examining the MSCs in an iron context could be useful for improving allogeneic and autologous MSC therapy. For example, when normal MSCs and MSCs from diabetics were administered into ischaemic rat limb, normal MSCs showed better perfusion than diabetic MSCs (H. Kim, Han, et al., 2015). Here, the ineffectiveness of diabetic MSCs was attributed to their poor angiogenic ability, which is likely an effect of excess iron because iron reduces MSC angiogenic potential (Y. Zhang, Zhai, et al., 2015). Thus, knowledge of the iron-induced damage to the MSCs can help evaluate the efficacy of MSC therapy, while substances that have shown to rescue MSCs from the toxic effects of iron-loading (Camiolo et al., 2019;Yang et al., 2016;Yao et al., 2019) Blanc et al., 2008). Successful HSC engraftment is essential in conditions like β-thalassaemia for which the only curative option is allogeneic HSC transplantation from a compatible donor (Crippa et al., 2019), often difficult to find. In current times, the understanding of MSCs from an iron perceptive is particularly useful because MSCs are now used in clinical trials to manage COVID-19 (Raza & Khan, 2020).
Collectively, viewing MSC physiology from an iron perspective can enhance our knowledge of disease pathology, aid in developing adjunctive therapeutic and management strategies for excess-ironrelated conditions, and improve MSC transplantation outcomes.
Accordingly, this review examines the role of iron and related proteins in MSC biology. It addresses several aspects, namely, the damaging effects of excess iron on MSC physiology, saviours of MSCs from the detrimental effects of iron-loading, contribution of iron and endogenous iron-related proteins in normal MSC processes, and lastly, the role of MSCs in clinical conditions of excess iron.

| EFFECT OF EXCESS IRON ON MSC COMPONENTS AND PROCESSES
The effects of iron on MSC components and MSC processes have been summarised in Tables 1 and 2, respectively. These alterations are mediated by iron-induced changes in cell signalling pathways and/or vice versa, eventually altering MSC characteristics, processes and functionality. For example, the expansion of LIP can increase ROS production ( Table 1) that can stimulate signalling pathways leading to MSC cycle arrest in G0/G1 phase, apoptosis and reduced proliferation of the MSCs (Table 2) (Y. Zhang, Zhai, et al., 2015).
There are some notable points about Tables 1 and 2. First, the in vitro studies cited in these tables used the NTBI ferric ammonium citrate as the iron source. Second, in most of these studies, the ironinduced alterations in MSC physiology were attributed to excess iron and/or excess-iron-induced ROS. Third, the majority of these studies were conducted on BM-MSCs. This is because BM is the most common source for MSCs. Although present as a small fraction in the BM, BM-MSCs can be easily expanded in vitro for experimental purposes (Pittenger et al., 2019). In addition, there is an important link between BM-MSCs and haematopoiesis, which involves iron. In the BM, while HSCs carry out haematopoiesis, MSCs provide the niche for this process. The niche is formed by the paracrine effector molecules secreted by the BM-MSCs and the cells into which the BM-MSCs differentiate, that is, BM-stromal cells, pericytes, myofibroblasts, osteoblasts and endothelial cells; some of which also provide physical support (Fei et al., 2014). Thus, BM-MSCs control the fate of HSCs. Although low levels of ROS are required for the renewal of HSCs (Brissot et al., 2020), iron overload increases ROS levels and negatively affects haematopoiesis (Gattermann et al., 2012;Okabe et al., 2014); directly by damaging HSCs and causing apoptosis and cell cycle arrest Jin et al., 2018), and indirectly by damaging the MSCs via ROS signalling (Lu et al., 2012). This hinders normal MSC contribution to haematopoiesis. Essentially, under iron-loaded conditions, while the fate of HSCs is directly affected by excess-ironinduced ROS, it is also affected by lack of sufficient support from the MSCs. This clearly explains the significance of iron regulation in the MSCs for normal haematopoiesis.

| EXCESS IRON INDUCES ALTERATIONS IN MSC SIGNALLING PATHWAYS
MSC survival, proliferation, differentiation, homoeostasis and functionality are mediated by myriads of signalling pathways. The activation of these pathways usually involves successive phosphorylation of several downstream mediators that ultimately modulate the expression of specific genes. Here, selected MSC pathways that show iron-induced alterations are discussed ( Figure 2). These alterations may or may not encompass major ROS involvement, but these modulate MSC functionality under ironloaded conditions and explain the excess iron and/or ROS-mediated detrimental effects on MSCs summarised in Table 2.

| Wnt signalling
Typically, Wnt signalling is initiated upon binding of extracellular Wnt proteins to the transmembrane frizzled receptors, and LRP5/6 co-receptors (canonical) or Frizzled receptors and Rho2 coreceptors (noncanonical). Canonical Wnt signalling is mediated by stabilising the cytoplasmic protein β-catenin followed by its entry into the nucleus. Here, it binds to transcription factors and activates Wnt target genes, which are involved in the regulation of cell development, stemness, fate, migration, polarity and differentiation (Lehoczky & Tabin, 2018). Accordingly, Wnt signalling regulates MSC fate, bone remodelling and homoeostasis, as evidenced by its role in the osteogenic differentiation of BM-MSCs in adult marrow (Houschyar et al., 2019;. Both canonical and noncanonical Wnt pathways are involved in iron-regulated osteogenic differentiation of MSCs. For example, iron chelation increased the osteogenic differentiation of MSCs by increasing β-catenin levels (Qu et al., 2008). This demonstrated the role of iron in MSC differentiation via canonical Wnt/β-catenin signalling. Similarly, Wnt5a

Cellular components Normal function
Effect of iron-loading on BM-MSCs or umbilical cord MSCs, as observed in vitro, in animal or clinical studies

Osteopontin
• Secreted extracellular matrix protein that has diverse biological functions including bone remodelling, immune functions and a role in cancers and inflammatory diseases • Expression inhibited (Yao et al., 2019) • Expression decreased in BM-MSCs of patients with low-risk myelodysplastic syndromes (Fei et al., 2014) • Expression stimulated by the iron-binding protein lactoferrin  Active β-catenin • Cell-cell adhesion, gene transcription and intracellular signal transducer in the Wnt signalling pathway • Expression inhibited (Yao et al., 2019) Adipogenic genes • Peroxisome proliferator-activated receptor gamma (PPARγ), adipsin and adipocyte protein-2 mediate adipogenesis and are adipogenic differentiation markers • Fatty acid-binding protein 4 (Fabp4) is an adipocyte specific protein • Expressions increased (Y. Zhang, Zhai, et al., 2015) • Expression of PPARγ increased • Regulates (inhibits) cell cycle progression at G1 phase by binding to and preventing the activation of cyclin E-CDK2 or cyclin A-CDK2 or cyclin D-CDK4. Thus, acts as a tumour suppressor (Chu et al., 2008) • Expression reduced (Borriello et al., 2016) Cyclins A and E • Group of proteins that bind to and activate the CDKs for regulating cell cycle progression. The two together phosphorylate target proteins that trigger a specific event within the cycle Also, Wnt/β-catenin signalling can enhance ROS production, induce aging in MSCs via p53 and p21  and thereby hamper the reparative functions of MSCs. This further adds to the recognised role of Wnt in cancer development and represents Wnt signalling as a target for cancer therapy.

| ROS signalling
Iron treatment increases ROS production in the MSCs (Table 1), which positively correlates with LIP Y. Zhang, Zhai, et al., 2015). ROS is not only a by-product  (Table 2). For example, a study showed that activation of the

Cellular components Normal function
Effect of iron-loading on BM-MSCs or umbilical cord MSCs, as observed in vitro, in animal or clinical studies Stem cell factor • A cytokine that can act as a transmembrane protein as well as a soluble protein and plays a role in haematopoiesis.
Wnt/β-catenin pathway in MSCs can repair lipopolysaccharideinduced injury to lung epithelium in mice (Cai et al., 2015). As excess ROS alters Wnt signalling in MSCs, the reparative function of this pathway in the MSCs may be hampered by excess-iron induced ROS
T A B L E 2 Impact of iron-loading on MSC processes/activities

| MAPK pathways
The MAPK cascades encompass major signal transduction pathways that maintain cell survival, growth, differentiation, cell cycle and Excess iron can increase ROS levels, which can stimulate p38-MAPK signalling (Son et al., 2011). Activation of this pathway includes phosphorylation of the protein p38 and promotion of cell death (Zarubin & Han, 2005). In mice BM-MSCs, iron treatment increased the levels of p38 and phospho-p38 (Shen et al., 2018;. Also, in human BM-mononuclear cells (a heterogenous group of cells that include the BM-MSCs), iron increased the levels of phospho-p38-MAPK and p38-MAPK, collectively suggesting iron-induced stimulation of this pathway , which could be one of the causes of iron-induced MSC death (Table 2). However, in iron-treated human umbilical cord MSCs, total p38-MAPK expression remained unaltered  and in iron-treated rat BM-MSCs, phosphorylation of p38 remained unaltered (Yao et al., 2019). Further clarification is needed on the iron activation of the p38-MAPK pathway in MSCs. Nonetheless, activation of p38-MAPK is associated with apoptosis (Zarubin & Han, 2005) and this explains the iron-induced apoptosis of MSCs (Table 2). Also, p38 controls the cell cycle via p21, a cyclin dependent kinase inhibitor that mediates cell cycle arrest, and which is a substrate of p38 (Zarubin & Han, 2005). Hence, the iron-induced MSC cycle arrest (Table 2) could be partly attributed to the iron-induced increment in p21 levels, as observed in mice BM-MSCs .

| p53 signalling
p53 is a tumour suppressor protein and a transcription factor that is recognised as the guardian of the human genome. Upon activation, p53 exerts several protective effects including cell cycle arrest, apoptosis, DNA-repair and regulates ROS levels and cell metabolism (Budanov, 2014). Iron-loading significantly increased p53 levels in human BM-mononuclear cells , in human umbilical cord MSCs , and in mice BM-MSCs ; likely a defensive response against the toxic effects of excess iron. This elevation in p53 levels may be due to iron-induced or ROS-induced stimulation of the p38-MAPK pathway , which involves p38-mediated phosphorylation and activation of p53 (Zarubin & Han, 2005). Elevation in p53 may also be due to iron/ROS-induced DNA breaks, which may stimulate the ataxia-telangiectasia-mutated (ATM)-Chk2-p53-p21 pathway. In this pathway, DNA damage induces phosphorylation (activation) of p53 and its stabilisation by ATM and ataxia telangiectasia and Rad3-related protein (ATR), which acts on p21. Thus, the iron-induced increment in p53 levels may have contributed to the cycle arrest of iron-loaded BM-mononuclear cells and umbilical cord MSCs, and led to apoptosis of iron-treated MSCs   (Table 2); probably a defence mechanism against the toxic effects of excess-iron-induced ROS.

| Exogenous iron-chelators and antioxidants
Some exogenously administered compounds such as iron chelators and antioxidants showed to offer protection to the iron-loaded

| Herbs
Usage of herbal ingredients in enhancing osteogenesis in MSCs has been discussed previously (Udalamaththa et al., 2016).

| Melatonin
Melatonin is a pineal-gland-secreted naturally occurring hormone that performs numerous physiological functions including the regulation of sleep patterns and acting as an endogenous antioxidant (Reiter et al., 2016). When added exogenously to iron-loaded mice BM-MSCs, it rescued osteogenic differentiation. It reversed the ironinduced reduction in the expressions of the osteoblast-specific genes alkaline phosphatase (ALPL), collagen-1 (COL-1), BMP-2 and BMP-4, and increased the expressions of osteogenic transcription factors RUNX-2 and BGLAP (osteocalcin). In addition, melatonin partially prevented iron-induced adipogenic differentiation by inhibiting the expressions of the adipogenic genes PPARγ and C/EBPα . This may have additionally favoured BM-MSC osteogenic differentiation because inhibition of adipogenic regulators can promote osteogenic differentiation (Atashi et al., 2015;Su et al., 2015). The results were replicated in in vivo studies . Melatonin also rescued cell proliferation, attenuated premature senescence and reduced depolarisation of mitochondrial membrane potential in iron-loaded BM-MSCs.
Mechanistically, it prevented the upregulation of p53, ERK and p38 proteins, thereby blocking p53/ERK/p38 activation . In another study involving mice BM-MSCs, melatonin inhibited iron-induced ROS accumulation , Pretreatment of BM-MSCs with melatonin has shown to prevent MSC death . Thus, addition of optimised concentrations of melatonin to the growth medium could regulate ROS levels and prevent the loss of MSC reparative potential.

| IRON AND IRON-RELATED ENDOGENOUS EFFECTORS OF MSCs
MSCs express the mediators of iron regulation, namely, transferrin receptor-1 (TfR-1), ferroportin, hepcidin, ferritin, divalent metal transporter 1 (DMT-1), Zrt-/Irt-like protein 14 (ZIP14) and ZIP8, the zinc transporters (Crippa et al., 2019;Esfandiyari et al., 2019). Their ability to uptake iron through the iron-uptake channels and store it as ferritin shows their direct contribution in managing excess iron, like the macrophages. However, prolonged iron (and ROS) exposure appears to damage the iron-sensing and iron-storing machinery in the MSCs. Although iron and iron-related proteins have a canonical role in mediating normal MSC physiology, these also have specific roles in MSC biology, as described in this section.

| Iron
At the physiological level, ascorbate promotes intestinal iron absorption (Sharp & Srai, 2007). At cellular level, iron and ascorbate are important for determining cell fate specification. Histone methylation/demethylation is an important epigenetic event that modulates the expression of genes by turning them off and on. Several MSC genes whose products contribute to the development of BM niche and provide hematopoietic supportive functions are subjected to H3K36 methylation, which in turn is regulated by lysine-specific demethylases (KDMs) that are responsive to iron (Crippa et al., 2019). This suggests that via modulation of KDMs, iron may play a crucial role in determining the MSC's function of supporting haematopoiesis. In vitro, a combination of iron and ascorbate regulated histone methylation in human skeletal MSCs, where the ascorbate-regulated histone demethylase KDM4B was found to be crucial and sufficient to promote the specification of MSCs from mesoderm progenitors. This ascorbate-induced promotion of MSC specification was iron-dependent but redox-independent, which clearly shows the significance of iron in MSC specification. Also, the ascorbate-iron combination promoted long-term MSC self-renewal and increased the osteochondrogenic potential (Liu et al., 2020). This approach could be utilised to extend the lifespan of MSCs in vitro during pretransplantation procedures and to increase the probability of repairing cartilage injuries posttransplantation.

| Hepcidin
MSCs can secrete the iron hormone hepcidin (Esfandiyari et al., 2019). Since excess iron is a risk factor for osteoporosis (Weinberg, 2006), it is possible that the MSC-secreted hepcidin reduce local tissue iron levels and thereby act as an endogenous protector against the development of osteoporosis (Zhang et al., 2018).
Hepcidin may also execute its anti-bone-loss function by modulating BM-MSC differentiation. Hepcidin treatment has been shown to enhance osteogenic differentiation and mineralisation, and increase the levels of ALP and osteocalcin in rat BM-MSCs. Simultaneous increments in the mRNA expressions of BMP-2 and SMADs 1, 5 and 8 indicate that hepcidin-induced osteoblastic differentiation is mediated via the BMP2/SMAD pathway (H. Lu, Lian, et al., 2015).
As BM-MSC osteogenic differentiation involves p38-MAPK-SMAD signalling too , the effect of hepcidin treatment on the components of MAPK signalling were examined. Data showed that hepcidin upregulated phospho-p38 levels, suggesting that hepcidin-induced osteogenic differentiation may additionally involve the p38-MAPK pathway (M. Lu, Xia, et al., 2015). Together, this supports the role of MSC-derived hepcidin in MSC differentiation into osteoblasts and bone metabolism.

| Transferrin receptor, ferroportin and iron transporters
Iron entry in the MSCs occurs via similar routes to other cell types, that is, via transferrin-dependent and transferrin-independent mechanisms (Borriello et al., 2016). Transferrin-dependent mechanism involves iron entry into the cells via the cell-surface protein TfR-1.
Cellular iron ejection is mediated via the transmembrane protein ferroportin. In response to alterations in cellular iron status, their mRNA transcripts are subjected to regulatory mechanisms to produce these proteins at optimal levels so that intracellular iron homoeostasis is maintained. For example, a normal regulatory response of proliferating nonerythroid cells to increased intracellular iron is TfR-1 mRNA degradation (to prevent further iron uptake) and ferroportin mRNA upregulation (to remove excess intracellular iron) (Muckenthaler et al., 2008). Such responses were observed in irontreated MSCs from healthy subjects (Crippa et al., 2019)  The mechanisms by which these iron-injured BM-MSCs divert from canonical regulation and acquire an iron-gaining and retaining phenotype should be investigated.
Notably, there are additional aspects to TfR-1 responses, and it is yet to be confirmed which of the following aspects are applicable to MSCs. First, TfR-1 response to intracellular iron concentration (direct or inverse association) is cell-type specific and is modulated by cell proliferation, differentiation, antigens and mitogens (Schäfer et al., 2007). For example, in proliferating nonerythroid cells, TfR-1 expression is negatively regulated by intracellular iron (Chan et al., 1994), whereas in cultured human monocytes-macrophages, iron treatment upregulates TfR-1 expression (Testa et al., 1989).
Also, in proliferating nonerythroid cells, TfR-1 numbers are positively correlated with proliferation, while in haemoglobin-synthesising cells, TfR-1 numbers elevate during differentiation, so these negatively correlate with proliferation (Chan et al., 1994), in light of the fact that usually, increased proliferation is linked with reduced differentiation.
With regard to transferrin-independent mechanisms of MSC iron uptake, levels of the NTBI transporters ZIP14, ZIP8 and DMT-1 in BM-MSCs from healthy and β-thalassaemia patients were similar, unlike the case with TfR-1. However, like TfR-1, these NTBI transporters were induced at higher levels in iron-treated BM-MSCs from β-thalassaemia patients compared to BM-MSCs from healthy subjects (Crippa et al., 2019). Collectively, this suggests that in βthalassaemia, the MSCs not only lose their iron-sensing ability (partly due to the impaired iron-sensing mechanism mediated via TfR-1 and ferroportin transcripts) but also acquire unregulated iron levels via two routes; transferrin receptor-dependent and independent mechanisms; the latter mediated via the aforementioned iron transporters that allow NTBI uptake from plasma (Knutson, 2019). As such, β-thalassaemia patients show high levels of NTBI in the serum (al-Refaie et al., 1992).

| Lactoferrin
Lactoferrin is an iron-binding glycoprotein found in body secretions such as saliva, tears, serum, colostrum and milk. It exhibits antimicrobial, anti-inflammatory and immunomodulatory activities and has an anabolic effect on the bone (Cornish et al., 2004). Lactoferrin proposed to enhance liver fibrosis in iron-loaded conditions (Mehta et al., 2018). These examples collectively demonstrate that ironinduced activation of signalling pathways is independent of cell type or purpose; physiological or pathological.
In the MSCs, lactoferrin activated p38 and thereby the p38-MAPK signalling, which was dependent on TGF-β/Smad-2 signalling . As the MAPK-pathway induces osteogenic differentiation in MSCs and bone formation, it is not surprising that in postmenopausal women, lactoferrin reduced bone resorption markers and was thus beneficial for bone turnover (Bharadwaj et al., 2009). As the MAPK-pathway reflects one of the noncanonical TGF-β signalling routes, the effect of lactoferrin on p38-MAPK signalling in the MSCs indicates that like canonical TGF-β signalling, noncanonical TGF-β signalling networks may also affected by iron.

| Lipocalin-2
Lipocalin-2 is a small iron-binding cytokine of innate immunity that is released by various cell types/tissues under physiological and pathological conditions. It exhibits iron regulatory and transporting ability. By scavenging iron, it not only reduces iron availability for pathogens and provides protection against bacterial infections but also protects from oxidative stress (Xiao et al., 2017).
Also, lipocalin-2 treatment primed BM-MSCs for osteogenesis and chondrogenesis (Tsai & Li, 2017) and promoted the generation of ROS, which is known to modulate BM-MSC differentiation (M. Lu, Xia, et al., 2015). However, adipogenesis was impaired via decrement in PPARγ expression (M. Lu, Xia, et al., 2015). This was expected because inhibition of adipogenic differentiation is usually accompanied by promotion of osteogenic differentiation of BM-MSCs (Su et al., 2015). Notably, this MSC response to lipocalin-2 contrasts the MSC response to iron, where iron diminishes osteogenic differentiation and in some instances, promotes adipogenic differentiation in MSCs (Table 1) (Atashi et al., 2015); rightly so because lipocalin-2 is an iron scavenger and is expected to trigger a response opposite to that induced by NTBI (Tables 1 and 2). Alongside, lipocalin-2 has shown to elevate the expressions of TGF-β, VEGF and BMP-2, (growth factors involved in regulating marrow environment) and increase the osteoblast products osteoprotegerin and collagen type-1. Collectively, this indicates a role of lipocalin-2 in BM-MSC differentiation and remodelling of the BM environment (M. Lu, Xia, et al., 2015).
In addition to understanding the physiological role of lipocalin-2 in MSC biology, it is important to study MSC responses to lipocalin-2 treatment because lipocalin-2 overexpression in MSCs is being explored as a means of protecting and strengthening the MSCs to tackle the adverse environments during their in vitro expansion for transplantation (Halabian et al., 2015).

| Bone morphogenetic proteins
BMPs are multifunctional circulatory growth factors that are essential for a variety of developmental and physiological processes.
Although BMP-2 is thought to mediate basal hepcidin induction in the liver, BMP-6 regulates (induces) hepatic hepcidin in response to increased tissue iron and thereby plays a central role in iron homoeostasis (Corradini et al., 2011;Silvestri et al., 2019). Since MSCs can produce hepcidin (Esfandiyari et al., 2019), it is likely that the MSCs may utilise this pathway for hepcidin production. BMPs also mediate osteogenic and chondrogenic differentiation of MSCs via the BMP/ SMAD pathway and aid in bone formation. Although BMP-6 is produced by BM-MSCs before differentiation into osteoblasts (Vukicevic & Grgurevic, 2009), BMP-2 promotes osteoblastic differentiation of MSCs and is essential for bone formation (Beederman et al., 2013). Thus, the usage of BMP-2 has been explored to enhance MSC therapy. When human BM-MSCs were exposed to pulsed electromagnetic fields (clinically used for bone fracture healing) in combination with BMP-2, the osteogenic activity of BMP-2 was enhanced via activation of SMAD1/5/8 and p38-MAPK signalling (Martini et al., 2020). This demonstrates the significance of BMP-2 in repair therapies (Taheem et al., 2018).
This overview of the MSC effectors clearly shows a link between iron regulation and MSC functionality and suggests that the mechanisms involved in these processes may be mediated via common signalling pathways and/or factors.

| CLINICAL CONDITIONS OF IRON EXCESS AND MSCS
Iron overload can occur due to mutations in the iron-regulatory genes, as characterised in hereditary hemochromatosis (Pietrangelo, 2016) Wang et al., 2016). Notably, some of these conditions are not independent of each other but are interconnected, often one leading to another ( Figure 5). For instance, excess iron is a risk factor for cancers (Fonseca-Nunes et al., 2014;Torti & Torti, 2013). Hereditary hemochromatosis patients are at risk of developing hepatocellular carcinoma, and also show increased risk for diabetes and metabolic syndrome (Simcox & McClain, 2013;Wessling-Resnick, 2017). Inevitably, hereditary hemochromatosis patients and β-thalassaemia major patients can develop diabetes (Barnard & Tzoulis, 2013;Barton & Acton, 2017). This section discusses the contribution of MSCs in some of these iron loading pathologies and envisages a role of MSCs in promoting interconnections between different pathologies in excess iron states.

| Hereditary hemochromatosis
Mutations in the genes that either encode hepcidin or modulators of hepcidin (HFE, HJV, TFR-2) or ferroportin lead to iron deposition in the body; a group of conditions called hereditary hemochromatosis (Pietrangelo, 2010). Approximately 40%-80% and 30% of hereditary hemochromatosis patients suffer from osteopenia and osteoporosis, respectively, involving a disbalance between bone resorption and bone formation (Jeney, 2017). Thus, in the context of MSCs in hereditary hemochromatosis, the most studied topic has been the impact of excess iron on BM-MSC differentiation and bone health. The ability of BM-MSCs to differentiate into osteoblasts is crucial for bone remodelling and healing (Jeney, 2017). Several secreted differentiation factors like TGF-β, BMPs, Wnt proteins and the Indian HedgeHog protein can activate respective signalling cascades to stimulate osteogenic differentiation of BM-MSCs. All these pathways in the BM-MSCs converge on the main osteogenic transcription factor RUNX-2, which mediates the transcription of the main bonespecific proteins osteocalcin, osteopontin collagen I-α1, and ALP (Jeney, 2017). Iron (and ferritin) selectively inhibits the osteogenic differentiation of BM-MSCs by preventing the upregulation of RUNX-2 and its target genes osteocalcin and ALP (Balogh et al., 2016). Such an effect of iron and ferritin is also seen in osteoblasts, where the iron-induced inhibition of osteoblast activity is mediated via ferritin's ferroxidase activity. Here, ferritin downregulates the osteoblast-specific markers ALP, osteocalcin and CBF-α1 and inhibits calcification (Zarjou et al., 2010). This makes elevated iron a risk factor for osteoporosis in hereditary hemochromatosis (Guggenbuhl et al., 2005;Jeney, 2017;Valenti et al., 2009). Also, iron-loading causes apoptosis of BM-MSCs (Yuan et al., 2019) and affects osteoblasts negatively, which reduces bone formation (Tsay et al., 2010). The increased risk for osteoporosis is linked with low osteogenic potential of the circulating MSCs, even though MSC numbers may be increased (Jeney, 2017). This explains bone loss and low bone mineral density in hereditary hemochromatosis patients (Jandl et al., 2020).

| Diabetes
Diabetes is characterised by prolonged elevation of blood sugar levels. A strong correlation between high iron levels and diabetes is well-established (Swaminathan et al., 2007). Contextually, two F I G U R E 5 Iron-related conditions and associated risks. The figure shows that iron-related conditions can cause predisposition to cancer, osteoporosis and diabetes. Moreover, an iron-related condition can cause predisposition to multiple clinical conditions. African iron-overload is a diet-related acquired condition. In hereditary haemochromatosis, iron-excess occurs due to mutations in iron-related genes whereas in β-thalassaemia and MDS, iron-loading occurs primarily due to repeated blood transfusions. The contribution of mesenchymal stem cells in causing this predisposition is unknown. It is highly probable that iron-damaged mesenchymal stem cells play a role in maintaining or promoting the predisposition, that is, enhance the pathology of a certain disease/condition and increase the risk for development of another disease/ condition. MDS, myelodysplastic syndromes factors increase the risk for type 2 diabetes; high iron stores, that is, increased ferritin levels, and a low transferrin receptor: ferritin ratio.
Inevitably, type 2 diabetes is frequently associated with high ferritin levels. Insulin resistance in these patients can be ameliorated by iron chelation (Lecube et al., 2004). In addition, decreased iron-binding antioxidant capacity (Van Campenhout et al., 2006) and the prevalence of NTBI in type 2 diabetic patients indicates the contribution of NTBI in diabetes (Lee et al., 2006). Together, this reiterates the significance of iron in the development of diabetes. On the other hand, physiological stress in diabetes causes functional defects in the MSCs. Here, MSCs show elevated oxidative stress, reduced proliferation and differentiation, increased apoptosis and altered cytokine profile (Fijany et al., 2019). Thus, it is likely that iron-damaged MSCs may partly contribute to diabetes pathology.
Normally, MSCs promote vasculogenesis, that is, de novo formation of blood vessels. However, in diabetes, MSCs show altered support towards vasculogenesis and angiogenesis (growth of newly formed blood vessels) by expressing substantially low levels of VEGF, a potent angiogenic factor, and αβ-crystallin, a chaperone for VEGF (Fijany et al., 2019). Simultaneously, iron-induced reduction of VEGF in the MSCs has been reported (Table 1). Therefore, altered vasculogenesis and angiogenesis in diabetes could be partly or fully attributed to the effects of excess iron on MSCs. In addition, hyperglycaemia decreases RUNX-2 expression, which in turn affects osteopontin, osteocalcin and osteoprotegerin (all osteogenic factors), and thereby reduces the osteogenic differentiation potential of MSCs (Fijany et al., 2019). Although this attenuation of MSC's osteogenic potential in diabetes could be due to the excess-iron induced effects on MSCs (Table 1), it could also be due to the reduced expression of BMP-2 (upstream activator of RUNX-2 expression) in high glucose states. Here, involvement of BMP-2, the mediator of basal iron regulation via basal hepcidin induction, and the restoration of RUNX-2 expression by addition of exogenous BMP-2 jointly reiterate the link between iron and glucose regulation. Furthermore, it presents BMP-2 as an important target for repairing impaired osteogenic differentiation potential of MSCs in diabetic states (Fijany et al., 2019

| Myelodysplastic syndromes
MDS are a group of blood disorders characterised by ineffective haematopoiesis in the BM leading to insufficient numbers and abnormality of blood cells. Here, an abnormal BM environment is accompanied by altered functions of BM-MSCs. Iron-loading in these patients begins when ineffective erythropoiesis represses hepcidin production in the liver leading to uncontrolled intestinal iron uptake and uncontrolled release of iron from splenic macrophages (Brissot et al., 2020;Gattermann, 2018).
Blood transfusion is one of the treatment options for MDS which adds to the existing iron load and becomes the main cause of iron overload in these patients. The consequent increase in ROS production and ironinduced toxicity (Angelucci et al., 2017) is bound to affect the BM-MSCs.
Accordingly, iron-loaded MDS patients show diminished BM-MSC quantity, proliferative ability and differentiation potential. In these patients, iron-loading promotes mitochondrial fragmentation in the MSCs , and downregulation of VEGF-A, CXCL-12 and TGF-β1

| β-Thalassaemia
β-Thalassaemia is an inherited disorder characterised by ineffective erythropoiesis and anaemia due to mutations in the β-globin gene, which results in the reduction or absence of β-globin chains of haemoglobin.
Thalassaemia major patients receive regular blood transfusions to maintain haemoglobin levels, which causes iron-loading (high iron and ferritin) despite the concomitant use of iron chelators (Jeney, 2017). Ironloading is aggravated due to ineffective-erythropoiesis-induced hepcidin dysregulation (Crippa et al., 2019). This predisposes the patients to multiple end-organ complications including a risk for osteoporosis (Tsay et al., 2010;Wong et al., 2016). Thus, more than 60% adult patients in this group have low bone mineral density (Jeney, 2017). An iron impact on bone health implies an impact on BM-MSCs in these patients.
BM-MSCs from β-thalassaemia patients showed insufficiently activated RUNX-2 osteogenic gene, limited osteogenic potential, and impaired differentiation into adipocytes; the latter occurred possibly due to lack of sufficient multipotent progenitors that would additionally form the adipose tissue (Crippa et al., 2019). Along with altered clonogenicity capacity, these cells showed reduced proliferation and reduced ability to attract the HSCs. The reason for impoverished BM-MSC quality and reduced frequency of primitive MSCs (CD271 + and CD146 + ) in these patients was attributed to elevated ROS (Crippa et al., 2019), which is one of the direct results of iron-loading. Other altered characteristics of these MSCs such as diminished antioxidative response and reduced expression of the genes whose products contribute to the formation of the bone-marrow niche (Crippa et al., 2019) could be partly attributed to elevated iron levels in these patients (Tables 1 and 2). The latter group includes reduced expressions of the following in BM-MSCs of βthalassaemia patients (Crippa et al., 2019): CXCL-12 and KITLG that are essential for the engraftment, retention, survival and proliferation of HSCs , CDH-2, encoding the cell adhesion molecule Ncadherin, which plays a role in regulating and retaining HSCs in the BM (Arai et al., 2012), FGF-2 and IL-6 that help stem cells self-renew and maintain stemness of BM-MSCs, respectively (Coutu & Galipeau, 2011;Pricola et al., 2009), and VEGF-A and ANGPT-1 that regulate HSC quiescence and engraftment during autologous HSC transplantation (Nowicki et al., 2017). In cases of iron overload, the resultant unregulated ROS levels can promote stem cell activation and therefore the HSCs are likely to exit quiescence (Chaudhari et al., 2014). Moreover, it is postulated that chronic iron exposure might alter the activities of the irondependent proteins that modulate chromatin exposure, alter histone methylation patterns, and cause epigenetic remodelling of MSCs in β-thalassaemia patients leading to altered functionality of the BM-MSCs diabetes, β-thalassaemia and MDS ( Figure 6).
F I G U R E 6 Involvement of iron and iron-related proteins in MSC-biology and its implications. MSCs facilitate osteogenesis, adipogenesis, chondrogenesis and provide bone-marrow support. These cells secrete both, the iron-hormone hepcidin and the iron-binding protein lipocalin-2. Non-transferrin-bound-iron transporters, transferrin receptor and ferroportin facilitate MSC iron entry and exit, respectively. Hypoxia inducible factors are additional iron-related players that influence MSC functionality. Excess iron alters MSC signalling pathways like ROS, PI3K/AKT, MAPK, p53, AMPK/MFF/DRP1 and Wnt. Systemic iron-binding proteins like transferrin and lactoferrin bind excess free iron, and this limits excess-iron induced damage to MSCs. In addition, endogenous hepcidin, ferritin and haem oxygenase-1 can shield the MSCs from the detrimental effects of iron-loading. Exogenous compounds like iron chelators (deferasirox and deferoxamine), herbs (Herba Epimedii and Astragalus membranaceus), antioxidants (NAC) and some naturally occurring compounds (melatonin and α-lipoic acid) can protect MSCs from excess-iron induced toxicity. Separate treatments of MSCs with exogenous BMP-2, hepcidin and lipocalin-2 have shown beneficial effects. Iron-loaded MSCs may facilitate or accelerate the pathological progression of excess-iron conditions like hereditary hemochromatosis, diabetes, β-thalassaemia and MDS.