Calcium regulation of stem cells

Abstract

Pluripotent and post‐natal, tissue‐specific stem cells share functional features such as the capacity to differentiate into multiple lineages and to self‐renew, and are endowed with specific cell maintenance mechanism as well as transcriptional and epigenetic signatures that determine stem cell identity and distinguish them from their progeny. Calcium is a highly versatile and ubiquitous second messenger that regulates a wide variety of cellular functions. Specific roles of calcium in stem cell niches and stem cell maintenance mechanisms are only beginning to be explored, however. In this review, I discuss stem cell‐specific regulation and roles of calcium, focusing on its potential involvement in the intertwined metabolic and epigenetic regulation of stem cells.

This review discusses how calcium regulates stem cell niches and stem cell maintenance and how it is possibly involved in the intertwined metabolic and epigenetic regulation of stem cells.

Glossary

5hmC

5 hydroxymethyl cytosine

5mC

5‐methylcytosine

ADP

adenosine diphosphate

AMPK

AMP‐activated protein kinase

AP‐1

activator protein 1

Atp2b1‐4

ATPase plasma membrane Ca²⁺ transporting 1‐4

ATPA2

ATPase sarcoplasmic/endoplasmic reticulum Ca²⁺ transporting 2

ATP

adenosine triphosphate

ATPC1

ATPase secretory pathway Ca²⁺ transporting 1

BM

bone marrow

CamK

calmodulin kinase

CaMKII

calcium/calmodulin‐dependent kinase 2

Cas9

CRISPR‐associated protein 9

CaSR

calcium‐sensing receptor

CoA

Co‐enzyme A

CRAC

calcium release‐activated calcium

CRISPR

clustered regularly interspaced short palindromic repeats

Crtc

Creb‐regulated transcriptional co‐activator

DHA

dihydroxyacetone phosphate

ER

endoplasmic reticulum

ESC

embryonic stem cell

ETC

electron transport chain

FAD

flavine adenine dinucleotide

FGF2

fibroblast growth factor 2

G3P

glycerol‐3‐phosphate

GEXC

gene expression commons

GGCX

gamma‐glutamyl carboxylase

Gla

gamma‐carboxyglutamate

GPCR

G protein‐coupled receptor

GPD2

FAD‐glycerol‐3‐phosphate dehydrogenase 2

GSK3

glycogen synthase kinase 3

HAT

histone acetyl transferase

HDAC

histone deacetylase

h

human

HIF

hypoxia‐induced factor

HSC

hematopoietic stem cell

IDH

isocitrate dehydrogenase

IP3R

IP3 receptor

iPSC

induced pluripotent stem cell

ISC

intestinal stem cell

JMJC

Jumanji C

KGD

a‐ketoglutarate dehydrogenase

LIF

leukemia inhibitory factor

MAM

mitochondria‐associated membrane

MAPK2

mitogen‐activated protein kinase 2

MCU

mitochondrial calcium uniporter

MeCP2

methyl CpG‐binding protein 2

MEF2C

myocyte enhancer factor 2C

MFN2

mitofusin 2

MGP

Matrix Gla protein

m

mouse

mTOR

mammalian target of rapamycin

NAADP

nicotinic acid adenine dinucleotide phosphate

NAD

nicotine adenine dinucleotide

NCLX

sodium calcium lithium exchanger

NCX

sodium calcium exchanger

NFAT

nuclear factor of activated T cells

OAA

oxaloacetic acid

ORAI1

ORAI calcium release‐activated calcium modulator 1

PDH

pyruvate dehydrogenase complex

PDK

pyruvate dehydrogenase kinase

PDP1

pyruvate dehydrogenase phosphatase 1

PHD

prolyl hydroxylase domain

PIP2

phosphatidylinositol 4,5‐bisphosphate

PLC

phospholipase C

PMCA

plasma membrane calcium efflux pump

POU5F

POU class 5 homeobox 1

PSC

pluripotent stem cell

PTH

parathyroid hormone

RISP

Riske iron‐sulfur protein

ROS

reactive oxygen species

RyR

ryanodine receptor

SERCA

sarcoplasmatic calcium pump

Slc8a1

solute carrier family 8 member a1

Slc8b1

solute carrier family 8 member a1

SOCE

store‐operated calcium entry

STIM1

stromal interaction molecule 1

TCA

tricarboxylic acid cycle

Tcf7l1

transcription factor 7‐like 1

TET

ten‐eleven translocated

TPC

two‐pore channel

TRPV6

transient receptor potential cation channel subfamily V member 6

VDAC1

voltage‐dependent anion‐selective channel 1

Introduction

Because of the promise of stem cells for the treatment of a wide variety of diseases, their biology has received enormous attention and major progress was made. The term “stem cell” covers pluripotent stem cells (PSCs), which either arise in the inner cell mass of preimplantation embryos (embryonic stem cells, ESCs) or can be generated by reprogramming of somatic cells (induced pluripotent stem cells, iPSCs) and give rise to all lineages of the embryo proper, totipotent stem cells that also generate extraembryonic tissues, and post‐natal stem cells that endow specific organs and tissues in which they reside with regenerative capacity 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13. These stem cell types share functional features such as multipotentiality and the capacity to self‐renew. However, while distinct from more mature cells and progenitors, marked differences exist in the transcriptional and epigenetic regulation and the response to external cues among the different types of stem cells. More recent studies have begun to investigate the specific cell biology of stem cells, focusing on organellar functions, cellular quality control, and metabolism 14, 15, 16, 17, 18, 19, 20. Here, too, stem cells appear endowed with maintenance mechanisms, organellar physiology, and metabolic regulation that differ from those in mature cells, but also vary among types of stem cells. An understudied area is the specific role of calcium in the biology of stem cells. As calcium is a highly versatile and ubiquitous second messenger that regulates a wide variety of cellular functions, stem cell‐specific functional roles and regulation of calcium are to be anticipated.

Calcium signaling and regulation

Calcium signaling and regulation is a very extensive field of research. Here, I point out some important features (Fig 1), mostly citing appropriate reviews. Calcium concentration in steady state is low in the cytoplasm (10⁻⁷ M). Most cellular calcium is stored in the ER, lysosomes, and mitochondria, where concentrations are up to 100‐fold higher. Calcium is removed against a steep gradient (extracellular calcium is ~1–2 M⁻³) in an ATP‐dependent process from the cytoplasm by plasma membrane efflux pumps (PMCAs) of which there are four isoforms (PMCA1‐4, encoded by Atp2b1‐4, respectively) 21 and, predominantly in excitable cells, the Na⁺/Ca²⁺ exchanger, NCX (Slc8a1) 22. ATP‐driven SERCA pumps transport calcium into the endoplasmic reticulum (ER). Calcium entry into the intermembrane space of the mitochondria is mediated by the VDAC1 channel in the outer membrane. The mitochondrial calcium uniporter (MCU) and its multiple regulatory accessory proteins transport calcium across the inner membrane, driven by the mitochondrial membrane potential gradient 23, 24. Finally, approximately 200 calcium‐binding proteins are known that, among others, buffer calcium and therefore reduce the concentration of free calcium 25. Calcium entry into the cytoplasm from the extracellular space is initiated by voltage‐gated channels or by receptor‐operated ionotropic channels and can be amplified by store‐operated calcium entry (SOCE) 26. SOCE is also induced by any signaling pathway that leads to the generation of inositol 1,4,5‐trisphosphate (IP3). IP3 arises by the action of phospholipase C family members, which catalyze the conversion of phosphatidylinositol 4,5‐bisphosphate (PIP2) to diacylglycerol and IP3. IP3 binds IP3 receptors (IP3Rs) on the ER, triggering their opening and causing influx of calcium from the ER to the cytoplasm. The resulting ER calcium depletion is sensed by the EF‐hand domain protein, STIM1, which associates with calcium channels on the plasma membrane, ORAI1, 2 and 3 (collectively called CRAC channels). The CRAC channels subsequently open, allowing influx of calcium from the extracellular space, thus amplifying the calcium transient and refilling ER calcium stores 26. In neurons and in skeletal, heart, and smooth muscle cells, three homologous ryanodine receptors (RyRs) function as calcium‐gated calcium channels on the ER 27, 28. Both RyRs and IP3Rs are subjected to positive (calcium‐induced calcium release) or negative feedback by calcium depending on local calcium concentration, while PLC activity itself is enhanced by calcium. Calcium can also be released from the mitochondria through the NCLX (Slc8b1) Ca²⁺/Na⁺ exchanger 29. Furthermore, acidic organelles such as lysosomes contain two‐pore channels (TPC) that can release calcium in response to the second messenger, nicotinic acid adenine dinucleotide phosphate (NAADP) 30.

Figure 1. Calcium regulation in cells.

Schematic representation of cellular calcium influx, efflux, and buffering mechanisms. DAG: diacylglycerol; IP3: inositol 1,4,5‐triphosphate; MCU: mitochondrial calcium uniporter; NCLX: mitochondrial sodium/calcium exchanger; NCX: plasma membrane sodium/calcium exchanger; ORAI1: calcium release‐activated calcium channel protein 1; PIP2: phosphoinositol 4,5‐biphosphate; PLC: phospholipase C; PMCA: plasma membrane calcium efflux pump; SERCA: sarcoplasmatic calcium pump; RyR: ryanodine receptor; STIM1: stromal interaction molecule 1; VDAC1: voltage‐dependent anion‐selective channel 1.

Cytoplasmic calcium regulates an enormous variety of cellular processes, ranging from excitation–contraction coupling in muscle cells to neurotransmitter release, intracellular signaling, transcription and translation, cell motility and shape, hormone release, and metabolism, often via binding of calmodulin, which regulates calmodulin kinases (CaMKs). Calmodulin also regulates nuclear factor of activated T cells (NFAT). An increase in cytoplasmic calcium activates the serine/threonine phosphatase, calcineurin, via calcium–calmodulin. Calcineurin dephosphorylates NFAT family members, thus allowing their migration into the nucleus, where they regulate transcription of a broad array of genes in a cell‐type‐dependent fashion in conjunction with AP‐1 31.

A fascinating question is how one second messenger can carry such a tremendously high information content resulting in the regulation of an astounding but coordinated diversity of cellular process. Part of the answer lies in cell‐intrinsic features that allow a specific response to a calcium signal. Only cells endowed with a contractile apparatus will show shortening in response to a calcium transient, and specific calcium transients will only occur in cells expressing the required calcium channels. The other part of the answer lies in spatiotemporal decoding, which links a calcium signal with a specific output within a given cell. Calcium increases in the cytoplasm can be sustained, transient or oscillatory, with associated and sometimes reciprocal changes occurring in organelles. The biological basis of calcium oscillations is still unclear, but involves complex feedback mechanisms between cytoplasmic calcium, intracellular calcium stores, and influx and efflux mechanism at the plasma membrane 32. Frequency, duration, and amplitude of periodic variation in calcium concentration are decoded by calcium‐regulated mechanisms 33, 34, 35, 36. Although several reports suggested a major role for frequency encoding 35, 36, more recent studies suggest that width, amplitude, and degree of sustainability of the spikes as well as baseline calcium are integrated to generate a specific functional output 32. The resulting responses are determined, among others, by the binding kinetics of calcium to specific effectors and by the kinetics of the enzymatic activity of these effectors on their targets. Among these effectors, calmodulin and CaMKs are the most intensely studied. Furthermore, cytoplasmic calcium concentrations are not uniform. Calcium concentrations can locally be very high in so‐called calcium microdomains, such as those that occur near CRAC channels, ryanodine receptors, and IP3Rs (Fig 1) 37. The importance of such microdomains is illustrated by the mitochondrial calcium transporter, MCU. The affinity of the MCU for calcium is too low to be active at ambient cytosolic calcium concentrations. However, the MCU complex is located near mitochondria‐ER contacts (mitochondria‐associated membranes or MAMs) where local concentrations are much higher during calcium release from the ER 38, 39. Finally, nuclear calcium is subjected to influx originating from calcium stores in the nuclear envelope and the nucleoplasmic reticulum, which is continuous with the ER. Differential roles and partially independent regulation of nuclear and cytoplasmic calcium have been reported 40, 41. These principles are nicely illustrated by NFAT. NFAT is preferentially activated by calcium transients of specific amplitude, duration, and frequency 35, while one isoform, NFAT1, is specifically activated in microdomains near CRAC channels, whereas NFAT4 requires an increase in both cytoplasmic and nuclear calcium for activation 42.

As stem cells reside in a specific niche that orchestrates their function, the field should also consider regulation of extracellular calcium in the stem cell microenvironment. Serum calcium concentration is tightly regulated by endocrine hormones (parathyroid hormone, calcitonin, and vitamin D) which regulate intestinal absorption, renal secretion, and organismal buffering of calcium in bones 43. Tissue calcium concentrations are likely locally regulated as well, although very little information is available on this technically challenging subject 44. Given the role of the niche in the regulation of stem cell function and maintenance in various tissues 8, 45, 46, 47, this area of research would be of major relevance, however. Although little research has been done on the regulation of tissue calcium, multiple extracellular proteins bind calcium, which not only regulates their function, but may also affect extracellular free calcium concentration 48. Deletion of matrix Gla protein (MGP), for example, causes ectopic calcification of arteries and cartilage 49. MGP is part of a family of proteins with γ‐carboxyl glutamic acid (Gla) domains, which relatively avidly bind and therefore potentially buffer calcium. Gla proteins include, in addition to MGP, coagulation factors, anti‐coagulation factors (Protein S and C), the osteoblast‐specific hormone osteocalcin, and periostin 50. Gamma‐carboxylation is performed by the vitamin K‐dependent enzyme, gamma‐glutamyl carboxylase (GGCX). This likely explains why vascular calcification, tracheal calcification, and osteoporosis are calcium‐related side effects of anticoagulant treatment with warfarin, an inhibitor of the vitamin K cycle 51, 52. Both specific regulation of and signaling mediated by intracellular calcium and the availability of calcium in the niche should therefore be considered when examining the calcium physiology of stem cells.

Calcium and stem cell regulation

I focus here on data suggesting specific roles of calcium in the maintenance, self‐renewal, and fate decisions of pluripotent and post‐natal stem cells. Studies on specific roles of calcium in stem cell maintenance and tissue regeneration have been published in PSCs, the hematopoietic system, the skin, and the fly intestine.

Hematopoietic stem cells

Hematopoietic stem cells (HSCs) are perhaps the most paradigmatic type of post‐natal stem cell. HSCs reside in the bone marrow (BM), are quiescent, can self‐renew, and generate all lineages of the hematopoietic system 9. If is not fully understood how steady‐state function and homeostatic responses of HSCs are regulated, however 9, 53, 54. HSCs display clonal heterogeneity in their differentiation potential ranging from rare lymphoid‐biased HSCs to balanced myeloid/lymphoid and myeloid‐dominant HSCs with low lymphoid potential. The latter become progressively preponderant with age 55, 56, 57, 58. The first evidence that calcium affects HSC function came from studies showing that hematopoiesis‐specific deletion of mitofusin 2 (Mfn2) impaired the function of HSCs with extensive lymphoid potential 59. MFN2 is one of two homologous mitofusins that promote fusion of mitochondria. In addition, MFN2 is located at MAMs where it promotes transfer of calcium from ER to mitochondria 60, 61. Mfn2 mRNA is relatively selectively expressed in adult HSCs, which have highly fused mitochondria. In the absence of Mfn2, intracellular calcium in HSCs was elevated, leading to the nuclear translocation of NFAT. Inhibition of NFAT signaling using a peptide that inhibits calcineurin‐NFAT interaction rescued the defect 59.

Further studies then showed that overall, HSCs are endowed with low intracellular calcium compared with progenitors conveyed, among others, by elevated expression and activity of PMCA pumps compared with other hematopoietic cells (Fig 1A). Culture of HSCs in vitro in the presence of cytokines leads to a rapid loss of stem cell potential. Culture of BM HSCs in low calcium, however, increased their maintenance as determined by phenotype, function, and single cell expression signature. Although quantitative expansion was not observed, maintenance over a 2‐week culture period in the presence of cytokines was improved 20‐fold. Furthermore, in particular HSCs with serial repopulation capacity were functionally maintained. Similar data were obtained with human cord blood cells. One of the underlying mechanisms is likely inhibition of differentiation and maintenance of a HSC single cell expression signature. Corroborating these beneficial effects of a low‐calcium environment on HSC maintenance, the calcium concentration in the BM interstitial fluid was approximately fourfold lower than that of serum (Fig 2A) 62. A requirement for the calcium‐sensing receptor (CaSR), a G protein‐coupled receptor (GPCR) that responds to increased extracellular calcium concentrations, has been reported in the homing of HSCs from fetal liver, where fetal hematopoiesis takes place, to the BM around birth 63. These data would suggest that HSCs are attracted to the BM because of higher calcium concentrations. Luchsinger et al, however, could not confirm expression of CaSR on adult BM cells 64. Analysis of Casr mRNA expression in the publicly available Gene Expression Commons (GEXC) database (https://gexc.riken.jp/) 65, which contains genome‐wide microarray expression profiles of hematopoietic populations, also revealed very low expression throughout the hematopoietic system, although temporary expression of CaSR in neonatal HSCs cannot be excluded. Although around osteoporotic osteoclastic erosion sites, calcium can be 20‐fold higher than in serum 66, and bone is the prime organismal calcium buffer. Trabecular bone may therefore absorb calcium in steady state in young individuals. Intriguingly in this context, a recent report indicates that HSCs localize near areas of bone turnover, which may affect local calcium concentrations 67. While a recent report showed expansion of HSCs in vitro, among other measures by replacing albumin with polyvinyl alcohol l 68, low‐calcium cultures can open avenues for HSC manipulation based on their specific physiological features. It could, for example, be envisaged that low‐calcium culture may be used for gene transfer and gene editing and for purging of BM from malignant cells as the capability to maintain HSCs in vitro may allow elimination of contaminating tumor cells using specific drugs or antibodies prior to autologous transplantation.

Figure 2. Calcium regullation in various stem cell types and niches.

Schematic representation of calcium concentration and its effect in HSCs (A), skin (B), in vitro cultured keratinocytes (C), mESCs (D), and fly ISCs (E).

Further supporting a role for intracellular calcium concentration in HSC function, Umemoto et al showed that after induction of HSC cycling in vivo, intracellular calcium rose in association with an increase in mitochondrial membrane potential, indicating a metabolic role of calcium in HSCs 69 (Fig 2A). In disagreement with these findings, however, Fukushima et al reported that the most potent and quiescent (so‐called dormant) HSCs are in fact endowed with higher intracellular calcium compared with cycling HSCs and progenitors 70. The specific calcium indicator dye used in that study, CaSIR1, differs from more classically used dyes, such Indo1 and FluoForte, and may potentially explain some of this discrepancy. While Umemoto et al 69 found that nifedipine, an antagonist of L‐type voltage‐gated calcium channels promoted maintenance of HSCs, Fukushima et al 70 observed no effect of another L‐type channel inhibitor, verapamil. Voltage‐gated calcium channels are, however, typically expressed in excitable cells. Furthermore, examination of the GEXC database 65 indicates that expression of voltage‐gated calcium channels is very low in HSCs.

Some clinical and experimental data suggest a role for endocrine regulation of serum calcium—and potentially of calcium in the BM space—on hematopoiesis. Single daily injection of parathyroid hormone (PTH1‐34, teriparatide) or expression of a constitutively active PTH receptor in osteoblasts promotes bone anabolism 71. In mice, both interventions increased HSC number and function 72, 73. The underlying mechanism was reported to involve activation of Notch signaling 73. Although it is now accepted that not osteoblasts but rather BM sinusoids are the niche for HSCs 45, single dose PTH may affect BM calcium, thus potentially explaining its effect on HSC maintenance. Further supporting this notion is the fact that PTH has been reported to predominantly stimulate trabecular and not periosteal osteoblasts 74. Trabecular bone is where most of the hematopoietic activity resides. On the other hand, primary and secondary hyperparathyroidism, where PTH production is sustained and mobilization of calcium occurs through predominant osteoclast activation leading to hypercalcemia, has been reported to be associated with hematological abnormalities that were corrected by treatment of PTH overproduction 75, 76, 77, 78, 79, 80, 81, 82, 83. Many of these reports indicate the presence of myelofibrosis, which is also associated with myeloproliferative disease. Myeloproliferative changes were also observed in a mouse model of osteogenesis imperfecta associated with high bone turnover 84. An association between myelodysplasia and osteoporosis has been suggested 85. Osteopontin likely inhibits hydroxyapatite crystal formation 86, thus preventing deposition of calcium into the bone, and is a negative regulator of HSC function in mice 87. Finally, it was recently reported that vitamin K antagonists, of which tissue calcification and osteoporosis are clinical side effects 51, 52, change bone composition and impair HSC function in mice and that their use correlates with increased incidence of myelodysplasia, although a direct role for BM calcium was not considered 88. Taken together, published evidence, though circumstantial, supports the hypothesis that bone dynamics affect hematopoiesis and HSC function in vivo and that calcium might be involved.

Epidermal stem cells

The upper layer of the skin (stratum corneum or cornified layer) forms a tight physical barrier that is continuously sloughed and replaced by cells differentiating from cycling epidermal stem cells in the basal layer that undergo cell division perpendicular or oblique to the underlaying basement membrane, rather than parallel as in self‐renewing divisions. These cells then stop dividing and move through successive layers and corresponding stages of differentiation (spinous layer and granular layer) before becoming cellular ghosts consisting of proteinaceous material enveloping insoluble keratin filaments and being embedded in cellular lipids. After wounding, more basal epidermal progenitors reenter the cell cycle and subsequently migrate to close the defect. Furthermore, mostly quiescent hair follicle stem cells, that in steady state are periodically activated during the hair cycle, may participate in wound healing as well 8, 12. In the intact skin, a calcium gradient exists with low concentrations in the basal layers, increasing concentrations in spinous and granular layers, and lower concentrations again in the cornified layer 89, 90, 91, 92 (Fig 2B). The contribution of intracellular and extracellular calcium to these measurements, which were performed on fixed samples and measured total and not only free calcium, is still unclear. More recent data, using two‐photon fluorescence lifetime imaging in unfixed tissues, suggest that intracellular calcium stores are the prime locale of this gradient, but that cellular calcium in the basal layer appears heterogeneous 93 (Fig 2B). This heterogeneity may be linked to heterogeneous composition of this layer in terms of proliferative potential and clone size of epidermal stem and progenitor cells 8, 12. This study also showed that the extracellular space in the skin is very small. A drawback, however, was that the dissociation constant of the dye used may have been below the range of the extracellular calcium concentration gradient, which would therefore have gone undetected.

There is a substantial amount of indirect evidence linking extracellular calcium concentration to skin homeostasis, however. The calcium gradient is rapidly dissipated in injured and diseased skin 94, 95, suggesting a role for the skin barrier in maintaining this calcium gradient and a role for calcium in maintaining and restoring skin barrier function 90. Furthermore, extracellular calcium regulates in vitro growth and differentiation of basal keratinocytes (Fig 2C). Culture protocols for keratinocyte expansion call for low (~0.05‐0.1 mM) extracellular calcium, whereas increasing extracellular calcium induces differentiation and expression of differentiation‐associated genes through both transcriptional and post‐transcriptional mechanisms 90 in human 96, 97 and in murine 98, 99 cultured keratinocytes. Furthermore, the differentiation‐inducing effect of calcium requires the CaSR in vitro 62, 100 and in vivo 101, and, as well as SOCE, at least in vitro 102. Several cation channels are expressed in a layer‐specific fashion in the skin 90. One of those, the selective calcium‐entry channel, TRPV6, increases keratinocyte intracellular calcium in high‐calcium conditions 103, thus promoting differentiation, and is itself induced by another calcium‐regulating factor, 1,25‐hydroxy‐vitamin D 104, that is also required for differentiation and acts in concert with CasR in this respect 105. Together, these findings indicate a role for regulated extracellular calcium concentrations in skin homeostasis. On the other hand, several genetic skin diseases are caused by mutation in genes encoding ATPases that pump calcium into intracellular stores, thus illustrating their importance in maintaining skin homeostasis. Mutation of the Golgi calcium ATPase, ATP2C1, in the blistering dermatosis, Hailey–Hailey disease 106, abrogates the calcium gradient, even in unaffected skin 107. Darier's disease or follicular keratosis is a dyskeratotic dermatosis caused by mutation in the SERCA pump, ATP2A2 108, and is associated with ER stress caused by chronic ER calcium depletion. Physiological ER stress, caused by activation of calcium release from the ER in keratinocytes, is associated with differentiation and skin repair, however 109. Altogether, these observations indicate a major role for both intracellular and extracellular calcium regulation in skin homeostasis.

Pluripotent stem cells

PSCs occur in two states that may be the extremes of a developmental continuum 7, 13. The first is the naïve or “ground” state, which has been best documented in mouse (m) ESCs and is still somewhat controversial in human (h) PSCs. The naïve state is characterized by a distinct colony morphology, culture requirements (typically GSK3 and MAPK2 inhibition in the presence of leukemia inhibitory factor (2i/LIF)), genome‐wide expression signature, genome‐wide hypomethylation, and by highly efficient contribution to the germline after blastocyst injection. “Primed” PSCs are, as the name implies, epigenetically primed for the expression of lineage‐specific genes, require FGF2 and activin for expansion, and do not contribute as efficiently to chimeras after blastocyst injection. They are believed to represent post‐implantation epiblast stem cells. hPSCs are most similar to the latter 7, 13.

Some early studies addressed levels and intracellular dynamics of calcium in mESCs. As these studies mostly predate a clear definition of naïve and primed mESCs, they used mESCs grown in serum+LIF, a condition where cellular potentials are heterogeneous and lineage‐specific genes are expressed. The functional machinery for calcium signaling typical of non‐excitable cells (i.e. the absence of ryanodine receptors and voltage‐gated channels) is present in mESCs 110, and signaling mediated by IP3 is required for preimplantation embryos 111, blastocyst formation 112, and mESC proliferation 111, 113. Oscillation of cytoplasmic calcium was observed that depended on release from the ER and refilling of ER stores through SOCE, occurred predominantly during the G1‐to‐S transition, and appeared to promote cell cycle progression 114. Furthermore, the TPC2 channel inhibited neural differentiation of mESC 115. Overall, however, manipulation of these mechanisms only had a limited effect on mESC self‐renewal in these conditions 116.

A critical role for calcium signaling was identified when naïve and primed mESCs were compared in defined culture conditions (Fig 2D). In a CRISPR/Cas9 screen for genes involved in the transition from the naïve to the primed state, MacDougall et al 117 found that one of the limited set of genes that appeared to be required for this transition was Atp2b1. Atp2b1 encodes the plasma membrane calcium efflux pump, PMCA1. During exit from the naïve state, intracellular calcium decreased. In the absence of Atp2b1, exit from naïve pluripotency was impaired, and baseline cytosolic calcium strongly increased. Nevertheless, changes in Atp2b1 expression were not documented. Furthermore, a proportional reduction in intracellular calcium was still observed in Atp2b1 ^−/− cells during exit from naïve pluripotency, suggesting that, while Atp2b1 may be responsible for relatively low‐calcium levels in mESCs overall, this gene likely does not explain the difference in intracellular calcium between the naïve and primed states. The role of other PMCAs was not explored, however. Combining deletion of both Atp2b1 and Tcf7l1, a transcriptional repressor of genes required for naïve pluripotency, allowed for the growth of mESCs in the absence of 2i/LIF for multiple passages with a gene expression signature close, but not identical to that of naïve mESC in 2i/LIF conditions. Combining GSK3 inhibition with calcitonin allowed for the maintenance of pluripotency as evidenced by the potential to generate embryoid bodies containing markers for all three germlayers, albeit that after more prolonged culture in these conditions, primitive streak and definitive endoderm markers were induced. These conditions are therefore unlikely to be useful for the long‐term maintenance of naïve mESCs. Furthermore, the most stringent test of naïve pluripotency, contribution to the germline after blastocyst injection, was not performed. Though intuitively attractive, the role of calcitonin is somewhat curious. Calcitonin is an endocrine hormone produced by parafollicular cells in the thyroid that can acutely reduce serum calcium by increasing the fractional excretion of calcium and phosphate by the kidney and by inhibiting the osteolytic activity of osteoclasts in the bone 118, 119. Its receptor is a GPCR that, similar to most GPCRs, induces IP3 production and release of calcium from the ER 118 resulting in calcium transients. While playing a role in regulating serum calcium, this receptor is not known to specifically regulate baseline intracellular calcium levels. It is therefore possible that the effect of calcitonin on the maintenance of naïve pluripotency could be recapitulated by agonism of other GPCRs. We do note, however, that calcitonin is also produced by the preimplantation endometrium in response to progesterone 120 and that calcitonin stimulates trophoblast growth prior to implantation 121. An effect on the inner cell mass, where ESCs reside, has not been demonstrated. The authors could also identify a role for calcineurin/NFAT signaling, as the calcineurin inhibitor, cyclosporine A (CsA), could partially compensate for the defect observed in Atp2b1 ^−/− cells and facilitate transition to a primed state. However, CsA could not maintain self‐renewal after withdrawal of LIF and the MAPK2 inhibitor 117. This finding contrasts with a previous report showing that calcineurin/NFAT signaling is required for differentiation of mESCs and that inhibiting this signaling pathway promoted maintenance of mESC in the absence of LIF 122. It is possible that differing culture conditions (e.g. the presence of serum in the report of Li et al 122) and the fact that MacDougal et al 117 examined naïve‐to‐primed transition while Li et al 122 examined differentiation after LIF withdrawal could play a role in these discrepancies. Taken together, the role of calcium in mESC biology is still unclear, and no studies have reported a role for calcium in human pluripotency.

Intestinal stem cells

In the fly midgut, absorption of l‐glutamate induced PLC activation mediated by the metabotropic l‐glutamate receptor and resulted in reduced calcium oscillation frequency but increased average cytosolic calcium (Fig 2E). Increased cytosolic calcium then activated Creb‐regulated transcriptional co‐activator (Crtc), a conserved substrate for calcineurin, thus promoting proliferation of intestinal stem cells (ISCs). Not only l‐glutamate feeding, but also a multitude of signaling pathways and injuries resulted in reduced oscillation and elevated cytosolic calcium. These observations indicate calcium as major integrator of signals inducing ISC proliferation 123. It is not known to what extent these findings hold in mammalian ISCs. It is interesting to note, however, that l‐glutamate also promoted proliferation and maintenance of the intestinal epithelium in mice 124. To what extent calcium‐induced ISC proliferation also enhanced differentiation and how this striking integration of signals impinging on cellular calcium dynamics occurs is not yet clear. It has been shown, however, that the mechanoreceptor, Piezo, expressed in a subset of unipotent precursors of enteroendocrine cells, promotes enteroendocrine differentiation by increasing intracellular calcium through ER calcium release and SOCE, leading to Notch inhibition 125.

Conclusion

Calcium appears to play a major and previously underappreciated role in the regulation of stem cell fate and maintenance. The preponderance of the evidence, at least in skin and in the hematopoietic system as well as in primed mESCs (and therefore likely also in hPSCs), suggests that low baseline calcium is associated with the stem cell state. Furthermore, in skin keratinocytes and in HSCs, a low‐calcium environment promotes maintenance and prevents differentiation.

Mechanisms of action of calcium in stem cells

Calcium is involved in a multitude of generic signaling pathways downstream of cytokines and morphogens and can affect an enormous range of cellular processes, including transcriptional activation of NFAT, all of which may be relevant to stem cells. However, many unique roles of calcium in stem cells may be explained by its effects on two closely linked, interacting processes that are regulated in a highly specific fashion in various types of stem cells and play a major role in maintaining stem cell identify and function: the epigenome and metabolism.

Calcium and epigenetic regulation

Cellular identity is determined and maintained by the epigenome, a set of covalent modifications comprising DNA methylation and hydroxymethylation 126, and now more than a dozen modifications of histones 127. These DNA and histone modifications determine accessibility to transcription factors and, therefore, gene expression. When stem cells differentiate, the chromatin landscape shifts dramatically: Genes involved in the maintenance of stem cell identity are repressed, and lineage‐specific genes are activated. Some of these lineage‐specific genes are endowed with both activating and inhibitory epigenetic marks, a configuration termed “poised”, in the stem cell state 128, 129.

Calcium regulates the activity of calpains, a family of 14 homologous proteolytic enzymes involved in the post‐translational regulation of multiple proteins 130. Calpains have been shown to target, among others, Ten‐eleven translocated (TET) enzymes in mESCs 131. The three TET homologues convert 5‐methylcytosine (5mC) to 5‐hydroxymethyl cytosine (5hmC) 132, 133. 5hmC is an epigenetic mark and a step in the processive demethylation of 5mC 132, 133, 134. TET enzymes play important roles in early development and differentiation of mPSCs 135. TET2 plays a major role in HSCs by restraining expression of epigenetically poised lineage‐specific genes 134. Tet2 ^−/− HSCs show progressive myeloid bias in vivo that ultimately leads to myeloid malignancy 136, 137, 138, 139, 140, 141. Heterozygous TET2 mutations are drivers in a large fraction of myeloid malignancies 142, 143, 144, 145 and are frequently encountered in clonal hematopoiesis of undermined significance, a condition where the contribution of one HSC clone to the blood increases with age and that is associated with elevated incidence of hematological malignancies 146, 147. Together, these findings suggest that TET2 may be an epigenetic master tumor suppressor in HSCs. TET2 is also required for the beneficial effect of low‐calcium conditions on HSC maintenance in vitro 64 (Fig 3). More specifically, calpain inhibition by small molecules or by overexpression of the ubiquitously expressed calpain inhibitor, calpastatin, mimicked the effect of a low‐calcium culture on HSC maintenance in vitro 64. Furthermore, low‐calcium culture increased TET protein expression and 5hmC deposition in the nucleus. Consistent with these findings, Tet2 ^−/− HSCs were impervious to the enhancing effect of low calcium on HSC maintenance. Interestingly, calcium may also regulate other tumor suppressors. There is, for example, evidence that p53 is also a target of calpains 148. Tumor suppressors in general have been suggested to be essential for steady‐state HSC maintenance 149. These findings therefore raise the hypothesis that BM might, by virtue of its low‐calcium milieu, provide a tumor suppressor environment for HSCs.

Figure 3. Calcium regulation of DNA and histone modifications.

Schematic representation of potential direct regulation of covalent histone and DNA modifications by calcium. 5hmC: 5 hydroxymethyl cytosine; 5mC: 5‐methyl cytosine; CaMKII: calcium/calmodulin‐dependent kinase 2; HDAC4: histone deacetylase 4; MeCP2: methyl CpG‐binding protein 2; TET: ten‐eleven translocated.

Although a direct role of calcium on chromatin remodeling in stem cells has not been demonstrated yet, it is likely that the aforementioned role of calcium in the transition from naïve‐to‐primed mESCs 117 directly or indirectly involves epigenetic modifications, as the naïve and primed PSC epigenome and transcriptome differ 7, 13. In other cell types, calcium is known to affect histone modifications. The most studied type of post‐translational histone modification, acetylation on lysine residues, which opens chromatin, is regulated by the balanced action of several classes of histone acetyl transferases (HATs) and deacetylases (HDACs). Calcium signaling through CaMKs mediates export of several class IIa histone deacetylases from the nucleus 150. In muscle progenitors, HDAC4 negatively regulates transcription by myocyte enhancer factor 2C (MEF2C). Calcium‐activated CaMKII phosphorylates HDAC4, which forms complexes with another class IIa HDAC, HDAC5 151, promoting transport of HDAC4 to the cytoplasm (Fig 3), thus allowing MEF2C to induce muscle differentiation. A similar mechanism plays a role in cardiac hypertrophy 152, 153. HDAC4 is induced upon differentiation of mESCs and is repressed by the pluripotency factor, POU5F1, potentially suggesting a negative regulatory role of HDAC4 in the maintenance of pluripotency 154.

Further evidence for a role of calcium in epigenetic regulation comes from a screen for small molecules that derepress epigenetically silenced tumor suppressor genes in a colon cancer cell line 155. Of the 12 hits observed, 3 were known DNA methylation or HDAC inhibitors, 5 were cardiac glycosides, and 3 were other compounds, including AsO₃, which is used in the treatment of acute promyelocytic leukemia. All, except for the known epigenetic modifiers, affected cellular calcium handling and increased the activity of CaMKII, leading to nuclear export of methyl CpG‐binding protein, MeCP2 (Fig 3). This resulted in derepression of silenced loci without epigenetically affecting those. The underlying mechanism is unclear, however, as cardiac glycosides reduced SOCE and ER calcium without affecting cytosolic calcium, suggesting that this effect may be mediated by (undetected) cytosolic calcium sparks. Other compounds, on the other hand, caused sustained increases in cytosolic calcium. Altogether, the data suggest that CaMKII, and therefore calcium, plays a role in epigenetic regulation by affecting the binding of epigenetic readers to CpG‐methylated DNA, a finding that may have relevance to epigenetic regulation of stem cells as well.

Calcium and metabolism in stem cells

Multiple metabolites act as substrates for or as competitive inhibitors of enzymatic reactions involved in covalent epigenetic modifications. Metabolism shapes the epigenome 156, 157, 158, 159 and therefore drives cell fate and plays a major role in the establishment and maintenance of pluripotency 15, 17. Through its profound effects on metabolism, calcium likely also indirectly affects the epigenome (Fig 4). Furthermore, metabolism may in turn regulate intracellular calcium (Fig 5).

Figure 4. Calcium regulation of cellular metabolism and its repercussion on epigenetic modifications.

Effect of calcium on mitochondrial function and on mitochondrial metabolites involved in chromatin remodeling. Enzymes activated by calcium are marked in red. DHA: dihydroxyacetone phosphate; ETC: electron transport chain; G3P: glycerol‐3‐phosphate; GPD2: glycerol‐3‐phosphate dehydrogenase 2; IDH: isocitrate dehydrogenase; KGD: α‐ketoglutarate dehydrogenase; OAA: oxaloacetic acid; PDH: pyruvate dehydrogenase complex; PDP1: pyruvate dehydrogenase phosphatase 1.

Figure 5. Interplay between calcium, glycolysis and TCA.

Interplay between regulation of intracellular calcium by glycolysis and its effect on the TCA and chromatin remodeling. ETC: electron transport chain; OAA: oxaloacetic acid; PMCA: plasma membrane calcium efflux pump; TCA: tricarboxylic acid cycle.

Stem cell metabolism

Most stem cells where metabolism could be assessed rely on glycolysis for ATP production to a larger extent than progenitors and mature cells 14, 160, 161, 162, 163, 164. This suggests that mitochondrial respiration is more dispensable for stem cells than it is for progenitors and mature cells, a notion supported by some experimental data in HSCs 165, 166. A similar phenomenon exists in tumor cells, where aerobic glycolysis and the pentose phosphate shunt provide anabolic building blocks to support cell proliferation and thus an increase in biomass. This effect, also called the Warburg effect 167, is unlikely to play a role in HSCs, as these are quiescent. Furthermore, both naïve and primed mESCs proliferate very rapidly yet show differential reliance on glycolysis 17, 168: Naïve mESCs are more oxidative despite having a more immature mitochondrial morphology, whereas primed mESCs are glycolytic, but possess more mature mitochondrial morphology 168. In addition, at least HSCs in fact have high mitochondrial mass 169, whereas within the HSC population, cells with high mitochondrial mass are more quiescent and possess higher repopulation capacity after transplantation 170. Taken together, the purpose of glycolytic ATP production and attenuated mitochondrial respiration in stem cells is unclear. As detailed below, however, calcium is regulated by glycolysis and in turn regulates mitochondrial activity, suggesting a link between calcium and stem cell metabolism.

Calcium and glycolysis

A prevailing idea in the field is that glycolysis and reduced mitochondrial respiration in quiescent HSCs are driven by hypoxia. HSCs are maintained more efficiently in hypoxic than in normoxic conditions in vitro, likely through reduced generation of reactive oxygen species (ROS) in hypoxic conditions 171, 172. Using oxygen‐sensitive probes and two‐photon live imaging, it was shown that BM is hypoxic particularly near vascular niches, where most HSCs reside 45, but that hypoxia may be caused by oxygen consumption in regions of high cellularity 173. HSCs showed enhanced staining in situ for the hypoxia marker, pimonidazole 174, 175, 176, 177, 178, but did so irrespective of their location, while many neighboring cells did not stain 179. Pimonidazole forms adducts with cellular constituents after reacting with electrons arising from the respiratory chain that do not find an oxygen acceptor. Pimonidazole staining therefore detects cells with low oxidative phosphorylation and not necessarily only hypoxic cells 179. Indeed, it was recently documented that HSCs do not reside in the areas with the deepest hypoxia 67. Thus, it is difficult to understand why HSCs stain relatively selectively for pimonidazole. Furthermore, even in normoxic conditions in vitro, glycolytic flux is increased and mitochondrial respiration and ATP production are reduced in HSCs compared with progenitors 164, 169. Altogether, these observations may suggest that glycolysis and attenuated respiration in HSC are not only a response to hypoxia, but may be to some extent hardwired.

This notion is further supported by studies on the role of hypoxia‐induced factor (HIF) in HSCs. Anaerobic glycolysis is driven by dimers of HIF1α or HIF2α and HIF1β. HIF1α and HIF2α are targeted for degradation through prolyl hydroxylation by oxygen‐sensitive dioxygenases, PHD enzymes 180. HSCs in mice with inducible deletion of HIF1α or with deletion of pyruvate dehydrogenase kinase (PDK) 2 and PDK4, both HIF targets that inhibit entry of pyruvate into the mitochondrial tricarboxylic acid (TCA) cycle thus enhancing glycolysis at the expense of respiration, were reported to lose quiescence and display defects after transplantation 161, 175. However, it was subsequently shown that deletion of HIF1α and HIF2α individually 181, 182, of HIF1β or of both HIF1α and HIF2α 183 resulted in no defects or in a subtle loss of HSC function and minimal changes in the expression of glycolytic enzymes 183. Furthermore, although, similar to Pdk2^−/−Pdk4^−/− mice, HSC function is impaired in Pdk1 ^−/− mice, these authors observed that conditional deletion of HIF1α had no effect on the expression any PDK isoforms in vivo 184. HIF1α may also not only be directly stabilized by hypoxia (except in very severe hypoxia), but also, and perhaps primarily, by mitochondrial ROS arising through as yet unclear mechanisms during hypoxia 185, 186, 187, 188. This conflicts with the notion that mitochondrial ROS are low in HSCs and are detrimental to their maintenance 189. It is therefore possible that glycolysis, while undoubtedly important for HSC maintenance as demonstrated by the consistent phenotypes of Pdk‐deleted mice 161, 184, is not primarily driven by a HIF‐mediated response to hypoxia.

A possible reason for the apparent glycolytic hardwiring of HSCs is that glycolytic ATP production is required to reduce intracellular calcium as PMCA pumps are fueled by glycolysis 190, 191, 192, 193, possibly because the glycolytic machinery is located near the plasma membrane, at least in erythrocytes 193. Indeed, inhibition of glycolysis increased intracellular calcium in HSCs 64. Furthermore, two ubiquitously expressed “housekeeping” PMCA pumps, PMCA1 and PMCA4, are expressed more highly in HSCs compared with progenitors. Pharmacological inhibition of these pumps increased cytoplasmic calcium 64, indicating their involvement in the reduced intracellular concentration in HSCs. Similarly, the transition of naïve‐to‐primed mESCs, which are more glycolytic 168, was promoted by expression of PMCA1 117. It is therefore possible that glycolysis is required for maintaining reduced intracellular calcium.

Calcium and TCA cycle

Mitochondria affect the epigenome through intermediates of the TCA cycle 157, 158, 159, 194. The importance of the TCA for epigenetic regulation is, for example, illustrated by the fact that the TCA cycle, but not the electron transport chain (ETC.) activity, regulates histone acetylation 194. Furthermore, conditional deletion of the Riske iron‐sulfur protein (RISP), a component of complex III of the electron transport chain, in the hematopoietic system severely impaired HSC function. RISP deletion, likely by increasing the NADH/NAD⁺ ratio, increased TCA metabolites that play a role in the epigenome, including α‐ketoglutarate, fumarate, succinate, and L(S)‐2‐hydroxyglurate, and was associated with histone hypermethylation and hypoacetylation and DNA hypermethylation in hematopoietic cells 195. At least in HSCs, therefore, attenuated production of these metabolites appears beneficial for their maintenance.

Calcium enhances mitochondrial activity to match ATP production to cellular demands. Furthermore, constitutive release of calcium from the ER to the mitochondria is required to maintain mitochondrial respiration in chicken DT40 B cells. In its absence, a bioenergetic crisis ensues that, through activation of AMPK and subsequent inhibition mTOR, leads to autophagy 196. This finding is consistent not only with low mitochondrial activity, but also with the relatively active autophagy in HSCs, which is required for their maintenance and function 19, 197. Similar to other cells, increased intracellular calcium also enhances respiration in HSCs. When cycling of HSCs was induced by the administration of 5‐fluorouracil, which eliminates cycling progenitors and forces HSCs out of quiescence to replace these, both cytoplasmic calcium and mitochondrial membrane potential increased 69. Furthermore, culture of HSCs in vitro in the presence of high‐calcium concentrations increased mitochondrial respiration and compromised their functional maintenance, while inducing increased expression of differentiation markers 64. Increased respiration has also been shown to be essential for exit from the stem cell state and differentiation of HSCs in vivo 166.

The mechanistic basis of the calcium‐mediated enhancement of mitochondrial respiration, which in most stem cells is associated with exit from the stem cell state, is most likely the well‐established fact that the activity of two enzymes of the TCA cycle, α‐ketoglutarate dehydrogenase (KDH) 198 and isocitrate dehydrogenase (IDH) 199, and of pyruvate dehydrogenase phosphatase I (PDP1) 200, 201 is increased by calcium 23 (Fig 4). It is compelling to note that most or all of the metabolites that were present in excess in RISP‐deleted mice 195 are generated or affected by these enzymes. PDP1 activates the pyruvate dehydrogenase complex. Activation of PDP1 (but not of the differentially expressed homologue PDP2) by calcium therefore promotes flux of pyruvate into the TCA cycle and leads to increased formation of acetylCoA, a co‐factor in acetylation reactions, from pyruvate. After condensation with oxaloacetate to form citrate, acetylCoa can be regenerated in the cytoplasm by ATP‐citrate lyase 202. In hPSCs, acetylCoA, generated from glycolytically derived pyruvate, delays exit from pluripotency 203. Mitochondrial IDH performs oxidative decarboxylation of isocitrate to α‐ketoglutarate, a co‐factor for dioxygenases such as TET enzymes and JMJC proteins involved in DNA and histone demethylation, respectively. KDH in turn converts α‐ketoglutarate to succinyl‐CoA. Succinate, and its downstream derivative in the TCA cycle, fumarate, both inhibit dioxygenases as competitive inhibitors of α‐ketoglutarate 204, whereas succinylation modifies histones 158. Interestingly, recent evidence indicates that some metabolic enzymes can be recruited to the nucleus, where they locally generate metabolites involved in chromatic remodeling 159. KDH is one of these enzymes, and nuclear KGH has been reported to be involved in histone succinylation 205. The role of nuclear calcium would be of major interest, but has not been addressed yet in stem cells. In addition to generating metabolites involved in epigenetic modifications, these three critical calcium‐regulated dehydrogenase complexes also generate reducing equivalents in the form of NADH. Calcium therefore is expected to reduce the mitochondrial NAD⁺/NADH ratio. Interestingly, calcium also increases the activity of FAD‐glycerol phosphate dehydrogenase (GPD2), which is involved in transferring reducing equivalents from glycolysis‐generated NADH to the mitochondria as FADH2 206, 207, regenerating cytoplasmic NAD⁺ (Fig 4). NAD⁺ in turn affects chromatin structure, as a class of HDACs, the sirtuins, and the histone poly(ADP)‐ribosylating enzyme, poly(ADP)ribose polymerase, are positively regulated by their co‐factor NAD⁺ 208.

Taken together, it appears very plausible that differential regulation of intracellular calcium and of calcium in the extracellular niche between stem cells and their progeny plays a major role in the maintenance or the dismantling of the stem cell epigenome and therefore in stem cell function, by tuning or balancing the concentrations of TCA‐generated metabolites that regulate the deposition of covalent modifications on DNA and DNA‐associated proteins.

The interplay between calcium, glycolysis and the TCA

In HSCs 64 and primed mESCs 117, intracellular calcium is reduced through the action of PMCA pumps. These are driven primarily by glycolysis‐derived ATP 190, 191, 192, 193. Furthermore, in many stem cells, mitochondrial activity is attenuated, a phenomenon that in HSCs has been shown to be associated with reduced intracellular calcium 64, 69. It is therefore plausible that calcium regulation contributes to the metabolic wiring of stem cells. More specifically, in HSCs glycolysis reduces intracellular calcium, which in turn attenuates mitochondrial activity 64, most likely by reducing the activity of the aforementioned dehydrogenases involved in the generation of substrates and/or of inhibitors of epigenetic modification (Fig 5). In HSCs, reduced calcium also stabilizes TET enzymes by inhibiting calpains 64, further reinforcing the notion that calcium regulates the epigenome. While these principles may be differentially operative in distinct types of stem cells, similar mechanisms could play a role in ESCs. In mESCs, calpains have been shown to cleave TET enzymes 131. Furthermore, naïve mESCs display higher intracellular calcium compared with primed mESCs, where intracellular calcium is reduced, although the purported role of PMCA1 in this decrease has not been convincingly demonstrated 117. Naïve mESC are also more oxidative, whereas primed mESCs are more glycolytic 168. Together, these observations might be consistent with repressed respiration in the presence of reduced cytoplasmic calcium concentrations regulated by glycolysis‐driven PMCA pumps in primed mESCs, a situation that would be remarkably similar to that observed in HSCs, but remains to be experimentally verified. Taken together, it is possible and even likely that the metabolism in stem cells, at least in part, regulates and is in turn regulated by intracellular calcium to shape the stem cell epigenome.

Box 1: In need of answers.

1. What are the exact calcium concentrations and dynamics and their functional significances in various stem cells in situ?

2. Stem cell‐specific genetically encoded calcium reporters need to be developed for in vivo imaging.

3. What are the calcium concentrations and how are they regulated in the stem cell microenvironment or niche?

4. What are the further downstream stem cell‐specific mechanisms regulated by calcium?

5. We need strategies to harness newly acquired insights into the calcium physiology of stem cells for applications in regenerative medicine.

Future prospects

Recent evidence begins to integrate calcium physiology and signaling into an array of mechanisms that determine stem cell identity, lineage potential, maintenance, and self‐renewal and suggests that calcium regulation, metabolism, and the transcriptional and epigenetic landscape of stem cells are closely interconnected. This area of research is still in its infancy, however. A detailed, granular understanding of how exactly metabolism regulates the epigenome is lacking, and insights into how calcium directly or indirectly through its metabolic effects regulates the epigenome are just beginning to emerge. Furthermore, we still do not know the exact calcium concentrations and dynamics of calcium oscillations as well as their significance in the cytoplasm and in intracellular calcium stores of stem cells in situ. Rather than measuring intracellular calcium using fluorescent dyes ex vivo, transgenic expression of genetically encoded calcium reporters targeted to various subcellular compartments 209 would be preferable for the determination of calcium ex vivo and, by intravital imaging, in situ, as was reported in the fly intestine 123. This will allow to not only asses baseline calcium, but also to examine calcium events, which may change coordinately with extracellular calcium and with stem cell function. Given the evidence of tissue‐specific calcium gradients in skin and bone, which affect stem and progenitor cell function, it would also be of major interest to determine extracellular calcium concentrations in stem cell niches, understand how calcium is regulated in the various stem cell niches, and how extracellular niche calcium impacts stem cell function. In the bone, for example, it is likely that proximity to either blood vessels or trabecular bone and the degree of ongoing remodeling of trabecular bone will determine local calcium concentration and affect HSC function. Despite the technical hurdles, such studies would be of major interest to the field.

Conflict of interest

The authors declare that they have no conflict of interest.

EMBO Reports (2020) 21: e50028