TRPM7 channels mediate spontaneous Ca2+ fluctuations in growth plate chondrocytes that promote bone development

See allHide authors and affiliations

Sci. Signal.  09 Apr 2019:
Vol. 12, Issue 576, eaaw4847
DOI: 10.1126/scisignal.aaw4847

Ca2+ fluctuations promote growth plate chondrogenesis

In the growth plates of long bones, ordered arrays of chondrocytes proliferate, mature, secrete cartilage matrix, and then undergo apoptosis. The cartilage matrix is subsequently replaced with trabecular bone. Because Ca2+ signaling is implicated in chondrogenesis in vitro, Qian et al. performed Ca2+ imaging of live chondrocytes in slices of embryonic mouse femurs. Growth plate chondrocytes generated spontaneous Ca2+ fluctuations that depended on the cation channel TRPM7. Experiments in bone slices and in cultured metatarsal bones in which Trpm7 was conditionally knocked out indicated that Trpm7-mediated Ca2+ fluctuations were required for the proper development of chondrocytes and bone outgrowth. Furthermore, chondrocyte-specific knockout of Trpm7 in mice caused defects in chondrocyte Ca2+ fluctuations, growth plate morphology, and bone outgrowth.


During endochondral ossification of long bones, the proliferation and differentiation of chondrocytes cause them to be arranged into layered structures constituting the epiphyseal growth plate, where they secrete the cartilage matrix that is subsequently converted into trabecular bone. Ca2+ signaling has been implicated in chondrogenesis in vitro. Through fluorometric imaging of bone slices from embryonic mice, we demonstrated that live growth plate chondrocytes generated small, cell-autonomous Ca2+ fluctuations that were associated with weak and intermittent Ca2+ influx. Several genes encoding Ca2+-permeable channels were expressed in growth plate chondrocytes, but only pharmacological inhibitors of transient receptor potential cation channel subfamily M member 7 (TRPM7) reduced the spontaneous Ca2+ fluctuations. The TRPM7-mediated Ca2+ influx was likely activated downstream of basal phospholipase C activity and was potentiated upon cell hyperpolarization induced by big-conductance Ca2+-dependent K+ channels. Bones from embryos in which Trpm7 was conditionally knocked out during ex vivo culture exhibited reduced outgrowth and displayed histological abnormalities accompanied by insufficient autophosphorylation of Ca2+/calmodulin-dependent protein kinase II (CaMKII) in the growth plate. The link between TRPM7-mediated Ca2+ fluctuations and CaMKII-dependent chondrogenesis was further supported by experiments with chondrocyte-specific Trpm7 knockout mice. Thus, growth plate chondrocytes generate spontaneous, TRPM7-mediated Ca2+ fluctuations that promote self-maturation and bone development.


Long bones develop through the process of endochondral ossification, in which cartilage is formed at both ends and then replaced by trabecular bone (1, 2). This growth plate cartilage is vascularized and subsequently replaced by an epiphyseal line after bone growth is complete, making it distinct from articular cartilage, which is not vascularized and persists in joints throughout postnatal life. During early embryogenesis, chondrogenic progenitor cells derived from the neural crest, somites, and somatopleures propagate and form precartilage condensations as primitive bone rudiments. The progenitor cells promptly differentiate into chondroblasts and organize a perichondrial layer at the periphery of the condensation. The chondroblasts underneath the perichondrium undergo differentiation to produce cartilage matrix components and gradually organize morphologically distinct zones, each of which contains homogeneous chondrocytes at the same stage of maturation. Round-shaped chondrocytes propagate in a growth plate at each of the rounded ends of developing bones (epiphyses) and produce the matrix proteins aggrecan and type II collagen. The round chondrocytes then morphologically change into flat chondrocytes that vigorously proliferate and are arranged in characteristic columnar arrays. The columnar chondrocytes subsequently stop proliferating and differentiate into hypertrophic chondrocytes exclusively producing type X collagen. The hypertrophic chondrocytes lastly become bloated and undergo apoptotic cell death while they are gradually replaced by trabecular bone, which forms the diaphysis, the shaft of a long bone, through the action of osteoblasts and osteoclasts. Remarkable progress has been made in understanding the molecular basis of the transcriptional networks and Wnt and Hedgehog signaling pathways that control chondrogenesis in growth plates (3). However, ion homeostasis, maintained by a defined set of cell surface channels, pumps, and transporters, is largely uncharacterized during chondrogenic maturation.

Ion channels play fundamental roles in cellular homeostasis, and alteration of ionic currents is associated with functional specialization during cellular differentiation and maturation. In particular, ion channels contributing to Ca2+ influx across the plasma membrane and Ca2+ release from intracellular stores constitute divergent Ca2+-handling machineries that control specific cellular responses (4). The members of transient receptor potential (TRP) channel superfamily are a group of cell surface cation channels that are classified into six subfamilies: TRPC, TRPV, TRPM, TRPA, TRPML, and TRPP (5, 6). Of eight members in TRPM subfamily, TRPM7 forms a divalent cation-permeable channel, and its gating is stimulated and suppressed by various factors, such as intracellular Mg2+, Mg2+-nucleotide complexes, extracellular pH, and extracellular oxygenation (5, 6). Trpm7 is moderately expressed in heart, bone, and adipose tissues, but it is weakly expressed ubiquitously (7). TRPM7 channels participate in important cellular processes, including cell growth and adhesion, and Trpm7 knockout mice die before embryonic day 7.5 (E7.5) (8). Moreover, irregular endochondral ossification has been reported in zebrafish bearing a Trpm7 mutation, although whether this mutation causes a loss or gain of TRPM7 function is not clear (9).

The functions of Ca2+-permeable channels have been systematically examined in cultured chondrocyte lines (1012), and Lu et al. (13) have proposed that TRPM7 channels participate in the maturation of chondrogenic mouse ATDC5 cultured cells. Therefore, TRPM7 channels may play an important role in chondrocytes, but TRPM7-mediated Ca2+ influx has not been examined in native chondrocytes in developing bones. We developed a method for Ca2+ imaging in slices of living bones harvested from mouse embryos and found that resting intracellular Ca2+ concentration ([Ca2+]i) fluctuated in native growth plate chondrocytes. We provide evidence that the spontaneous Ca2+ fluctuations were predominantly generated by TRPM7 channels and were critical for proper chondrocyte maturation and bone development. Knocking out Trpm7 in embryonic bones cultured ex vivo or specifically in chondrocytes in vivo impaired chondrocyte maturation and bone growth.


Ca2+ fluctuations in growth plate chondrocytes

Ca2+ signaling and gene expression profiles for Ca2+-handling proteins in several cultured chondrocyte lines have been reported (1012), but Ca2+ signaling has not been reported in native chondrocytes freshly prepared from developing bones. To better understand the Ca2+-handling features of growth plate chondrocytes, we isolated femoral bones from perinatal mice at E17.5 and prepared bone slices that were used for Ca2+ imaging analysis in physiological saline. In longitudinal bone slices loaded with the Ca2+ indicators Fluo-4 and Fura-2, fluorescence microscopy distinguished round, columnar, and hypertrophic chondrocytes with characteristic morphological features, which were clustered in their proper zones of the epiphyseal cartilage tissue (Fig. 1A). Therefore, the slice preparations provide an ideal system for imaging studies of living growth plate chondrocytes in conditions closely approximating their native context.

Fig. 1 Growth plate chondrocytes exhibit spontaneous Ca2+fluctuations.

(A) Representative fluorescence image of a Fluo-4–loaded femoral growth plate slice prepared from an E17.5 mouse. The zones of round (R), columnar (C), and hypertrophic (H) chondrocytes are indicated. Scale bar, 100 μm. (B) Fura-2 imaging of round and columnar chondrocytes under basal conditions. Representative images of recorded cells and Fura-2 ratiometric recordings derived from individual Ca2+ fluctuation–positive cells (red and green traces) and Ca2+ fluctuation–negative cells (blue and black traces) are shown. Data are representative of ≥19 cells from ≥5 mice. Scale bar, 20 μm. (C) Quantification of the fraction of round and columnar chondrocytes exhibiting Ca2+ fluctuations and the resting [Ca2+]i in fluctuation-negative and fluctuation-positive round and columnar chondrocytes. Ca2+ fluctuation–positive cells were defined as growth plate chondrocytes exhibiting at least one spontaneous event (>0.025 in Fura-2 ratio) during a 270-s observation window under basal conditions. Statistical analysis of resting [Ca2+]i revealed no difference between fluctuation-positive and fluctuation-negative populations in either round or columnar chondrocytes (t test). (D) Quantification of the amplitude, duration, and frequency of spontaneous Ca2+ events in fluctuation-positive round and columnar chondrocytes. The fura-2 ratio (F340/F380) was measured and calibrated for estimating [Ca2+]i as described in Materials and Methods. Data represent means ± SD. The numbers of cells and mice examined are shown in parentheses in the keys and above the graph bars, respectively.

In ratiometric Fura-2 imaging of round and columnar chondrocytes, the resting [Ca2+]i frequently exhibited weak increases and decreases in a physiological bathing solution, and these Ca2+ fluctuations occurred autonomously and randomly under basal conditions (Fig. 1B). During the 270-s recording period, the Ca2+ fluctuations were detected in approximately 20% of round chondrocytes and 10% of columnar chondrocytes (Fig. 1C). The cells undergoing Ca2+ fluctuations and the cells not undergoing Ca2+ fluctuations exhibited similar resting [Ca2+]i (Fig. 1C) and were morphologically indistinguishable from one another and properly positioned in their appropriate zones of the growth plate. The repeated Ca2+ fluctuations were unsteady and nonperiodic events, characterized by small peak amplitudes ranging from 25 to 45 nM in estimated [Ca2+]i and for durations ranging from 40 to 80 s (Fig. 1D). On the other hand, similar Ca2+ fluctuations were occasionally detected in hypertrophic chondrocytes, but the fluctuation-positive cell population was too small to characterize their spontaneous Ca2+ events (fig. S1A). Moreover, we did not detect spontaneous Ca2+ events in the perichondrium, a connective tissue that contains fibroblasts and chondrocyte precursor cells (fig. S1A).

These observations suggest that proliferating chondrocytes in the growth plate are potentially capable of generating spontaneous Ca2+ fluctuations (Fig. 1). However, this property may gradually disappear in hypertrophic chondrocytes over the course of their maturation as they swell and eventually undergo apoptotic cell death (fig. S1A). Several research groups have reported spontaneous and oscillatory Ca2+ events in chondrocytes prepared from articular cartilage (1418), and they have also proposed that purinergic receptors (15) and TRPV4 channels (18) are involved in these Ca2+ events. Spontaneous Ca2+ events in growth plate chondrocytes, such as those we describe here, have not been previously reported.

Ca2+ fluctuations and Ca2+ entry

Either Ca2+ entry through plasma membrane channels or Ca2+ release from intracellular stores could account for the spontaneous Ca2+ fluctuations observed in growth plate chondrocytes. To evaluate the contribution of Ca2+ influx to the Ca2+ fluctuations, we performed Ca2+ imaging in bathing solutions that either lacked Ca2+ or included gadolinium ions (Gd3+), which nonspecifically inhibit Ca2+ channels. Both treatments essentially abolished the spontaneous Ca2+ events in round chondrocytes (Fig. 2A).

Fig. 2 Ca2+influx is essential for Ca2+fluctuations in chondrocytes.

(A) Fura-2 imaging of round chondrocytes in Ca2+-free bathing solution or in the presence of the calcium channel blocker Gd3+. Representative traces from individual fluctuation-positive cells are shown in the recordings, and the observations are summarized in the bar graph. (B) Fura-2 imaging of round chondrocytes in the presence of TG or CPA. Representative traces recorded from individual fluctuation-positive (red and green) and fluctuation-negative (black) cells are shown, and the observations are summarized in the bar graphs. Data represent means ± SEM, and the numbers of cells and mice examined are shown in parentheses in the keys and above the graph bars, respectively. Significant differences between values before and after the treatments are marked with asterisks [*P < 0.05 and **P < 0.01 in one-way analysis of variance (ANOVA) and Dunnett’s test]. DMSO, dimethyl sulfoxide.

To examine the contribution of Ca2+ release from intracellular stores, we applied the sarco/endoplasmic reticulum Ca2+–adenosine triphosphatase (SERCA) inhibitors, thapsigargin (TG) and cyclopiazonic acid (CPA). In general, both inhibitors evoke Ca2+ leak responses and deplete intracellular Ca2+ stores, leading to sustained [Ca2+]i increases through activated store-operated Ca2+ entry (SOCE) mediated by Orai channels (19). In our measurements, the Ca2+ fluctuations were increased rather than depressed under both TG- and CPA-treated conditions, although their inhibitory effects were confirmed by initial Ca2+ leak responses and subsequent [Ca2+]i increases (Fig. 2B). In response to the SERCA inhibitor treatments, fluctuation-negative cells often converted to fluctuation-positive cells, and the fluctuation amplitudes were continually enhanced. The observations suggest that most chondrocytes are potentially capable of generating Ca2+ fluctuations. Moreover, Ca2+ influx seemed to primarily evoke the Ca2+ fluctuations, whereas it is unlikely that Ca2+ release was directly responsible for their generations.

Gene expression analysis in growth plate

We next set out to identify cell surface Ca2+ channels responsible for the Ca2+ fluctuations. Cultured chondrocytes contain several Ca2+-permeable channels, including purinergic P2X and voltage-gated Ca2+ channels (1012). However, inhibitors of these types of channels did not affect the Ca2+ fluctuations (fig. S1B), suggesting that the spontaneous Ca2+ events were mediated by channels not previously characterized in growth plate chondrocytes. Therefore, we sought to identify transcripts encoding other Ca2+ channels in chondrocytes.

To define the gene expression profile in growth plate chondrocytes, we isolated the terminal region containing round chondrocytes and the adjacent region that is enriched with columnar and hypertrophic chondrocytes from the femoral epiphysis and subjected to total RNA preparations for microarray analysis. The profiling data thus obtained roughly reflected accurate gene expression of the chondrocyte subsets (fig. S2A). For example, the round chondrocyte marker Pthlh and the hypertrophic chondrocyte marker Col10a1 were preferentially detected in the corresponding preparations. We surveyed the transcriptome for cell surface Ca2+-permeable channels and identified several TRP channel family genes (Trpc1, Trpc2, Trpm4, Trpm5, Trpm7, and Trpv4). Reverse transcription polymerase chain reaction (RT-PCR) analysis confirmed the expression of these genes in both of the chondrocyte-enriched preparations (fig. S2B). Inhibitors that target TRPC, TRPM4, TRPM5, or TRPV4 channels exerted no obvious effects on the spontaneous Ca2+ fluctuations (fig. S1B). However, 2-aminoethoxydiphenyl borate (2-APB), which inhibits the inositol trisphosphate receptor (IP3R), calcium release-activated calcium channel protein (Orai), and several TRP family channels, including TRPM7 (6), was the only inhibitor that abolished the Ca2+ fluctuations (fig. S1B). Thus, we focused on TRPM7 channels.

Ca2+ fluctuations and TRPM7 modulators

The small molecules FTY720, an immunosuppressant used to treat multiple sclerosis, and NS8593, originally reported as a blocker for small-conductance Ca2+-activated K+ channels, are widely used TRPM7 inhibitors; both chemicals potently reduce the channel open probability (20, 21). FTY720 and NS8593 attenuated the spontaneous Ca2+ fluctuations at their reported effective doses (Fig. 3A). Both the inhibitors markedly suppressed the generation of the Ca2+ events and reduced their amplitudes in round chondrocytes (Fig. 3B).

Fig. 3 TRPM7 modulators alter Ca2+fluctuations in chondrocytes.

(A) Fura-2 imaging of round chondrocytes in the presence of TRPM7 inhibitors (FTY720 and NS8593) or activators (naltriben and NNC550396). Representative traces recorded from individual round chondrocytes are shown. (B) Quantification of the effects of TRPM7 modulators on the fraction of cells that exhibited Ca2+ fluctuations and the amplitude of the fluctuations. Data represent means ± SEM, and the numbers of cells and mice examined are shown in parentheses in the keys and above the graph bars, respectively. Significant differences between the values before and after the modulator treatments are marked with asterisks (*P < 0.05 and **P < 0.01 in one-way ANOVA and Dunnett’s test).

The δ-opioid receptor antagonist naltriben and the T-type Ca2+ channel inhibitor NNC550396 are TRPM7 activators (22). Both naltriben and NNC550396 greatly facilitated the Ca2+ fluctuations at their effective concentrations in round chondrocytes (Fig. 3A). After the activator treatments, both the populations of Ca2+ fluctuation–positive cells and the amplitudes of the Ca2+ events significantly increased (Fig. 3B). Moreover, FTY720 pretreatment mostly canceled the effects of naltriben (fig. S2C), suggesting that both chemicals precisely target TRPM7 channels.

Ca2+ fluctuations and TRPM7 channels

Trpm7 knockout mice die at an early embryonic stage (8); therefore, Trpm7 loss-of-function studies require tools for conditional gene modification, such as the fusion protein between Cre recombinase and the estrogen receptor (Cre-ER) that renders Cre activity inducible by ER ligands (23). We generated Trpm7 conditional knockout mice, in which the Cre-binding loxP sequences were inserted into the Trpm7 locus flanking exon 22 (floxed allele, Trpm7fl) and the Cre-ER transgene was ubiquitously expressed (fig. S3A). Mice pregnant with the conditional knockout pups were repeatedly injected with tamoxifen, and then the E17.5 homozygous floxed mutants with or without the transgene, Trpm7fl/fl (Cre-ER+/−) and Trpm7fl/fl (Cre-ER−/−) embryos, respectively, were recovered (fig. S3B). RT-PCR analysis of RNAs prepared from femoral epiphyseal specimens showed that Trpm7 was efficiently inactivated in the Cre-ER+/− embryos; the Trpm7 mRNA contents were reduced to 5 to 15% of that in the Cre-ER−/− control embryos (fig. S3C). Therefore, growth plates from Trpm7fl/fl (Cre-ER+/−) embryos were used as a source of Trpm7 knockdown chondrocytes. By Ca2+ imaging, the proportions of cells positive for Ca2+ fluctuations were reduced, and the fluctuation amplitudes were dampened in Trpm7 knockdown chondrocytes compared to chondrocytes prepared from Trpm7fl/fl (Cre-ER−/−) embryos (Fig. 4A). Moreover, Trpm7 knockdown reduced the sensitivity of chondrocytes to naltriben, which increased Ca2+ fluctuations in wild-type and Trpm7fl/fl (Cre-ER−/−) cells but not in Trpm7fl/fl (Cre-ER+/−) cells (Fig. 4B). Trpm7 knockdown did not affect the Ca2+ responses of chondrocytes to adenosine 5′-triphosphate or CPA (fig. S3D).

Fig. 4 Trpm7 knockdown reduces Ca2+fluctuations in chondrocytes.

(A) Representative Fura-2 recording traces from fluctuation-positive (red and green) and fluctuation-negative (blue and black) round chondrocytes from tamoxifen-treated Trpm7fl/fl (Cre-ER−/−) and Trpm7fl/fl (Cre-ER+/−) mice. The cells were examined under basal conditions and in the presence of the TRPM7 activator naltriben. (B) Quantification of the effects of Trpm7 knockdown on Ca2+ fluctuations under basal and naltriben-treated conditions. Data represent means ± SEM, and the numbers of cells and mice examined are shown in parentheses in the keys and above the graph bars, respectively. Significant differences between the Cre-ER−/− and Cre-ER+/− cells are marked with asterisks (*P < 0.05 and **P < 0.01 in t test). (C) Scatterplot to evaluate the correlation between Ca2+ fluctuation amplitudes under basal and naltriben-treated conditions in round chondrocytes from Trpm7 knockdown (filled circles) and control (open circles) embryos. Each circle represents one cell. In this diagram, the regression line based on all plots from both cell populations is drawn according to the linear least-squares method, and the Pearson R value was determined.

On the basis of the experiments using the chemical modulators and the Trpm7 knockdown chondrocytes (Figs. 3 and 4), we concluded that TRPM7 channels were primarily responsible for generating the spontaneous Ca2+ fluctuations. This conclusion is further supported by clear correlations between the Ca2+ fluctuation amplitudes under basal and naltriben-treated conditions in both Cre-ER+/− and Cre-ER−/− cells (Fig. 4C). The Pearson’s correlation coefficient (R) for the data recorded from these cells, in which cell surface TRPM7 densities were highly divergent, was 0.54. The clustered distribution of Cre-ER+/− cell– and Cre-ER−/− cell–derived plots suggested that the fluctuation amplitudes were predominantly determined by cellular TRPM7 channel density.

Spontaneous Ca2+ entries and oscillations mediated by TRPM7 channels have been proposed in several cell types (2427). For example, nematode intestinal cells generate Ca2+ oscillations accompanied by Ca2+ conductance that are biophysically similar to TRPM7-mediated currents (25), and Ca2+ oscillations in mouse fertilized eggs are inhibited by the TRPM7 inhibitor NS8593 (24). However, the molecular mechanisms underlying the spontaneous activation of TRPM7 channels in these cell types have not been described.

Maintenance of Ca2+ fluctuations

We applied various chemical modulators of channels, transporters, and signaling enzymes to the bone slices (Figs. 2 and 3 and fig. S1B) and additionally found that two chemicals exerted notable effects on the Ca2+ fluctuations in chondrocytes. The first chemical was the phospholipase C (PLC) inhibitor U73122. This inhibitor effectively attenuated the Ca2+ fluctuations in a dose-dependent manner (Fig. 5A) and inhibited the increased Ca2+ fluctuations in naltriben-treated cells (fig. S4A), whereas its inactive homolog U73343 exerted no effects (Fig. 5B). Moreover, the Ca2+ fluctuations were also suppressed by the other PLC inhibitors compound 48/80 and manoalide (fig. S4B), both of which are less specific than U73122 (28). These observations suggest that the Ca2+ fluctuations were evoked downstream of phosphatidylinositol 4,5-bisphosphate hydrolysis. This is consistent with previous reports that PLC-mediated phosphatidylinositol (PI) turnover activates TRPM7 channels in several cell types (5, 29), implying that TRPM7-mediated Ca2+ fluctuations were likely evoked by basal PLC activity in growth plate chondrocytes.

Fig. 5 PLC and BK channel modulators alter Ca2+fluctuations in chondrocytes.

(A) Representative Fura-2 recording traces from fluctuation-positive and fluctuation-negative round chondrocytes in the presence of the PLC inhibitor U73122, the inactive analog U73343, the BK channel blocker paxilline, or the BK channel opener NS1619. (B) Quantification of the effects of U73122, U73343, paxilline, and NS1619 on Ca2+ fluctuations. Data represent means ± SEM, and the numbers of cells and mice examined are shown in parentheses in the keys and above the graph bars, respectively. Statistical differences between DMSO (vehicle control) and modulator treatments are marked with asterisks (*P < 0.05 and **P < 0.01 in one-way ANOVA and Dunnett’s test). (C) Proposed functional coupling of TRPM7 channel activity with PLC-mediated PI turnover and cellular K+ efflux through BK channels in growth plate chondrocytes. TRPM7-mediated Ca2+ entry seems to be essential for growth plate chondrogenesis to promote bone outgrowth. PIP2, phosphatidylinositol 4,5-bisphosphate; DAG, diacylglycerol; IP3, inositol 1,4,5-trisphosphate.

The second chemical that notably affected the Ca2+ fluctuations in chondrocytes was the big-conductance Ca2+-activated K+ (BK) channel opener NS1619. The microarray and RT-PCR analysis indicated that genes encoding BK channel subunits (Kcnma1) were differentially expressed between round and columnar growth plate chondrocytes (fig. S2A) and that NS1619 markedly facilitated the Ca2+ fluctuations under basal and naltriben-treated conditions in round chondrocytes (Fig. 5, A and B, and fig. S4A). In contrast, we observed no inhibitory effect of the BK channel blocker paxilline under basal conditions (Fig. 5), but the facilitated Ca2+ fluctuations under naltriben-treated conditions were inhibited by paxilline (fig. S4A). Furthermore, as expected, the NS1619-induced facilitation was abolished by pretreatment with the TRPM7 inhibitor FTY720 (fig. S2C). BK channel–mediated K+ efflux induces hyperpolarization to enhance the driving force for Ca2+ influx and thus often stimulates Ca2+ entry mediated by TRP and Orai channels in various cell types (30). On the basis of the measurements using PLC and BK channel modulators, we propose that TRPM7 and BK channels may constitute a positive feedback loop for generating and maintaining Ca2+ fluctuations in growth plate chondrocytes (Fig. 5C).

Ca2+ fluctuations and store depletion

The link between TRPM7 channels and SOCE has been debated, and we observed an increase in Ca2+ fluctuations in round chondrocytes under store-depleted conditions (Fig. 2B). To determine whether store depletion or sustained [Ca2+]i increase facilitated the Ca2+ events, we used the Orai channel inhibitor YM58483 (also known as Pyr2 and BTP2) (31), which did not affect TRPM7-mediated Ca2+ fluctuations in our measurements when applied alone (fig. S1B). In TG-treated chondrocytes exhibiting facilitated Ca2+ fluctuations and increased [Ca2+]i, YM58483 efficiently decreased [Ca2+]i baselines, likely because of disruption of Orai-mediated Ca2+ entry, and also dampened the fluctuation amplitudes (fig. S5, A and B). These observations suggested that [Ca2+]i increases primarily activated TRPM7 channels to facilitate the Ca2+ fluctuations under store-depleted conditions. Although TRPM7 channels, together with Orai channels, were activated in TG-treated chondrocytes, our data indicate that it is unlikely that TRPM7 channels directly sense store depletion themselves, consistent with another study using chemical inhibitors and genetic manipulations in various cell types (32).

In addition to BK channels, major PLC subtypes containing EF-hand Ca2+-binding motifs are activated in a Ca2+-dependent manner (28). To assess the possibility that BK channels and PLC subtypes are activated in a Ca2+-dependent manner to facilitate the Ca2+ fluctuations under store-depleted conditions, we applied the PLC and BK channel inhibitors to TG-treated chondrocytes. In response to the PLC inhibitor U73122, the TG-induced increase in Ca2+ fluctuations was inhibited in terms of both the fluctuation amplitude and the proportion of the cells that exhibited fluctuations (fig. S5, A and B). In contrast, the BK channel blocker paxilline did not affect TG-induced Ca2+ fluctuations (fig. S5, A and B). Therefore, activated PLC subtypes seem to predominantly contribute to the facilitation of the Ca2+ fluctuations under store-depleted conditions. BK channels become Ca2+-dependently activated to generate transient outward K+ currents and immediately undergo inactivation with rapid time constants (<10 ms), whereas recovery from the inactivated state requires [Ca2+]i reduction (<100 nM) and hyperpolarization (33, 34). In chondrocytes with depleted Ca2+ stores, BK channels might be largely inactivated because activated Orai and TRPM7 channels likely interrupt the recovery process by increasing both [Ca2+]i and membrane potential.

Cultured Trpm7-deficient bones

Metatarsal organ culture offers a model system for studying endochondral bone development (35). To examine the role of the Ca2+ fluctuations in Trpm7 knockout growth plates, we cultured third metatarsal bone rudiments from Trpm7 conditional knockout embryos. During 8 days in the ex vivo culture, the control Trpm7fl/fl (Cre-ER−/−) bones underwent outgrowth and increased in length by 1.8-fold, regardless of 4-hydroxy-tamoxifen (4OH-Tam) application (Fig. 6A). In Trpm7fl/fl (Cre-ER+/−) bones, 4OH-Tam treatment reduced Trpm7 mRNA content by nearly two orders of magnitude (fig. S6) and reduced outgrowth, which was apparent as a reduction in the length of the diaphyses (Fig. 6A). The results indicated that Trpm7 deficiency interfered with growth plate chondrogenesis.

Fig. 6 Defective outgrowth and chondrogenesis in cultured Trpm7-deficient bones.

(A) Outgrowth of cultured Trpm7-deficient metatarsal bones. Representative images of metatarsals from Trpm7fl/fl (Cre-ER+/−) embryos before and after 8 days of culture in the presence or absence of 4OH-Tam are shown. Scale bar, 0.3 mm. In the right graph, longitudinal growth of metatarsals from Trpm7fl/fl (Cre-ER+/−) and control [Trpm7fl/fl (Cre-ER−/−)] embryos cultured in the presence or absence of 4OH-Tam was quantified over the culture period. The bone sizes on culture day 8 were statistically analyzed between groups. (B) Histological analysis of Trpm7-deficient metatarsal bones. Cross sections of bones from Trpm7fl/fl (Cre-ER+/−) mice cultured in the presence or absence of 4OH-Tam show the growth plate images and high-magnification views of the round (R), columnar (C), and hypertrophic (H) chondrocyte zones. Scale bars, 100 μm. (C) Summary of graphical representations of zonal areas containing round, columnar, and hypertrophic chondrocytes and their size of cells in each zone. Data represent means ± SEM, and the numbers of mice examined are shown in parentheses in the keys. Significant differences between the groups are marked with asterisks (*P < 0.05, **P < 0.01 in t test).

We histologically analyzed growth plates in mid-longitudinal sections from the cultured bones (Fig. 6B). Trpm7-deficient bones maintained normal cross-sectional area and cell densities in round chondrocyte zones, suggesting that TRPM7 channels were not essential for the fundamental functions of round chondrocytes. However, the cross-sectional area of the total epiphyseal region was reduced in Trpm7-deficient bones, as expected from the observed retardation in outgrowth. This reduction was mainly due to decreased sizes of the columnar and hypertrophic chondrocyte zones, with the reduction in the size of the columnar zone more pronounced than that of the hypertrophic zone (Fig. 6C). Therefore, the morphological transition from round to columnar chondrocytes or the proliferation of columnar chondrocytes was most likely attenuated in Trpm7-deficient bones.

When we examined cell sizes in Trpm7-deficient bones, hypertrophic chondrocytes decreased in size, whereas both round and columnar chondrocytes maintained normal sizes (Fig. 6C). During the developmental progression of growth plate chondrocytes, columnar cells differentiate into prehypertrophic cells and then into mature hypertrophic cells, with an approximately 20-fold increase in cell volume, before undergoing apoptotic cell death (1). This hypertrophic maturation is likely interrupted in Trpm7-deficient bones. Considering the histological and cytological defects, TRPM7 channels seem to contribute to maturation processes in growth plate chondrocytes during bone development.

TRPM7 and Ca2+-dependent signaling

We attempted to address the signaling cascades activated downstream of TRPM7-mediated Ca2+ fluctuations in growth plate chondrocytes. Ca2+ entry generally triggers intracellular signaling by activating Ca2+-dependent enzymes, and previous studies suggest the involvement of Ca2+/calmodulin-dependent protein kinase II (CaMKII) and calcineurin (CaN) in chondrogenesis (3638). In our metatarsal culture system, the CaMKII inhibitors KN93 and cell-permeable autocamtide-2–related inhibitory peptide II (pAIP-II) dose-dependently inhibited bone outgrowth, whereas the CaN inhibitor FK506 exerted no significant effects at doses typically used for cell culture experiments (fig. S7A). In addition, both KN93 and pAIP-II caused histological abnormalities that roughly mimicked Trpm7 deficiency; clear reductions in both the size of the columnar chondrocyte zone and the size of hypertrophic cells were observed in the CaMKII inhibitor–treated bones (fig. S7, B and C). These defects suggested that CaMKII signaling may be interrupted in Trpm7-deficient growth plate chondrocytes.

Ca2+-dependent CaMKII activation results in autophosphorylation at Thr286 in the α subtype or at Thr287 in the β, γ, and δ subtypes, and the amount of CaMKII activity is reflected by the relative abundance of the phosphorylated forms compared to the total abundance (39). The microarray analysis suggested that the genes encoding the CaMKII α, γ, and δ subtypes were expressed in both round and columnar growth plate chondrocytes (fig. S2A). We prepared growth plate extracts from cultured metatarsal bones to detect the amounts of total and autophosphorylated CaMKII isoforms using antibodies that are specific for the nonphosphorylated and phosphorylated forms but do not distinguish between the different CaMKII isoforms (Fig. 7A). By immunoblot analysis, the total amounts of CaMKII subtypes were similar between Trpm7-deficient and control samples, but the abundance of the phosphorylated forms was reduced in extracts from Trpm7-deficient growth plates. As a reference experiment, we also examined the transcription factor cAMP (cyclic adenosine 3′,5′-monophosphate) response element–binding protein (CREB), which is phosphorylated mainly by cAMP-dependent protein kinase (PKA). Western blotting suggested that total and phosphorylated CREB were similar in abundance between Trpm7-deficient and control growth plates (Fig. 7A). Therefore, Trpm7 deficiency seemed to inhibit CaMKII activation without affecting PKA signaling in growth plate chondrocytes. On the other hand, mRNA expressions of Ccnd1 and Axin2 are inhibited by the CaMKII, β-catenin, and T cell factor (TCF) signaling cascade (40). Consistent with the proposed CaMKII inactivation, Ccnd1 and Axin2 expressions were significantly increased in Trpm7-deficient growth plates (Fig. 7B). Furthermore, Prg4 expression is controlled by CREB (41), and similar Prg4 mRNA contents were observed in Trpm7-deficient and control growth plates (Fig. 7B). These results suggested that TRPM7-mediated Ca2+ entry activated CaMKII and its downstream signal cascade in the proliferating chondrocytes.

Fig. 7 Inactivated CaMKII signaling in cultured Trpm7-deficient bones.

(A) Immunoblotting for and quantification of CaMKII, phosphorylated CaMKII (pCaMKII), CREB, and phosphorylated CREB (pCREB) in growth plates from cultured Trpm7-deficient bones. Growth plate extracts were prepared from Trpm7fl/fl (Cre-ER+/−) femoral bones cultured for 8 days in the presence or absence of 4OH-Tam and subjected to Western blot analysis and subsequent quantification by densitometry. The antibodies used for detecting total CaMKII and phosphorylated CaMKII react with all CaMKII subtypes. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) is a loading control. Cultured Trpm7-deficient bones from six mice were analyzed. (B) Expression of the indicated transcripts in cultured Trpm7-deficient bones. Total RNAs were prepared from Trpm7fl/fl (Cre-ER+/−) metatarsal bones cultured for 8 days in the presence or absence of 4OH-Tam and subjected to RT-PCR analysis using the primer sets listed in table S1. The cycle threshold (Ct) value determined for each reaction was normalized to the internal control Gapdh value using the comparative Ct method. Data represent means ± SEM, and the numbers of mice examined are shown in parentheses. Significant differences between the groups are marked with asterisks (*P < 0.05 and **P < 0.01 in t test).

Chondrocyte-specific Trpm7-deficient mice

To specifically knock out Trpm7 in chondrocytes, we noted Col11a2, which encodes collagen type XI α2 and behaves as a chondrocyte-specific marker gene. In the 11Enh-Cre transgenic mice, Cre recombinase activity is controlled by the Col11a2 promoter and enhancer and thus deletes loxP-flanked genes in a chondrocyte-specific manner (42). We crossed Trpm7fl/fl mice with 11Enh-Cre mice to generate Trpm7fl/fl (11Enh-Cre+/−) mice that survived fetal life (fig. S8A), in contrast to the embryonic lethal phenotype of systemic Trpm7 knockout (8). In E17.5 Trpm7fl/fl (11Enh-Cre+/−) mice, the loxP-flanked segment of the Trpm7fl locus was deleted in femoral growth plates, but such recombination events were not detected in other tissues (fig. S8B). In the femoral growth plates of Trpm7fl/fl (11Enh-Cre+/−) mice, Trpm7 mRNA expression was significantly decreased (fig. S8C). From a gross anatomical perspective, the chondrocyte-specific Trpm7 deficiency seemed to systemically inhibit long bone development because Trpm7fl/fl (11Enh-Cre+/−) mice had smaller limbs than their Trpm7fl/fl (11Enh-Cre−/−) littermates (fig. S8A). Moreover, it would be worthy to note our preliminary observation that two chondrocyte-specific Trpm7-deficient mice that survived postnatal development for at least 3 weeks displayed a generalized growth failure (fig. S8D).

In E17.5 femoral bones, round chondrocytes from Trpm7fl/fl (11Enh-Cre+/−) mice exhibited spontaneous Ca2+ fluctuations that were weaker than those from Trpm7fl/fl (11Enh-Cre−/−) controls and were less sensitive to naltriben-induced facilitation of the Ca2+ events (Fig. 8A). The femoral bones of the mutant mice were shortened compared to bones from control littermates (Fig. 8B). By histological analysis, the femoral growth plates of Trpm7fl/fl (11Enh-Cre+/−) mice displayed the derangement of columnar cell arrays and reduced swelling of hypertrophic cells that were observed in the conditional knockout metatarsal ex vivo model, and the size of round and columnar chondrocytes was also reduced (Fig. 8C). Furthermore, in the Trpm7fl/fl (11Enh-Cre+/−) growth plates, CaMKII autophosphorylation was inhibited (Fig. 8D), and the abundances of Ccnd1 and Axin2 mRNAs were increased (Fig. 8E). The observations suggested that CaMKII signaling was likely weakened in the Trpm7fl/fl (11Enh-Cre+/−) growth plates. In contrast, the expression of the CREB target gene Prg4 in Trpm7fl/fl (11Enh-Cre+/−) was similar to that in Trpm7fl/fl (11Enh-Cre−/−) controls. Consistent with results from cultured metatarsal Trpm7-deficient bone, analysis of bones from Trpm7fl/fl (11Enh-Cre+/−) embryos suggested that TRPM7-mediated Ca2+ fluctuations activate CaMKII and its downstream signaling, thus promoting growth plate chondrogenesis.

Fig. 8 Abnormal features of femoral growth plates in chondrocyte-specific Trpm7-deficient mice.

(A) Ca2+ fluctuations in round chondrocytes in femoral bone slices prepared from chondrocyte-specific Trpm7 knockout [Trpm7fl/fl (11Enh-Cre+/−)] and control [Trpm7fl/fl (11Enh-Cre−/−)] embryos at E17.5. Recordings were performed in the absence and presence of the TRPM7 activator naltriben. Typical recording traces from individual Ca2+ fluctuation–positive (red, purple, and green) and Ca2+ fluctuation–negative (black) cells are shown. The proportion of cells that showed Ca2+ fluctuations and the amplitudes of the fluctuations were quantified and statistically analyzed. (B) Mid-longitudinal sections of femoral specimens from Trpm7fl/fl (11Enh-Cre+/−) and Trpm7fl/fl (11Enh-Cre−/−) embryos show the zones of round (R), columnar (C), and hypertrophic (H) chondrocytes. Scale bar, 1 mm. Quantification of total bone length and sizes of the entire growth plate (Total) and each chondrocyte zone. (C) Representative images of the round, columnar, and hypertrophic cell zones in femurs from Trpm7fl/fl (11Enh-Cre+/−) and Trpm7fl/fl (11Enh-Cre−/−) embryos. Scale bar, 100 μm. Cell sizes in the growth plate zones were measured and statistically analyzed. (D) Immunoblotting and quantification of CaMKII autophosphorylation in femoral growth plates of Trpm7fl/fl (11Enh-Cre+/−) and Trpm7fl/fl (11Enh-Cre−/−) embryos. (E) RT-PCR analysis of the indicated transcripts in growth plates of Trpm7fl/fl (11Enh-Cre+/−) and Trpm7fl/fl (11Enh-Cre−/−) embryos. Data represent means ± SEM, and the numbers of mice and cells examined are shown in parentheses. Statistical differences between the genotypes are indicated with asterisks (*P < 0.05 and **P < 0.01 in t test). (F) Model of signaling downstream of TRPM7-mediated Ca2+ entry in chondrocytes. This scheme is proposed on the basis of the similarity of defects between Trpm7-deficient and CaMKII inhibitor–treated bones and is further supported by insufficient CaMKII autophosphorylation and altered gene expression in Trpm7-deficient growth plates.

The femoral bones from the chondrocyte-specific Trpm7-deficient mice (Fig. 8), the cultured Trpm7-deficient metatarsals (Figs. 6 and 7), and the cultured metatarsals treated with CaMKII inhibitors (fig. S7) showed similar outgrowth retardation, cytological defects, and biochemical abnormalities. In terms of severity, the chondrocyte-specific Trpm7-deficient femurs exhibited more obvious impairments than the Trpm7-deficient metatarsals generated by conditional knockout. The more severe defects likely reflect the continuous effects of Trpm7 ablation throughout chondrogenesis in Trpm7fl/fl (11Enh-Cre+/−) mice, although the gene ablation seemed incomplete in the growth plates (fig. S9, A and B). The in vivo Trpm7 ablation using the 11Enh-Cre transgene fully confirms the observations from cultured conditional knockout bones and further supports the view that the metatarsal culture system reproduces physiological endochondral ossification processes (35).


TRPM7 channels open in response to various metabolic changes and control important cellular functions including proliferation, differentiation, and apoptosis in various cell types (5, 6). Our present study demonstrates that TRPM7 channels predominantly contribute to generating spontaneous Ca2+ fluctuations (Figs. 1 to 4), which are likely triggered by basal PLC activity and stimulated upon BK channel activation, in growth plate chondrocytes (Fig. 5). In chondrocytes with depleted Ca2+ stores, SOCE increases [Ca2+]i and thus likely stimulates PLC activity, which facilitates the Ca2+ fluctuations (Fig. 2 and fig. S5). Furthermore, our observations indicate that TRPM7-mediated Ca2+ fluctuations promote the maturation of growth plate chondrocytes and thus contribute to bone outgrowth (Figs. 6 and 8). Longitudinal outgrowth of developing bones mainly depends on successive extension of the columnar chondrocyte zone (1), and this process and the subsequent hypertrophic maturation were severely impaired in cultured Trpm7-deficient bones (Figs. 6 and 8). Our conclusion is in accordance with a report that TRPM7 channels are involved in hypertrophic differentiation of cultured chondrogenic ATDC5 cells in vitro (13) and is further supported by the observation that bone outgrowth is systemically perturbed in zebrafish bearing a Trpm7 mutation (9). The bone phenotypes in this mutant zebrafish are consistent with the histological and cytological defects associated with the reduced bone outgrowth that we observed in cultured Trpm7-deficient metatarsals and in femurs from chondrocyte-specific Trpm7 knockout mice.

TRPM7 is a bifunctional protein containing a divalent cation-conducting channel and a C-terminal Ser and Thr kinase. TRPM7 kinase activity is unlikely to contribute to bone development because apparently normal body growth has been reported in two independent lines of Trpm7 knock-in mice in which endogenous TRPM7 was replaced with a mutant form in which the kinase domain was either deleted or mutated but the channel function was not affected (43, 44). TRPM7 channels take part in not only cellular Ca2+ signaling but also Mg2+ homeostasis, and it has been reported that Trpm7-deficient growth defects in cultured cell lines are restored by high Mg2+-containing media (4547). In our ex vivo bone culture model, the regular medium contained 0.8 mM Mg2+, and bone outgrowth was severely inhibited in media supplemented with high Mg2+ (>3 mM; fig. S9A). We did not observe restoration of the Trpm7-deficient outgrowth defects by increasing Mg2+ from 0.8 to 2.8 mM (fig. S9B). Therefore, it is still unclear whether TRPM7-mediated Mg2+ conduction is essential for bone outgrowth.

Similar outgrowth and histological defects were observed in Trpm7-deficient bone and CaMKII inhibitor–treated bones (Figs. 6 and 8 and fig. S7), and reduced CaMKII autophosphorylation was found in Trpm7-deficient growth plates (Figs. 7A and 8D). Therefore, TRPM7-mediated Ca2+ fluctuations seem to activate CaMKII at least partly and significantly, and insufficient CaMKII signaling likely leads to the histological and cytological defects observed in Trpm7-deficient growth plates. The proposed CaMKII activation by TRPM7-mediated Ca2+ fluctuations is consistent with our RT-PCR data indicating that Trpm7 deficiency increased both Ccnd1 and Axin2 expressions without obviously affecting the transcription of major CREB-dependent genes and chondrogenic marker genes in both cultured and femoral bones (Figs. 7B and 8E and fig. S6). Under the canonical Wnt signaling cascade controlling chondrogenesis, inactivated glycogen synthase kinase 3 inhibits phosphorylation-dependent β-catenin degradation and thus stimulates the transcription of β-catenin and TCF-dependent genes such as Ccnd1 and Axin2 (40). In contrast, the β-catenin– and TCF–mediated transcriptions are attenuated by Ca2+-dependent activation of CaMKII (40, 48). In Trpm7-deficient bones, disrupted Ca2+ entry may inhibit CaMKII activation to enhance Ccnd1 and Axin2 expressions (Fig. 8F). However, in addition to CaMKII and CaN, other Ca2+-dependent signaling cascades are often activated by Ca2+ entry in various cell types. Detailed experiments are required to reach to a definitive conclusion about the downstream signaling of autonomous TRPM7-mediated Ca2+ fluctuations in growth plate chondrocytes.

Our study predicts that modulations of both TRPM7-mediated Ca2+ fluctuations and downstream Ca2+-dependent signaling events can control cell fate decision in chondrocytes. Autologous chondrocyte implantation is becoming a popular clinical treatment for damaged articular cartilage tissues using chondrocytes that are recovered from individual patients, expanded and differentiated in vitro, and then transplanted into defective sites (49). We speculate that chemical modulation of TRPM7-mediated signaling may improve the generation of autografts by stimulating the proliferation or differentiation of chondrocytes in culture. For example, perhaps TRPM7 inhibitors could be used to stimulate expansion of chondrocytes in vitro by interrupting cellular differentiation, and later, stimulating TRPM7 activity might promote the proper chondrocyte differentiation before implantation. In addition, it might be valuable to explore the beneficial effects of TRPM7 modulators in animal models of bone and articular disorders. The present study, together with further biological characterizations of TRPM7-mediated Ca2+ signaling, is expected to have implications for therapeutic development in the orthopedic field.


Bone slices

Femoral bones were isolated from E17.5 C57BL mice and immersed in a physiological salt solution (PSS) containing 150 mM NaCl, 4 mM KCl, 1 mM MgCl2, 2 mM CaCl2, 5.6 mM glucose, and 5 mM Hepes (pH 7.4). Longitudinal bone slices (thickness, ~60 μm) were prepared using a vibrating microslicer (model NVSL, World Precision Instruments) essentially as described previously (50). The bone slices thus prepared maintained enough mechanical stability for handling in Ca2+ imaging and DNA and RNA preparations. All experiments in this study were conducted with the approval of the Animal Research Committee according to the regulations for animal experimentation at Kyoto University.

Ca2+ imaging

Fura-2 imaging was performed as described previously (51). For indicator loading, bone slices placed on glass-bottom dishes (Matsunami) were incubated in PSS containing 15 μM Fura-2 AM (Dojindo) for 60 min at 37°C. For ratiometric imaging, excitation lights of 340 and 380 nm were alternately delivered, and emission light of >510 nm was detected by a cooled electron multiplying charge-coupled device camera (model C9100-13; Hamamatsu Photonics) mounted on an upright fluorescence microscope (DM6 FS, Leica) carrying a water immersion objective (HCX APO L 40×, Leica). The measured Fura-2 ratios (F340/F380 values) were calibrated in terms of Ca2+ concentration as described previously (52). Ca2+-EGTA solutions containing 0 to 39 μM free Ca2+ were made in a buffer designed to mimic the intracellular milieu (100 mM KCl and 10 mM Hepes); the Ca2+-EGTA and EGTA amounts required for the preparations were calculated using MaxChelator software ( After adding Fura-2 (5 μM in final concentration), the Ca2+-EGTA standard solutions were used for generating a calibration curve in the Ca2+ imaging system. All experiments were carried out at room temperature (23° to 25°C).

In typical measurements, we randomly selected 20 to 30 cells in each slice preparation and determined Ca2+ fluctuation–positive populations that generated spontaneous Ca2+ events (>0.025 in Fura-2 ratio) to analyze the fluctuation parameters using commercial software (Leica Application Suite X). The recorded traces were also analyzed using our custom-written macro in Fiji/ImageJ (Find Peaks plugin, National Institutes of Health).

Gene expression analysis

Total RNAs were prepared from mouse tissues using a commercial kit (Isogen, Nippon Gene). RNA preparations from femoral cartilage plate sections enriched with round chondrocytes or columnar and hypertrophic chondrocytes were reverse-transcribed and analyzed using the GeneChip Mouse Genome 430 2.0 (Affymetrix) according to the manufacturer’s instructions. The data obtained have been deposited in the National Center for Biotechnology Information–Gene Expression Omnibus (NCBI-GEO) database ( under accession number GSE105256. The array probe intensities were analyzed with the robust multiarray analysis expression algorithm, which represents the log transformation of intensities (background corrected and normalized) from the GeneChips (53), and were visualized in the heat maps (fig. S2A). To further analyze gene expression, mRNA contents were determined by quantitative RT-PCR as described previously (54). Total RNAs were reverse-transcribed using the ReverTra ACE qPCR-RT Kit (Toyobo), and the resulting complementary DNAs were examined by real-time PCR (LightCycler 480 II, Roche). The Ct was determined from the amplification curve as an index for relative mRNA content in each reaction. The RT-PCR primer sets used in this study are listed in table S1.

Trpm7 conditional knockout mice

C57BL6 mice bearing the Trpm7-floxed gene were generated by an outsourced research company (TransGenic Inc.) as illustrated (fig. S3A). In the conditional Trpm7-targeting vector designed according to the recombineering-based construct (55), two loxP sites were inserted into the introns immediately upstream and downstream of exon 22 such that a flippase recognition target (FRT)–containing neomycin resistance cassette was included in the loxP-flanked region. The targeting vector was introduced into embryonic stem cells, and homologous recombination–positive clones were isolated and used for generating germline chimeric mice by the embryo aggregation method. Heterozygous mutant mice bearing the homologous mutation were produced and bred with the Flp mice (the Jackson Laboratory) to remove the neomycin resistance cassette. Trpm7-floxed (Trpm7fl) mice were maintained under our conventional housing conditions.

For tamoxifen-induced Trpm7 gene ablation, the floxed mice were crossed with CAGG-CreER transgenic mice (the Jackson Laboratory). Female Trpm7fl/fl (Cre-ER−/−) mice were crossed with male Trpm7fl/fl (Cre-ER+/−) mice during night time, and noon time of the next day was denoted as E0.5 if there was positive vaginal plug. The pregnant mice were intraperitoneally injected with tamoxifen (50 μg/g of body weight) and then euthanized on E17.5 for preparing femoral Ca2+-imaging and histological specimens from the embryos. For ex vivo bone culture, E15.5 embryos were isolated from the nontreated pregnant mice for preparing metatarsal rudiments.

For chondrocyte-specific Trpm7 gene ablation, the floxed mice were crossed with 11Enh-Cre transgenic mice (National Institutes of Biomedical Innovation, Health and Nutrition, Japan). Oocytes collected from Trpm7fl/fl (11Enh-Cre−/−) females and sperm obtained from Trpm7fl/fl (11Enh-Cre+/−) were subjected to in vitro fertilization. The resulting fertilized eggs were incubated in vitro until the two-cell stage and then transferred into the Institute of Cancer Research (ICR) mice to generate E17.5 embryos for preparing femoral specimens.

Metatarsal organ culture

Metatarsal bone rudiments were cultured as previously described (56). Briefly, the three central metatarsal rudiments were dissected from E15.5 embryos and cultured in minimum essential medium Eagle alpha modification (α-MEM) containing ascorbic acid (5 μg/ml), 1 mM β-glycerophosphate pentahydrate, penicillin (100 U/ml), streptomycin (100 μg/ml), and 0.2% bovine serum albumin (fatty acid free). The rudiments from Trpm7fl/fl (Cre-ER−/−) and Trpm7fl/fl (Cre-ER+/−) embryos were individually placed into wells of 24-well plates containing the culture medium (300 μl) in the presence or absence of 1 μM 4OH-Tam (Sigma-Aldrich) and incubated in a CO2 incubator for 8 days (Fig. 6 and fig. S9B). For the pharmacological survey, bone rudiments from wild-type embryos were incubated in the same culture medium supplemented with CaMKII and CaN inhibitors (fig. S7) or MgCl2 (fig. S9). The cultured explants were analyzed under a photomicroscope (BZ-X710, Keyence) for total length measurements using ImageJ software. For histological analysis, the explants at the eighth day of the organ culture were fixed in 4% paraformaldehyde, embedded in Super Cryoembedding Medium (Section-lab), and frozen in liquid nitrogen. Serial cryosections (thickness, 6 μm) were prepared from the fixed specimens and stained with hematoxylin and eosin. In the sectional images, round, columnar, and hypertrophic chondrocytes were distinguished by their characteristic morphological features.


Femoral rudiments from E15.5 embryos were cultured for 8 days in the presence or absence of 1 μM 4-OH-Tam as described above. Growth plate parts were prepared from the cultured bones and homogenized in the lysis buffer containing 50 mM Hepes (pH 7.0), 150 mM NaCl, 10% glycerol, 1% Nonidet P-40, 2 mM MgCl2, 1 mM EDTA, 100 mM NaF, 10 mM Na3PO4, 1 mM Na3VO4, and 20 mM β-glycerophosphate. After sonication (Astrason Ultrasonic Processor XL, Misonix), the homogenates were centrifuged (17,000g, 30 min) to remove insoluble tissue debris. The resulting growth plate extracts were electrophoresed through 7.5% SDS–polyacrylamide gels, and the separated proteins were electrophoretically transferred to nylon membranes (polyvinylidene difluoride, Merck Millipore). After treatments with a blocking reagent (Blocking One-P solution, Nacalai Tesque), the membranes were probed with primary antibodies and then exposed to horseradish peroxidase–labeled secondary antibody (sc2357, Santa Cruz Biotechnology). Immunoreactivities were visualized using a chemiluminescence reagent (GE Healthcare Life Sciences) and an image analyzer (Amersham Imager 600, GE Healthcare Life Sciences) and were quantitatively analyzed by means of ImageJ software. The primary antibodies were anti-CaMKII (EP1829Y, Abcam), anti–phospho-CaMKII (#12716, Cell Signaling Technology), anti-CREB (#4820, Cell Signaling Technology), anti–phospho-CREB (#4276, Cell Signaling Technology), and anti-GAPDH (G9545, Sigma-Aldrich).


Fig. S1. Ca2+ imaging in hypertrophic and perichondrial cells and pharmacological characterization of Ca2+ fluctuations.

Fig. S2. Gene expression analysis and effects of TRPM7 modulators on growth plate chondrocytes.

Fig. S3. Conditional knockout of Trpm7.

Fig. S4. Effects of PLC and BK channel modulators on naltriben-facilitated Ca2+ fluctuations.

Fig. S5. Characterization of facilitated Ca2+ fluctuations under store-depleted conditions.

Fig. S6. Gene expression analysis in cultured Trpm7-deficient metatarsal bones.

Fig. S7. Effects of CaMKII and CaN inhibitors on cultured metatarsal bones.

Fig. S8. Generation and characterization of chondrocyte-specific Trpm7-deficient mice.

Fig. S9. Extracellular Mg2+ and cultured Trpm7-deficient bones.

Table S1. PCR primers used in the study.


Acknowledgments: We thank K. Yamamoto (Doshisha University), R. Nishimura (Osaka University), and Y. Imaizami (Nagoya City University) for valuable suggestions and comments. Funding: This work was supported in part by the MEXT/JSPS (KAKENHI and Platform for drug discovery, informatics and structural life science), Vehicle Racing Commemorative Foundation, Salt Science Research Foundation, Mochida Memorial Foundation for Medical and Pharmaceutical Research, Takeda Science Foundation, Kobayashi International Scholarship Foundation, and Japan Foundation for Applied Enzymology. Author contributions: N.Q., A.I., D.T., and Y. Miyazaki conducted the Ca2+ imaging analysis. A.I. conducted the microarray analysis. N.Q., A.I., A.K., R.N., M.T., S.K., and M.N. conducted the biochemical and cell physiological analyses. R.S., T.N., and Y. Mori generated the tamoxifen-induced Trpm7 knockout mice. N.Q., A.K., R.N., and M.T. conducted the metatarsal organ culture analysis. H.M. conducted the in vitro fertilization of Trpm7 conditional knockout mice. N.Q., A.I., and H.T. drafted the manuscript. H.T. oversaw this project. Competing interests: The authors declare that they have no competing interests. Data and materials availability: The data from the microarray analysis have been deposited in the NCBI-GEO database under accession number GSE105256. All other data needed to evaluate the conclusions in the paper are present in the paper or the Supplementary Materials.
View Abstract

Navigate This Article