Research ArticleBONE DISEASE

The PTH/PTHrP-SIK3 pathway affects skeletogenesis through altered mTOR signaling

See allHide authors and affiliations

Science Translational Medicine  19 Sep 2018:
Vol. 10, Issue 459, eaat9356
DOI: 10.1126/scitranslmed.aat9356

Skeletal signaling sleuthing

Skeletal dysplasias are rare genetic disorders affecting bone and cartilage growth during development. Csukasi et al. identified two patients with developmental delay and a skeletal phenotype similar to Jansen metaphyseal chondrodysplasia, a disorder caused by altered parathyroid hormone signaling. They identified a genetic mutation in SIK3 as the cause of the of the patients’ dysplasia. The SIK3 mutation altered mTOR signaling, and parathyroid hormone signaling was found to regulate SIK3 activity. This study identifies a common signaling pathway underlying two distinct skeletal disorders, suggesting it plays an important role during skeletal development.


Studies have suggested a role for the mammalian (or mechanistic) target of rapamycin (mTOR) in skeletal development and homeostasis, yet there is no evidence connecting mTOR with the key signaling pathways that regulate skeletogenesis. We identified a parathyroid hormone (PTH)/PTH-related peptide (PTHrP)–salt-inducible kinase 3 (SIK3)–mTOR signaling cascade essential for skeletogenesis. While investigating a new skeletal dysplasia caused by a homozygous mutation in the catalytic domain of SIK3, we observed decreased activity of mTOR complex 1 (mTORC1) and mTORC2 due to accumulation of DEPTOR, a negative regulator of both mTOR complexes. This SIK3 syndrome shared skeletal features with Jansen metaphyseal chondrodysplasia (JMC), a disorder caused by constitutive activation of the PTH/PTHrP receptor. JMC-derived chondrocytes showed reduced SIK3 activity, elevated DEPTOR, and decreased mTORC1 and mTORC2 activity, indicating a common mechanism of disease. The data demonstrate that SIK3 is an essential positive regulator of mTOR signaling that functions by triggering DEPTOR degradation in response to PTH/PTHrP signaling during skeletogenesis.


Multiple signaling pathways act in concert to regulate skeletogenesis, including WNT, transforming growth factor–β/bone morphogenetic protein, hedgehog, fibroblast growth factor, and parathyroid hormone (PTH)/PTH-like peptide (PTHLH) [or PTH-related peptide (PTHrP)], among others (1). PTH/PTHLH governs chondrocyte proliferation and hypertrophy through the binding of PTH or PTHrP to a common receptor, PTH receptor-1 (PTH1R). Pthlh knockout mice show decreased chondrocyte proliferation, premature chondrocyte maturation, and accelerated bone formation (2, 3). Similarly, in Pth1r knockout mice, proliferative chondrocytes prematurely leave the cell cycle and differentiate (4), demonstrating the importance of this pathway in regulating coordinated chondrocyte proliferation and differentiation. Activation of PTH1R, a heterotrimeric guanine nucleotide–binding protein, also initiates a cascade of events that result in the inhibition of two key transcription factors and regulators of chondrocyte hypertrophy, MEF2 and RUNX2 (5, 6). In humans, heterozygosity for an activating mutation in PTH1R produces metaphyseal chondrodysplasia Jansen type (JMC) (7), a disorder characterized by hypercalcemia with suppressed concentrations of PTH and PTHrP, short-bowed limbs, and a distinctive radiographic pattern (8).

The mammalian (or mechanistic) target of rapamycin (mTOR) is an evolutionary conserved serine-threonine (Ser/Thr) kinase that integrates diverse environmental stimuli, including nutrients and growth factors, and translates them into a wide variety of central cellular responses (9). mTOR forms the catalytic subunit of two functionally distinct multiprotein complexes, mTOR complex 1 (mTORC1) and mTORC2. mTORC1, composed of the proteins RAPTOR, PRAS40, mLST8/GβL, and DEPTOR, plays a major role in the regulation of cell growth and proliferation in response to nutrients, amino acids, growth factors, and energy sufficiency. The best characterized substrates of mTORC1 are the ribosomal S6 kinases (S6Ks). mTORC2 is less characterized than mTORC1; it is involved in cell survival in response to growth factors, mostly by promoting the activation of AKT. mLST8/GβL and DEPTOR are found in both mTOR complexes, whereas RICTOR, mSIN1, and PROTOR are unique components to mTORC2 (10). mTORC1 and mTORC2 signaling is essential for chondrocyte proliferation and differentiation, as well as bone development. Mouse mutants that lack Raptor or Rictor show delayed chondrocyte hypertrophy and bone formation (11, 12), and recently, mTORC1 activity was shown to coordinate chondrocyte proliferation and hypertrophy in part through activation of PTHrP transcription (13).

Salt-inducible kinase 3 (SIK3) is a member of the adenosine 5′-monophosphate (AMP)–activated Ser/Thr protein kinase (AMPK) family that comprises 14 members (14, 15). In mammals, the SIK subfamily is composed of SIK1, SIK2, and SIK3 kinases that share a conserved N-terminal catalytic domain but differ in sequence and function at their C terminus (15). SIKs have been mostly characterized in relation to the regulation of hepatic glucose metabolism after their phosphorylation and activation by LKB1 (liver kinase B1), a master kinase for the AMPK family and a negative regulator of mTORC1 signaling (1618). Although the function of SIK3 is less precisely understood, it has been shown to be important for normal skeletal development; Sik3 knockout mice show growth plate abnormalities associated with delayed chondrocyte hypertrophy and primary bone formation (19). Although there is no evidence of a connection between SIK3 and PTH/PTHrP signaling, deletion of SIK2 in osteocytes demonstrated that SIK2 is required for the cellular response to PTH; PTH signaling inhibits SIK2 by triggering its phosphorylation (20). Conversely, SIK3-deficient osteocytes showed normal responsiveness to PTH, suggesting that, at least in osteocytes, PTH specifically targets SIK2 (20). To date, mutations in SIK2 and SIK3 have not been associated with human disease, whereas heterozygosity for mutations in SIK1 is associated with early infantile epileptic encephalopathy (21).

Here, we describe a previously uncharacterized recessively inherited skeletal disorder sharing radiographic similarities to JMC. The disorder resulted from homozygosity for a missense mutation in the catalytic domain of SIK3. Loss of SIK3 kinase activity triggered accumulation of the mTOR negative regulator DEPTOR, leading to downstream impairment of mTORC1/mTORC2 activity. Constitutive activation of the PTH1R (as seen in JMC) inhibited SIK3 activity and caused down-regulation of mTOR signaling through increased DEPTOR accumulation. Our results support a common mechanism of disease in two differing metaphyseal chondrodysplasias and demonstrate a key role for a newly described PTH1R-SIK3-DEPTOR pathway in skeletal development.


A distinct skeletal disorder is caused by a recessively inherited mutation in SIK3

We identified siblings (International Skeletal Dysplasia Registry numbers R07-429A and R07-429B) with radiographic characteristics similar to dominantly inherited JMC but with unique features that did not match any of the currently classified skeletal disorders. The affected siblings came from a consanguineous family, suggesting a recessively inherited disorder due to homozygosity for a mutation, identical by descent. The clinical and radiographic phenotypes are detailed in table S1. In addition to the skeletal phenotype, the siblings demonstrated significant developmental delay with brain magnetic resonance imaging abnormalities, a severe unclassified immunodeficiency, and normal PTH concentration with mild hypercalcemia. Radiographic findings included widened/flared metaphyses with irregular ossifications, moth-eaten long bones, fragmentation of the proximal metacarpals, rounded vertebral bodies, and a distinctive transverse gap seen in the tibias (Fig. 1, A to F, and table S1). Affected individual R07-429A had a more severe phenotype, particularly in her immune system, and died of an Epstein-Barr virus–induced small muscle cancer at 10 years of age. The affected sibling, R07-429B, is still alive at age 14.

Fig. 1 Homozygosity for a missense mutation in SIK3 produces a novel skeletal disorder.

(A to F) Radiographs of R07-429A and R07-429B showing a large separation of the irregular epiphyses from metaphyses, widened/flared metaphyses with irregular ossifications and irregular ossification front (A and B), brachydactyly with irregular metacarpal and phalangeal metaphyses and delayed epiphyseal ossification (C and D), absence of pubic bone ossification (E), and delayed ossification of the basal occipital bone and subjectively increased density of the skull (F). Arrows point to different defects described. (G and H) Western blots and quantification of SIK3 protein concentrations in patient-derived fibroblasts compared to control cells. Graphs represent means ± SEM. Student’s t test, *P < 0.05. Control and R07-429A, n = 5; R07-429B, n = 3. (I) SIK3 gene expression as determined by quantitative polymerase chain reaction (qPCR). Graphs represent means ± SEM. Control and R07-429A, n = 5; R07-429B, n = 3. Individual values of quantifications are provided in table S2.

Because the radiographic phenotype overlapped with JMC, the coding exons of PTH1R were sequenced; however, no mutations were identified. Exome sequencing was carried out, and the data were filtered under a recessive model. This analysis identified homozygosity for a missense variant c.385C>T in exon 2 of SIK3 (NM_025164), predicting the amino acid substitution p.R129C (fig. S1A). The unaffected brother and parents were heterozygous for the variant. Fibroblast cells derived from the two affected individuals were used to assess the effect of the SIK3 variant on mRNA and protein amounts. The expression of the gene was unaffected, but SIK3 protein was decreased more than 60% in cells derived from both affected individuals (Fig. 1, G to I), indicating that the variant compromises the stability and/or synthesis of the protein. Further supporting the human phenotypic and molecular findings, Sik3 knockout mice show skeletal defects (19).

SIK3R129C shows decreased kinase activity

To determine whether the mutation affected SIK3 kinase activity, we first performed three-dimensional (3D) in silico homology modeling of the SIK3 catalytic domain and found that mutated Arg129 adjoins Asp130, a conserved residue directly involved in substrate binding (Fig. 2, A to C). The position of Arg129 in the structure, along with its evolutionary conservation, indicated its importance for maintaining the catalytically active conformation of the kinase. The p.R129C substitution exchanges a large positively charged amino acid for a small neutral amino acid, possibly altering local conformation in the vicinity of the substrate binding site and/or interfering with proper positioning of the substrate binding site with respect to the activation loop (Fig. 2, B and C).

Fig. 2 Impaired kinase function in SIK3R129C mutant.

(A) Alignment of SIK3 orthologs demonstrating evolutionary conservation of p.R129 (conserved residues in gray; R129 in red). (B) 3D homology model of SIK3 kinase domain with indicated adenosine 5′-triphosphate (ATP)–binding site (green), substrate binding site (yellow), and activation loop (blue); arrow indicates position of R129. (C) Magnified view of the mutant p.R129 (red) surroundings. (D) C-terminally FLAG-tagged wild-type (WT) and R129C SIK3 variants transfected into human embryonic kidney (HEK) 293 T cells and purified by FLAG IP. The quantities of three independent transfections for each SIK3 variant were determined by Western blot (WB). Cells transfected with empty plasmid were labeled empty. (E to G) Purified SIK3 in cell-free kinase assays with radioactive ATP and AMARA or CHKtide peptide substrates. Phosphorylation signal was determined by ATP[32P] autoradiography of kinase reaction spotted on blotting paper (E and F) or by scintillation (G). Recombinant active SIK3 was used as a positive control for kinase activity; samples with ATP omitted served as negative controls. (F) ATP[32P] signal shown in (E) normalized to concentration of immunoprecipitated SIK3 in each reaction. Graphs represent means ± SEM. AMARA, n = 7; CHKtide, n = 11. Student’s t test, *** P < 0.0001. (G) ATP[32P] scintillations expressed as percentages of signal obtained in kinase assays with WT SIK3. Signal in immunocomplexes isolated from cells transfected with empty plasmid instead of SIK3 (as denoted by “E,” red dashed lines) represents the background activity. Graphs represent means ± SEM. AMARA (from left to right), n = 14, n = 15, n = 6, n = 6, n = 6, n = 6, and n = 6; CHKtide (from left to right), n = 28, n = 28, n = 10, n = 10, n = 10, n = 10, and n = 10. Student’s t test, ***P < 0.0001. Individual values of quantifications are provided in table S2. C.P.M., counts per minute.

To test this prediction, we compared the catalytic activity of the mutant SIK3 (SIK3R129C) with WT SIK3. WT SIK3 and SIK3R129C constructs were expressed in HEK293T cells, immunoprecipitated, and used in cell-free kinase assays with AMARA or CHKtide peptides as substrates (Fig. 2D). The SIK3-mediated phosphorylation of both AMARA and CHKtide was significantly decreased (P < 0.0001) in the presence of SIK3R129C compared with WT SIK3 (Fig. 2, E to G). These data demonstrated that the p.R129C mutation led to impairment of SIK3 Ser/Thr kinase activity.

SIK3 regulates mTOR activity

SIK3 is a member of the AMPK family of proteins. Because AMPK inhibits mTORC1 through direct phosphorylation of tuberous sclerosis complex 2 (TSC2) (22) and Raptor (23), we tested whether SIK3 also regulated mTOR signaling. Using patient cells to measure the phosphorylation of S6K1 and S6, two downstream effectors of mTORC1 and AKT, as a readout for mTORC2 complex activity under serum starvation (inactivation of mTOR), we observed a decrease in the phosphorylation of S6K1, S6, and AKT compared to control (Fig. 3A and fig. S2), indicating that SIK3 is a positive regulator of mTORC1 and mTORC2 activity and that mTOR signaling is in impaired in SIK3R129C cells.

Fig. 3 SIK3-deficient cells show decreased mTORC1 and mTORC2 activity due to accumulation of DEPTOR.

(A) Western blots showing amounts of indicated proteins in patient and control cells starved overnight (O/N) before protein extraction. GAPDH, glyceraldehyde-3-phosphate dehydrogenase. (B) Co-IP of SIK3 with different mTORC1/mTORC2 components. C-terminally FLAG-tagged WT (SIK3) and R129C SIK3 variants were transfected into HEK293T cells and purified by FLAG IP. Co-IP of endogenous RICTOR, RAPTOR, and GβL was detected by Western blot. (C) C-terminally FLAG-tagged WT (SIK3), R129C SIK3 variants, and N-terminally hemagglutinin (HA)–tagged DEPTOR were cotransfected into HEK293T cells and purified by HA IP. Co-IP of SIK3 was detected by Western blot. (D) Western blots of control and patient-derived fibroblasts starved overnight and then treated with 10% fetal bovine serum (FBS) for 4 to 7 hours. (E) Western blots of control fibroblasts transfected with 10 nM small interfering RNA (siRNA) and control siRNA (control). Forty-eight hours after transfection, cells were starved overnight and then treated with 10% FBS for 4 hours or overnight. (F) N-terminally HA-tagged DEPTOR was transfected into control and patient-derived fibroblasts, and 24 hours later, cells were starved overnight before harvest or treatment with 10% serum + 15 μM MG132 for 10 hours. DEPTOR was immunopurified using HA beads, and its interaction with endogenous β-transducin repeats-containing protein (β-TRCP) was detected by Western blot. EV, empty vector (negative control). WCL, whole-cell lysates. Quantification of Western blots is shown in figs. S2 and S4. Individual values of quantifications are provided in table S2.

We next investigated whether SIK3 directly interacted with components of the mTORC1 and mTORC2 complexes. Western blot analysis of WT SIK3 immunoprecipitated from transfected HEK293 T cells detected Co-IP of RICTOR (mTORC2) and GβL (mTORC1/mTORC2) and a weak interaction with RAPTOR (mTORC1) (Fig. 3B). SIK3R129C also interacted with these respective proteins. Because HEK293T cells have very low expression of the mTORC1/mTORC2 component DEPTOR (24), we cotransfected DEPTOR with WT SIK3 and SIK3R129C and showed that SIK3 can also interact with DEPTOR (Fig. 3C).

DEPTOR proteasome degradation is impaired in SIK3-deficient cells

Because SIK3 deficiency induced a decrease in mTORC1 and mTORC2 activity, we reasoned that one of the components common to both complexes (GβL and DEPTOR) might be responsible for the reduced activity on substrates. We measured GβL and DEPTOR protein amounts in mutant and control cells under three different conditions: serum starvation to inactivate mTOR, 30 min of serum treatment to activate mTOR, or 30 min of insulin treatment to activate mTOR. Under all of these conditions, there was a greater accumulation of DEPTOR in mutant versus control cells, whereas no change was observed in GβL concentration (Fig. 3A and figs. S2 and S3).

DEPTOR is a negative regulator of mTORC1 and mTORC2 and is phosphorylated and degraded after several hours of serum stimulation (24). We therefore investigated DEPTOR amounts in patient cells after longer exposures to serum and found that DEPTOR remained higher in patient versus control cells after 4 and 7 hours of treatment but completely degraded after overnight exposure (Fig. 3D and fig. S4A). These results demonstrate that SIK3 contributes to DEPTOR degradation upon serum stimulation and that, although this process is delayed in mutant cells, DEPTOR is ultimately degraded because it is phosphorylated by other kinases, including mTOR (2528).

To confirm that SIK3 promoted mTOR activity by triggering DEPTOR degradation, we measured the phosphorylation of S6K1, S6, and AKT in the same prolonged exposures to serum (4- and 7-hour and overnight treatments). Decreased expression of pS6K1, pS6, and phosphorylated AKT (pAKT) observed in mutant cells under starvation and after 30 min of serum or insulin treatments similarly recovered under these prolonged exposure conditions, showing that, even in mutant cells, progressive degradation of DEPTOR occurred over time, restoring mTOR signaling (Fig. 3D and figs. S3 and S4A). siRNA-mediated knockdown of SIK3 also induced an accumulation of DEPTOR and a decrease in pAKT, pS6K1, and pS6 under starvation, replicating the results observed in patient cells, whereas no difference was observed after 4-hour or overnight serum exposure (Fig. 3E and fig. S4B).

Phosphorylated DEPTOR is recognized by the F-box protein β-TRCP for subsequent ubiquitination and proteasome degradation by the SCF E3 ubiquitin ligase complex (2527). To determine whether SIK3 is necessary for DEPTOR interaction with β-TRCP, we transfected control and patient cells with HA-tagged DEPTOR, immunoprecipitated DEPTOR, and detected its interaction with endogenous β-TRCP. The lack of SIK3 kinase activity in the patient cells decreased the interaction between DEPTOR and β-TRCP (Fig. 3F), demonstrating that SIK3 contributes to the phosphorylation-dependent DEPTOR degradation by the proteasome.

SIK3 is coexpressed with mTOR components in the growth plate

On the basis of the clinical phenotype, we explored the role played by SIK3 in skeletal development in relation to mTOR components. Expression patterns of SIK3, pS6, and DEPTOR were determined in P1 and P21 mouse cartilage growth plates by immunohistochemistry. At P1, both SIK3 and pS6 were expressed in early proliferating and prehypertrophic chondrocytes (Fig. 4, A, B, D, E, and G, arrows). At P1, DEPTOR was primarily expressed in resting and proliferating chondrocytes (Fig. 4, C and F). At P21, SIK3, DEPTOR, and pS6 showed overlapping expression patterns, particularly in proliferating and prehypertrophic chondrocytes (Fig. 4, I to K and M to O), indicating an ongoing role for SIK3-mTOR activity in chondrocyte behavior in the maturing growth plate.

Fig. 4 SIK3 is coexpressed with mTOR components in the growth plate.

Immunolocalization of SIK3, DEPTOR, and pS6 in P1 (A to H) and P21 (I to P) mouse femur growth plates. (A) Picrosirius red hematoxylin staining of P1 cartilage growth plate chondrocytes. (B to D) Immunolocalization of SIK3 (B), DEPTOR (C), and pS6 (D) at P1. (E to H) Magnification of boxed regions in (A) to (D). (I) Picrosirius red hematoxylin staining of P21 growth plate. (J to L) Immunolocalization of SIK3 (J), DEPTOR (K), and pS6 (L) at P21. (M to P) Magnification of boxed regions in (I) to (L). Dashed lines represent separation of proliferative and hypertrophic zones. Arrows point to zones of expression detected with the different antibodies. Scale bars, 50 [(A to H) and (M to P)] and 200 μm (I to L).

PTH/PTHrP signaling inhibits mTOR activity through SIK3

Because of the similar skeletal abnormalities between the patients with deficient SIK3 and the JMC patients with overactivation of PTH1R, we hypothesized that SIK3 was regulated by PTH/PTHrP signaling. Using chondrocytes derived from a patient with JMC (R93-393), we measured the concentration of total cellular DEPTOR and the activation of mTORC1 and mTORC2 substrates under serum starvation and after 4 hours of serum stimulation. DEPTOR protein was elevated in the JMC patient cells compared to the controls under serum deprivation (Fig. 5A and fig. S5). As seen in SIK3-deficient cells, JMC cells showed persistence of DEPTOR protein after serum treatments relative to controls, indicating that the PTH/PTHrP/PTH1R pathway is also involved in DEPTOR degradation. Concordant with this finding, the phosphorylation of S6 and AKT was also decreased (Fig. 5A and fig. S5). To determine whether PTH/PTHrP/PTH1R signaling might regulate DEPTOR accumulation by controlling the activity of SIK3, we measured the phosphorylation of SIK3 at a known activating residue (T163) (15, 29) and found it decreased in the JMC patient cells (Fig. 5A and fig. S5). Because SIK2 is inhibited by PTH in osteocytes (20), we questioned whether SIK2 phosphorylation was also affected in JMC; JMC chondrocytes showed a decrease in pSIK2 at T175, known to be important for its activity (Fig. 5A and fig. S5) (30). Treatment of primary human control chondrocytes with PTHrP showed a decrease in SIK3 activity measured by pSIK3T163, consistent with our results using JMC chondrocytes (fig. S6) and further suggesting that the PTHrP-PTHR pathway is a negative regulator of SIK3. No significant change was observed in the activity of SIK2 after the treatment (fig. S6).

Fig. 5 PTH/PTHrP signaling regulates mTOR activity through SIK3-mediated DEPTOR degradation.

(A) Western blot of JMC-derived patient and control chondrocytes starved overnight and then treated with 10% FBS for 4 hours. Quantification shown in fig. S5. Individual values of quantifications are provided in table S2. (B to E) Picrosirius red hematoxylin staining of control (B and C) and JMC patient (D and E) proximal phalange growth plates. Magnifications of boxed regions in (B) and (D) are shown in (C) and (E). (F to I) Immunolocalization of DEPTOR in control (F and H) and patient growth plates (G and I). Magnifications of boxed regions in (F) and (G) are shown in (H) and (I). Arrows indicate expression of DEPTOR in proliferative chondrocytes and absence in hypertrophic cells in control compared to constant expression throughout the growth plate in the patient. Scale bars, 50 μm.

Histological analysis of a growth plate from a patient with JMC showed a severe disorganization with a hypocellular reserve zone, a reduced proliferative region with clusters of late-proliferating and diminished number of hypertrophic chondrocytes that fail to organize into columns, and invasion of bone territories (spicules or mineralization front) into the cartilaginous growth plate, as observed by intense picrosirius red staining (Fig. 5, B to E). The absence of hypertrophic chondrocytes has also been described in the Sik3 knockout mouse, and the authors concluded that the major effect of Sik3 deficiency in skeletal development was disruption of chondrocyte hypertrophy (19). Immunolocalization of DEPTOR in human cartilage growth plate showed that its expression is higher in resting and proliferative chondrocytes in control tissues, decreasing upon commitment to the hypertrophic program (Fig. 5, F and H). In contrast, JMC patient growth plate showed a constant expression of DEPTOR throughout the poorly organized growth plate (Fig. 5, G and I), concordant with Western blot analyses that showed increased amounts of DEPTOR (Fig. 3A). These results support that, in WT chondrocytes, PTH/PTHrP inhibits SIK3 phosphorylation and acts as a negative regulator of mTORC1/mTORC2 activity (and PTH/PTHrP inhibits SIK2 in osteocytes but not in chondrocytes). In JMC, because of constitutive activation of the PTH/PTHrP pathway, there is exaggerated and persistent inhibition of SIK3, leading to DEPTOR accumulation and subsequent decrease in mTOR activity (fig. S7).


Here, we identified a signaling pathway connecting PTH/PTHrP, SIK3, and mTOR and demonstrated that it has a key role in the regulation and maintenance of skeletal development. These findings were determined by uncovering a new recessively inherited skeletal dysplasia, similar to autosomal dominantly inherited JMC but with several distinctions. The causative gene encodes SIK3, and the missense mutation is within the highly conserved kinase domain, resulting in decreased Ser/Thr kinase activity. Supporting the human phenotypic findings, Sik3 knockout mice show similar skeleton defects, including metaphyseal expansion of the limb joint region, shortened long bones, hypoplastic pelvis, and delayed membranous ossification of the skull bones (large fontanelle remains open from juvenile stage until adulthood) (19).

Addressing whether the identified family represents an isolated finding, review of the literature revealed a report of a 1-year-old male from a consanguineous mating with similar radiographic and clinical findings to our patients, particularly the findings of the frontal bossing, hypertelorism, prominent joints, psychomotor delay, hypercalcemia, a transverse gap in the ulna and marked epiphyseal delay with absent ossification of the pubis, and a highly irregularly distorted metaphyses (31). JMC was considered in that report but was excluded as final diagnosis based on the unique constellation of findings. Interrogation of the International Skeletal Dysplasia Registry identified another case with similar findings, particularly the finding of a transverse gap in the long bones from another consanguineous mating. Attempts were made to obtain DNA from these cases but failed because of passage of time and loss of contact. However, these other cases support that this new disorder is likely not isolated solely to this family.

We demonstrated that SIK3 promotes the activity of mTOR, a central regulator of mammalian metabolism and physiology, and showed how mTOR is rapidly switched from a low- to a high-activity state through SIK3-dependent DEPTOR degradation in response to nutrient sufficiency. Several lines of evidence support this model. First, SIK3 directly interacts with DEPTOR. Second, mutant SIK3 patient cells showed a high accumulation of DEPTOR protein under starvation and after short-term serum treatments; DEPTOR amounts returned to normal after long-term serum treatments, corresponding with recovery of normal activity of mTORC1 and mTORC2. Third, DEPTOR protein in patient cells was incapable of interacting with β-TRCP, the F-box subunit of an SCF E3 ubiquitin ligase complex, further explaining why it remains high in patient cells. Finally, knockdown of SIK3 increased DEPTOR concentration, phenocopying mutant kinase-deficient SIK3. Recently, SIK3 was proposed as a downstream target of mTORC2; however, our results position SIK3 upstream of mTOR, suggesting the existence of a feedback mechanism between them. DEPTOR and mTOR also form a feedback loop, and mTOR kinase activity is partially responsible for DEPTOR phosphorylation and degradation (2527), supporting this hypothesis. Other kinases that phosphorylate DEPTOR include casein kinase 1 (25) and the mitogen-activated protein kinases p38γ and δ (28)—all of them act in cooperation to promote DEPTOR degradation in response to upstream signals. Our work adds a new component to the already complex and tightly regulated mechanism of activation of mTOR.

Deregulation of mTOR activity contributes to many types of diseases including cancer and diabetes (32, 33). Previous work showed that about 28% of multiple myelomas overexpress DEPTOR (24). Here, we linked a genetic disease, and particularly, a skeletal dysplasia, with reduced mTOR activity. Although no mutations in any mTOR components have been identified in humans, Raptor, Rictor, and TSC1 loss-of-function mice show skeletal defects (1113), demonstrating the importance of mTOR activity in cartilage and bone formation. We showed that SIK3 is present in the growth plate where it is coexpressed with DEPTOR and the mTORC1 substrate S6 in proliferative and early hypertrophic chondrocytes. Previous studies have shown that, although mTORC1 activity is required for chondrocyte growth, its inactivation is essential for chondrocyte terminal differentiation, establishing the dynamic regulation of mTORC1 activity as crucial for skeleton development (11, 13). Although mTORC2 signaling is much less characterized in cartilage and bone, high amounts of phosphorylated Akt have been found in proliferative chondrocytes in mice (34). Akt signaling is known to enhance chondrocyte proliferation and delay hypertrophy (35), suggesting that pathogenicity due to diminished SIK3 activity and DEPTOR accumulation on chondrocyte proliferation and differentiation may also result from perturbed mTORC2 activity. Our findings point to the importance of a SIK3-dependent activation of mTORC1 and mTORC2 during chondrocyte development. This activation might also be dependent on the changing upstream signals that surround chondrocytes during differential stages of development.

JMC results from constitutive activation of PTH1R, the PTH/PTHrP receptor (7), and presents with skeletal similarities to those patients with mutant SIK3 (8). The findings that DEPTOR is also overexpressed in JMC, resulting in lower mTORC1 and mTORC2 signaling, and that this correlates with low activity of SIK3 support that a common mechanism of disease is partially responsible for both skeletal disorders.

Although no cartilage growth plates from the SIK3-affected patients were available to analyze, the growth plate from the patient with JMC showed grossly abnormal terminal chondrocyte differentiation in the hypertrophic zone with few cells, lack of normal hypertrophy, and poor column formation. This supports the role of DEPTOR-regulated mTOR activity in the development of the human cartilage growth plate. Furthermore, mTORC1 was recently shown to control chondrocyte proliferation and differentiation through the regulation of PTHrP transcription (13). Our results support a model in which active PTH-PTHrP/PTH1R–meditated signaling in early proliferative chondrocytes inhibits SIK3 activity, leading to an accumulation of DEPTOR and therefore low mTOR activity. Upon commitment to the hypertrophic program, decrease of PTH-PTHrP/PTH1R relieves SIK3 inhibition, triggering DEPTOR degradation and activation of mTOR signaling to promote the initiation of chondrocyte hypertrophy (fig. S7).

Our results demonstrate that aberrant PTH/PTHrP-SIK3-mTOR signaling causes a skeletal dysplasia similar to but distinct from JMC. Although there are clinical distinctions based on neurologic and immune systems abnormalities that underscore the multifunctional role of SIK3, we uncovered a detailed mechanism of overlapping disease in which hyperactivation of PTH1R or SIK3 deficiency leads to decreased mTOR signaling due to impaired DEPTOR turnover. Our results identify mTOR components as potential therapeutic targets for treatment of these and perhaps other skeletal disorders. Recognizing the importance of mTOR signaling in cancer and given that some cancers express aberrant concentrations of DEPTOR (24), which could be recovered by modification of SIK3 activity, the implications of this work extend to identifying SIK3 as a target for cancer drug development.


Study design

The objective of this study was to identify the gene responsible for a rare skeletal phenotype observed in two patients. After the gene was identified, we focused on characterizing the molecular mechanism underlying the disease. Patients and their unaffected family members were studied under an approved human subjects protocol. Clinical information and imaging were obtained from a review of all available medical records. Experiments performed in cells were conducted at least three independent times, and all replicates were included in our data analyses.

Cell culture

Dermal fibroblast cultures were established from explanted skin biopsies from the two individuals affected with the missense mutation in SIK3 (International Skeletal Dysplasia Registry reference numbers R07-429A and R07-429B) and controls. Primary chondrocytes were isolated from distal femurs of the individual affected with JMC (International Skeletal Dysplasia Registry reference number R93-393) or age-matched normal controls by incubation of fragmented cartilage with 0.03% of bacterial collagenase II. All cells were grown in Dulbecco-Vogt–modified Eagle medium supplemented with 10% FBS. For serum treatments, cells were starved overnight (no serum) and then treated with the same media supplemented with 10% FBS. For protein analyses, cells were collected in IP lysis buffer (catalog no. 87787, Thermo Fisher Scientific) supplemented with proteinase inhibitors.

SIK3 overexpression experiments were performed by introducing a vector containing the SIK3 or SIK3R129C coding sequences tagged with DDK-flag; SIK3-DDK vector was obtained from OriGene (catalog no. RC223406). Knockdown of SIK3 was performed using three siRNAs (catalog no. SR308237, OriGene) and compared to a siRNA-scrambled control (catalog no. SR30004, OriGene). Electroporation was performed in a Nucleofector X system (Lonza) using Amaxa P1 primary cell kit and program DS-150 for fibroblasts and SE kit and program CM-104 for the immortal chondrocytes (Lonza).

Exome analysis

DNA was isolated, and library preparation and exome sequencing were performed as previously described (36). The samples were barcoded, captured using the NimbleGen SeqCap EZ Exome Library version 2.0 probe library targeting 36.5 million base pairs of genome, and sequenced on the Illumina GAIIx platform with 50-base pair bidirectional reads. NovoAlign was used to align the sequencing data to the human reference genome [National Center for Biotechnology Information (NCBI) build 37], and the Genome Analysis Toolkit (GATK) was used for post-processing and variant calling according to GATK best practices recommendations. For each sample, at least 90% of targeted bases were covered by at least 10 independent reads. Variants were filtered against database single nucleotide polymorphism build 137 (dbSNP137), National Institute of Environmental Health Sciences Environmental Genome Project exome samples (version 0.0.8), exomes from the National Heart, Lung, and Blood Institute Exome Sequencing Project (ESP6500), 1000 genomes (release version 3.20120430), and in-house exome samples. Mutations were further compared with known disease-causing mutations in Human Gene Mutation Database (version 2012.2). Variants were annotated using VAX34, and mutation pathogenicity was predicted using the programs PolyPhen35, Sift36, Condel37, CADD38, and MutationTaster. The mutations reported in this work were confirmed by bidirectional Sanger sequencing of amplified DNA from the probands and the parents. Primer sequences for exon 2 were 5′-CCCAGCTGGATGAAGAAAAC-3′ (forward) and 5′-GCACACAAGCACGTAGAGGA-3′ (reverse).

Western blot and IPs

For Western blot analyses, protein lysates were separated by electrophoresis on 10% or gradient (4 to 20%) SDS–polyacrylamide gels, transferred to polyvinylidene difluoride membranes, blocked in 5% milk, and probed with primary antibodies [anti-SIK3 antibody (1:2000; catalog no. 88495, Abcam), anti-pSIK2/SIK3 antibody (1:1000; catalog no. PA5-64607, Thermo Fisher Scientific), anti-pAKT (1:1000; catalog no. 4060, Cell Signaling Technology), anti-pS6K1 (1:1000; catalog no. 9205, Cell Signaling Technology), anti-pS6 (1:1000; catalog no. 2211, Cell Signaling Technology), anti-AKT (1:1000; catalog no. 4685, Cell Signaling Technology), anti-S6K1 (1:1000; catalog no. 9202, Cell Signaling Technology), anti-S6 (1:1000; catalog no. 2317, Cell Signaling Technology), and anti-GAPDH (1:2000; catalog no. 2118S, Cell Signaling Technology)]. Peroxidase-conjugated secondary antibodies (catalog nos. 7071 and 7072, Cell Signaling Technology) were used, and immunocomplexes were identified using the ECL (enhanced chemiluminescence) Detection Reagent (catalog no. 7003, Cell Signaling Technology). Fiji was used to quantify bands after gel analysis recommendations from ImageJ and (37) (, and the Mann-Whitney test was performed for statistical analysis using Prism software. Experiments were replicated at least three times to perform statistical analysis.

For SIK3 and SIK3R129C IPs, HEK293T cells were transfected with the appropriate FLAG-tagged vectors (OriGene RC223406 and its mutated version, which was generated by site-directed mutagenesis). Twenty-four hours after transfections, cells were harvested and protein extracts were immunopurified using anti-FLAG antibody (catalog no. F1804, Sigma-Aldrich) collected on Protein A/G agarose (Santa Cruz Biotechnology). For DEPTOR IP, HEK293T cells were transfected with HA-tagged DEPTOR and harvested 24 hours after transfection. Proteins were immunoprecipitated using anti-HA magnetic beads (Thermo Fisher Scientific).

Histological analyses and immunohistochemistry

For histology and immunocytochemistry, human (control and JMC) and mouse (WT, P1, and P21) tissues were fixed in 4% paraformaldehyde, decalcified using immunocal decalcification solution (catalog no. 1414-1, StatLab), and then paraffin-embedded. Paraffin blocks were sectioned at 10 μm. For histological analyses, sections were stained with Picrosirius red. For Picrosirius red staining, deparaffinized and rehydrated sections were stained in a 0.1% Direct Red 80 (catalog no. 43665, Sigma-Aldrich)/saturated picric acid (catalog no. P6744, Sigma-Aldrich) solution, followed by counterstaining Hematoxylin QS.

For immunohistochemistry, paraffin sections were boiled for 20 min in Antigen Unmasking Solution (Vector Laboratories) and subsequently stained using the Rabbit-specific HRP/DAB (ABC) Detection IHC Kit (Abcam). Primary antibodies used were anti-SIK3 antibody, (1:50; catalog no. 88495, Abcam), anti-DEPTOR (1:500; catalog no. 20985-1-AP, Proteintech), and anti-pS6 (1:400; catalog no. 2211, Cell Signaling Technology).

Site-directed mutagenesis

Site-directed mutagenesis was performed to generate SIK3R129C plasmid using the QuikChange II XL Site-Directed Mutagenesis Kit (catalog no. 200521, Agilent). Mutagenesis primers were designed using the QuikChange Primer Design Program ( Primers used were 5′-TTTTGTCACTGTCGGAACATTGTTCATTGTGATTTAAAAGCTGAAA-3′ (forward) and 5′-TTTCAGCTTTTAAATCACAATGAACAATGTTCCGACAGTGACAAAA-3′ (reverse).

RNA extraction and qPCR

RNA was extracted from fibroblasts using TRIzol reagent (Life Technologies). Complementary DNA (cDNA) was prepared from 1 μg of RNA using RevertAid First strand cDNA synthesis kit (Thermo Fisher Scientific) and amplified using Maxima SYBR Green/ROX qPCR Master Mix (Thermo Fisher Scientific). Gene expression was calculated using the 2ΔΔCT method of analysis against the stable housekeeping gene β-2-microglubulin (β2M). Three biological replicates were performed with two technical replicates each. qPCR primers were (i) SIK3, 5′-TCGGCTACTACGAGATCGAC-3′ (forward) and 5′-TCCCGGAAAATCTTCTTCAA-3′ (reverse) and (ii) B2M, 5′-TGACTTTGTCACAGCCCAAG-3′ (forward) and 5′-AGCAAGCAAGCAGAATTTGG-3′ (reverse).

3D in silico modeling and kinase assay

For SIK3 structural modeling, the 3D model for WT human SIK3 was obtained via template-based homology modeling using the Phyre software (38). The SIK3-specific functional elements, predicted using the NCBI Conserved Domain Database (39), were mapped onto a 3D model of the SIK3 using Chimera software (40).

For kinase assays, the C-terminally FLAG-tagged SIK3 was expressed in 293T cells and purified by FLAG IP 24 hours later. Kinase assays were performed using immunopurified SIK3 or 200 ng of recombinant active SIK3 (SignalChem), together with 4.6 μg of RBER-CHKtide peptide (ProQinase) or 10 μg of AMARA peptide (obtained from SignalChem or Abcam) as a substrate, respectively. Kinase assays were carried out in kinase assay buffer (SignalChem), according to the manufacturer’s instructions, supplemented with 1 μCi of 32P-ATP (Hartmann Analytic) for 30 min at 30°C. The reactions were terminated by spotting 4 μl from each reaction onto a nitrocellulose membrane (for autoradiography) or by spotting 10 μl onto a 1-cm2 rectangular piece of nitrocellulose membrane (for scintillation). Nitrocellulose membranes were allowed to dry for 10 min at room temperature and washed three times for 10 min in 1% phosphoric acid and 0.5 M NaCl. The SIK3 kinase activity was determined using 1214 RackBeta Liquid Scintillation Counter (LKB Instruments).

Statistical analysis

GraphPad Prism was used for statistical analysis. All values are means ± SEM, as indicated in figure legends. All comparisons in the study were performed using Student’s t test. Individual subject-level data are presented in table S2.


Fig. S1. Patients show a missense mutation in the catalytic domain of SIK3.

Fig. S2. Quantification of protein amounts of Western blots in Fig. 3A.

Fig. S3. Serum and insulin treatment in control and patient cells.

Fig. S4. Quantification of protein amounts of Western blots in Fig. 3 (D and E).

Fig. S5. Quantification of protein amounts of Western blots in Fig. 5.

Fig. S6. PTHrP treatment in primary chondrocytes.

Fig. S7. Model of action of SIK3.

Table S1. Clinical findings of affected individuals with a mutation in SIK3.

Table S2. Individual subject-level data.


Acknowledgments: We thank the families for their participation in this study. Funding: D.K. and D.H.C. are supported by NIH grants RO1 AR066124, R01 AR062651, and RO1 DE019567. I.D. is supported by a Geisman Fellowship award from the Osteogenesis Imperfecta Foundation. P.K. is supported by the Agency for Healthcare Research of the Czech Republic (15-33232A, 15-34405A, and NV18-08-00567), Czech Science Foundation (GA17-09525S), and Ministry of Education, Youth and Sports of the Czech Republic (LQ1605 NPU II). Author contributions: F.C., I.D., M.B., T.B., I.G., L.T., J.H.M., and H.L performed the experiments. C.Y.K., J.W., and D.K. examined the patients. F.C., P.K., D.H.C., and D.K. designed the experiments and analyzed the results. F.C. and D.K. wrote the manuscript. All authors read and approved the manuscript. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data associated with this study are present in the paper or the Supplementary Materials.

Stay Connected to Science Translational Medicine

Navigate This Article