1. Introduction
Salt-inducible kinases (SIK1, SIK2, SIK3) are members of the AMPactivated protein kinase (AMPK) family [1]. SIK2 was shown to be highly abundant in both mouse [2] and in human [3] adipose tissue, and we found it to be downregulated in adipose tissue and adipocytes from individuals with obesity and/or insulin resistance [3]. Moreover, we and others demonstrated that pharmacological inhibition or silencing of SIK isoforms results in reduced insulin signalling, observed as a reduction in activity-controlling phosphorylation of PKB/Akt, and glucose uptake, in both rodent and human adipocytes [3–5]. Collectively, this suggests that downregulation of SIK2 in adipocytes might play a causal role in the development of obesity-induced insulin resistance. However, the mechanisms underlying regulation of insulin signalling by SIK2 are poorly understood – especially in humans.The most well-known function of SIK isoforms is the regulation of gene expression through phosphorylation and cytosolic retention of transcriptional (co)-regulators, such as class IIa histone deacetylases (HDAC4/5/7), and CREB-regulated transcriptional co-activators (CRTCs) [1,6–8], which we have shown to be substrates of SIK2 also in adipocytes [4]. In mice, SIK2 regulates the expression of GLUT4, the glucose transporter responsible for insulin-induced glucose uptake in adipocytes [9], possibly through HDAC4 and CRTC2 [4,5]. However, in human adipocytes, GLUT4 expression was shown to be, if anything,
increased by SIK inhibition [3]. There is thus a need to identify additional genes regulated downstream of SIK2, which might mediate the effect of SIK2 on insulin signalling and glucose uptake.
JUP, also known as junction plakoglobin or γ-catenin, is a cytosolic component of adherens junctions and desmosomes, and links cadherins to actin and intermediary filaments [10]. It is a member of the catenin family, being highly homologous to the more extensively studied βcatenin [10]. Recently, JUP was shown to promote insulin signalling in muscle cells [11]. The proposed mechanism involved interaction of JUP with PI3K (phosphatidylinositol 3-kinase) p85 and the insulin receptor, proteins in the insulin signalling pathway [11]. The same study [11] showed that JUP downregulation inhibits insulin signalling, while JUP overexpression enhances it, as evaluated by analysis of activating phosphorylation of PKB/Akt. JUP was also reported to possess weak transcriptional activity, and to regulate nuclear presence of its homologue β-catenin, which is a well-established transcriptional co-activator [12]. To our knowledge, there are no previous studies of JUP in adipocytes, but its transcription was shown to be controlled by HDAC7, a member of class II HDACs, and highly homologous to the known SIK2 substrates HDAC4 and HDAC5 [1,13,14].The aims of this study were to evaluate whether changes in JUP expression affects insulin signalling and glucose uptake in adipocytes, and if, and by which mechanisms, SIK2 regulates JUP expression in these cells.
2. Materials and methods
2.1. Materials
3T3-L1 cells were from American Type Culture Collection. Cells were routinely found to be negative for mycoplasma contamination. DMEM, fetal bovine serum, penicillin/streptomycin, trypsin, PBS, dexamethasone, insulin, isobutyl-methyl-xanthine were all from Sigma. Pre-cast Novex SDS polyacrylamide Bis-Tris gels and LDS sample buffer were from Invitrogen. SuperSignal West Pico and Femto Chemiluminescent Substrates were from Thermo Scientific. RNeasy Mini Kit and Qiazol lysis reagent were from Qiagen. JUP and 18S QuantiTect primers (Qiagen) were used for PCR. HG9-91-01 and MC1568 were from MedChemExpress. Cuvettes for electroporation were from BioRad. 3H-2-deoxyglucose for glucose uptake was obtained from Perkin Elmer.
2.2. Antibodies
Primary antibodies used in this study were from Cell Signaling Technology (JUP, phospho-HDAC4 (Ser246)/5(Ser259)/7(Ser155), PKB, acetylated-lysine, CRTC3), Thermo Fisher (phospho-AS160 (Thr642)), Invitrogen (phospho-AS160 (Thr642)), Calbiochem (CRTC2), BD (HSP90), Sigma (actin), Innovagen (SIK2, generated inhouse as described in [15]).
2.3. Expression profiling in human adipose tissue
Transcriptome analysis of human subcutaneous white adipose tissue was performed by microarray in 56 women with varying body mass index (BMI) and metabolic status. Description of the cohort and microarray data has been published previously [16].
2.4. Culture, siRNA treatment, and stimulation of adipocytes
Murine 3T3-L1 fibroblasts were cultured and differentiated into adipocytes as previously described [17].Gene silencing in 3T3-L1 adipocytes was performed on day 8–12 after differentiation, by electroporation of 1–2 nmol of scrambled (Scr), SIK2, CRTC2 or JUP siRNA, all from Ambion, per 10 cm dish [4]. Seventy hours after electroporation, cells were stimulated as indicated in the figures, and lysed in LDS sample buffer including 75 mM DTT.Rat adipocytes isolated from 5-week old male Sprague Dawley rats (Taconic, Denmark) were stimulated and lysed as described previously [18]. Experiments with samples isolated from rats were ethically approved by Regional Ethical Committee on Animal Experiments in Malmö/Lund (ethical permit number
5.8.18–18,569/2018). Rats were housed in a humidity controlled room, with 12/12 h light/dark cycle and free access to food and water, in conventional shoebox cages with wood chip bedding
(2–4 rats per cage). For experiments, epididymal adipose tissue was isolated from rats at the local animal facility, after the animals were euthanized via CO2-sedation followed by decapitation. Adipose tissue was then transferred to the laboratory and used for isolation of adipocytes.Basal and insulin-stimulated uptake of 3H 2 deoxyglucose after JUP silencing for 96 h in 3T3-L1 adipocytes was measured as described previously [19].
2.5. RT qPCR
Cells were lysed in Qiazol lysis reagent (Qiagen) and RNA was extracted using the RNeasy Mini kit (Qiagen), according to the manual. One step realtime PCR was performed by using Quantifast SYBR Green RT-PCR kit (Qiagen) in a StepOnePlus Real Time thermal cycler (Applied Biosystems). The results were normalized to the endogenous reference gene 18S ribosomal RNA (18S) and to the Scr siRNA-treated cells. The results were calculated by using the 2T(-ΔΔC) method [20].
2.6. Western blot analysis
Equal volumes of cell lysates in LDS sample buffer were heated for 1–2 min at 95 °C and loaded on pre-cast Novex Bis-Tris 4–12% polyacrylamide gels. Proteins were electrotransferred to nitrocellulose membrane, which was blocked for 30 min in 50 mM Tris·HCl pH 7.6, 137 mM NaCl, and 0.1% (w/v) Tween 20 (TBS-T) containing 10% (w/v) skimmed milk, and then incubated with primary antibodies for 16 h at 4 °C, in TBS-T containing 5% (w/v) BSA. Detection was performed with horseradish peroxidase-coupled secondary antibodies and ECL, using a BioRad Chemidoc CCD camera and the Image Lab software. Signal intensities were quantified and normalized against a loading control (actin or HSP90) on the same membrane.
2.7. Data presentation and statistics
Data are presented as means −/+standard deviation (SD) throughout. In order to quantitate effects and to compare and present results from multiple independent experiments, western blot data was normalized to a loading control, and then expressed as fold or % of a specific condition within each experiment. The statistical significance of differences was tested by unpaired or paired t-test, or by one-way ANOVA with Holm-Sidak’s correction for multiple testing – as indicated in the legends. Correlations between mRNA expression of selected genes were
investigated by Pearson correlation test.
Fig. 1. Pharmacological inhibition or siRNA silencing of SIK2 leads to a reduction in JUP mRNA and protein expression. A. Correlation between mRNA levels of JUP and SIK2 in subcutaneous adipose tissue from 56 female individuals [16]. B. mRNA level of JUP in 3T3-L1 adipocytes treated with HG-9-91-01 (10 μM). CeF. Protein expression of JUP, HDAC4 (phospho-Ser246) and SIK2 in 3T3-L1 adipocytes treated with HG-9-91-01 (C; 24 h, D; 10 μM) or with SIK2 siRNA (E), and in primary rat adipocytes treated with HG-9-91-01 for 24 h (F). Hepatic progenitor cells Actin and HSP90 were used as loading controls. The means +SD from multiple independent experiments (B, D, F; n = 3, C, E; n = 4) are shown. Statistical significance of differences was analysed by unpaired t-test (E) or one-way ANOVA (B-D, F).
3. Results and discussion
3.1. SIK2 is required for the expression of JUP in adipocytes
To test whether there is a link between SIK2 and JUP, correlation analysis of mRNA levels for these two genes in human adipose tissue was performed. Indeed, JUP expression correlated positively with that of SIK2 in subcutaneous adipose tissue from 56 female human subjects (Fig. 1A, cohort previously described in [16]. Interestingly, both JUP and SIK2 mRNA expression displayed a positive correlation with Akt1/2 (PKBα/β) and TBC1D4 (encoding Rosiglitazone chemical structure the Rab-GAP AS160) – critical mediators of insulin-induced glucose uptake in adipocytes (Supplementary (S) Fig. 1A-E) [19,21].To explore if JUP is a downstream target for SIK2, we modulated the activity and expression of SIK2 in the 3T3-L1 adipocyte model. As shown in Fig. 1B,JUP mRNA levels were
reduced following 3 and 6 h of treatment with 10 μM of the selective SIK inhibitor HG-9-91-01, which inhibits the activity of all three SIK isoforms [22]. Furthermore, JUP protein
expression was reduced by HG-9-91-01 (24 h) in a dose-dependent manner (Fig. 1C), and the downregulation of JUP after treatment with 10 μM of the inhibitor was evident already after 6 h (Fig. 1D). A reduction of the phosphorylation of HDAC4 at Ser246 demonstrated the efficiency of HG-9-91-01 to inhibit SIK activity (Fig. 1C,D, F). To evaluate the specific contribution of SIK2 to the effect observed after inhibition of all three SIK isoforms, SIK2 was silenced by siRNA and JUP expression was monitored. Indeed, SIK2 silencing resulted in a marked decrease in JUP protein levels by almost 90% compared to Scr siRNA-treated cells (Fig. 1E). When analyzing JUP mRNA expression after SIK2 siRNA, there was on average no effect (data not shown). However, we found that the level of remaining SIK2 protein expression after SIK2 silencing tended to correlate with JUP mRNA (Fig. S2). This indicates that SIK2 might regulate JUP transcription but that JUP mRNA levels are maximally affected by SIK2 silencing at a time point earlier than the one we investigated (72 h). It is also possible that SIKs, in addition to controlling JUP transcription, regulate JUP at a post-transcriptional level. The reduction in JUP expression following HG-9-91-01 (10 μM, 24 h) or SIK2 siRNA varied somewhat (HG-9-91-01; 65–92% inhibition, SIK2 siRNA; 82–92%), but was nevertheless in a similar range. This lead us to conclude that SIK2 is likely to be the main SIK isoform responsible for regulation of JUP expression. This is in line with SIK2 being highly expressed, relative to SIK1 and SIK3, in adipocytes [3].Lastly, to address whether JUP expression is regulated by SIK2 in a more physiologically relevant cell model, we analysed JUP protein levels in primary rat adipocytes treated with 3 and 10 μM of HG-9-91-01 for 24 h. As shown in Fig. 1F, JUP expression was decreased by approximately 50%, when inhibiting SIKs in these cells.Together, these data indicate that JUP expression is regulated by SIK2 in adipocytes, most likely at the transcriptional level. The fact that SIK2 and JUP mRNA expression were positively correlated in human subcutaneous adipose tissue is interesting, and compatible with SIK2 being upstream of JUP also in humans. In the future, it would be important to test this idea by manipulating SIK2 expression/activity in human adipocytes.
Fig. 2. Pharmacological inhibition of SIK isoforms results in reduced insulin signalling. Protein expression (A) and analysis (B-E) of PKB (phospho-Ser473), PKB and HDAC4 (phosphoSer246) in 3T3-L1 adipocytes treated with HG-9-9101 (10 μM, 24 h) and stimulated with insulin (Ins). Equal amounts of protein lysates were loaded on two identical gels, and one gel was probed for phosphorylated proteins and the second gel was probed for the total corresponding proteins. For each gel, HSP90 was probed and used as a loading control for analysis. The means +SD from multiple independent experiments (n = 5) are shown. Statistical significance of differences was analysed by unpaired ttest (C, D) or one-way ANOVA (B, E).
3.2. SIK2 inhibition results in decreased activation of PKB and attenuated insulin signalling in 3T3-L1 adipocytes.
We have previously shown that modulation of SIK2 affects insulin signalling and glucose uptake in primary rat and in in vitro differentiated human adipocytes [3,4]. In line with this, treatment of 3T3-L1 adipocytes with the SIK inhibitor HG-9-91-01 for 24 h resulted in reduced phosphorylation of PKB at the activity-controlling site Ser473, both in the basal state and in
the presence of insulin (Fig. 2A, B). As a result of the reduced basal phosphorylation, the fold induction in response to insulin was increased after SIK inhibition (Fig. 2C).
Phosphorylation of PKB on both Ser473 and on Thr308 is critical for maximal activation of PKB. Ser473 phosphorylation stabilizes Thr308 phosphorylation and ablation of Ser473 phosphorylation, will drastically decrease PKB activity [23]. In this study, we evaluated phosphorylation status of Ser473 site in this study. Interestingly,we found the protein expression of PKB, measured by western blot with an antibody detecting all three PKB isoforms, to be significantly reduced by 30% in HG-9-91-01-treated adipocytes (Fig. 2A, D). This downregulation most likely accounts for the reduction in phosphorylated PKB because when normalizing against PKB expression, there was no significant difference in the phosphorylation of PKB following HG-9-9101, compared to control cells (Fig. 2E). The efficiency of HG-9-91-01 to inhibit SIK activity was again confirmed by a reduction of HDAC4 Ser246 phosphorylation (Fig. 2A). Since the treatment of these cells with HG-9-91-01 for 24 h resulted in a significant reduction in actin expression (data not shown), HSP90 was used as loading control in these experiments.
In our earlier work, we observed a marked reduction in the phosphorylation of PKB, without effects on expression, when inhibiting SIKs in in vitro differentiated human adipocytes [3]. In the current study, however, employing the SIK inhibitor HG-9-91-01 in 3T3-L1 adipocytes, the reduction in PKB phosphorylation appeared to be due to a downregulation of PKB expression. Apart from the difference in species of the cell models used, another reason to these slightly contrasting results may be related to the duration of SIK inhibition; 16 h in our previous study [3] and 24 h in the current. Interestingly, when analyzing the expression of SIK2 in human subcutaneous adipose tissue, we found a positive correlation between SIK2 and Akt1 (PKBα), suggesting that SIK2 might regulate PKB expression also in humans, and that further studies in human adipocytes are warranted. Collectively, these data indicate that SIK2 affects insulin signalling in adipocytes, via regulation of insulin signalling genes, such as PKB/Akt.
3.3. JUP is required to maintain insulin-induced signalling and glucose uptake in adipocytes
To our knowledge,our study is the first to address the role of JUP in adipocytes. Based on its well-established function in cell-cell junctions, JUP has mainly been studied in relation to adhesion and metastasis of tumour cells [10]. However, a recent study reported that JUP overexpression promoted insulin signalling in muscle cells and the proposed mechanism involved direct interaction of JUP with the insulin receptor (IR) and the regulatory p85 subunit of phosphatidylinositol 3-kinase (PI3K) – leading to an activation of the pathway [11]. To explore the role of JUP in adipocytes, we silenced JUP by siRNA in 3T3-L1 adipocytes (Fig. 3). The treatment with JUP siRNA resulted in a 70% silencing of JUP protein levels (Fig. 3A, B) as well as a significant reduction in the phosphorylation of PKB in the presence of insulin – without affecting basal phosphorylation (Fig. 3A, C). As a result, the fold induction of PKB phosphorylation in response to insulin was significantly reduced in JUP-silenced cells (Fig. 3D). Similar to what was observed after SIK inhibition,JUP silencing was associated with a significant 35% decrease in PKB protein levels (Fig. 3A, E). There were no differences in the specific PKB phosphorylation, i.e. phosphorylated PKB/total PKB, between Scr and JUP siRNA-treated cells (Fig. 3F). To explore whether the reduction in PKB expression and insulin-induced phosphorylation was translated into effects downstream of PKB, we monitored the phosphorylation of the Rab-GAP AS160 on Thr642 – a site known to be phosphorylated by PKB and to be involved in the regulation of glucose uptake by insulin [21]. Indeed,the phosphorylation of AS160 in the presence of insulin was markedly reduced after JUP silencing (Fig. 3A, G), as was the fold insulin-induced increase (Fig. 3H). Furthermore, AS160 protein expression was reduced by 28% in JUP siRNA-treated cells compared to Scr (Fig. 3A, I). When normalized to total AS160, the phosphorylation of AS160 Thr642 was slightly decreased in JUP-silenced cells, however without a statistically significant difference (Fig. 3J). Lastly, we investigated whether reduced insulin signalling following JUP siRNA resulted in effects on insulin-induced glucose uptake. Indeed, glucose uptake in response to 0.1 nM insulin was blunted by JUP silencing and significantly lower than that in Scr siRNAtreated cells (Fig. 3K). These results collectively show that JUP is required to maintain normal expression, and thus activity, of certain insulin signalling components and thereby insulin-induced glucose uptake.
An outstanding remaining question is whether regulation of JUP expression by SIK2 mediates, partly or fully, the effect of SIKs on insulin signalling. JUP silencing to a large extent mimicked the effect of HG-991-01 on insulin signalling, which is in line with JUP being a mediator of SIK effects. However, a finding that suggests additional mechanisms downstream of SIKs, at least in human cells, is that 1 h of HG-9-91-01 treatment was sufficient to inhibit insulin-induced glucose uptake in cultured human adipocytes, as shown in our previous work [3]. This effect is not likely to be mediated through changes in JUP transcription levels, but instead indicates that SIK isoforms regulate insulin signalling through several mechanisms, some of them on the post-transcriptional level. In the future, it would be interesting to overexpress JUP in HG-991-01-treated cells, to see if this leads to a reversal of the effect of SIK inhibition on insulin signalling.
A future challenge is to identify the mechanism(s) whereby JUP regulates PKB and AS160 expression. Interestingly, we found a correlation between JUP and Akt1/2 (PKBα/β) mRNA and TBC1D4 (AS160) mRNA expression in human adipose tissue. This indicates that the effect of JUP on PKB and AS160 expression maybe transcriptional and that it also applies to humans. Several studies have reported the involvement of JUP in the regulation of transcription. This function is believed to be due to its homology with β-catenin and thus its ability to compete with β-catenin for binding to various proteins, including the transcription factor TCF/LEF [12,24,25]. The binding of JUP to TCF/LEF can either activate or suppress gene transcription, depending on the targets and cellular context [12,24,25]. The β-catenin/TCF/LEF pathway is extensively studied in relation to its ability to regulate the transcriptional program of pre-adipocytes and to inhibit their differentiation into mature adipocytes [26,27]. The role and target genes of TFC/LEF in mature adipocytes are much less clear, but deletion of β-catenin specifically in adipocytes resulted in changes in the expression of a number of adipogenic, lipogenic and lipolytic genes [28]. Collectively, this indicates that the effects downstream of JUP could be mediated via changes in β-catenin signalling. It would thus be interesting to study whether β-catenin localization or expression is altered, and if modulation of β-catenin could rescue the downregulation of PKB, following SIK2 inhibition or JUP silencing.
3.4. The role of class IIa HDACs and CRTC2 in the regulation of JUP by SIK2
Class IIa HDACs (4/5/7) and CRTCs function as transcriptional coregulators, and are inhibited by SIK2-mediated phosphorylation in adipocytes [3,4]. To address whether any of these SIK2 substrates mediate the effect of SIK2 on JUP expression, we employed a combination of pharmacological inhibition and silencing of class IIa HDACs, CRTC2 and SIK(2) in 3T3-L1 adipocytes.To selectively inactivate all class IIa HDACs, we used the compound MC1568, which was previously demonstrated to inhibit only the activity of class II HDACs, and have no effect on class I HDACs [29]. Although the deacetylase activity of class II HDACs is very low compared to the one of class I HDACs [14], this activity might be important for its function. For example, deacetylase activity of HDAC4 was shown to be required for the recruitment of HDAC3, a major histone deacetylase, to the histone deacetylation complex [30]. The efficiency of MC1568 was investigated using an antibody recognizing acetylated lysine residues. The most prominent band on the western blot migrated at a size comparable with that of histone 3, which has been used before as an indicator of MC1568 efficiency [31]. The signal intensity of this band increased in the presence of MC1568 for 24 biogas slurry h (Fig. 4A, S3a). In line with our previous data (Fig. 1C, D), SIK inhibition resulted in reduced phosphorylation of HDAC4 (Fig. 4A) and a significant decrease in JUP protein expression (Fig. 4A, B). Treatment with MC1568 did not affect JUP levels in control cells (without HG-9-91-01), however, in HG9-91-01-treated cells, it resulted in a partial reversal of JUP expression (Fig. 4A, B). Class IIa HDAC inhibition did not reverse JUP downregulation in cells treated with SIK2 siRNA (Fig. S3b). The different effect of MC1568 in the context of inhibition of all SIK isoforms (SIK1, SIK2, SIK3), induced by HG-9-91-01 [22] versus SIK2 silencing, suggests that SIK1 or SIK3 to some extent contribute to the regulation of JUP, and that these isoforms may act through class IIa HDACs. The effect of SIK2 however, appears to be mediated via other mechanisms, such as for example changes in CRTC2.
The contribution of CRTC2 to SIK2-mediated regulation of JUP was investigated by treating 3T3-L1 adipocytes with CRTC2 siRNA, which resulted in a 69% reduction of CRTC2 protein levels (Fig. 4C, D). As observed in our previous experiments, SIK2 silencing significantly reduced JUP expression, and interestingly, so did siRNA against CRTC2 (Fig. 4C). When simultaneously silencing SIK2 and CRTC2, JUP expression was further reduced. We also noted that in SIK2-silenced cells,there was a consistent and significant reduction in CRTC2 protein levels, indicating that SIK2 is required to maintain CRTC2 protein expression (Fig. 4C, D). In line with this, inhibition of SIK isoforms with HG-9-91-01 resulted in a marked reduction in CRTC2 expression (Fig. 4E, S3c). These results suggest that, while SIK2 inhibits CRTC2 activity, at the same time it appears to be required to maintain its expression. We also noted downregulation of CRTC3 following HG-9-9101, indicating that SIKs also regulates this isoform (Fig. S3d). Collectively, this data suggests that SIK2 regulates JUP indirectly, through changes in CRTC2 and/or CRTC3 expression. Interestingly, according to computational prediction of CREB binding sites, JUP is expected to be a CREB target gene [32]. In the future, over-expression of CRTC2, or constitutively active CREB, could be employed to see if this prevents JUP downregulation induced by SIK2 silencing.In summary, our data indicate that CRTC2 or CRTC3 might mediate effects downstream of SIK2, to control JUP expression. Furthermore, class IIa HDACs appear to partially mediate the effect of SIK isoforms on JUP expression, but not the SIK2-specific one.
Fig. 4. Class IIa HDACs and CRTC2 contribute to the effect of SIK2 on JUP expression. A-B. Protein expression (A) and analysis (B) of JUP, HDAC4 (phospho-Ser246), and acetylated lysine (Ac-Lys) in 3T3-L1 adipocytes treated with HG-9-91-01 (10 μM, 24 h) and MC1568 (10 μM, 24 h). C-E. Protein expression and analysis of JUP (C), SIK2 (C), and CRTC2 (C, D, E) in 3T3-L1 adipocytes treated with SIK2 siRNA, CRTC2 siRNA (C, D) or with HG-991-01 (10 μM, 24 h) (E). HSP90 was used as a loading control in HG-9-91-01-treated cells (A, B, E), and actin in siRNA experiments (C, D). The means +SD from n = 4 independent experiments are shown. Statistical significance of differences was analysed by unpaired t-test (B, E) or one-way ANOVA (C, D).
4. Conclusions
In conclusion,we have shown that JUP is a novel gene controlled by SIK2 in adipocytes, and that it plays a role in the regulation of insulin signalling and glucose uptake in these cells. The effect of SIK isoforms on insulin signalling might be partially mediated by alterations of JUP expression, but this needs to be further explored. Moreover, our data indicate that CRTC2 is a possible mediator of SIK2-induced regulation of JUP (Fig. 5).The role of JUP in insulin signalling and glucose uptake in adipocytes makes JUP an interesting molecule in relation to obesity-induced insulin resistance. Future studies should aim to characterize the expression of JUP in human adipose tissue and cells, as well as mechanisms upstream and downstream of JUP in adipocytes.Salt-inducible kinases (SIK1, SIK2, SIK3) regulate JUP expression via CRTCs and possibly class II HDACs. JUP is required for maintaining expressionand phosphorylation levels of the insulin signalling proteins PKB and AS160, and thus for insulin-induced glucose uptake. JUP might be a mediator of the effects of SIK2 on insulin signalling in adipocytes.