Introduction

Type 2 diabetes mellitus (T2DM) is a chronic disease, characterized by hyperglycemia and relative insulin deficiency that affects more than 400 million people worldwide [1]. Despite the wide range of available antidiabetic drugs, side effects, drug resistance, and patient heterogeneity highlight the urgent need for new therapeutics with novel mechanisms to address clinical needs [2]. Eukaryotes possess a conserved signaling pathway that responds to low cellular ATP (due to reduced oxygen or glucose) by activating adenosine monophosphate-activated protein kinase (AMPK) [3], which then phosphorylates key targets to regulate lipid and glucose metabolism [4]. Given that AMPK is essential for glucose homeostasis [3], it represents a key therapeutic target for diabetes and related metabolic diseases. Mechanistically, AMPK is activated when its α-subunit is phosphorylated at Thr172 by upstream kinases via energy- or Ca²⁺-dependent pathways [3, 5]. Once activated, AMPK suppresses anabolic (energy-consuming) processes and promotes catabolic (energy-producing) ones [6]. In nutrient-rich conditions, ATP binds to the γ subunit of AMPK and keeps it inactive. Under energetic stress, AMP replaces ATP, triggering liver kinase B1 (LKB1) to phosphorylate the α subunit of AMPK.

Notably, cytoplasmic Ca²⁺ can activate AMPK independently of cellular energy status [5, 7]. Lysosomes, mitochondria and endoplasmic reticulum (ER) are the main intracellular Ca²⁺ stores [8] and some connections between distinct cytosolic Ca²⁺ sources and AMPK activation have been investigated [9]. Nevertheless, the involvement of lysosomes in AMPK activation and its implications for treating diabetes and metabolic diseases have yet to be elucidated. Transient Receptor Potential Mucolipin 1 (MCOLN1/TRPML1) is the best characterized lysosomal Ca²⁺ channel, mediating Ca²⁺ release from the lumen to the cytosol. TRPML1 dysfunction underlies various metabolic diseases, lysosomal storage diseases and neurodegenerative disorders [10–12]. For example, mutations in the gene encoding TRPML1 cause mucolipidosis type IV (MLIV), a lysosomal storage disease characterized by early-onset progressive neurodegeneration [13]. Lysosomal dysfunction is closely associated with multiple metabolic diseases, including obesity, T2DM, non-alcoholic fatty liver disease, and atherosclerosis [14, 15]. Impaired lysosomal function disrupts lipid metabolism, autophagy, and inflammation, leading to insulin resistance and β-cell failure [16].

Previous studies have shown that the synthetic TRPML1 agonist MK6-83 activates AMPK in human fibroblasts [9]. However, whether the TRPML1–AMPK axis regulates glucose metabolism, particularly in the context of T2DM, remains unclear. To address this, the current study investigates how TRPML1 activation regulates glucose metabolism, and ML-SA8, a TRPML1 agonist with higher in vitro potency (ML-SA8>ML-SA5 > MK6-83) [17–19], was applied to elucidate the therapeutic effects and molecular mechanisms of TRPML1 activation in T2DM models.

In the current proof-of-concept study, we found that both genetic and pharmacological activation of the lysosomal TRPML1 channel are sufficient to activate AMPK and alleviate diabetic phenotypes in vitro and in vivo. Therefore, we propose that pharmacological activation of TRPML1 may serve as a potential therapeutic strategy for T2DM.

Results

Pharmacological activation of lysosomal TRPML1 induces AMPK activation

Cytosolic Ca²⁺ reportedly activates AMPK [20, 21]. Lysosome, mitochondria and ER are the major compartmentalized Ca²⁺ stores in cells [8]; the mechanistic links between cytosolic Ca²⁺ and AMPK activation lead us to consider the discrete Ca²⁺ sources as a potential specificity determinant of AMPK activation [22, 23]. In this study, we investigated whether lysosome-released ions can activate AMPK, whose activation was measured by the phosphorylation status of threonine 172 (referred as Thr172) [24].

TRPML1 is the principal Ca²⁺-permeable channel that releases Ca²⁺ from the lumen into the cytosol upon cellular stimulation [25, 26].Thus we then investigated whether TRPML1 contributes to AMPK activation. Although the synthetic TRPML1 agonist MK6-83 was previously shown to activates AMPK in human fibroblasts [9], here we used a much more potent TRPML1 agonist, ML-SA8 [27] (EC₅₀=0.086 ± 0.013 μM for endogenous TRPML1; Suppl. Fig. 1a, b). As shown in Fig. 1a, b, ML-SA8 (0.3-3 μM, 2 h) induces robust AMPK phosphorylation in a dose-dependent manner in HepG2 cells as well as several other cell lines including HeLa (Suppl. Fig. 1c, d) and HAP1 (Suppl. Fig. 1e, f). And this activation by ML-SA8 was completely blocked by dorsomorphin/compound C (CC, 10 μM, an AMPK inhibitor [28]) (Fig. 1c, d). Similarly, AMPK can be activated by another less potent TRPML1 agonist ML-SA5 (potency SA8> SA5>MK6-83 [18, 27, 29]) in HepG2, HeLa and HAP1 cells (Suppl. Fig. 2a-c). Note that the specificities of ML-SAs have been previously validated using TRPML1 knockout (KO) cells [18, 29] and confirmed in the atomic-resolution co-structures [30, 31]. In addition, we investigated the role of another lysosomal Ca²⁺ channel, TPC2, using its specific agonist riluzole [25]. As shown in Suppl. Fig. 3a, b; AMPK activation required 800 μM of riluzole, a concentration exceeding its physiological range [32–35]. Therefore, we focused our research on ML-SA8-induced AMPK phosphorylation.

Figure 1.

Glucose transporters (GLUTs) at the plasma membrane reportedly facilitate glucose uptake by an inward diffusion gradient for glucose [36]. In particular, GLUT4 is known to participate in AMPK-mediated glucose uptake by translocating from cytoplasm to plasma membrane [37]. Although GLUT4 is not the predominant GLUT isoform in the liver, it is nevertheless expressed in this organ [38–40]. Therefore, we next investigated whether ML-SA8 regulates the translocation of GLUT4. As shown in Fig. 1e, ML-SA8 (1 μM) induced a substantial increase in GLUT4 protein at the plasma membrane (PM) of HepG2 cells by immunofluorescence, indicating that ML-SA8 promotes GLUT4 translocation to the PM. Moreover, we isolated PM and post-PM fractions (supernatants devoid of PM) from HepG2 cells treated with ML-SA8 or vehicle control. Quantification of GLUT4 levels in the PM fraction confirmed that ML-SA8 stimulated GLUT4 translocation (Fig. 1f, g). Moreover, ML-ML-SA8 upregulated the expression level of GLUT4 in HepG2 cells (Suppl. Fig. 4). Collectively, these finding indicate that ML-SA8 promotes both PM translocation and the expression of GLUT4.

We then studied the effect of ML-SA8 on glucose levels by measuring both cellular glucose uptake and glucose consumption in the medium [41]. First, we constructed a palmitic acid (PA)-induced insulin resistance (IR) cell model [42, 43]. As shown in Suppl. Fig. 5a, b, insulin (100 nM, 0.5 h) induced glucose influx in HepG2 cells measured by 2-deoxy-2-[(7-nitro-2,1,3-benzoxadiazol-4-yl)amino]-d-glucose (2-NBDG), a fluorescent glucose analog, which is transported into cells and phosphorylated by hexokinase, trapping it intracellularly and allowing fluorescence-based quantification of glucose uptake [44]. In contrast, in PA (0.25 mM, 12 h)-treated HepG2 cells, insulin (100 nM, 0.5 h)-induced glucose influx was almost completely blocked, suggesting that PA-treated HepG2 cells are less sensitive to insulin (Thereafter IR-HepG2 represents PA-induced insulin resistant HepG2 cells). ML-SA8 (1 μM, 2 h)-induced AMPK phosphorylation was validated in IR-HepG2 cells by western blot analysis (Suppl. Fig. 5c). In IR-HepG2 cells, we observed that upon ML-SA8 (0.1-3 μM, 2-6 h) treatment, the intensities of 2-NBDG were dramatically increased compared to control in a dose- (Fig. 1h, i) and time-dependent (Suppl. Fig. 5d, e) manner. Correspondingly, glucose consumption in the culture medium was dramatically decreased with ML-SA8 (1 μM, 4 h) treatment in IR-HepG2 cells (Fig. 1j). Likewise, ML-SA8 (1 μM) was sufficient to induce cellular glucose uptake in various cells such as human hepatocellular carcinoma cells (Huh7) and mice myoblast cells (C2C12) following IR (PA-treated) induction (Suppl. Fig. 5f, g). Taken together, these data demonstrated that stimulation of TRPML1 activates AMPK signaling cascades, which facilitates GLUT4 PM translocation, leading to the increases of cellular glucose uptake.

TRPML1 is essential in ML-SA-induced AMPK activation

We then investigated whether increasing the expression of TRPML1 in HepG2 cells is sufficient to increase the cells’ sensitivity to ML-SA8 treatment. Remarkably, a low dose of ML-SA8 (0.03 μM) can activate AMPK in HepG2 cells overexpressing GFP-TRPML1, but not in wild type (WT) cells (Fig. 2a). Consistently, ML-SA8 (0.03 μM) can also promote cellular glucose uptake in TRPML1-overexpressing IR-HepG2 cells (Fig. 2b, c), while no such effects were observed in IR-HepG2 cells. Collectively, these data demonstrate that the TRPML1 overexpression can promote AMPK activation.

Figure 2.

Next, we investigated whether TRPML1 is required for ML-SA-induced AMPK activation using the TRPML1 specific synthetic inhibitor-ML-SI5, whose specificities have been previously validated [17, 27]. Pretreatment with ML-SI5 (3 μM) completely abolished ML-SA8-triggered (1 μM, 2 h) AMPK phosphorylation, while no effect on metformin (MET)-induced AMPK activation (Fig. 2d, e). Moreover, in IR-HepG2 model, ML-SI5 substantially inhibited ML-SA8-, but not MET-induced cellular glucose uptake and extracellular glucose consumption (Fig. 2f-h). To further validate the TRPML1-dependent mechanism, we utilized a TRPML1 knockout (KO) HAP1 cell line constructed by CRISPR-Cas9 as previously described [45]. In TRPML1 KO HAP1 cells, ML-SA8 (1 μM, 2 h)-induced AMPK phosphorylation and cellular glucose uptake were almost completely abolished compared to wild-type (WT) HAP1 cells (Fig. 2i-l). Collectively, these data demonstrate that TRPML1 is both sufficient and necessary for ML-SA-induced AMPK activation in IR model cells.

Ca²⁺ -CaMKKβ -dependence of ML-SA-induced AMPK activation

Lysosomal TRPML1 is a non-selective cation channel that is permeable to Ca²⁺, as well as heavy metal ions such as Fe²⁺ and Zn²⁺ [46]. Given the well-established role of Ca²⁺ in AMPK activation [47], we first examined whether Ca²⁺ is required for ML-SA’s AMPK activation upon TRPML1 stimulation. BAPTA-AM (20 μM), a membrane-permeable Ca²⁺ chelator (fast-acting) [48] readily blocked ML-SA8-evoked AMPK activation (Fig. 3a, b) and cellular glucose uptake (Fig. 3c, d). In contrast, EGTA-AM, another Ca²⁺ chelator (slow-acting), failed to block ML-SA8’s effects (Fig. 3a-d). The inhibitory effects of BAPTA-AM (20 μM) and EGTA-AM (20 μM) on cytosolic Ca²⁺ was confirmed in HEK293 cells overexpressing GCaMP7-TRPML1 (Suppl. Fig. 6a). Considering that BAPTA has much faster Ca²⁺ association and dissociation kinetics than EGTA [49], this disparate Ca²⁺-dependent mechanism suggests that ML-SA8-regulated AMPK cascades through a fast and local Ca²⁺ signal from lysosomes. Next, we examined whether Fe²⁺ and Zn²⁺ involve in ML-SA-triggered AMPK cascades. TPEN (10 μM), a high-affinity Zn²⁺ chelator [50], and deferoxamine (DFO) (10 μM), a Fe²⁺ chelator [51], both failed to block ML-SA8-induced AMPK activation and cellular glucose uptake (Suppl. Fig. 6b-e). Together, these data suggest that TRPML1 activation regulates AMPK-GLUT4-glucose pathway via a fast Ca²⁺ release from lysosomes.

Figure 3.

We next investigated how ML-SA8 induces AMPK phosphorylation. A key mechanism of AMPK activation involves phosphorylation of its subunit at Thr172 site by upstream kinases, including CaMKKβ (calcium/calmodulin-dependent protein kinase kinase β which responds to elevated cytosolic Ca²⁺ [52, 53] and LKB1 [7]. Considering our finding that TRPML1-dependent Ca²⁺ release is essential for AMPK activation (Fig. 3a-d), we hypothesize that CaMKKβ serves as the critical upstream kinase bridging lysosomal Ca²⁺ signaling and ML-SA8-stimulated AMPK activation. Notably, STO609 (a CaMKKβ inhibitor, 10 μM) [54] significantly suppressed ML-SA8-induced AMPK phosphorylation in HepG2 cells (Fig. 3e, f). Furthermore, the effects of ML-SA8 on cellular glucose uptake (2-NBDG) were completely blocked by STO609 or CC (Fig. 3g, h). To further validate these findings, a genetic knockdown approach was employed. Consistent with the pharmacological inhibition, CaMKKβ knockdown markedly attenuated ML-SA8-induced AMPK phosphorylation (Fig. 3i, j) and abolished the associated increases in glucose uptake and extracellular glucose consumption (Fig. 3k–m). Collectively, these findings establish CaMKKβ as an essential upstream mediator required for AMPK activation and the improved glucose metabolism following ML-SA8 treatment.

Overall, these in vitro experiments demonstrate that pharmacological activation of TRPML1 stimulates AMPK via a Ca²⁺-CaMKKβ-dependent pathway, thereby promoting GLUT4 translocation to the PM. This process enhances intracellular glucose uptake and helps maintain glucose homeostasis in a hepatic IR model (Fig. 3n).

TFEB-independence of ML-SA-induced AMPK activation

TRPML1 activation is known to upregulate lysosomal biogenesis through TFEB activation [48, 55, 56]. While AMPK reportedly mediates phosphorylation of TFEB, and its transcriptional activity [57]. TFEB belongs to the microphthalmia (MiT/TFE) family, which also contains TFE3 and MITF, contributing to autophagic and lysosomal biogenesis [55, 58]. Hence, we examined whether TFEB/TFE3/MITF participates in TRPML1 stimulation-induced AMPK activation by using TFEB/TFE3/MITF triple KO (TKO) cells. The efficiency of TFEB/TFE3/MITF TKO was confirmed by Western blotting with anti-TFEB, TFE3 and MITF specific antibodies (Suppl. Fig. 6a). We found that in TKO cells, upon ML-SA8 (1 μM, 2 h) treatment, the increase of intensity of p-AMPK/total AMPK was not affected (Suppl. Fig. 7b, c), suggestive of TFEB/TFE3/MITF-independence of ML-SA8-triggered AMPK activation.

Next, we also assessed the role of AMPK in TRPML1 activation-mediated TFEB activity. Consistent with previous studies [18], ML-SA8 treatment (1 μM, 4 h) led to GFP-TFEB translocation from cytosol to nuclei in HeLa GFP-TFEB stable cells (Suppl. Fig. 7d, e), and this effect can be blocked by ML-SI5 treatment. However, CC or STO609 treatment failed to block ML-SA8-induced TFEB nuclear translocation (Suppl. Fig. 7d, e), suggesting that AMPK is dispensable for TRPML1 activation-induced subcellular TFEB translocation. Taken together, these results suggest that ML-SA8 -induced TFEB nuclear translocation is AMPK-independent.

Administration of ML-SA8 ameliorates diabetic phenotypes in db/db mice

Hepatic steatosis is strongly associated with T2DM and is considered as an exacerbating factor of hepatic insulin resistance. We next investigated whether the small-molecule TRPML1 agonist ML-SA8 could alleviate diabetic phenotypes in C57BLKS/J-Lepr KO (db/db) mice, a well-established diabetic mouse model [59]. To determine whether ML-SA8 targets/activates AMPK in vivo, 9-week-old db/db mice were intraperitoneally (i.p.) injected with ML-SA8 (4 mg/kg) or vehicle (5% DMSO, 90% PEG300, 5% ddH₂O). After 24 h, liver tissues were collected and pT172-AMPK/AMPK levels were measured by western blotting. As shown in Fig. 4a, b, ML-SA8 treatment significantly increased pT172-AMPK protein levels in liver tissues compared to vehicle, demonstrating that ML-SA8 activates AMPK in the liver of db/db mice.

Figure 4.

We next evaluated the in vivo therapeutic efficacy of ML-SA8. 9-week- old db/db mice were randomly divided into three groups and treated daily for 6 weeks: 1) Vehicle group (DB+V): i.p. injection of vehicle (5% DMSO, 90% PEG300, 5 % ddH₂O); 2) ML-SA8 group (DB+SA8) : i.p. injection of ML-SA8 at 2, 4, 8 mg/kg. We further validated the targeting of AMPK in various tissues by 6-week treatment of ML-SA8. ML-SA8 induced AMPK phosphorylation levels in liver (Suppl. Fig.8a), as well as in adipose (Suppl. Fig.8b) and skeletal muscle tissues (Suppl. Fig.8c) in db/db mice. Moreover, GLUT4 protein expression significantly increased in liver of db/db mice treated with ML-SA8 by western blot (Fig. 4c, d) and immunofluorescence analysis (Suppl. Fig.8d). Furthermore, administration of SA8 (4 and 8 mg/kg) significantly reduced fasting blood glucose (FBG) levels (Fig. 4f), HbA1c levels (Fig. 4g) and random blood glucose (RBG) levels (Suppl. Fig.8e), oral glucose tolerance tests (OGTT) (Fig. 4h) and insulin tolerance test (ITT) (Fig. 4i). These data suggests that ML-SA8 reduced plasma glucose levels and increased the insulin sensitivity in db/db mice. Histological staining analysis showed that liver tissues in DB+V mice exhibited hepatic steatosis, lipid droplet accumulation, hepatocyte swelling and ballooning (Fig. 4j). In contrast, administration of SA8 (4 mg/kg) in DB mice was sufficient to restore a normal, well-defined, typical hepatocyte morphology (Fig. 4j). Furthermore, in the Periodic Acid-Schiff (PAS) staining used for glycogen deposition analysis, the glycogen levels in liver tissues of DB+SA8 group displayed a significant elevation compared with DB+V group (Fig. 4j), indicating that SA8 prevented reduced glycogen synthesis in db/db mice. Furthermore, ML-SA8 (4 mg/kg) was able to reduce serum alanine aminotransferase (ALT) (Fig. 4k) and aspartate aminotransferase (AST) levels (Fig. 4l). Notably, ML-SA8 treatment had no effects on body weight (Suppl. Fig.8f), total cholesterol (TC) (Suppl. Fig.8g), serum triglyceride (TG) (Suppl.Fig.8h) and serum creatinine (Suppl.Fig.8i). Collectively, these findings indicate that ML-SA8 alleviates the hallmarks of diabetic phenotypes in db/db mice.

Discussion

Lysosomes, a cellular hub in metabolism, protein degradation, and nutrient sensing, play an indispensable role in numerous metabolic pathways and maintaining cellular homeostasis [60]. Emerging evidence implicates lysosomal calcium signaling as a pivotal yet understudied modulator of metabolic homeostasis. In the current study, we found that genetic or pharmacological activation of the lysosomal Ca²⁺ channel-TRPML1 exhibited glucose-lowering effects in both in vitro and in vivo diabetic models. TRPML1 activation led to AMPK phosphorylation via a Ca²⁺-CaMKKβ-dependent mechanism and promoted GLUT4 translocation, thereby facilitating glucose influx to reduce hyperglycemia and other phenotypes of T2DM (Fig. 4m). Hence, our proof-of-concept study highlighted TRPML1 as a functional therapeutic target for T2DM and provided evidence that small-molecule TRPML1 agonists may have broad therapeutic potential.

Lysosomal Ca²⁺ is a key regulator for various lysosomal functions i.e., lysosomal exocytosis, membrane trafficking, catabolite export, ion homeostasis and nutrient sensing [61]. Ca²⁺ in the lysosomal lumen is about ∼ 0.5 mM, approximately 5,000-fold higher than cytosolic Ca²⁺ (∼100 nM). TRPML1 is predominantly localizes on the membranes of late endosomes and lysosomes (LELs) in all mammalian cell types [62]. The related TRPML2 and TRPML3 channels are also permeable to Ca²⁺, but are more restrictively expressed [62]. TRPML channels play essential roles in various lysosomal functions and respond to a variety of stimuli, such as pH, nutrients and cellular stress, as well as to small molecules [46], suggesting that TRPML1 activities can be specifically modulated. The synthetic agonists (ML-SAs) and inhibitors (ML-SIs) that specifically targeting TRPML1 directly bind to TRPML1 protein in the atomic-resolution co-structures [30, 31]. Notably, the concentration of ML-SAs used in this study has no cytotoxicity towards the cell lines we used including HepG2, HeLa and HAP1 cells (Suppl. Fig.9a, b).

Although GLUT4 is abundant in skeletal muscle, fat, and heart but low in the liver (where GLUT2 predominates) [38], it regulates hepatic glucose homeostasis [63]. Under basal conditions, GLUT4 resides intracellularly and translocates to the PM upon insulin or other stimuli to drive glucose uptake[63]. Given that GLUT4 mediates AMPK-dependent glucose uptake via translocation [37], we examined the effect of ML-SA8 on GLUT4 localization and expression. While our findings support ML-SA-mediated regulation of GLUT4, a GLUT2 - dominant contribution cannot be excluded.

The MiT/TFE family includes TFEB, TFE3, MITF and TFEC. TFEB, TFE3 and MITF, but not TFEC are master regulators of lysosome and autophagy by controlling the “coordinated lysosomal expression and regulation” (CLEAR) network, covering genes associated to lysosomal exocytosis and biogenesis, and autophagy [55, 64, 65]. TFEB/TFE3/MITF are phosphorylated by MTOR (mechanistic target of rapamycin kinase) and then bind to YWHA/14-3-3 in the cytosol inactively [66, 67]. Whereas, under stress conditions i.e. starvation, TFEB/TFE3/MITF are dephosphorylated and actively translocate into nucleus, promoting the expression of genes [56, 68, 69]. TRPML1-released Ca²⁺ release reportedly activates calcineurin, which binds and dephosphorylates TFEB, thus promoting its nuclear translocation [69]. Interestingly, we found that AMPK is not required for TFEB nuclear translocation induced by ML-SA8 (Suppl. Fig.7d, e). Furthermore, by using TFEB/TFE3/MITF TKO cells we found that knockout TFEB/TFE3/MITF has no effect on ML-SA8-induced AMPK activation (Suppl. Fig.7a-c). These results suggest that TRPML1-mediated activation of TFEB and AMPK may be through distinct independent pathways.

In summary, we show a novel AMPK activation pathway via the lysosomal TRPML1-Ca²⁺-CaMKKβ axis. Pharmacological activation of TRPML1 can efficiently diminish diabetic phenotypes in both in vitro and in vivo models by targeting AMPK. Hence, targeting lysosomal Ca²⁺ channels i.e. TRPML1 may represent a promising approach to treat T2DM and related metabolic diseases.

Methods and materials

Cell lines and cell culture

HeLa (RRID: CVCL-0030), HepG2 (RRID:CVCL_A8FT) and C2C12 (RRID: CVCL-0188) cells were purchased from American Type Culture Collection (ATCC). Huh7 (RRID: CVCL-0336) cells were purchased from Thermo Scientific. HepG2, Huh7 and C2C12 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Thermo Scientific, 11195-065) supplemented with 10% fetal bovine serum (Thermo Scientific, 10091148) and 1% Penicillin-Streptomycin (PS; Solarbio, P1400). HeLa cells were maintained in modified Eagle’s medium (MEM; Thermo Scientific, 1964643) supplemented with 10% FBS and 1% PS. GFP-TFEB HeLa stable, HAP1 (RRID: CVCL-Y019, Horizon Company), TRPML1 KO HAP1, HEK293-GCaMP7-TRPML1 and ATG5 KO HAP1 cells were maintained in our laboratory as previously described [45, 56, 70, 71] and cultured in Iscove’s Modified Dulbecco’s Medium (IMDM; Gibco, 6123179) supplemented with 10% FBS and 1% PS. All cells were used after in-house authentication for mycoplasma contamination and maintained at 37 °C in a humidified 5% CO₂ incubator. Cells with passage number less than 20 were used in this study.

Construction of insulin resistance (IR) model

PA (Sigma-Aldrich, P0500) was applied to construct the IR model [72]. Briefly, 9.2 mg PA was completely dissolved in 1ml of 0.1 M NaOH by warming to 80 ℃ until becoming transparent and clear. The PA solution was further diluted in fatty acid-free bovine serum albumin solution (BSA, 0.325 g BSA dissolved in 8 ml of 0.9 % NaCl by heating at 45℃). The two solutions were then mixed with gentle agitation to get final PA stock solution (4 mM), which was further sterilized by passing through 0.22 μm filters, aliquoted and stored at -20 ℃. Cells were cultured at approximately 70% confluency and treated with PA (0.25 mM, 12 h) to construct IR model, followed by chemical treatment as required conditions.

Plasmid transfection

Plasmids TRPML1-mCherry and GFP-TRPML1 were maintained in our laboratory [73]. Cells were seeded into 6-well plates and transfected with plasmids using Lipofectamine 3000 (Thermo Scientific, 2163785). The efficiency of the transfection was validated by Western blotting or Q-PCR.

RNA interference

The siRNA sequences targeting human CaMKKβ (5ʹ-GGA CCA UCU GUA CAU GGU GUU CGA A-3ʹ) were chemically synthesized by GenePharma. HepG2 cells were seeded into 6-well plates and transfected with siRNA using Lipofectamine RNAi-Max reagent (Thermo Fisher Scientific, 13778150). The efficiency of the siRNA knockdown was assayed by Q-PCR.

RNA extraction and Q-PCR

Total RNA was isolated from cells using Trizol (Thermo Scientific, 191002), and cDNA was generated using the GoScript Reverse Transcription System (Promega, 0000316057) according to the manufacturer’s protocol. Q-PCR was performed using the SYBR Green Mix (TOYOBO, 563700) in CFX Connect Optics (Bio-Rad). The changes in the mRNA expression of target genes were normalized to that of the housekeeping gene HPRT. Primer sequences used in this study are listed below:

HPRT; For 5ʹ-TGGCGTCGTGATTAGTGATG-3ʹ, Rev 5ʹ-CTGTTCTCGTCCAGCA GACACT-3ʹ

CaMKKβ; For 5ʹ- TCCAGACCACGACATAG -3ʹ, Rev 5ʹ-CAGGGGTGCAGCTTGATTTC-3ʹ

GCaMP7 Ca²⁺ imaging

GCaMP7 Ca²⁺ imaging was performed in HEK293 cells stably expressing GCaMP7-TRPML1, a lysosome-targeted genetically-encoded Ca²⁺ sensor [70]. Cells were excited at 488 nm, and the emission fluorescence was recorded at 510 nm using a FlexStation 3 scanning fluorometer. Fluorescence intensity at 488 nm excitation (F₄₈₈) was acquired at regular intervals under controlled temperature (e.g., 37 °C) and continuous perfusion with appropriate buffer. Data were normalized to baseline fluorescence (F₀) and presented as ΔF/F₀.

Immunofluorescence and confocal imaging

For immunofluorescence detection of GLUT4, cells were grown and stained on coverslips in 24-well plates. 4% paraformaldehyde (Solarbio, P1110) was added to each well and cells were fixed for 15 min at room temperature. 0.3% Triton X-100 (Sigma-Aldrich, T8787) was used for permeabilization for 10 min, followed by 1 h blocking with 1% bovine serum albumin (Sigma-Aldrich, WXBC2612V) in phosphate-buffered saline (PBS; Thermo Scientific, 21600069). The cells were incubated overnight at 4°C with anti-GLUT4 antibodies (Abcam, AB33780) (1:1000). After washing for 5 times with PBS, secondary antibodies conjugated to Alexa Fluor -488 (Thermo Scientific, 1705869) (1:1000) were added and incubated for 2 h in a dark place. DAPI (Sigma-Aldrich, D9542) was added to stain nuclei for 15 min and briefly washed for 3 times. Coverslips were then mounted with Fluoromount-G (Southern Biotech) and images were acquired with an Olympus or Zeiss confocal microscope.

Measurement of cellular glucose uptake

Cellular glucose uptake was measured using a fluorescence-labeled glucose analogue-2-NBDG (MCE, 186689-07-6) [44]. Briefly, cells were seeded about 70% confluency and treated with chemicals as required conditions. Cells were then washed and fresh DMEM (no glucose, no FBS, no phenol red, Solarbio, D6540) containing 50 μM 2-NBDG was added for 30 min, followed by PBS wash three times. The fluorescence was visualized with an Olympus IX-73 inverted microscope. Fluorescence intensity was quantified using Image J software.

Measurement of extracellular glucose consumption

Glucose consumption in the medium was measured using a glucose consumption kit (Nanjing Jiancheng Bioengineering Institute, A154-1-1) [74]. Cells were seeded in a 6-well plate at a density of 1 × 10⁶ cells/ml and treated with chemicals as required conditions in DMEM (no FBS, no phenol red) with 10 mM glucose. After treatment, an aliquot of supernatant medium was added to a 96-well plate. The absorbance was measured at 505 nm with a spectrophotometer VARIOSKAN LUX (Thermo Scientific). Glucose concentration was calculated as: [Medium Glucose] = [(Sample_absorbance-blank_absorbance) / (Standard_absorbance-blank_absorbance)] × 5.55 mmol/L × X, where 5.55 mmol/L is the concentration of the glucose standard, and X is the dilution factor.

Protein extraction and Western blotting

Cells or tissues were lysed in RIPA buffer (Solarbio, R0010) supplemented with 1x protease inhibitors cocktail (Sigma-Aldrich, P8340) (1: 100) and 1 x phosphatase inhibitor cocktail (Abcam, GR304037-28) (1: 100) on ice for 20 min. The membrane protein fraction was extracted using a membrane and cytoplasm protein extraction kit according to manufacturer’s protocol (Beyotime, P0033, Shanghai, China). Protein samples were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis gels, followed by transferring to polyvinylidene difluoride membranes (Merck, R7DA8778E). The membranes were blocked with 5% nonfat dry milk for 1 h and incubated with antibodies against AMPK (Cell Signaling Technology, 2532) (1:1000), p-AMPK (Cell Signaling Technology, 2535) (1:1000), GLUT4 (Abcam, AB33780) (1:1000), Vinculin (Abcam, AB8226) (1:2000), TFEB (Cell Signaling Technology, 4240S) (1:1000), TFE3 (HPA02300881) (1:1000), MITF (Abcam, AB20663) (1:1000) and GAPDH (Sigma-Aldrich, G9545) (1:10,000), respectively. Bound antibodies were detected horseradish peroxidase-conjugated anti-rabbit (Abcam, ab6721) or anti-mouse secondary antibodies (Abcam, ab6789) (1:10,000) in 5 % nonfat dry milk for 2 h at room temperature with agitation. Protein bands were visualized using enhanced chemiluminescence reagents (Thermo Scientific, 203-17071) in a Li-COR Biosciences Odyssey Fc system. Band intensities were quantified using the Image J software.

Animals

Specific pathogen free (SPF) grade male C57BLKS/J mice (8-week-old) and male C57BLKS/J-Lepr KO (db/db) mice were purchased from GemPharmatech Limited Company (Nanjing, China). Throughout the manuscript, “DB” denotes genetically diabetic db/db mice. Animals were handled according to approved institutional animal care and use committee (IACUC) protocols (20241226208) of Zhejiang University of Technology. Mice were acclimated for one week prior to the experiment. After the acclimation period, mice were randomly assigned to experimental groups. At the end of the experiments, all mice were euthanized by CO₂ inhalation followed by cervical dislocation to confirm death. Blood and liver tissues were rapidly collected for subsequent biochemical and molecular analyses. All experimental procedures were conducted in accordance with the Guide for the Care and Use of Laboratory Animals in the Zhejiang University of Technology, Hangzhou, China, and conformed to the National Institutes of Health Guide for Care and Use of Laboratory Animals.

Histopathological analysis (H&E and PAS staining)

Liver sections were fixed in 4 % paraformaldehyde buffer and embedded in paraffin. Liver slices were then deparaffinized, rehydrated, and stained with H&E (Solarbio, G1120) or PAS (Solarbio, G1281). All staining procedures were performed according to standard protocols followed by microscopy examination. The section images were obtained using the Olympus/Zeiss Fluorescence inverted microscope.

Blood glucose and insulin measurements

Oral Glucose Tolerance Test (OGTT) and Insulin Tolerance Test (ITT) were performed on each group of mice. For OGTT, mice fasted for 10 h received oral glucose (1g/kg). For ITT, mice fasted for 4 h received i.p. injection of insulin (0.75 U /kg). Tail blood glucose was measured at 0, 30, 60, 90, and 120 min post-administration. Random blood glucose (RBG) was measured every four days (9:00–11:00 AM) to assess non-fasted status, minimizing circadian variation. Fasting blood glucose (FBG) was measured at baseline (Week 0) and endpoint (Week 6) after 6 h fast. Blood glucose was determined using a Sinocare glucometer, and plasma insulin by an ultrasensitive insulin ELISA kit (Solarbio, SEKM-0141). Area under the curve (AUC) was calculated for glucose response. Blood samples were collected through mice tail. After sterilizing the tail with an alcohol swab, a small incision was made at the tail vein tip after discarding the first droplet.

Reagents

Chemicals used in this study including BAPTA-AM (Thermo Scientific, 1824047), EGTA-AM (Sigma-Aldrich, E3478), DFO (MCE, HY-B0988), TPEN (MCE, HY-100202), GPN (Santa Cruz Biotech, sc-252858), riluzole (Sigma-Aldrich, R116), PA (Sigma-Aldrich, P0500), insulin (Beyotime, P3376), metformin (Beyotime, S1741), CC (MCE, HY-13418A), STO609 (MCE, HY-19805), TG assay (Solarbio, BC0625), TC assay (Solarbio, BC1985), ALT assay (Solarbio, BC1555), AST assay (Solarbio, BC1565), HbA1c kit (Solarbio, BC5615), Creatinine assay (Nanjing Jiancheng Company, C011-2-1). ML-SA5, ML-SA8 and ML-SI5 were custom synthesized (available upon MTA request).

Statistical analysis

Data are presented as mean ± s.e.m. from at least 3 independent experiments. Statistical comparisons were performed using a two tailed Student’s t-test (paired or unpaired as appropriate) or one-way/two-way ANOVA followed by Tukey’s or Dunnett’s post hoc test for multiple comparisons. A P value < 0.05 was considered statistically significant.

Data availability

All data generated or analyzed during this study are included in the manuscript and supporting files

Acknowledgements

This work was supported by a grant from National Natural Science Foundation of China (31600823 to D. L.). Additional support was provided by Calygene Biotechnology Inc. (XT [2016]008@) and Lysoway Therapeutics, Inc. (KYY-HX-20210129). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. We appreciate the encouragement and assistance provided by Dr. Haoxing Xu and his laboratory members. The authors thanks Experimental Animal Center of Zhejiang University of Technology for providing the animal facilities and technical support for this study.

Additional information

Author contributions

J.Z., Data curation, Investigation, Methodology and Writing original draft; Z.P., Data curation, Validation, Investigation, Methodology and Writing original draft, review and editing; S.W., Data curation, Investigation, Methodology and Writing review and editing; H.Z., Data curation, Methodology, Writing review and editing; Y.W., Z.D., Q.W., Investigation, Methodology; D.L., Conceptualization, Supervision, Funding acquisition, Validation, Investigation, Methodology, Project administration, Writing original draft, review and finalization.

Abbreviations

• ALT: alanine aminotransferase

• AMP: adenosine monophosphate

• AMPK: AMP-activated protein kinase

• ANOVA: analyses of variance

• AST: aspartate aminotransferase

• BAPTA-AM: 1,2-bis(2-aminophenoxy)ethane-N,N,N’,N’-tetraacetic acid tetrakis-acetoxymethyl ester

• CaMKKβ: calcium/calmodulin-dependent protein kinase kinase β

• DFO: deferoxamine mesylate

• EGTA-AM: EGTA Acetoxymethyl ester

• ER: endoplasmic reticulum

• FBG: fasting blood glucose

• FBS: fetal bovine serum

• GLUT4: Glucose Transporter 4

• GPN: Glycyl-L-phenylalanine 2-naphthylamide

• HbA1c: hemoglobin A1c

• H&E: hematoxylin-eosin

• ITT: insulin tolerance test

• IR: insulin resistance

• KO: knockout

• LKB1: liver kinase B1

• MCOLN1/TRPML1: mucolipin 1

• MET: metformin

• ML-SA8: mucolipin-specific synthetic agonist 8

• ML-SI5: mucolipin-specific synthetic inhibitor 5

• OGTT: oral glucose tolerance test

• PA: palmitic acid

• PAS: periodic acid-schiff

• PBS: phosphate-buffered saline

• Q-PCR: quantitative real time polymerase chain reaction

• RBG: random blood glucose

• TFEB: transcription factor EB

• TG: triglyceride

• TC: total cholesterol

• TPCs: two-pore channels

• TPEN: N,N,N’,N’-tetrakis (2-pyridylmethyl) ethylene-diamine

• T2DM: type 2 diabetes mellitus

• ULK1: Unc like autophagy activating kinase 1

• WT: wild-type

• 2-NBDG: 2-deoxy-2-[(7-nitro-2,1,3-benzoxadiazol-4-yl)amino]-d-glucose

Funding

MOST | National Natural Science Foundation of China (NSFC) (31600823)

• Dan Li

Additional files

Supplementary figures