Impaired degradation of PLCG1 by chaperone-mediated autophagy promotes cellular senescence and intervertebral disc degeneration

ABSTRACT

Defects in chaperone-mediated autophagy (CMA) are associated with cellular senescence, but the mechanism remains poorly understood. Here, we found that CMA inhibition induced cellular senescence in a calcium-dependent manner and identified its role in TNF-induced senescence of nucleus pulposus cells (NPC) and intervertebral disc degeneration. Based on structural and functional proteomic screens, PLCG1 (phospholipase C gamma 1) was predicted as a potential substrate for CMA deficiency to affect calcium homeostasis. We further confirmed that PLCG1 was a key mediator of CMA in the regulation of intracellular calcium flux. Aberrant accumulation of PLCG1 caused by CMA blockage resulted in calcium overload, thereby inducing NPC senescence. Immunoassays on human specimens showed that reduced LAMP2A, the rate-limiting protein of CMA, or increased PLCG1 was associated with disc senescence, and the TNF-induced disc degeneration in rats was inhibited by overexpression of Lamp2a or knockdown of Plcg1. Because CMA dysregulation, calcium overload, and cellular senescence are common features of disc degeneration and other age-related degenerative diseases, the discovery of actionable molecular targets that can link these perturbations may have therapeutic value.

Abbreviation: ATRA: all-trans-retinoic acid; BrdU: bromodeoxyuridine; CDKN1A/p21: cyclin dependent kinase inhibitor 1A; CDKN2A/p16-INK4A: cyclin dependent kinase inhibitor 2A; CMA: chaperone-mediated autophagy; DHI: disc height index; ER: endoplasmic reticulum; IP: immunoprecipitation; IP3: inositol 1,4,5-trisphosphate; ITPR/IP3R: inositol 1,4,5-trisphosphate receptor; IVD: intervertebral disc; IVDD: intervertebral disc degeneration; KD: knockdown; KO: knockout; Leu: leupeptin; MRI: magnetic resonance imaging; MS: mass spectrometry; N/L: NH₄Cl and leupeptin; NP: nucleus pulposus; NPC: nucleus pulposus cells; PI: protease inhibitors; PLC: phospholipase C; PLCG1: phospholipase C gamma 1; ROS: reactive oxygen species; RT-qPCR: real-time quantitative reverse transcription PCR; SA-GLB1/β-gal: senescence-associated galactosidase beta 1; SASP: senescence-associated secretory phenotype; STV: starvation; TMT: tandem mass tag; TNF: tumor necrosis factor; TP53: tumor protein p53; UPS: ubiquitin-proteasome system.

Introduction

Cells undergo senescence in response to exogenous and endogenous stresses, including inflammatory stimuli, organelle stress, and persistent DNA damage [1]. The major feature of cellular senescence is stable cell cycle arrest accompanied by the senescence-associated secretory phenotype (SASP) and activation of senescence-associated GLB1/β-galactosidase (SA-GLB1/β-gal) [2]. Cellular senescence plays a key role in the pathology of many age-related diseases, such as neurodegenerative diseases, cancer, and intervertebral disc degeneration [1,3]. Therefore, it is crucial to understand how senescence is controlled at the molecular and cellular levels. Senescence-associated cell cycle arrest is mainly induced by TP53 (tumor protein p53) and its downstream targets CDKN1A/p21 (cyclin dependent kinase inhibitor 1A) and CDKN2A/p16-INK4A (cyclin dependent kinase inhibitor 2A) [4]. Recently, abnormally elevated calcium signaling was also found to trigger cellular senescence [5,6]. Increased calcium flux to the cytoplasm can come from various calcium channel receptors on the cell membrane or from ITPR/IP3R (inositol 1,4,5-trisphosphate receptor)-mediated calcium release from the endoplasmic reticulum (ER) [5,7,8]. Calcium overload not only induces oxidative stress to accelerate senescence [9], but also independently mediates cellular senescence associated with organelle dysfunction [10,11]. Therefore, fine regulation of calcium homeostasis by cells under physiological and pathological conditions plays a non-negligible role in preventing the senescence process.

Chaperone-mediated autophagy (CMA) is a selective lysosomal autophagy process. The protein substrate of CMA is characterized by a KFERQ-like motif that is recognized by HSPA8 and carried to the lysosomal surface, where the substrate is transported to the lysosomal lumen for degradation through the action of LAMP2A [12]. LAMP2A is the rate-limiting protein of CMA, and its expression level is associated with the activity of CMA [13]. CMA is involved in protein quality control and proteome remodeling [14], thereby broadly regulating processes such as cell growth, differentiation, metabolism, and transcription [15,16]. Notably, CMA was found to be strongly associated with senescence [17]. Age-dependent decreases in CMA activity occur in virtually all cell types and tissues in both rodents and humans [12]. CMA deficiency is associated with various age-related degenerative diseases, including retinal degeneration [18], atherosclerosis [19,20], Parkinson disease [21] and Alzheimer disease [22]. In cells deficient in CMA activity, SA-GLB1 activation and accumulation of CDKN1A, CDKN2A, and lipofuscin were found, which are hallmarks of cellular senescence [17]. However, the pathway involved in CMA deficiency-induced cellular senescence remains unclear.

Cellular senescence is a common feature of many age-related degenerative diseases. Here, we focused on intervertebral disc degeneration (IVDD), an age-related degeneration of the motor system that is strongly associated with chronic disabling low back pain [23]. The nucleus pulposus (NP) is the central structure of the intervertebral disc. Senescence of NP cells (NPC) induced by inflammatory factors is the main cause of IVDD [24]. We found that CMA inhibition mediated TNF-induced NPC senescence and disc degeneration. Through sequencing and pathway analysis, we found that CMA inhibition-induced senescence of NPC was associated with abnormal activation of calcium signaling pathways. Both calcium overload and CMA deficiency were previously identified as triggers of cellular senescence, but the molecular targets linking these perturbations remained poorly understood. Based on the structure and function of CMA, we performed unbiased proteomic screening and programmed filtering, and identified PLCG1 (phospholipase C gamma 1) as a potential target of CMA deficiency-induced calcium overload. PLCG1, a major member of the PLC (phospholipase C) family, catalyzes the generation of inositol 1,4,5-trisphosphate (IP3) and leads to the increase in cellular calcium concentration by activating ITPR in the ER [25]. We found that PLCG1 was efficiently degraded by the CMA pathway, but not by macroautophagy or the ubiquitin-proteasome system (UPS). Importantly, we found that TNF-induced cellular senescence was associated with insufficient degradation of PLCG1 by CMA, which led to persistent calcium overload. Collectively, this study revealed the mechanism of CMA inhibition-induced cellular senescence and provided therapeutic ideas for inflammation-induced intervertebral disc senescence and degeneration.

Results

CMA attenuates TNF-induced cellular senescence and IVDD

CMA is active in the brain, liver, and other organs, and is involved in inhibiting their age-related degeneration [26,27]. We examined LAMP2A levels in several important tissues of rats, including the brain, liver, lung, kidney, heart, bone marrow, and intervertebral disc (IVD) (Figure S1A). Rat LAMP2A expression in IVD was at an intermediate level, higher than that in the heart and bone marrow, suggesting a potential non-redundant role of CMA in IVDD. IVDD is a senescent process. Previous studies have identified the relationship between CMA deficiency and cellular senescence. To explore the potential association between CMA and IVDD, we first examined the expression levels of LAMP2A in degenerated or healthy human IVD specimens. Immunofluorescence and immunoblotting of tissues showed that LAMP2A levels were downregulated in human degenerated IVD compared to the control group (Figure 1A,B). Meanwhile, the mRNA level of LAMP2A was also decreased in degenerated IVD (Figure 1C). In contrast, the levels of senescence markers, including TP53, CDKN1A, and CDKN2A, were significantly elevated in the degenerated IVD compared with the control group (Figure 1A,B). Inflammatory factors trigger the initiation and progression of IVDD [29]. We found that in TNF-treated NPC, LAMP2A levels were downregulated, while SA-GLB1 levels were increased (Figure 1D). The KFERQ-like motif is the targeted sequence of CMA. We used KFERQ-PA-mCherry-1 to track CMA activity, which allows for increased number of visualized puncta associated with lysosome surface when CMA is activated [30] (Figure S1B). We found that the number of visualized spots was significantly reduced after TNF stimulation in NPC (Figure 1E), suggesting that TNF inhibited the CMA activity. In contrast, NPC treated with serum starvation or AR7 (a known CMA activator) showed more puncta associated with the lysosomal surface [31] (Figure 1E). In addition, immunoblotting showed that TNF treatment reduced the level of LAMP2A in NPC (Figure 1F). These results suggested that CMA levels in NPC were suppressed under TNF-mediated inflammatory stimulation.

Figure 1.

CMA attenuates TNF-induced cellular senescence and IVDD. (A) tissue immunofluorescence showing differential expression of LAMP2A, TP53, CDKN1A, and CDKN2A in healthy and degenerated human intervertebral discs. Three tissue samples each from healthy and degenerated discs were used to stain the corresponding senescence indicators. For each indicator, the average cell fluorescence intensity in at least three fields of view was selected for quantification. (B) immunoblot showing the levels of LAMP2A, TP53, CDKN1A, and CDKN2A in healthy and degenerated human intervertebral discs. Three specimens were used for testing in each group. (C) three normal and degenerated human disc tissue samples were collected to measure LAMP2A mRNA levels. (D) fluorescence levels of LAMP2A and SA-GLB1 in NPC after treatment with TNF (20 ng/mL) for 48 h. The average cell fluorescence intensity of at least 3 fields was used for quantification. (E) CMA reporter KFERQ-PA-mCherry1 was used to visualize CMA activity in NPC under TNF (20 ng/mL, 48 h), serum (24 h), or AR7 (25 μg/mL, 24 h) treatment. The average number of intracellular red puncta in at least 3 fields of view per group was calculated to reflect CMA activity. (F) immunoblot showing levels of senescence-associated proteins (TP53, CDKN1A, and CDKN2A) in NPC after TNF alone treatment (20 ng/mL, 48 h) or combined with LAMP2A overexpression. The right panel represents the quantification of each group normalized to NC. (G) immunofluorescence showing the levels of SA-GLB1 and TP53 in NPC after TNF alone treatment (20 ng/mL, 48 h) or combined with LAMP2A overexpression. Quantification of the mean cellular fluorescence intensity of at least 3 fields is provided in (fig. S1C). (H) flow cytometry results of the cell cycle. The proportion of cells in S phase was used for comparison between groups. (I) in vivo modeling method of IVDD. The SD rats used were male and 2 months old. Local injection of lentivirus (1×10⁸ TU/mL, 2 μL) was performed on the first day, and TNF injection (20 ng/mL, 2 μL) was performed on the third day. Injections were repeated weekly for one month and maintained for another month, followed by physical examination, imaging, and sampling at week 8. (J) MRI of the indicated disc segment in the rat tail after injection of TNF or combined with LAMP2A lentivirus. The administration concentration and dosage are as described above. (K) the disc height index (DHI; mm) was used for quantification of MRI results of each group after local injections. The DHI is inversely related to the degree of IVDD. (L) histological staining of indicated rat discs after treatments, including H&E, safranin O-fast green staining, and Masson staining. (M) histological scores of IVDD in rats. Higher scores (0-15) represent more severe degeneration. The scoring criteria is as described before [28]. (N) Western blot showing protein levels of LAMP2A, TP53, CDKN1A, and CDKN2A in NPC treated with DMSO, TNF (20 ng/mL, 48 h), or ATRA (10 μM, 48 h). All figures show mean ± SEM of at least three independent experiments. */#p < 0.05, **/##p < 0.01, ***/###p < 0.001, ns means not significant. The p-value symbols represent comparisons to the left group of the labeled line (A, C, D, E, H, and N) or to the vector+tnf group (F, L, and M).

Given the relationship between CMA inhibition and senescence, we wanted to know whether TNF-induced NPC senescence and IVDD are associated with inhibited CMA activity. We stably overexpressed LAMP2A in NPC using lentivirus. After 48 h of TNF treatment, the levels of senescence markers TP53, CDKN1A, and CDKN2A were significantly increased, accompanied by decreased expression of LAMP2A. In contrast, overexpression of LAMP2A inhibited TNF-induced upregulation of the above senescence-associated proteins (Figure 1F, 1G and S1C). Furthermore, LAMP2A overexpression inhibited TNF-induced activation of SA-GLB1 in NPC (Figure 1G and S1C). Another prominent feature of cellular senescence is cell cycle arrest. We found that the proportion of NPC in S phase and the level of bromodeoxyuridine (BrdU) were significantly reduced by TNF, which were reversed by LAMP2A overexpression (Figure 1H and S1D). As described previously [29], we performed intradiscal injection of TNF in rats (Figure 1I), with or without local lentiviral injection of LAMP2A. Physical examination, MRI imaging, specimen collection and subsequent staining were performed at week 8 to assess the degree of IVDD. We found that local overexpression of LAMP2A attenuated TNF-induced disc degeneration, which was manifested by reduced caudal stiffness (Figure S1E), decreased Pfirrmann score (Figure S1F), reduced disc height collapse (Figure 1J,K), as well as decreased histological scores (Figure 1L,M) [28]. Meanwhile, immunofluorescence of tissue sections showed that overexpression of LAMP2A suppressed TNF-induced upregulation of TP53, CDKN1A, and CDKN2A (Figure S1G and S1H). Collectively, overexpression of LAMP2A attenuated TNF-induced NPC senescence both in vitro and in vivo. Furthermore, we treated NPC with the CMA inhibitor all-trans retinoic acid (ATRA) [31]. It was of note that, similar to TNF treatment, ATRA increased the levels of senescence-associated proteins (Figure 1N), consistent with a previous study [32]. Taken together, these findings suggested that CMA inhibition mediated TNF-induced NPC senescence.

CMA inhibition promotes calcium overload-induced cellular senescence

To investigate the underlying mechanism by which CMA inhibition affects cellular senescence, we performed RNA sequencing on CMA inhibitor-treated NPC. KEGG pathway enrichment analysis revealed that CMA inhibition significantly activated the calcium signaling pathway (Figure 2A). This was confirmed by subsequent measurement of Ca²⁺ levels (Figure 2B). Ca²⁺ fluxes from the ER to mitochondria are known to be the impetus of cellular senescence [5], and Ca²⁺ overload has been proven to be the trigger of NPC senescence and disc degeneration [33,34]. Therefore, we speculated that CMA inhibition-induced cellular senescence could be related to Ca²⁺ overload, while the role of CMA in calcium homeostasis has not been formally reported.

Figure 2.

CMA deficiency promotes calcium overload-induced cellular senescence. (A) RNA sequencing analysis of NPC treated with or without the CMA inhibitor ATRA (10 μM, 48 h). KEGG pathway analysis of the differential genes shown in the volcano plot suggested that the calcium signaling pathway was significantly enriched. (B) levels of Ca²⁺ flux in NPC treated with LAMP2A-KO or ATRA (10 μM, 48 h). Values in each group were normalized to that of the NC group for comparison. (C) immunofluorescence showing levels of Ca²⁺ and SA-GLB1 after LAMP2A-KO, and Ca²⁺ was visualized by fluorescent probe fluo-4 AM. The average cellular fluorescence intensity of at least 3 fields of view was selected for quantification. (D) immunoblot showing protein levels of LAMP2A, TP53, CDKN1A, and CDKN2A in LAMP2A-KO NPC treated with or without BAPTA-AM (5 μM, 24 h). TNF-treated NPC (20 ng/mL, 48 h) served as the positive control. The right panel shows the normalized quantification of the band of interest. (E) immunofluorescence showing Ca²⁺, TP53, and CDKN2A levels in LAMP2A-KO NPC treated with or without BAPTA-AM (5 μM, 24 h). The right panel represents quantification of the mean fluorescence intensity of cells in each group derived from at least 3 fields of view. (F) flow cytometry results of cell cycle of NPC treated with LAMP2A-KO or combined with DMSO or BAPTA-AM (5 μM, 24 h). The proportion of cells in S phase was used for comparison between the indicated groups. (G) levels of Ca²⁺ flux in tnf-treated NPC (20 ng/mL, 48 h) detected by fura-2, normalized to the value of the NC group. (H) Ca²⁺ levels and CMA activity were assessed at 24 h, 36 h, and 48 h after TNF treatment (20 ng/mL). NPC were transfected with the KFERQ-PA-mCherry-1 plasmid and then photoconverted to monitor the activity of CMA. The fold change (/control) of visualized puncta number of each group represents the relative CMA level. The fluo-4 AM fluorescent probe was used to detect Ca²⁺ levels. Representative images for statistics are provided in (fig. S2B). Mean fluorescence intensities were used for quantification, normalized to the control group. (I) Ca²⁺ and SA-GLB1 fluorescence images of NPC treated with TNF (20 ng/mL, 48 h) or combined with LAMP2A overexpression. The right panel represents the quantification of mean cell fluorescence intensity for each group. At least 3 fields of view were used for quantitative assessment. (J) Western blot of LAMP2A and senescence markers (TP53, CDKN1A, and CDKN2A) of NPC treated with TNF (20 ng/mL, 48 h) or combined with LAMP2A overexpression or BAPTA-AM (5 μM, 24 h). The right panel represents normalized quantification of band intensities of interest. (K) BrdU levels of NPC treated with TNF in the presence or absence of BAPTA-AM, detected at 450 nm, normalized to that of the control group. The cells were administered as described above. All data show mean ± SEM of at least three independent experiments. */#p < 0.05, **/##p < 0.01, ***/###p < 0.001, ns means not significant. p-value symbols represent comparisons to the left group of the labeled line (B, F, G, I, and J) or to the LAMP2A-KO+DMSO group (D and E) or to the DMSO+TNF group (K).

We knocked out the LAMP2A gene in human NPC using the CRISPR-Cas9 system. LAMP2A is derived from three different splice variants (A, B, and C) of exon 9 of the LAMP2 gene (Figure S1I). They share a common luminal domain but contain distinct cytoplasmic and transmembrane regions. We designed multiple sgRNA-Cas9 lentiviral systems targeting the differential regions of exon 9 and screened sgRNA#3 as an effective target for LAMP2A knockout (KO), which can simultaneously maintain the normal expression of other LAMP2 isoforms (Figure S1J and S1K). We confirmed that the expression of major lysosomal proteases was not affected after LAMP2A-KO (Figure S1L). In LAMP2A-KO NPC, Ca²⁺ levels were significantly increased (Figure 2B,C). Meanwhile, LAMP2A-KO also increased the level of the senescence marker SA-GLB1 (Figure 2C). Subsequently, we treated NPC with the Ca²⁺ chelator BAPTA-AM, which significantly attenuated LAMP2A-KO-induced Ca²⁺ overload. Notably, the addition of BAPTA-AM also attenuated LAMP2A-KO-induced upregulation of senescence-associated proteins, including TP53, CDKN1A, and CDKN2A (Figures 2D,E). LAMP2A-KO also led to cell cycle arrest in NPC, as evidenced by the reduced proportion of the S phase (Figure 2F) and decreased BrdU levels (Figure S1M). With the addition of BAPTA-AM, LAMP2A-KO-induced cell cycle arrest was also alleviated (Figure 2F and S1M). Collectively, these findings supported that CMA deficiency promoted senescence of NPC in a calcium-dependent manner.

We have demonstrated that CMA inhibition mediated TNF-induced cellular senescence. However, the role of Ca²⁺ overload caused by CMA inhibition in TNF-induced senescence has not been characterized. By Ca²⁺ probe detection, we found that TNF-treated NPC also showed elevated Ca²⁺ levels (Figure 2G and S2A). To explore whether TNF-induced Ca²⁺ overload was derived from CMA inhibition, we tracked the alterations of CMA activity with a CMA reporter and Ca²⁺ levels with a Fluo-4 AM probe [30]. After 24 h of TNF treatment, the activity of CMA in NPC was inhibited, while the Ca²⁺ level was not significantly increased (Figure 2H and S2B). At 36 and 48 h, Ca²⁺ levels increased with inhibition of CMA activity (Figure 2H and S2B). This suggested that CMA inhibition preceded the increase in Ca²⁺ levels in response to TNF, a finding that supported the role of CMA inhibition in mediating TNF-induced Ca²⁺ overload in NPC. Further, we stably overexpressed LAMP2A in NPC with lentivirus. We found that LAMP2A-overexpressing NPC were resistant to TNF-induced Ca²⁺ overload and senescence (Figure 2I,J). Both LAMP2A overexpression and BAPTA-AM treatment inhibited TNF-induced expression of senescence-associated proteins, including TP53, CDKN1A, and CDKN2A (Figure 2J). Inhibition of calcium signaling by BAPTA-AM also rescued the cell cycle arrestinduced by TNF (Figure 2K). Taken together, these lines of evidence supported that Ca²⁺ overload caused by CMA inhibition mediated the TNF-induced NPC senescence.

PLCG1 is screened as a putative substrate of CMA

CMA is known to regulate cellular pathophysiological processes by selectively degrading specific substrates. We speculated that calcium overload resulted from the disturbed degradation of certain substrates due to CMA inhibition (Figure 3A). Several features of CMA substrates were identified based on previous studies, including containing KFERQ-like sequences, increasing after LAMP2A knockdown, and interacting with HSPA8 and LAMP2A [12]. To search potential substrates, we generated LAMP2A-KO NPC and performed quantitative mass spectrometry (MS) analysis of three independent replicate samples based on tandem mass tag (TMT) markers (Figure 3B,C). We analyzed all upregulated proteins after LAMP2A-KO (Figure 3D) and searched for potential CMA substrates containing KFERQ-like sequences according to the KFERQ Finder [35]. The reliability of the TMT-MS results was verified by the decrease in LAMP2A (sg/Scb = 0.112) and the increase of several known CMA substrates (IDH2, ACSL4, and NFKB2) (Figure 3E). HSPA8 and LAMP2A are responsible for substrate transport and transmembrane transfer separately during the CMA process [12]. Structurally, the CMA substrate interacts with HSPA8 and LAMP2A. To obtain another collection of reliable CMA substrates, we performed immunoprecipitation of HSPA8 and LAMP2A in NPC (Figure 3F), followed by label-free MS detection of interactors of HSPA8 or LAMP2A. We focused on high-scoring proteins interacting with HSPA8 and LAMP2A (Figure 3G,H). To search for the potential substrate of CMA affecting calcium homeostasis, we performed programmatic filtering of motif features and cross-analyzed TMT-MS results, shared interacting proteins of HSPA8 and LAMP2A, and gene sets of calcium signaling pathways, and identified PLCG1 as a putative target that met all requirements (Figure 3I,J). We then verified the interaction of HSPA8 or LAMP2A with PLCG1 in NPC (Figure 3K). Meanwhile, we found that the level of PLCG1 was significantly elevated in LAMP2A-KO NPC (Figure 3L). These findings supported that PLCG1 could be a potential mediator of excessive Ca²⁺ flux induced by CMA inhibition.

Figure 3.

PLCG1 is screened as a putative substrate of CMA. (A) schematic representation of the mechanisms underlying calcium overload-induced cellular senescence in the context of CMA failure. (B) schematic diagram of collecting protein supernatants from LAMP2A-KO NPC for TMT labeling and MS detection. (C) statistical results of protein and peptide identification and quantification by TMT-MS. (D) volcano plot showing all differential proteins in MS results after LAMP2A-KO. (E) validation of the LAMP2A level in TMT-MS results, and fold change of several known CMA substrates, including IDH2, ACSL4, and NFKB2. (F) schematic diagram of MS-based immunoprecipitation proteomics analysis. Immunoprecipitates of the bait proteins LAMP2A and HSPA8 were first isolated by magnetic beads and then the interactors were detected by MS. (G and H) volcano plot analysis of the IP-MS results showed that several interacting proteins were significantly enriched in the immunoprecipitates of HSPA8 and LAMP2A. (I) venn diagram of proteins overlapping between upregulated proteins by TMT-MS, interacting proteins pulled by bait proteins HSPA8 and LAMP2A, and gene sets of calcium signaling pathway. PLCG1 was screened as a putative CMA substrate in regulating calcium homeostasis. (J) secondary spectrum of PLCG1 in IP-MS results. (K) Co-immunoprecipitation blot confirming endogenous interactions of PLCG1 with HSPA8 and LAMP2A in NPC. (L) immunoblot showing PLCG1 levels in NPC after LAMP2A-KO. The right panel shows the PLCG1 levels of each group, normalized to the value of the NC group. All data show mean ± SEM of at least three independent experiments. */#p < 0.05, **/##p < 0.01, ***/###p < 0.001, ns means not significant. p-value symbols represent comparisons to the sg-Scb group (L).

PLCG1 is regulated by CMA, but not macroautophagy or the UPS

Next, we aimed to explore and verify the degradation pathway of PLCG1. To determine whether PLCG1 was degraded by lysosomes, NPC were treated with leupeptin (Leu) or NH₄Cl or a combination of both (N/L). Elevation of SQSTM1 confirmed the inhibitory effects of these treatments on lysosomal enzymes (Figure 4A,B). Western blot showed that the degradation of PLCG1 was inhibited by Leu or NH₄Cl (Figure 4A). These inhibitory effects could be superimposed and time-dependent (Figure 4A,B). The lysosomal pathway and the UPS are two major pathways for protein degradation. Compared with N/L treatment, MG132 treatment did not affect the expression of PLCG1 in NPC (Figure S2C). Likewise, MG132 treatment did not significantly affect PLCG1 levels with increasing time or concentration (Figure 4C and S2D). These findings suggested that PLCG1 was degraded via the lysosomal pathway, rather than the UPS. Immunofluorescence also showed that PLCG1 colocalized with lysosomes (Figure 4D). We have verified that PLCG1 is elevated in LAMP2A-KO NPC. Notably, there was no significant change in the mRNA level of PLCG1 after LAMP2A-KO (Figure 4E). In the sg-Scb group, the level of PLCG1 was increased by N/L, but not affected by MG132. However, in the LAMP2A-KO group, the level of PLCG1 was not significantly affected by either MG132 or N/L treatment (Figure 4F), indicating that the lysosomal degradation of PLCG1 requires the presence of LAMP2A. Subsequently, we overexpressed LAMP2A in NPC to promote CMA. It resulted in decreased level of PLCG1, which was rescued by co-treatment with N/L (Figure 4G).

Figure 4.

PLCG1 is regulated by CMA, but not macroautophagy or the UPS. (A) immunoblotting showing the level of PLCG1 in NPC after treatment with Leu (10 μM), NH₄Cl (20 mM), or their combination for 12 h. Right panels represent normalized (/NC) band intensities of interest. (B) Western blot showing PLCG1 levels in NPC treated with a combination of Leu and NH₄Cl (N/L) for 3, 6, and 9 h. Right panels represent normalized (/NC) band intensities of interest. (C) immunoblot of PLCG1 in NPC treated with different concentrations (2, 5, and 10 μM) of MG132 for 12 h. Ubiquitin was used as the positive control. The right panel shows relative levels of PLCG1 of each group, normalized to NC group. (D) immunofluorescence showing colocalization of PLCG1 with lysosomes. LAMP1 is used as a lysosomal marker. (E) RT-qPCR showing mRNA levels of PLCG1 after LAMP2A-KO, normalized to the sg-Scb group. (F) immunoblot showing the expression levels of PLCG1 in LAMP2A-KO NPC treated with MG132 or N/L for 12 h. The right panel represents normalized (/DMSO) band intensities of interest. (G) the levels of PLCG1 in NPC under LAMP2A overexpression or combined with N/L treatment. The right panel shows relative levels of PLCG1 of each group, normalized to the vector group. (H) immunoblotting showing PLCG1 and SQSTM1 levels in NPC after ATG7 knockdown using si-ATG7. The relative level of PLCG1 of each group was normalized to that of si-Scb group. (I) Western blot showing PLCG1 levels after 3 MA treatment (5 mM) for 3, 6, and 9 h. The right panel shows relative levels of PLCG1 of each group, normalized to the NC group. (J) Western blot showing the levels of PLCG1 in NPC after serum starvation (STV) for 24 h and 48 h. The right panel shows normalized PLCG1 levels of each group. (K) effect of STV treatment or combined with LAMP2A-KO on the expression of PLCG1. The relative levels of PLCG1 of each group were normalized to that of the NC group. (L) PLCG1 levels in NPC treated with different concentrations of AR7 (10, 25, 50, and 100 μM) for 24 h. The right panel shows normalized PLCG1 levels of each group (/NC). (M) lysosomal uptake assay. After co-incubation with lysosomes from HEK293T cells in the presence or absence of protease inhibitor (PI), PLCG1-GST was taken up and degraded by lysosomes. A lane with empty lysosomes was used as control, and substrate without lysosomes (20 ng) was used as input. LAMP1 was used to demonstrate equal lysosomal loading. The right panel shows normalized PLCG1 levels of each group. All figures show mean ± SEM of at least three independent experiments. */#p < 0.05, **/##p < 0.01, ***/###p < 0.001, ns means not significant. p-value symbols represent comparisons to the left group of the labeled line (A, B, C, E, F, H, I, J, K, L, and M) or to the LAMP2A group (G).

Lysosomal degradation pathways include nonselective macroautophagy and selective autophagy represented by CMA [17]. It was previously reported that several CMA substrates are also degraded by macroautophagy [36]. We wondered whether PLCG1 was also affected by macroautophagy in NPC. ATG7 is essential for autophagosome formation in the macroautophagy process [37]. After knockdown of ATG7 with si-ATG7, macroautophagy was inhibited, which was manifested by the increase of SQSTM1 (Figure 4H). However, the level of PLCG1 was not increased, but decreased. This could arise from crosstalk between macroautophagy and CMA [12]. Furthermore, treatment with either macroautophagy inhibitor 3 MA or its activator rapamycin did not affect PLCG1 levels in NPC (Figure 4I and S2E). Serum starvation (STV) is a recognized condition for CMA activation [15]. After the serum was removed, the expression of LAMP2A was increased in NPC, while the level of PLCG1 was decreased (Figure 4J). Immunofluorescence also showed enhanced colocalization of PLCG1 and lysosomes after starvation for 24 h (Figure S2F). In contrast, LAMP2A-KO abolished starvation-induced downregulation of PLCG1 (Figure 4K). Although reports suggest that starvation also activates macroautophagy, downregulation of PLCG1 by starvation was not reversed by 3 MA but by N/L (Figure S2G). Furthermore, the CMA activators, including AR7, QX77, and AKT inhibitor, also decreased PLCG1 levels in a concentration-dependent manner (Figure 4L and S2H-I). Finally, we co-incubated GST-PLCG1 with lysosomes purified from HEK293T cells in the presence or absence of protease inhibitor (PI). The membrane integrity of active lysosomes isolated from HEK293T cells was demonstrated using electron microscopy (Figure S2J), and lysosome integrity was quantified by a β-Hexosaminidase assay in the presence and absence of the detergent Triton X-100 (Figure S2K). Immunoblotting of recovered lysosomes clearly showed that PLCG1-GST was taken up and degraded by lysosomes (Figure 4M). Meanwhile, we identified an endogenous interaction between HSPA8 and PLCG1 (Figure S2L). Collectively, these lines of evidence supported PLCG1 as the target of CMA, rather than macroautophagy or UPS.

PLCG1 is a bona fide substrate of CMA

For a protein to be validated as a CMA substrate, it must contain at least one KFERQ-like motif, which is required for recognition by HSPA8 and subsequent lysosomal translocation [38]. We analyzed the full-length amino acid sequence of PLCG1 and found two typical KFERQ-like sequences: ₅₂QVKLE₅₆ and ₂₆₈VDRLQ₂₇₂ (Figure 5A). These KFERQ motifs are highly conserved in several species searched (Figure S3A and S3B). Furthermore, we verified the lysosomal translocation of PLCG1. PLCG1 was detected in lysosomes isolated from NPC, and it was increased with starvation (Figure 5B). Knockdown of LAMP2A, but not ATG7, reduced PLCG1 levels in lysosomes (Figure 5C). Further, we focused on the interaction between PLCG1 and LAMP2A or HSPA8. Immunofluorescence showed colocalization of PLCG1 with HSPA8 or LAMP2A, which was enhanced by serum starvation (Figure 5D). Co-immunoprecipitation results also showed that the association of PLCG1 with LAMP2A was increased by serum starvation (Figure 5E). To further identify the sequences on which the capture of PLCG1 by HSPA8 depends, we mutated glutamine (Q) to arginine (R) in the two KFERQ sequences we found, and obtained two mutants of PLCG1: PLCG1^Q52R and PLCG1^Q272R (Figure S3A and S3B). We found that compared with WT PLCG1 and PLCG1^Q52R, PLCG1^Q272R displayed missing binding ability to HSPA8 (Figure 5F). In turn, after stable transfection in NPC with WT or MUT plasmids, co-immunoprecipitation of HSPA8 showed binding to WT PLCG1 or PLCG1^Q52R, but not PLCG1^Q272R (Figure 5G). Similarly, the Q>A mutation at position 272, rather than 52, also eliminated the interaction between PLCG1 and HSPA8 (Figure S3C and S3D). This suggested that ₂₆₈VDRLQ₂₇₂ is a target KFERQ-like motif that binds to HSPA8 for CMA degradation. For responses to serum starvation, WT PLCG1 and PLCG1^Q52R groups showed decreased levels of PLCG1-HA, which in the PLCG1^Q272R group were not significantly affected (Figure 5H). In the absence of starvation, the basal level of PLCG1-HA in the PLCG1^Q272R group was also higher than that in the WT PLCG1 and PLCG1^Q52R groups (Figure 5H). In both WT PLCG1 and PLCG1^Q52R overexpression groups, co-immunoprecipitation blots for PLCG1-HA showed binding with LAMP2A, and it was enhanced by starvation. However, as in the vector group, LAMP2A was not detected in the PLCG1-HA immunoprecipitates of the PLCG1^Q272R transfection group (Figure 5I). In addition, we investigated the responsiveness of the Q mutation of PLCG1 to N/L. In WT PLCG1 and PLCG1^Q52R transfection groups, the level of PLCG1-HA was increased by N/L treatment, as clearly demonstrated by the fold change of PLCG1-HA. In the PLCG1^Q272R transfection group, the expression of PLCG1-HA was not affected by N/L (Figure 5J). Finally, when we co-incubated protein lysates from WT PLCG1, PLCG1^Q52R, or PLCG1^Q272R transfected NPC with active lysosomes extracted from HEK293T in the presence or absence of PI, only the PLCG1^Q272R group showed resistance to lysosomal uptake (Figure 5K). Taken together, these results supported PLCG1 as a bona fide substrate of CMA, and the degradation of PLCG1 by CMA was highly dependent on its ₂₆₈VDRLQ₂₇₂ sequence recognized by HSPA8.

Figure 5.

PLCG1 is a bona fide substrate of CMA. (A) graphical representation of KFERQ-like motifs in the human PLCG1 sequence. (B) immunoblot of PLCG1 in lysosomes (lyso) of control (NC) or STV (serum starvation for 24 h) NPC. LAMP1 was used as an internal reference, and GAPDH was used as a positive control. The right panel represents PLCG1 levels in lysosomes, normalized to the NC group. (C) levels of PLCG1 in lysosomes of NPC after transfection of si-LAMP2A or si-ATG7. LAMP1 was used to demonstrate equal lysosomal loading. The right panel represents relative PLCG1 levels in lysosomes, normalized to the si-Scb group. (D) immunofluorescence imaging showing colocalization of PLCG1 with HSPA8 or LAMP2A in NPC treated with or without STV (24 h). The quantitative analysis of colocalization was assessed using ImageJ-colocalization Finder, and the mean Pearson correlation index is as indicated. (E) LAMP2A levels detected in immunoprecipitates of PLCG1 from NPC treated with STV (24 h) or not. (F) immunoblot of HSPA8 that bound with PLCG1-HA in NPC transfected as indicated. (G) levels of PLCG1-HA in immunoprecipitates of HSPA8 in NPC transfected with WT PLCG1, PLCG1^Q52R, or PLCG1^Q272R plasmids. (H) effect of STV on PLCG1-HA levels in NPC transfected as indicated. The right panel represents relative levels of PLCG1-HA normalized to the NC group. (I) NPC transfected as indicated were treated with STV or not, followed by immunoprecipitation of HA and immunoblot detection of LAMP2A. The right panel represents the relative levels of LAMP2A, normalized to the WT-NC group. (J) NPC transfected as indicated were treated with or without N/L for 12 h, and the PLCG1-HA levels in each group were detected by immunoblotting. PLCG1-HA levels in each group were normalized to the WT-NC group for comparison. Fold changes of PLCG1-HA levels in each group after N/L treatment are shown in the lower left panels. (K) effect of Q52R and Q272R mutations on the association between PLCG1 and lysosomes. Lysates of transfected NPC as indicated were incubated with lysosomes purified from HEK293T cells. Levels of PLCG1-HA in recovered lysosomes were determined using immunoblot in the presence or absence of PI. The right panel shows relative levels of PLCG1-HA normalized to the NC group. All figures show mean ± SEM of at least three independent experiments. */#p < 0.05, **/##p < 0.01, ***/###p < 0.001, ns means not significant. p-value symbols represent comparisons to the left group of the labeled line (B, C, H, I, J, and K).

PLCG1 induces cellular senescence in a calcium-dependent manner

PLCG1 is a major member of the PLC family that increases intracellular Ca²⁺ flux through IP3 signaling. As a specific substrate of CMA, PLCG1 was likely to mediate the Ca²⁺ overload in the context of CMA inhibition, which triggered cellular senescence. To explore the relationship between PLCG1, cellular Ca²⁺ levels, and senescence, we stably overexpressed PLCG1 in NPC with lentivirus. We found that PLCG1 overexpression significantly promoted Ca²⁺ levels (Figure 6A), and the expression of senescence-associated proteins TP53, CDKN1A, and CDKN2A, consistent with LAMP2A-KO (Figure 2D) or TNF treatment groups (Figure 6B). The protein level of IL1B/IL-1β, a central factor of SASP, was also increased by PLCG1 (Figure 6B). Meanwhile, immunofluorescence imaging showed Ca²⁺, TP53, and SA-GLB1 levels were uniformly promoted by overexpression of PLCG1 (Figure 6C). These findings suggested that PLCG1 promoted Ca²⁺ overload and senescence of NPC. To explore the role of Ca²⁺ in PLCG1-induced senescence, we treated PLCG1-overexpressed NPC with the Ca²⁺ chelator BAPTA-AM. The results showed that BAPTA-AM significantly inhibited PLCG1-induced expression of TP53, CDKN1A, CDKN2A, and IL1B (Figures 6D,E). Furthermore, PLCG1 also upregulated the mRNA levels of TP53, CDKN1A, and CDKN2A, as well as the mRNA levels of SASP factors, including CCL3, IL1B, IL6, and CXCL8 (Figure S4A and S4B). Notably, the upregulated transcription of these senescence-associated proteins by PLCG1 was comprehensively suppressed by BAPTA-AM (Figure S4A and S4B). Previous studies have reported that PLCG1 is required for Gasdermin D-dependent IL1B release [40]. We proved that PLCG1 increased the mRNA and protein levels of IL1B. After PLCG1 overexpression, we also detected a significant decrease both in the proportion of NPC in the S phase (Figure 6F) and in the level of BrdU (Figure S4C). In contrast, the addition of BAPTA-AM restored the proliferative capacity of NPC (Figure 6F and S4C). In the process of senescence induction, PLCG1 also increased reactive oxygen species (ROS) levels in a Ca²⁺-dependent manner (Figure S4D), which could be associated with mitochondrial damage caused by Ca²⁺ overload [41]. As suggested by previous studies, the presence of oxidative stress may induce a ROS-Ca²⁺ positive feedback loop leading to accelerated cellular senescence [9]. Taken together, these data supported that PLCG1 promoted NPC senescence in a Ca²⁺-dependent manner.

Figure 6.

PLCG1 induces cellular senescence in a calcium-dependent manner. (A) the effect of PLCG1 overexpression on the Ca²⁺ level in NPC. The values of each group were normalized to the NC group for comparison. (B) immunoblotting showing the levels of PLCG1, TP53, CDKN1A, CDKN2A, and IL1B in NPC after PLCG1 overexpression. The TNF treatment (20 ng/mL, 48 h) group was used as a positive control. The right panel represents relative levels of PLCG1 in each group. (C) immunofluorescence showing levels of Ca²⁺, TP53, and SA-GLB1 after PLCG1 overexpression. The lower panel shows the mean cell fluorescence intensity for each group. At least 3 fields of view in each group were used for quantitative evaluation. (D) immunofluorescence showing CDKN2A and TP53 levels in NPC after PLCG1 overexpression or combined with BAPTA-AM treatment (5 μM, 24 h). The right panel represents the average fluorescence intensity from at least 3 fields for each group. (E) immunoblot showing the levels of TP53, CDKN1A, CDKN2A, and IL1B in NPC treated as indicated. The right panel represents the relative levels of TP53, CDKN1A, CDKN2A, and IL1B in each group, normalized to the NC group. (F) flow cytometry results of the cell cycle of NPC treated as indicated. The proportion of cells in S phase was used for comparison. (G) levels of Ca²⁺ and SA-GLB1 in NPC treated as indicated, quantified by the mean fluorescence intensity. Representative fluorescence images are provided in (fig. S4E). (H) immunoblotting showing the levels of PLCG1, TP53, CDKN1A, CDKN2A, and IL1B in NPC treated as indicated. The cell administration method is as described above. The right panel represents relative band intensities of interest, normalized to the NC group. (I) quantification of the mean cellular fluorescence intensity of PLCG1 in paraffin sections from normal or degenerated human disc specimens. Representative fluorescence images are provided in (fig. S5A). (J) MRI of the indicated rat tail discs after injection of TNF (20 ng/mL, 2 μL) or combined with sh-Plcg1 lentivirus (1×10⁸ TU/mL, 2 μL). The SD rats used were male and 2 months old, and the treatment methods were consistent with (fig. 1I). (K) histological staining of the indicated rat discs injected with TNF or combined with sh-Plcg1 lentivirus, including H&E, Safranin O-fast green, and Masson staining. (L) the intervertebral disc height index (DHI, mm) of the treated disc segments. DHI is negatively correlated with the degree of degeneration. (M) histological scoring of treated rat intervertebral discs. The score (0–15) is positively correlated with the degree of disc degeneration [39]. (N) quantitative results of the average cellular fluorescence intensity of PLCG1, TP53, CDKN1A, and CDKN2A in paraffin sections of rat intervertebral discs of each group. Representative fluorescence images were provided in (fig. S5D). All figures show mean ± SEM of at least three independent experiments. */#p < 0.05, **/##p < 0.01, ***/###p < 0.001, ns means not significant. p-value symbols represent comparisons to the left group of the labeled line (A, B, C, F, H, I, and N) or to the PLCG1+DMSO group (D and E) or to the TNF+si-scb group (G) or to the sh-Scb+TNF group (L and M).

Given that TNF induced CMA inhibition, elevated PLCG1 may also be involved in TNF-induced NPC senescence. Immunoblotting showed that PLCG1 levels in NPC were elevated after TNF stimulation (Figure 6B). Next, we knocked down PLCG1 with si-PLCG1, which attenuated TNF-induced Ca²⁺ overload (Figure 6G and S4E-F). Meanwhile, si-PLCG1 also inhibited TNF-induced activation of SA-GLB1 (Figure 6G and S4E) and downregulation of BrdU levels (Figure S4G). Similarly, TNF-triggered upregulation of TP53, CDKN1A, CDKN2A, and IL1B was also inhibited by si-PLCG1 (Figure 6H). Compared with human normal discs, elevated PLCG1 levels were found in human degenerated disc samples (Figure 6I and S5A-B). Next, we evaluated the role of PLCG1 in TNF-induced disc senescence and degeneration in rats. We generated IVDD rat models with TNF intradiscal injection, and knocked down Plcg1 locally with lentivirus injection. We found that knockdown of Plcg1 ameliorated TNF-induced IVDD, as manifested by increased disc height (Figure 6J-K), reduced Pfirrmann grades (Figure S5C), and decreased histological scores (Figure 6M). Tissue immunofluorescence confirmed the knockdown efficiency of Plcg1 (Figure S5D). Compared with the TNF alone treatment group, the indicated discs co-treated with Plcg1-knockdown (KD) showed larger NP area, clearer boundaries, and more ordered structures (Figure 6K). Meanwhile, Plcg1-KD decreased the levels of senescence markers, including TP53, CDKN1A, and CDKN2A in degenerated discs (Figure 6N and S5D). Taken together, these findings supported that PLCG1 mediated TNF-induced NPC senescence in IVDD.

Abnormal accumulation of PLCG1 mediates CMA inhibition-induced cellular senescence

CMA inhibition triggered Ca²⁺ overload, which constituted the impetus of TNF-induced NPC senescence. After a proteomic screen, we identified PLCG1 as a substrate of CMA and established the role of PLCG1 in Ca²⁺ overload and senescence of NPC. Next, we performed transient transfection of si-PLCG1 in LAMP2A-KO NPC to observe the effect of PLCG1 knockdown on CMA inhibition-induced Ca²⁺ overload and senescence. We found that PLCG1-KD significantly inhibited the LAMP2A-KO-induced increase of Ca²⁺ levels (Figure 7A,B). At the same time, si-PLCG1 or the PLCG1 inhibitor (U-73122) abolished LAMP2A-KO-induced activation of SA-GLB1 (Figure 7A and S5E). Like BAPTA-AM treatment, PLCG1-KD or the PLCG1 inhibitor counteracted the upregulation of TP53, CDKN1A, CDKN2A, and IL1B by LAMP2A-KO in NPC (Figure 7C and S5F-G). Meanwhile, both PLCG1-KD and the PLCG1 inhibitor reversed LAMP2A-KO-induced cell cycle arrest of NPC (Figure 7D,E and S5H-I). At the transcriptional level, PLCG1-KD uniformly reduced the mRNA levels of TP53, CDKN1A, CDKN2A, and SASP factors (IL1B, CCL3, IL6, and CXCL8) increased by LAMP2A-KO (Figure 7F,G). Taken together, these findings supported the blocked degradation of PLCG1 as the trigger of cellular senescence in the context of CMA inhibition. To further determine the role of this mechanism in the context of TNF-induced senescence, we performed lentiviral co-transfection of LAMP2A and PLCG1 in NPC. Fluorescent probes revealed that overexpression of LAMP2A alone attenuated TNF-induced Ca²⁺ overload and SA-GLB1 activation (Figure 7H,I), as previously shown. However, this protective effect of CMA was abolished when PLCG1 was co-expressed (Figure 7I). Furthermore, LAMP2A-overexpressing NPC were resistant to TNF-induced expression of senescence-associated proteins (TP53, CDKN1A, CDKN2A, and IL1B), which was reversed by co-expression of PLCG1 (Figure 7J-K and S5J). Consistently, PLCG1 also abolished the protective effect of LAMP2A on TNF-induced cell cycle arrest (Figure 7L). In short, these findings supported that the impaired degradation of PLCG1 by CMA mediated TNF-induced senescence of NPC.

Figure 7.

Abnormal accumulation of PLCG1 mediates CMA deficiency-induced cellular senescence. (A) fluorescence levels of Ca²⁺, SA-GLB1, and TP53 in NPC treated with LAMP2A-KO or combined with PLCG1-KD. The Ca²⁺ levels were visualized by fluo-4 AM probe and quantified by the mean fluorescence intensity. The lower panel represents the mean fluorescence intensity of each group. At least 3 fields of view were used for quantitative assessment. (B) Ca²⁺ levels in NPC treated with LAMP2A-KO or combined with PLCG1-KD, measured by fura-2. Values of each group were normalized to that of the NC group. (C) immunoblotting showing the levels of TP53, CDKN1A, CDKN2A, and IL1B in NPC treated as indicated. Right panel represents between-group comparisons of band of interest, normalized to the sg-Scb group. (D) flow cytometry results of the cell cycle of NPC treated as indicated. The cell proportion in S phase was used for comparison between the indicated groups. (E) absorbance of BrdU measured at 450 nm in NPC treated with LAMP2A-KO or combined with si-PLCG1, normalized to the value of sg-Scb group. (F and G) rt-qPCR results showing the mRNA levels of TP53, CDKN1A, CDKN2A, IL1B, CCL3, IL6, and CXCL8 of NPC treated as indicated. (H) Ca²⁺ levels of NPC treated with TNF (20 ng/mL, 48 h) or combined with vector, LAMP2A, and LAMP2A+PLCG1 overexpression. The Ca²⁺ levels were measured by fura-2 at 340/510 nm. (I) fluorescence of Ca²⁺ and SA-GLB1 in NPC treated as indicated, the right panel represents the mean fluorescence intensity of each group. At least 3 fields of view were used for quantitative assessment. (J) immunoblotting showing the levels of TP53, CDKN1A, CDKN2A, and IL1B in NPC treated as indicated, and the relative band intensity of each group was normalized to that of the NC group. (K) the mean fluorescence intensity of TP53, CDKN1A, and CDKN2A in NPC treated as indicated. Representative fluorescence images are presented in (fig. S5J). (L) normalized BrdU levels measured at 450 nm from NPC treated as indicated. All figures show mean ± SEM of at least three independent experiments. ^*/#p <0.05, ^**/##p <0.01, ^***/###p <0.001, ns means not significant. p-value symbols represent comparisons to the left group of the labeled line (A, C, D, and H) or to the LAMP2A-KO+si-Scb group (B, E, F, and G) or to the TNF+LAMP2A group (I, J, K, and L).

Discussion

The issue of low back pain, which has been widely associated with the aging population, is mainly caused by IVDD and is closely linked to the senescence of NPC induced by inflammatory factors [29,42]. Previous studies have shown that combined treatment with TNF and IFN-γ inhibits CMA activity at multiple levels, including reducing mRNA and protein levels of LAMP2A, increasing cathepsin A-dependent degradation of LAMP2A, and activating AKT signaling [43]. This is consistent with our finding that TNF inhibited CMA. CMA activity decreases with age and in response to persistent inflammatory stimuli [43]. Both conditions are associated with IVDD. Here, we found that CMA inhibition mediated TNF-induced NPC senescence. Previously, CMA deficiency has been shown to cause accumulation of toxic proteins, decreased defense against oxidative stress, and increased inflammasomes in age-related diseases [26,36,44–48]. Through sequencing analysis, we found that CMA inhibition significantly promoted the calcium signal in NPC. Ca²⁺ is a ubiquitous messenger that plays an important role in processes such as cell proliferation, differentiation, and senescence [49]. Importantly, dysregulation of Ca²⁺ signaling in response to stimuli, including mechanical stress, oxidative stress, and inflammation, is considered to be a key event in the development of disc senescence, a notion known as the “calcium hypothesis” of IVDD [33,50,51]. By blocking Ca²⁺, we found that CMA inhibition- or TNF-induced senescence of NPC was attenuated. Meanwhile, restoring the expression of LAMP2A also alleviated TNF-induced Ca²⁺ overload. Although at the basal level, the regulation of Ca²⁺ flux by CMA activation was redundant, the disruption of Ca²⁺ homeostasis by CMA inhibition was significant and sufficient to initiate Ca²⁺ overload-induced cellular senescence. Our data supported that the fine regulation of Ca²⁺ homeostasis to maintain a youthful state in NPC was dependent on the proper function of CMA.

The role of abnormal autophagy in senescence has received increasing attention [52]. Previous studies have reported that CMA activity is constitutively upregulated or downregulated in senescent cells, which may be due to differences in the role of CMA as a compensatory response to stress in different cells [17]. Another study reported that CMA helps maintain the normal function of the KEAP1-NRF2 pathway to counteract the accumulation of oxidative stress during cellular senescence [44]. Given the diversity of proteins that CMA can selectively degrade, and the fact that some of these cellular pathways are activated or repressed over time during cellular senescence [12], we hypothesized that CMA could be involved in NPC senescence through the timely degradation of Ca²⁺ regulators. According to several characteristics of CMA substrates, including possession of KFERQ-like sequences, increase after CMA inhibition, and interaction with HSPA8 and LAMP2A [16], we performed proteomic analysis and obtained a set of proteins that were upregulated after LAMP2A-KO and another set of proteins that interacted with HSPA8 and LAMP2A. Excessive Ca²⁺ flux in cells can be mediated by Ca²⁺ channel receptors on the cell membrane in response to external stimuli, or from ITPR/IP3R-mediated Ca²⁺ release in the ER. These uncontrolled Ca²⁺ fluxes have been shown to be important triggers of cellular senescence. By cross-analyzing the MS-predicted proteins and the gene sets of calcium signaling pathways, we speculated that PLCG1, which meets the characteristics of CMA substrates, may be the mediator through which CMA affected Ca²⁺ levels. The expression of PLCG1 and its corresponding homologues is highly conserved; however, we did not observe the presence of other members of the PLC family, including PLCB/PLC-β and PLCD/PLC-δ, in our MS-based screen for CMA substrates, which could be attributed to different conditions and specific cell types.

PLCG1 mediates IP3 generation, one of the major signals for cellular Ca²⁺ mobilization. IP3 triggers Ca²⁺ mobilization from two Ca²⁺ stores (the ER and the extracellular medium) into the cytosol [53]. Binding of IP3 to the ITPR/IP3R on the ER not only results in a rapid release of Ca²⁺ from the ER, but also a sustained Ca²⁺ influx from the extracellular medium mediated by Ca²⁺-activated channels in the plasma membrane [54]. This explains the finding of the persistent increase in Ca²⁺ levels in PLCG1-overexpressing NPC. Emerging evidence suggests that PLCG1 is involved in cancer [25], immune disorders [55], pyroptosis, and neurodegenerative diseases [56]. PLCG1 has been regarded as an important therapeutic target, and several small molecule pharmacological inhibitors of PLCG1, including U73122, ATA, and ritonavir, have been developed [57,58]. However, the development of new PLCG1 inhibitors faces high cytotoxicity caused by off-target effects [59], highlighting the importance of exploring the endogenous inhibition of PLCG1. Little is known about the endogenous regulation of PLCG1. Here, we unexpectedly found significant evidence for the degradation of PLCG1 via CMA. Further, we systematically verified that PLCG1 was a specific substrate of CMA and identified ₂₆₈VDRLQ₂₇₂ as the KFERQ-like sequence through which HSPA8 recognized PLCG1 and targeted CMA for its degradation. Our results suggested that macroautophagy and proteasome were not involved in the degradation of PLCG1 under basal conditions, which requires further investigation in other conditions or cell types. Active CMA helped maintain Ca²⁺ homeostasis in NPC. However, in CMA-inactive NPC, PLCG1 degradation was blocked, which consistently promoted cytoplasmic Ca²⁺ overload, thereby inducing premature NPC senescence and IVDD progression (Figure 8). PLCG1 accumulation caused by CMA inhibition also mediated TNF-induced senescence in NPC, which was alleviated by PLCG1 knockdown or BAPTA-AM. Interestingly, the association of the PLC family with senescence has been reported in several diseases and models. Recently, a role for PLC in shortening Drosophila lifespan by promoting mitochondrial and lysosomal Ca²⁺ overload was identified [7]. PLC is also involved in the senescence of bone marrow stromal cells, and selective inhibition of PLC activity reduces the levels of SA-GLB1 [60]. Collectively, these lines of evidence support the role of PLCG1 as a key hub linking Ca²⁺ overload, CMA deficiency, and senescence.

Figure 8.

Schematic representation of the mechanisms by which CMA inhibition mediates NPC senescence and IVDD. Active CMA contributes to the maintenance of Ca²⁺ homeostasis in NPC. In CMA-inactive NPC, PLCG1 degradation is blocked, which continuously promotes cytoplasmic Ca²⁺ flux and Ca²⁺ overload, thereby inducing premature NPC senescence and IVDD progression.

We found that PLCG1 was elevated in degenerated IVD and involved in the inflammation-induced senescence of NPC. However, in addition to promoting the release of IP3-mediated Ca²⁺ signals, PLCG1 also mediates the generation of the second messenger diacylglycerol, which participates in the processes of cell proliferation, differentiation, senescence, and apoptosis [61–64]. Therefore, the role of PLCG1 in IVDD needs more comprehensive research. In fact, studies have reported that inhibition of PLCG1 can also enhance autophagy to increase collagen type II and aggrecan levels in chondrocytes [65], suggesting that inhibition of PLCG1 can be beneficial in promoting extracellular matrix regeneration in IVDD, which needs further experimental evidence. Furthermore, inhibition of PLCG1 reduces mitochondrial ROS formation and prevents lipid peroxidation, thereby antagonizing apoptosis and increasing cell proliferation [66]. This is consistent with our study that PLCG1 increased ROS levels and promoted cell cycle arrest in a Ca²⁺-dependent manner. However, the response mechanism of PLCG1-induced oxidative stress and the involved pathological processes of disc degeneration need to be further elucidated. PLCG1 has also been reported to play a major role in VEGF-driven angiogenesis [54]. As an important feature of IVDD, angiogenesis can aggravate the inflammatory input and nerve invasion of the IVD, which is related to the aggravation of low back pain [67,68]. We have detected increased expression of PLCG1 in degenerated IVD, but it is unclear whether this is the cause of vascular invasion in IVDD, which requires further study.

Materials and methods

Human sample collection and ethics approval

Human degenerative disc tissues (n = 78, age 39–72 years, female n = 36, male n = 42) were obtained from patients undergoing discectomy for disc herniation or spinal stenosis. The grade of intervertebral disc degeneration in all patients was evaluated using the Pfirrmann scoring standard. Control specimens were obtained from patients undergoing surgery for thoracolumbar fractures or acquired scoliosis (n = 72, age 16–42 years, female n = 34, male n = 38). The study was conducted in accordance with the Declaration of Helsinki, and written informed consent was obtained from all included patients according to the committee’s recommendations. The research protocol using patient samples in this experiment was approved by the Ethics Committee of Tongji Medical College, Huazhong University of Science and Technology (No. S341).

Isolation, culture and treatment of nucleus pulposus cells (NPC)

Healthy human nucleus pulposus (NP) samples were isolated and cut into small pieces of 1 mm³, then digested with 0.2% type II collagenase (BioFroxx, 2275GR001) at 37°C for 8 h until the cells were completely detached from the tissue. The digest was centrifuged at 300 ×g for 5 min and then cultured in medium (Dulbecco’s Modified Eagle Medium/Nutrient Mixture F-12; Thermo Fisher Scientific 11,320,033) containing 10% fetal bovine serum (Thermo Fisher Scientific 10,099–141), 1% penicillin-streptomycin (Sigma-Aldrich, P4333) at 37°C, 5% CO₂. When the cells reached 75% confluence, they were digested with 0.25% trypsin (Thermo Fisher Scientific 25,200,072) and passaged. NPC of passage 2 were used in subsequent experiments. For cell treatment, NPC were treated with 20 ng/mL TNF (Sigma-Aldrich, T6674) in culture medium for 48 h to induce senescence. NPC were treated with ATRA (MedChemExpress, HY-14649) at a concentration of 10 μM for 48 h to inhibit CMA activity. NPC were treated with BAPTA-AM (MedChemExpress, HY-100545) at a concentration of 5 μM for 24 h to inhibit calcium overload. NPC were treated with 20 mM NH₄Cl (Sigma-Aldrich 326,372) and 10 μM leupeptin (MedChemExpress, HY-18234) separately or combined (N/L) for 3–12 h to inhibit lysosomal autophagy. Different concentrations of MG132 (MedChemExpress, HY-13259) were used to inhibit the ubiquitin-proteasome system of NPC. NPC were treated with 5 mM 3-MA (MedChemExpress, HY-19312) for 3–12 h to inhibit macroautophagy activity. AR7 (MedChemExpress, HY-101106) and QX77 (Selleck, S6797) were used to activate CMA. 3CAI (MedChemExpress, HY-16666) was used to inhibit the AKT signaling pathway. NPC were treated with U-73122 (Selleck, S8011) at a concentration of 2 μM for 24 h to inhibit PLCG1 activity. For the dosing group, DMSO-treated cells were used as controls. For the transfection group, cells transfected with si-Scb, sh-Scb, or vector were used as controls.

Senescence associated GLB1/β-galactosidase (SA-GLB1/β-gal) staining

Senescence β-galactosidase staining kit (Beyotime, C0602) or fluorescein di(β-D-galactopyranoside) (FDG; MedChemExpress, HY-101895) was used to stain SA-GLB1. The cells in the plate were washed with PBS (Thermo Fisher Scientific 10,010,023), and then were fixed with β-galactosidase staining fixative at room temperature for 15 min. After aspirating the cell fixative, the cells were washed with PBS for three times, 5 min each time. Then the cells in each well were incubated with staining working solution at 37°C without CO₂ for 12 h. An inverted phase-contrast optical microscope (Olympus, Japan) was used for observation. For fluorescent staining of SA-GLB1, an aliquot of reaction buffer was added to each well. Then, 2 mM FDG was added per well and the plate was kept in the dark at 37°C for 24 h without supply of CO₂. Images were visualized and captured using a fluorescence microscope (Olympus, Japan). All experiments were repeated at least three times.

Western blot

The RIPA lysis buffer (Biosharp, BL504A) pre-mixed with protease and phosphatase inhibitor cocktail (Beyotime, P1049) and PMSF (Beyotime, ST506) was used to lyse the sample at 4°C. The samples were centrifuged (12000 ×g, 15 min) after sonication, and the supernatant was collected. The protein concentration of the supernatant was quantified using the BCA protein quantification kit (Sevenbio, SW201–02). For protein denaturation, the protein supernatant was boiled for 10 min after adding SDS loading buffer (Epizyme, LT103). After electrophoresis in SDS-polyacrylamide gels, proteins were transferred to PVDF membranes and incubated in blocking buffer for 1 h. Membranes were then incubated overnight at 4°C with specific primary antibodies (primary antibodies used are listed in Table 1; the anti-LAMP2C antibody was produced as described previously [69]). Subsequently, the bands were washed in TBST (Tris-buffered saline with Tween 20 [Sigma-Aldrich, T9039]) for 30 min and incubated with HRP-conjugated secondary antibodies (ABclonal, AS014 and AS003) for 1 h at room temperature. Enhanced chemiluminescence reagent (MedChemExpress, HY-K1005) and ChemiDoc MP System (Bio-Rad, USA) were used to visualize proteins. ImageJ was used to quantify the protein abundance of the bands.

Table 1.

Primary antibodies.

Product name	Catalog	Manufacturer	Dilution Ratio
LAMP2A	ab125068	Abcam	1:1000 (WB); 1:200 (IF)
LAMP2B	ab18529	Abcam	1:1000
PLCG1	R27162	ZENBIO	1:2000 (WB); 1:200 (IF)
TP53	10442-1-AP	Proteintech	1:5000 (WB); 1:200 (IF)
CDKN1A	10355-1-AP	Proteintech	1:2000 (WB); 1:200 (IF)
CDKN2A	ab270058	Abcam	1:2000 (WB); 1:200 (IF)
ATG7	10088-2-AP	Proteintech	1:2000
IL1B	16806-1-AP	Proteintech	1:2000
SQSTM1	18420-1-AP	Proteintech	1:2000
TUBA	T40103	Abmart	1:10000
GAPDH	60004-1-Ig	Proteintech	1:10000
LAMP1	21997-1-AP	Proteintech	1:2000 (WB); 1:200 (IF)
HA	51064-2-AP	Proteintech	1:3000
CTSB	R381757	ZENBIO	1:1000
CTSD	R380946	ZENBIO	1:1000
CTSZ	R23743	ZENBIO	1:1000

Immunoprecipitation

Cell lysis buffer (Beyotime, P0013) was used to lyse the control or transfected cells after premixing with protease and phosphatase inhibitor cocktail (Beyotime, P1049) and PMSF (Beyotime, ST506). After centrifugation (12,000 ×g, 15 min) at 4°C, the protein supernatant was collected. Protein in the supernatant was immunoprecipitated with anti-protein A/G magnetic beads (MedChemExpress, HY-K0202) and the target antibody. The collected target proteins were separated by SDS-PAGE gel electrophoresis, and the bands were incubated with the designated primary antibodies and HRP-conjugated secondary antibodies successively. Proteins were then visualized using enhanced chemiluminescence reagent (MedChemExpress, HY-K1005) and the ChemiDoc MP System (Bio-Rad, USA). Protein abundance was quantified using ImageJ.

Real-time quantitative reverse transcription PCR (rt-qPCR)

The RNA extraction kit (Vazyme, RC101) was used to extract RNA from cultured cells or NP tissues. According to the manufacturer’s instructions, reverse transcription was performed using the cDNA synthesis kit (TransScript, AE311–02), followed by RT-qPCR using qPCR SuperMix reagent (TransScript, AQ601–01). The 7500 Fast RT-PCR System (Thermo Fisher Scientific, USA) was used for RT-qPCR. Table 2 provides the primer sequences used.

Table 2.

Primer sequences.

Gene	Primers
TP53	Forward 5′-CAGCACATGACGGAGGTTGT-3′
	Reverse 5′-TCATCCAAATACTCCACACGC-3′
CDKN1A	Forward 5′-TGTCCGTCAGAACCCATGC-3′
	Reverse 5′-AAAGTCGAAGTTCCATCGCTC-3′
CDKN2A	Forward 5′-GATCCAGGTGGGTAGAAGGTC-3′
	Reverse 5′-CCCCTGCAAACTTCGTCCT-3′
IL1B	Forward 5′-AGCTACGAATCTCCGACCAC-3′
	Reverse 5′-CGTTATCCCATGTGTCGAAGAA-3′
CCL3	Forward 5′-CAGAATCATGCAGGTCTCCAC-3′
	Reverse 5′-GCGTGTCAGCAGCAAGTG-3′
IL6	Forward 5′-ACTCACCTCTTCAGAACGAATTG-3′
	Reverse 5′-CCATCTTTGGAAGGTTCAGGTTG-3′
CXCL8	Forward 5′-AGACAGCAGAGCACACAAGC-3′
	Reverse 5′-ATGGTTCCTTCCGGTGGT-3′
LAMP2A	Forward 5′-AACTTCCTTGTGCCCATAGC-3′
	Reverse 5′-AGCATGATGGTGCTTGAGAC-3′
LAMP2B	Forward 5′-AGAGTGTTCGCTGGATGATG-3′
	Reverse 5′-TGCCAATTACGTAAGCAATCA-3′
LAMP2C	Forward 5′-AAGGGTTCAGCCTTTCAATG-3′
	Reverse 5′-ACAATTATAAGGAAGCCCAAGG-3′
PLCG1	Forward 5′-GGAAGACCTCACGGGACTTTG-3′
	Reverse 5′-GCGTTTTCAGGCGAAATTCCA-3′
GAPDH	Forward 5′-ACAACTTTGGTATCGTGGAAGG-3′
	Reverse 5′-GCCATCACGCCACAGTTTC-3′

Immunofluorescence staining

First, cells or tissues were fixed with 4% paraformaldehyde for 30 min and the cell membrane was permeabilized with 0.5% Triton X-100 (BioFroxx, 1139ML100). Sections were washed with PBS and incubated with 2% bovine serum albumin (BSA; Sigma-Aldrich, A2153) for blocking at room temperature for 2 h, followed by overnight incubation at 4°C with the corresponding primary antibodies. The primary antibodies used were Anti-LAMP2A (1:200 dilution; Abcam, ab125068), Anti-TP53 (1:200 dilution; Proteintech 10,442–1-AP), Anti-CDKN1A/p21 (1:200 dilution; Proteintech 10,355–1-AP), Anti-CDKN2A/p16-INK4A (1:200 dilution; Abcam, ab270058), Anti-PLCG1 (1:200 dilution; ZENBIO, R27162). Sections were washed three times and incubated with the corresponding fluorochrome-conjugated secondary antibodies for 1 h at room temperature. Lysosomes were stained with LysoTracker kit (Beyotime, C1046). Nuclei were counterstained with 4’,6-diamidino-2-phenylindole (DAPI; Beyotime, C1002). The slides were imaged under a fluorescence microscope (Olympus, Japan). ImageJ was used to quantify the mean fluorescence intensity.

Animal model of intradiscal injection

The surgical method was as described previously [29]. Sprague-Dawley rats (250 ± 20 g) were reared under specific-pathogen-free (SPF) conditions at a room temperature of 24°C and 50% humidity. The rats were anesthetized using pentobarbital (3%, w: v, 2 mL/kg). Without knowing the group allocation, the operator used needles (33 G) to vertically inject pre-prepared drugs, including PBS, TNF, vector-loaded lentivirus, Lamp2a-loaded lentivirus, sh-Scb-loaded lentivirus, and sh-Plcg1-loaded lentivirus into the Co7/8 and Co8/9 intervertebral discs of each rat. The volume of each injection was 2 μL, and the depth was about 5 mm. Injections were administered once a week for a month, followed by a maintenance period of another month. All animals were allowed free movement without weight restrictions.

Imaging analysis and histological evaluation of the rat IVDD model

At the eighth week after surgery, all rats were photographed by magnetic resonance imaging (MRI) after anesthesia. During imaging, rats were maintained in a prone position with their tails in a straight line. The disc height index (DHI) and the degree of disc degeneration were determined based on MRI results and Pfirrmann grading as previously described [39]. Rats were sacrificed after MRI examination, and disc specimens were collected. After the specimens were fixed, decalcified, dehydrated, and embedded in paraffin, they were cut into 4 μm thick sections. Sections were then subjected to H&E staining, Safranin O-fast green staining, and Masson staining for histological assessment of the degree of disc degeneration [28].

Cell cycle analysis

The Cell Cycle and Apoptosis Analysis Kit (Beyotime, C1052) was used for cell cycle analysis. Sample preparation was performed according to the manufacturer’s instructions. Briefly, cells were harvested and resuspended in 70% ethanol and fixed at 4°C for 12 h. Then, they were washed with pre-cooled PBS and incubated with the prepared propidium iodide staining solution. The red fluorescence of the sample was detected at an excitation wavelength of 488 nm by a flow cytometer (Sony Biotechnology, Japan), and the light scattering was analyzed. Cell cycle analysis was performed using FlowJo V10 (BD Biosciences, USA) analysis software, and the proportion of cells in the S phase was quantified.

BrdU detection

According to the manufacturer’s instructions, the BrdU cell proliferation assay kit (Abcam, ab287841) was used to detect the cell proliferation marker BrdU. Control or treated NPC were cultured in 96-well plates at 5000 cells per well. The BrdU solution was added to the wells followed by incubation at 37°C for 4 h. The medium was removed and the cells in each well were fixed with 100 µL of fixation solution for 30 min at room temperature. Then the solution was carefully aspirated and 100 μL of BrdU detection solution was added to each well, followed by incubation at room temperature with gentle shaking for 1 h. After washing, 100 μL anti-mouse HRP-linked antibody solution was added per well and the plate was incubated at room temperature for another 1 h. After that, 100 µL TMB substrate was added into each well and the absorption was measured at 650 nm for 30 min. 100 µL of stop solution was added into each well to stop the color development, and the absorption was measured at 450 nm.

Determination of intracellular calcium levels

For fluorescent staining of calcium ions in cells, Fluo-4 AM (Beyotime, S1060) was used for detection according to the manufacturer’s instructions. The cells were washed with PBS three times, followed by incubation with Fluo-4 AM working solution at 37°C for 60 min for fluorescent probe loading. Next, the plate was washed three times with PBS, and incubated for another 30 minutes to ensure complete conversion of Fluo-4 AM to Fluo-4 in the cells. Fluorescence of Fluo-4 was detected with a fluorescence microscope to determine changes in intracellular calcium concentration. In addition, calcium levels were measured in 96-well microplates using the Fura-2 Calcium Flux Assay Kit (Abcam, ab176766). Briefly, Fura-2 solution was added to the cells. The cells were incubated at 37°C for 1 h, then incubated at room temperature for 20 min. Calcium levels were analyzed and quantified using a Victor Nivo microplate reader (PerkinElmer, USA) at Ex/Em 340/510 nm, and each group of values was normalized to the control group.

RNA extraction, sequencing and analysis

Total RNA was extracted from NPC using TRIzol reagent (Thermo Fisher Scientific 15,596,026). Nanodrop ND-2000 (Thermo Fisher Scientific, USA) and Agilent Bioanalyzer 4150 (Agilent Technologies, USA) were used to check the A260/A280 absorbance ratio and RIN value of RNA samples. The PE library was constructed using the mRNA-sequencing Lib Prep Kit (ABclonal, RK20350). The mRNA was purified from 1 μg of total RNA using oligo (dT) magnetic beads, followed by mRNA fragmentation in first strand synthesis reaction buffer. Subsequently, using the mRNA fragment as a template, the first strand of cDNA was synthesized using random primers and reverse transcriptase (RNase H), and then the second strand of cDNA was synthesized using DNA polymerase I, RNase H, buffer, and dNTPs. Synthetic double-stranded cDNA fragments were ligated with adapter sequences for PCR amplification. After PCR products were purified, Agilent Bioanalyzer 4150 was used for library quality assessment. Finally, the NovaSeq 6000 sequencing platform (Illumina, USA) was used for sequencing with paired-end 150 bp reads. Data generated from the Illumina platform were used for bioinformatics analysis.

Tmt-based mass spectrometry (TMT-MS)

NPC were infected with a lentiviral vector targeting LAMP2A (sg-LAMP2A) or a nonspecific control (sg-Scb). The cells were harvested and lysed with SDT buffer (4% SDS, 100 mM Tris-HCl, 1 mM DTT, pH 7.6). The amount of protein was quantified using the BCA Protein Assay Kit (Bio-Rad 5,000,002). Peptide mixes (100 µg per sample) were labeled using TMT reagent according to the manufacturer’s instructions (Thermo Fisher Scientific 90,110). MS analysis was performed on a Q Exactive mass spectrometer (Thermo Fisher Scientific, USA) that was coupled to Easy-nLC (Thermo Fisher Scientific, USA) for 60/90 min. Protein identification was performed by searching spectra against the UniProt database, and the ratio of 114:113 was calculated to analyze the relative expression levels of proteins after LAMP2A knockout [70].

Mass spectrometry-based immunoprecipitation proteomics analysis

To identify interacting proteins of LAMP2A and HSPA8, cell lysates were collected from NPC overexpressing LAMP2A or HSPA8 and immunoprecipitated with protein A/G magnetic beads conjugated to the corresponding antibodies. The immunoprecipitated bead samples were washed three times with pre-cooled PBS, and then incubated in the reaction buffer (1% SDC, 100 mM Tris-HCl, pH 8.5, 10 mM TCEP, 40 mM CAA) at 95°C for 10 min for protein denaturation, cysteine reduction and alkylation. The eluate was trypsinized overnight at 37°C. The peptides were purified using a homemade SDB desalting column. MS data acquisition was performed on a Q Exactive HF-X mass spectrometer (Thermo Fisher Scientific, USA) and Easy-nLC 1200 system (Thermo Fisher Scientific, USA) according to the manufacturer’s instructions. MS raw data were analyzed with MaxQuant (v1.6.6) using the Andromeda database search algorithm. The “proteingroups.txt” file generated by MaxQuant was used for further analysis. The mean and SD of the protein fold change were calculated, and the mean ±1.64 SD was set as the cutoff. The fold change of proteins beyond the cutoff line was identified as significant.

RNA interference and plasmid transfection

The wild-type plasmid WT PLCG1, and mutant plasmids PLCG1^Q52R and PLCG1^Q272R used to overexpress PLCG1-HA were designed and provided by Miaoling company (Wuhan, China). Lentiviral encapsulation and centrifugal concentration of the plasmids were carried out using the Lenti-Easy packaging mix (GeneChem, LPK001) according to the manufacturer’s instructions. After verifying the infection efficiency, NPC were used for subsequent processing. Knockdown of ATG7, LAMP2A, and PLCG1 in NPC was achieved by transient transfection of siRNAs. SiRNAs against human ATG7 (si-ATG7), human LAMP2A (si-LAMP2A), human PLCG1 (si-PLCG1), and control siRNA (si-Scb) were provided by RIBOBIO (Guangzhou, China) and were delivered using Lipofectamine 2000 (Thermo Fisher Scientific 11,668,027) according to standard protocols. CRISPR-Cas9-based knockout of human LAMP2A was achieved by lentiviral-delivered sgRNA provided by GeneChem (Shanghai, China). Sg-LAMP2A#3 was used for subsequent experiments. After verifying the efficiency of gene silencing or knockout, NPC were used for subsequent treatments. Knockdown of rat Plcg1 was achieved by lentiviral-delivered shRNA. The plasmid used to overexpress human LAMP2A was constructed by GeneChem (Shanghai, China) and packaged in lentivirus. The lentivirus-packaged rat Lamp2a was from FengHui company (ChangSha, China). The sequences of si-RNA, sh-RNA and sg-RNA are listed in Tables 3 and 4.

Table 3.

Si-RNA and sh-RNA sequences.

Sequence name	target-sequence
homo-si-ATG7	GAACGAGTATCGGCTGGAT
homo-si-PLCG1	GCTTCTATGTAGAGGCAAA
homo-si-LAMP2A	CTGGGATGTTCTTGTACAA
rat-sh-Plcg1	CCGGGCCAGCTTGTAGCACTCAATTCTCG AGAATTGAGTGCTACAAGCTGGCTTTTT

Table 4.

Sg-RNA sequences.

Sequence name	target-sequence
homo-LAMP2A sgRNA#1-F	5’-CACCGCAACTTCCTTGTGCCCATAG-3’
homo-LAMP2A sgRNA#1-R	5’-AAACCTATGGGCACAAGGAAGTTGC-3’
homo-LAMP2A sgRNA#2-F	5’-CACCGGGTCTCAAGCACCATCATGC-3’
homo-LAMP2A sgRNA#2-R	5’-AAACGCATGATGGTGCTTGAGACCC-3’
homo-LAMP2A sgRNA#3-F	5’-CACCGGTGTTGCTGGCTTATTTTAT-3’
homo-LAMP2A sgRNA#3-R	5’-AAACATAAAATAAGCCAGCAACACC-3’
homo-LAMP2A sgRNA#4-F	5’-CACCGTAGAATAAGTACTCCTGCCA-3’
homo-LAMP2A sgRNA#4-R	5’-AAACTGGCAGGAGTACTTATTCTAC-3’
homo-LAMP2A sgRNA#5-F	5’-CACCGCTGCAGTCTTGAGCTAGATG-3’
homo-LAMP2A sgRNA#5-R	5’-AAACCATCTAGCTCAAGACTGCAGC-3’

Assay of CMA activity using PA-mCherry-KFERQ reporter

As previously described [30,71], the level of CMA was detected by photoconverting the reporter plasmid, pSIN-PAmCherry-KFERQ-NE, which was obtained from Addgene (102365; deposited by Shu Leong Ho). PA-mCherry-KFERQ-NE was transfected into NPC using lentivirus. For CMA activity assays, transfected cells were photoactivated with a 405/20 nm LED array and then subjected to subsequent treatments. More than 90% of cells survived photoconversion. Cells were fixed and co-stained with DAPI before imaging. Images were acquired using a confocal fluorescence microscope (Olympus, Japan). CMA activity was quantified by counting the number of red puncta per cell in at least three fields of view under a 60× objective.

Lysosome isolation

Lysosomal fractions were extracted from cell homogenates by sequential differential centrifugation followed by density gradient centrifugation using the Lysosomal Extraction Kit (Sigma-Aldrich, LYSISO1) according to the manufacturer’s protocol. Briefly, cell homogenates were centrifuged at 1,000 ×g for 10 min at 4°C to remove nuclei, and the supernatant fraction was centrifuged at 20,000 ×g for 20 min at 4°C to pellet lysosomes and other organelles. The resulting supernatant fraction was collected as the cytoplasmic fraction, and the pellet was collected as the crude lysosomal fraction and resuspended in 1× extraction buffer. Next, OptiPrep (Sigma-Aldrich, D1556) and sucrose (Sigma-Aldrich, S5016) were added to construct gradient medium solutions of discrete densities (27%, 20%, 18%, 16%, 12%, 8%) and centrifuged at 15,000 ×g for 4 h at 4°C. Fractions were collected, added with buffer and centrifuged at 20,000 ×g for 15 min to obtain the final lysosomal pellet. The presence of lysosomes was determined using the acid phosphatase assay kit (Sigma-Aldrich, CS0740). Lysosomal integrity was verified by measuring the activity of β-Hexosaminidase, and transmission electron microscopy was used to visualize lysosomal structural integrity.

Lysosomal binding and uptake assays

Assays were performed as previously described [70]. Briefly, purified GST-PLCG1 or cell lysates derived from WT- and MUT-PLCG1 overexpressing NPC were incubated with freshly enriched lysosomes from HEK293T (obtained from ATCC [CRL-3216]) cells in a reaction buffer (10 mM 3-[N-morpholino] propane sulfonic acid [Sigma-Aldrich, M1254], 0.3 M sucrose, pH 7.4) for 20 min at 37°C. The lysosomes were then recovered by centrifugation at 14,000 ×g for 10 min and washed three times with the reaction buffer. The lysosomes were resolved in SDS loading buffer for immunoblot analysis. In the presence of lysosomal protease inhibitors, the detected protein represents the amount of protein bound to the lysosomal membrane and taken up by the lysosome.

Statistical analysis

Results represent mean ± standard error of the mean (SEM). Significant differences were analyzed using a two-tailed Student’s t-test and one-way or two-way ANOVA. Statistical comparisons and graph generation were performed using GraphPad Prism 9 software. All data were obtained from at least three independent experiments. Meanings of significant identifiers: ^*/#p <0.05, ^**/##p <0.01, ^***/###p <0.001. ns means not significant.

Funding Statement

The work was supported by the National Natural Science Foundation of China (NSFC) [No.82172497; No. 81974348].

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data and materials availability

There are no restrictions on data availability in the manuscript. All data are available in the main text or the supplementary materials. All main and supplementary figures have associated source data. All data, code, and materials used in the analyses are available to any researcher for purposes of reproducing or extending the analyses.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/15548627.2024.2395797