3.1 Generation of wild-type and ER retained missense variants of LDLR by site-directed mutagenesis and lentiviral stable transduction of HepG2 cells
The missense variants of LDLR- p.D482H and p.C667F were generated by site-directed mutagenesis of FLAG-tagged Wild Type (WT) LDLR plasmid DNA and subsequently sequenced using Sanger’s Direct DNA Sequencing to confirm the successful introduction of nucleotide substitutions (Supplemental Figure 1). HepG2 cells were stably transduced with the generated plasmids using lentiviruses. Hereafter, stable HepG2 cells expressing FLAG-tagged empty vector, WT-LDLR, and the missense variants of LDLR (p.D482H and p.C667F) are denoted as HepG2mock, HepG2WT+, HepG2D482H+, and HepG2C667F+ in the manuscript. The cellular localization of these plasmids in cells stably expressing WT (HepG2WT+) and variants (HepG2D482H+ and HepG2C667F+) as well as the efficiency of transduction was confirmed by immunostaining with FLAG-tagged antibody (Figure 1A). WT-LDLR was expressed predominantly on the plasma membrane of HepG2WT+, colocalizing with the plasma membrane marker- Na+K+ ATPase, and in the ER, along with the ER-resident chaperone- Calnexin. In HepG2D482H+ and HepG2C667F+, LDLR failed to reach the plasma membrane and is retained in the ER, co-expressing with Calnexin. The partially glycosylated precursor form of LDLR is synthesized in the ER. This immature protein transits from the ER to the Golgi complex where it gets fully glycosylated to the mature protein and gets translocated to the plasma membrane to perform its functions as the receptor of LDL. To elucidate the expression pattern of LDLR in HepG2WT+, HepG2D482H+, and HepG2C667F+, a western blot was performed by probing with anti-FLAG antibody and bands of LDLR precursor (120 kDa) and mature (160 kDa) forms were identified (Figure 1B(i)). The absence of the slowly migrating band that corresponds to the fully glycosylated mature LDLR of 160 kDa in HepG2 cells expressing the missense variants D482H and C667F reiterates the ER retention of these missense variants. The expression of fully glycosylated mature LDLR was significantly reduced in mutants HepG2D482H+ and HepG2C667F+. Non transduced HepG2 and Mock-transduced HepG2 represent the control cells used for the experiments, wherein FLAG tag of 25 amino acids is not expressed in the non-transduced lysates and undetected in the mock-transduced control cells (Figure 1Bi). The differences between WT LDLR expression patterns in HepG2WT+ and mutants in HepG2D482H+ and HepG2C667F+ were significant (Figure 1B(ii)). Overexpression of wildtype LDLR in HepG2 cells causes excessive protein accumulation in the ER, awaiting to be properly folded until it exits the ER for further maturation. This explains the expression of immature protein that is retained in the ER in HepG2WT+ cells.
3.2 Activation of UPR in HepG2 cells expressing mutant forms of LDLR
ER retention of p.D482H-LDLR and p.C667F-LDLR triggers ER stress response as demonstrated by upregulated transcripts of spliced X-box protein (XBP-1s), the downstream target of the IRE1A arm of UPR (19). To investigate whether the ER stress response is initiated by all three arms of the UPR, the relative mRNA expressions of these targets were evaluated using Real-Time PCR (Figure 2A). In conjunction with our previous reports that describe the cellular mechanism of pathogenicity of FH-causing LDLR variants, ER retention of the missense variants induced ER stress and upregulated the UPR arms (19). We assessed the transcripts of all three UPR arms and selected targets that play a vital role in ER Associated Degradation (ERAD) and ER Quality Control (ERQC). Non-transduced HepG2 cells were used as the control for the experiment to evaluate the basal level expression of all the targets studied and to confirm that the changes in transcript profile were not caused by the genetic manipulation employed. Compared to the non-transduced/ HepG2mock, XBP-1(s), ATF4, CHOP, and ATF6 representing the immediate targets of the three branches of UPR were upregulated in HepG2WT+, HepG2D482H+, and HepG2C667F+, with higher fold change in ER retained cells (Figure 2A(ii-v)). Interestingly, P58IPK- the PERK inhibitor, was downregulated in HepG2WT+, HepG2D482H+, and HepG2C667F+ (Figure 2A(vi)). In addition to the surge in transcripts of the UPR signal activators, overexpression of WT-, D482H-, and C667F-LDLR resulted in a simultaneous enrichment of EDEM1 (ER-degradation enhancing α-mannosidase-like Protein-1), suggesting the activation of the cell’s survival mechanisms to combat ER stress by recruiting ERAD components ((Figure 2A(vii)). In spite of the activation of all three ER sensors and ERAD components, there was a riveting reduction in the transcripts of GRP78/BiP- the ER stress master regulator.
Figure 2B (i-viii) describes the activation of the three ER stress sensors- IRE1A/spliced XBP-1, PERK, and ATF6 assessed by immunoblotting. Comparisons were made as described in the Methods section. The corresponding graphs representing GAPDH normalized protein expression depict the fold change in ER-retained variants and indicate the activation of all three ER stress sensor arms- IRE1A/XBP1-s, PERK/eIF2A, and ATF6 in HepG2WT+, HepG2D482H+, and HepG2C667F+ compared to the HepG2mock. The retention of the LDLR missense variants did not cause a significant fold change in the expression of phosphorylated IRE1A compared to WT-LDLR/empty vector-expressing cells. However, ER retention of these variants resulted in the activation of the RNase domain of IRE1A as indicated by an evident five-fold increase in XBP-1(s) protein. Likewise, the activation of the PERK/eIF2A arm was evident from the enhanced expression of phosphorylated eukaryotic Initiation Factor Alpha (eIF2A), with a two-to-four-fold upregulation of phosphorylated eIF2A in HepG2D482H+ and HepG2C667F+ compared to the HepG2WT+ and HepG2mock. Proteolytic cleavage of ATF6 was detected in LDLR-WT/variant-expressing cells. In addition to the inactive, glycosylated, or non-glycosylated isoforms of the transmembrane protein of 90kDa (pATF6-P), the cleaved N-terminal cytosolic domain of ~50 kDa (pATF6-N), together with intermediate isoforms of ATF6, as reported by Jin et al (20), were detected in the total protein lysates of HepG2WT+, HepG2D482H+, and HepG2C667F+, with predominant expression of pATF6-N in HepG2WT+ and HepG2D482H+ than in HepG2C667F+.
To confirm the ER-Golgi-nuclear trafficking of activated ATF6, HepG2mock, HepG2WT+, HepG2D482H+, and HepG2C667F+ were stained with antibodies to organelle-specific markers (Figure 2C). ATF6 is expressed along the membrane of the ER and colocalizes with Calnexin. Further, HepG2 cells retaining the LDLR variants in the ER exhibited altered subcellular localization of ATF6 in the Golgi complex and nuclei, wherein it co-expresses with Golgin-97 and the nuclear protein Histone H3, respectively. The ER-to-Golgi transition, together with the nuclear translocation of the cytosolic domain of ATF6, which migrates to 50 kDa as seen on the immunoblot (21), demonstrates the ATF6 arm-mediated ER stress response elicited by ER-retained missense variants of LDLR. To a lesser extent, HepG2mock exhibited basal expression of ATF6 within the Golgi and nuclei. As demonstrated in figures 2B and C, the proteolytic cleavage of ATF6 in the absence of ER stress propounds the potential importance of ATF6 in maintaining tissue homeostasis, in addition to the canonical ER stress response (22,23). Figure 2D represents the zoomed images of the merged panel. As for GRP78/BiP, expression of GRP78/BiP remained suspiciously low, in like manner as its transcripts, in HepG2 cells stably expressing WT or variants of LDLR and were further investigated (Figures 2A(i) and 2B (iii)).
3.3 Oxidized LDL aggravates UPR in ER stress caused by ER retained variants
To characterize the molecular mechanisms of cellular stress response and hepatotoxicity induced by oxLDL in ER-stressed cells, HepG2mock, HepG2WT+, HepG2D482H+, and HepG2C667F+ were administered with 100 µg per ml of oxidized LDL for 24 hours (24,25). Tunicamycin was used at a concentration of 2 µg per ml as the positive control to induce ER stress and DMSO represented the vehicle control for TM treated cells. Commercially available oxLDL is dissolved in PBS and hence oxLDL treated samples were compared with the respective non-treated, non-transduced HepG2 or mock-transduced HepG2. All comparisons were carried out between the respective non-treated controls and oxLDL- treated cell types, considering the fact that the lysates for different cell types were immunoblotted on different membranes. The change in expression of target proteins that are key players of IRE1A/XBP1-, PERK/eIF2A/ATF4/CHOP-, and ATF6- pathways were evaluated by Western blotting as represented in Figure 3A. The corresponding histograms representing target protein expression normalized to total protein or loading control are depicted in Figure 3B.
Tunicamycin treatment resulted in enhanced expression of all ER stress sensors and served as the positive control for ER stress induction. In comparison to the respective non-treated conditions, oxLDL treatment of HepG2WT+ increased IRE1A-P by two folds, which remained high in oxLDL treated - HepG2D482H+ and HepG2C667F+. In the presence of oxLDL, spliced XBP-1 protein expression in HepG2WT+ was at par with oxLDL-treated HepG2C667F+ and higher in HepG2D482H+ (Figures 3A (iii -v) and 3B (ii and iv). At the 24-hour time point, we were unable to see a significant change in phosphorylated eIF2A in oxLDL- treated versus non-treated HepG2WT+ and HepG2C667F+, but a reduction was noted in HepG2D482H+ (Figures 3A (viii and ix), 3B (iv)). An apparent depletion of pATF690 protein expression in oxLDL-treated HepG2WT+ was detected, with a more severe loss of the inactive ATF6 protein in oxLDL-treated HepG2D482H+ and HepG2C667F+ (Figures 3A (vii) and 3B (vi). However, we noticed the abundant expression of the intermediate isoforms in HepG2C667F+, with a lesser but consistent expression of the nuclear ATF6 (pATF6-N) even in oxLDL treated cells, compared to the respective non-treated cells. This may be attributed to the delay in processing and trafficking cleaved ATF6, as there was a significant increase in the levels of pATF6 intermediates than the pATF6 (90) isoforms in HepG2C667F+. Activation of all three branches of the UPR, as evident in ER-stressed cells expressing ER-retained LDLR variants, and to a lesser extent in HepG2WT+ cells, may be attributed to the cell’s pro-survival mechanisms of maintaining cellular homeostasis. To explore the involvement of prolonged ER stress, the expression of C/EBP homologous protein (CHOP) (Figures 3A (xi) and 3B (vii)), which is the central mediator of ER stress-induced apoptosis, and the apoptotic effector- cleaved caspase 3 (Figures 3A (xii and xiii) and 3B (viii)) were assessed. While non-treated HepG2WT+, HepG2D482H+, and HepG2C667F+ displayed negligible levels of CHOP with no significant upregulation in treated conditions, exposure to oxLDL substantially increased the expression of cleaved caspase 3 in HepG2WT+and HepG2C667F+. In addition to the endonuclease activity of its activated RNase domain, the kinase domain of stimulated IRE1A propagates inflammatory and apoptotic responses along the IRE1A-TRAF2-JNK pathway. Stress-dependent signaling along the IRE1A-JNK-BIM axis extends its deleterious effects on the mitochondrial membrane by disrupting mitochondrial membrane potential. This in turn leads to the release of pro-apoptotic protein- Cytochrome c into the cytosol, which further instigates the course of apoptotic cell death and brings about mitochondrial dysfunction. To explore the cumulative effect of ER- and oxidized LDL-mediated stress in FH conditions and its likelihood of inducing inflammatory and apoptotic responses that radiate along the ER-mitochondrial interphase, the expression profiles of phosphorylated JNK and BIM were examined. In HepG2WT+ and HepG2D482H+, oxLDL treatment was capable of actuating phosphorylation of JNK-1 and JNK-2 of 45 kDa and 54 kDa, respectively. Nonetheless, the expression of these phosphorylated isoforms in oxLDL-treated HepG2C667F+ was lesser than in non-treated HepG2C667F+ (Figures 3A (xiv and xv) and 3B (ix and x)). Similar to cleaved caspase 3, phosphorylated BIM expression was exalted several folds in oxLDL treated HepG2D482H+, and HepG2C667F+, compared to the respective non-treated conditions (Figures 3A (xvi and xvii) and 3B(xi)). Although we were unable to detect significant induction of CHOP expression in treated and non-treated HepG2WT+, HepG2D482H+, and HepG2C667F+, it was noted that oxLDL treatment augmented the ER stress response in cells carrying ER-retained LDLR variants. The increased levels of cleaved caspase 3 and phosphorylated BIM reveal the initiation of apoptotic signaling and progression of the UPR towards irreversible ER stress response that culminates in cell death. Enriched expression of phosphorylated JNK1 and JNK2 in HepG2WT+ and HepG2D482H+ marks the onset of inflammatory responses initiated by oxLDL in these cells. Next, we assayed the mitochondrial membrane potential (ΔΨm), cytochrome c release, and the extent of cytotoxicity in oxLDL treated and non-treated HepG2mock, HepG2WT+, HepG2D482H+, and HepG2C667F+. These results are elaborated in Figure 5(A-C).
GRP78/ BiP is the master regulator of the ER stress response, wherein it precludes activation of the UPR signal transducers IRE1A, PERK, and ATF6 by remaining bound to these proteins in the normal state. Previously, a significant upregulation of GRP78/BiP along with other ER chaperones was reported in D482H- and C667F-LDLR expressing HEK293T cells than in WT-LDLR transfected cells, which confirmed the induction of ER stress by these missense variants. Despite evident changes in the UPR branches initiated by ER stress, oxLDL, or in combination, like the response in tunicamycin-treated cells, the expression of BiP was surprisingly low after 24 hours of oxLDL treatment in HepG2WT+, HepG2D482H+, and HepG2C667F. In the respective non-treated controls, minimal levels of BiP expression were detected, with almost no expression in HepG2C667F+. Instead, tunicamycin treatment enhanced BiP expression in HepG2WT+, HepG2D482H+, and HepG2C667F+ (Figures 3A (vi) and 3B (iii)).
3.4 BiP is severely down-regulated in oxLDL treated HepG2 cells overexpressing WT- or missense variants of LDLR
The interesting finding of lower expression of BiP in oxLDL treated cells at a dose of 100 µg per ml for 24 hours urged us to track the expression pattern of BiP at various time points ranging from 8, 12, 16, 24, and 48 hours post oxLDL treatment. Figure 4 describes the expression pattern of BiP at various time points in oxLDL (100 µg per ml) treated and non-treated HepG2. Non-transduced control HepG2 (HepG2) cells were used as ‘control’ to confirm the endogenous BiP expression in oxLDL-treated (HepG2+) and non-treated conditions (HepG2-). Tunicamycin-treated cells represent the positive control for BiP induction caused by ER stress. oxidized LDL does not alter the expression of BiP in HepG2 cells at 8 and 12 hours post oxLDL treatment in HepG2WT+, HepG2D482H+, and HepG2C667F+ (Supplemental Figure 2(A-B)). At 16 hours (Figure 4A), oxLDL-treated HepG2WT+, HepG2D482H+, and HepG2C667F+ consistently expressed BiP protein, similar to the respective non-treated cells. At 24 hours (Figure 4B), although HepG2-, HepG2+ and HepG2WT+ expressed significant levels of BiP in non-treated and treated cells, its expression was severely downregulated in oxLDL treated HepG2D482H+ and HepG2C667F+. After 48 hours of oxLDL treatment, the upregulation of BiP expression in the presence of oxLDL is an example of the classic cellular response to stress. Nevertheless, at this time point, BiP expression reappeared in oxLDL-treated HepG2D482H+ and HepG2C667F+ but was intriguingly lower than in the corresponding non-treated conditions (Figure 4C). It cannot be ruled out that overloading the hepatic cells with LDLR exerts ER stress, irrespective of being the wildtype- or the missense variants and explains the absence of BiP in oxLDL treated HepG2WT+. Contrary to the pattern of upregulated targets of all three branches of UPR described in figures 3(A-B), the downregulation of BiP in oxLDL treated, ER stressed conditions emphasize the pathogenic involvement of oxLDL in sabotaging BiP from attending to its function as an ER chaperone and ER stress regulator, particularly in cells expressing excess LDLR. The comparatively lower expression of BiP in non-treated HepG2C667F+ at 24 hours and its re-appearance at 48 hours may be due to the delayed response of these cells to activate the ER stress signaling. While BiP maintains a low profile of expression in normal cells, BiP is a bonafide cancer stem cell marker. HepG2 cells are hepatosarcoma cells and hence the stress response will be different from non-cancerous cells. However, at this point, these explanations lack substantiating evidence from similar experiments conducted in non-cancerous cells.
3.5 Overexpression of BiP in HepG2WT+ and ER- stressed HepG2D482H+/HepG2C667F+ restores cellular homeostasis and reduces hepatotoxicity induced by oxLDL
As evident from the results described above, we observed subtle upregulation of apoptotic and inflammatory markers and severely depleted levels of BiP/GRP78 in HepG2mock, HepG2WT+, HepG2D482H+, and HepG2C667F+, treated with oxLDL at a dose of 100 µg per ml for 24 hours. To understand the effect of prolonged exposure to oxLDL in HepG2 overexpressing WT or ER retained LDLR variants, we increased the duration of treatment to 48 hours. This time point was chosen as we observed the reappearance of GRP78/BiP in oxLDL treated conditions than at 24 hours post-treatment. To further investigate whether it is the absence of optimal levels of BiP that facilitates the pernicious progression of UPR to apoptosis and cytotoxicity, BiP was overexpressed in HepG2mock, HepG2WT+, HepG2D482H+, and HepG2C667F+, thereafter treated with 100 µg per ml of oxLDL for 48 hours. Comparisons were made as described in Section 2.6. As evident from the Western blots for spliced XBP-1, CHOP, and cleaved caspase 3 in Figures 5A-1(i - iv) and 5A-2(i-iii), exposure to oxLDL treatment for 48 hours accentuated the activation of UPR and apoptotic pathways in HepG2mock, HepG2WT+, HepG2D482H+, and HepG2C667F+ which could be abated by the overexpression of BiP. A steep decline in spliced XBP-1 and CHOP was manifested in oxLDL treated, BiP-overexpressing HepG2mock, HepG2WT+, HepG2D482H+, and HepG2C667F+. In addition to the aforementioned apoptotic markers, P-JNK, and P-BIM were compared to corroborate the involvement of inflammatory and apoptotic responses spanning the ER-mitochondrial interphase (Figures 5A (v-viii) and 5A-2 (iv-vi)). Activated caspase 3 decreased by two-fold in BiP- overexpressing, oxLDL-treated HepG2C667F+, and a significant drop was observed in HepG2D482H+. On the other hand, it remained high in HepG2WT+ and showed no significant change in the oxLDL- treated conditions in HepG2mock cells. In all the evaluated conditions, oxLDL induced the phosphorylation of BIM, which was significantly reduced in BiP-overexpressing, oxLDL-treated conditions. HepG2mock showed a dramatic decline in P-JNK 45 and P-JNK 54 in oxLDL- treated cells overexpressing BiP. In HepG2C667F+, 48 hours of oxLDL treatment accelerated the phosphorylation of JNK more than it had at 24 hours. Besides, P-JNK45 protein outweighed the P-JNK-54 isoforms in HepG2C667F+ and portrayed no significant change in expression in the presence of excess BiP. At the same time, surplus amounts of BiP could bring about a two-fold reduction in P-JNK 54 in oxLDL-treated cells. HepG2WT+ and HepG2D482H+ cells did not show a significant difference in expression in the presence of auxiliary amounts of BiP at the 48-hour time point. Thus, BiP helped to keep inflammation and apoptosis under control and prevented the progression toward unfavorable cellular stress and apoptosis. An evident downregulation of ER stress sensors, apoptotic and inflammatory markers as depicted in Figures 5A-1 (i-viii) and 5A-2 (i-vi) reinforces the protective function of BiP in cells severely affected by oxLDL and ER stress.
The damages caused to the plasma membrane result in the release of the cytosolic enzyme lactate dehydrogenase (LDH) into the extracellular space. LDH assay was performed to compare the extent of cytotoxicity induced by oxLDL in BiP-expressing HepG2WT+, HepG2D482H+, and HepG2C667F+. The amount of LDH released was quantified colorimetrically and plotted as bar diagrams in Figure 5B (i-iv). While a two-fold increase in LDLH release was displayed in oxLDL-treated HepG2mock cells that express endogenous BiP, oxLDL-induced cytotoxicity was reduced by slightly over one-fold in cells overexpressing BiP. In non-treated conditions, LDH released by cells that carry an excess of BiP was lower than cells expressing endogenous BiP. A similar trend in LDH release was observed in oxLDL-treated, BiP overexpressing HepG2WT+, HepG2D482H+, and HepG2C667F+cells, whereby the cytotoxicity elicited by oxLDL was greatly reduced by BiP. In comparison to HepG2mock, LDH released by non-treated HepG2WT+, HepG2D482H+ and HepG2C667F+ with endogenous BiP expression was higher, suggesting the toxicity imposed on cells overexpressing LDLR. In this regard, toxicity was higher in HepG2D482H+ and HepG2C667F+ than in HepG2WT. Again, tunicamycin treatment induced a positive ER stress response as evident from the Western blotting and LDH assays.
Disrupted mitochondrial membrane potential (MMP) is an evident sign of mitochondrial dysfunction that simultaneously opens mitochondrial transition pores to release cytochrome c into the cytosol and triggers apoptotic cell death. Flow cytometric analysis of JC-1-stained cells was carried out to measure the mitochondrial membrane potential (ΔΨm) in oxLDL treated or non-treated HepG2WT+, HepG2D482H+ and HepG2C667F+. CCCP dye treatment disrupts the membrane potential of the cells and was used as the positive control for the assay (Figure 5C (i-xxiv)). In healthy cells, JC-1 dye exhibits potential-dependent accumulation as aggregates in the mitochondria that emit red fluorescence at 590 nm. In apoptotic cells, the membrane potential collapses, and the JC1 aggregates dissociate into monomers as indicated by the fluorescence shift from red (~590 nm) to green (~529 nm). The ratio of red to green cells indicates the percentage of polarized (Q2) to depolarized cells (Q4) in JC-1 dye-treated cells assayed using FACS. The histogram represents the Q2/Q4 ratio in oxLDL-treated and non-treated cells that either overexpress BiP or exhibit endogenous BiP (Figure 5C(xxv)). ER stress induced by missense LDLR variants did not collapse the MMP as evident from the Q2/Q4 ratio of ~ 4 in HepG2mock, HepG2WT+ or HepG2C667F+, which was two-timed higher in HepG2D482H+. In accordance with the findings elaborated above, the severity of disrupted MMP in HepG2mock, HepG2WT+, HepG2D482H+, and HepG2C667F+ was magnified in oxLDL treated cells, as evident by almost 50% reduction in healthy HepG2D482H+ and >90% in HepG2WT+ and HepG2C667F+. The protective function of BiP is apparent in non-treated HepG2mock, HepG2WT+, HepG2D482H+, and HepG2C667F+ that overexpress BiP. A transient enhancement of BiP in oxLDL-treated cells rescued these cells from the ill effects of oxLDL and prevented disruption of MMP in HepG2mock, HepG2WT+, HepG2D482H+, and HepG2C667F+. Furthermore, oxLDL-treated HepG2WT+, HepG2D482H+ and HepG2C667F+ overexpressing BiP released lesser amounts of cytochrome c compared to the oxLDL- treated cells that endogenously express BiP (Figure 5A-1(ix) and 5A-2 (vii)). Thus, BiP imparts protective functions in non-treated HepG2mock, HepG2WT+, HepG2D482H+, and HepG2C667F+, whereas in oxLDL-treated cells, it fortifies the cell’s survival mechanisms by minimizing ER stress and hepatotoxicity.