|Home | About | Journals | Submit | Contact Us | Français|
We have been using polarized hepatic WIF-B cells to examine ethanol-induced liver injury. Previously, we determined microtubules were more highly acetylated and more stable in ethanol-treated WIF-B cells. We proposed that the ethanol-induced alterations in microtubule dynamics may explain the ethanol-induced defects in membrane trafficking that have been previously documented. To test this, we compared the trafficking of selected proteins in control cells and cells treated with ethanol or with the histone deacetylase 6 inhibitor trichostatin A (TSA). We determined that exposure to 50 nM TSA for 30 minutes induced microtubule acetylation (~3-fold increase) and stability to the same extent as did ethanol. As shown previously in situ, the endocytic trafficking of the asialoglycoprotein receptor (ASGP-R) was impaired in ethanol-treated WIF-B cells. This impairment required ethanol metabolism and was likely mediated by acetaldehyde. TSA also impaired ASGP-R endocytic trafficking, but to a lesser extent. Similarly, both ethanol and TSA impaired transcytosis of the single-spanning apical resident aminopeptidase N (APN). For both ASGP-R and APN and for both treatments, the block in trafficking was internalization from the basolateral membrane. Interestingly, no changes in transcytosis of the glycophosphatidylinositol-anchored protein, 5′-nucleotidase, were observed, suggesting that increased microtubule acetylation and stability differentially regulate internalization. We further determined that albumin secretion was impaired in both ethanol-treated and TSA-treated cells, indicating that increased microtubule acetylation and stability also disrupted this transport step.
These results indicate that altered microtubule dynamics explain in part alcohol-induced defects in membrane trafficking.
Alcoholic liver disease is a major biomedical health concern in the United States. Although the progression of this disease has been well described clinically, little is known about the molecular basis for liver injury. Research aimed at identifying the molecules and pathways that promote alcohol-induced hepatotoxicity has been hampered by the lack of good in vitro models. Our recent studies have been performed in WIF-B cells, an emerging model system for the study of alcohol-induced liver injury. These hepatic cells are highly differentiated and form discrete apical and basolateral membrane domains in culture. Domain-specific membrane proteins are localized in WIF-B cells as they are in hepatocytes in situ, and many liver-specific functions are maintained.1 Importantly, WIF-B cells metabolize ethanol using alcohol dehydrogenase (ADH) and cytochrome P4502E1.2 Like hepatocytes, ethanol-treated WIF-B cells display a reduced redox state and increased triglycerides.2
More recently, we discovered that microtubule repolymerization was impaired in ethanol-treated WIF-B cells,3 consistent with reports performed in vitro and in isolated hepatocytes.4–6 We also found that steady-state microtubules in alcohol-treated WIF-B cells were more stable and were acetylated 2–3-fold more than were those in control cells.3 This posttranslational modification is characteristic of stable microtubule populations (see Westermann and Weber7 and the Discussion section). We confirmed these results in livers from ethanol-fed rats, indicating the findings have physiologic importance.3 We further determined that increased microtubule acetylation and stability in WIF-B cells was dependent on ethanol metabolism and was likely mediated by acetaldehyde. Thus, alcohol consumption and metabolism alter hepatic microtubule dynamics.
Because microtubules are central to multiple cellular processes, changes in their dynamics will likely alter hepatic function. This study was aimed at understanding the relationship between protein trafficking and alterations in microtubule dynamics. Not only is protein trafficking microtubule dependent, it is also selectively impaired by ethanol.8–10 Two transport pathways appear to be affected: transport of newly synthesized secretory or membrane proteins from the Golgi to the basolateral membrane and receptor-mediated endocytosis from the sinusoidal surface. Can the defects in secretion and endocytosis be explained by increased microtubule acetylation and stability?
To answer this question, we examined the trafficking of selected proteins in WIF-B cells treated with ethanol or trichostatin A (TSA), an inhibitor of histone deacetylase 6 (HDAC6). HDAC6 is the major tubulin deacetylase in liver and WIF-B cells (Grozinger et al.12 and manuscript in preparation), and when inhibited with TSA, increased microtubule acetylation and stability is observed.13–15 We determined that 50 nM TSA for 30 minutes induced microtubule acetylation and stability to the same extent in WIF-B cells as did ethanol, allowing us to test our hypothesis. This result also suggested that acetylation induced microtubule stability, a topic that is currently controversial (for example, see Tran et al.15 and Palazzo et al.16 and the Discussion section). As shown in situ,8–10 ethanol impaired albumin secretion and internalization of asialoglycoprotein receptor (ASGP-R), further confirming that WIF-B cells are a good model for studying alcoholic liver injury. Similarly, TSA treatment impaired albumin secretion and ASGP-R internalization, but to a lesser extent. Both agents impaired internalization of the single-spanning apical resident aminopeptidase N (APN), but distribution of the glycophosphatidylinositol (GPI)–anchored protein 5′-nucleotidase (5′-NT) was not altered, suggesting that microtubule acetylation and stability selectively regulate internalization. We concluded that increased microtubule acetylation and stability in part explain the alcohol-induced defects in membrane trafficking.
F12 (Coon's) medium, 4-methylpyrazole (4-MP), nocodazole, and TSA were purchased from Sigma-Aldrich (St. Louis, MO). Fetal bovine serum came from Gemini Bio-Products (Woodland, CA) and Hepes from HyClone (Logan, UT). Alexa-466-conjugated and Alexa-568-conjugated secondary antibodies were purchased from Invitrogen (Carlsbad, CA). HRP-conjugated secondary antibodies and the monoclonal antibodies against α-tubulin and acetylated α-tubulin came from Sigma-Aldrich. Polyclonal antibodies against rat serum albumin were from ICN Kappel (Costa Mesa, CA). Antibodies against ASGP-R, APN, HA321, and 5′-NT were all kindly provided by Dr. Ann Hubbard (Johns Hopkins School of Medicine, Baltimore, MD).
WIF-B cells were grown in a humidified 7% CO2 incubator at 37°C as described.1 Briefly, cells were grown in F12 medium (pH 7.0) supplemented with 5% FBS, 10 μM hypoxanthine, 40 nM aminopterin, and 1.6 μM thymidine. Cells were seeded on glass coverslips at 1.3 × 104 cells/cm2 and grown for 8–12 days until they reached maximum density and polarity. In general, cells were treated on day 7 with 50 mM ethanol in medium buffered with 10 mM Hepes (pH 7.0) for 72 hours as described.2 Cells were additionally treated with 4-MP or TSA as described in the figure captions.
WIF-B cells were fixed on ice with chilled phosphate-buffered saline (PBS) containing 4% paraformaldehyde for 1 minute and permeabilized with ice-cold methanol for 10 minutes as described.17 When staining for acetylated or total α-tubulin, cells were fixed and permeabilized with methanol at −20°C for 5 minutes. Cells were processed for indirect immunofluorescence as described.17 Alexa-conjugated secondary antibodies were used at 5 μg/mL.
Labeled cells were visualized by epifluorescence using an Olympus BX60 Fluorescence Microscope (OPELCO, Dulles, VA). Images were taken with a Coolsnap HQ2 digital camera (Photometrics, Tucson, AZ) and IPLabs image analysis software (Biovision, Exton, PA) or with a SPOT digital camera (Diagnostic Instruments, Sterling Heights, MI) and SPOT Advanced software. Adobe Photoshop (Adobe Systems Inc., Mountain View, CA) was used to compile images.
To quantitate the relative distributions of APN, 5′-NT, and ASGP-R, random fields from each slide were visualized by epifluorescence and digitized. From micrographs, the average pixel intensity of selected regions of interest (ROIs) placed at the apical or basolateral surface of the same WIF-B cell (APN and 5′-NT) were measured using the Measure ROI tool of ImageJ imaging software (National Institutes of Health). For ASGP-R, the average pixel intensity at the intracellular compartment or plasma membrane of the same cell was measured. The averaged background pixel intensity was subtracted from each value, and the ratio of apical to basolateral membrane (APN and 5′-NT) or intracellular to plasma membrane fluorescence intensity (ASGP-R) was determined.
ASGP-R was surface-labeled in control or treated cells for 15–30 minutes at 37°C with antibodies (diluted 1:25) specific to an extracellular epitope as described.17 Cells were washed 3 times with PBS, fixed as described above, and labeled with Alexa-488-conjugated secondary antibodies.
To analyze whole-cell lysates, cells grown on coverslips were washed 3 times in prewarmed PBS and lysed directly in 200 μL of SDS-PAGE sample buffer and boiled. To assess the relative amounts of soluble and insoluble tubulin, microtubules were depolymerized with 33 μM nocodazole for increasing times as described in the figure captions. Cells were washed 3 times with prewarmed PEM (80 mM PIPES, 2 mM EGTA, 1 mM MgCl2 [pH 6.8]) and extracted in 300 μL of PEM containing 0.15% Triton X-100 at 37°C for l minute. The supernatant (containing soluble tubulin) was collected, and SDS-PAGE sample buffer was added. The cells (containing polymeric tubulin) were washed 3 times in prewarmed PEM without detergent and lysed in SDS-PAGE sample buffer. Samples were immunoblotted with anti-α-tubulin (1:7,500) or anti-acetylated α-tubulin (1:4,000). Immunoreactivity was detected using enhanced chemiluminescence (PerkinElmer, Waltham, MA). The relative levels of tubulin were determined by densitometric analysis of immunoreactive bands.
Cells were pretreated in the absence or presence of 50 mM ethanol for 72 hours or in the absence or presence of 50 nM TSA for 15 minutes. Cells were rinsed 5 times with prewarmed serum-free medium and then reincubated in serum-free medium in the continued absence or presence of either agent. Either 0, 15, 30, or 60 minutes after reincubation, aliquots of the media were collected and analyzed for albumin secretion by immunoblotting. The cell lysates were also collected by solubilization directly into SDS-PAGE sample buffer. Samples were processed for immunoblotting and densitometry as described above. At each time, the percentage of secreted albumin was determined and plotted. A rate of albumin secretion (percentage of albumin secreted/minute) was calculated for each experiment.
The results are expressed as means ± SEMs. Comparisons between control, TSA- or ethanol-treated cells were made using the Student t test for paired data. P values ≤ 0.05 were considered significant.
To test our hypothesis that microtubule acetylation is responsible for alcohol-induced defects in membrane trafficking, we established TSA conditions that promoted equivalent levels of microtubule acetylation in the absence of alcohol. As shown in Fig. 1A, TSA promoted the concentration-dependent acetylation of microtubules, saturating at 500 nM. From this analysis and from time-dependence studies (data not shown), we chose 50 nM TSA for 30 minutes for subsequent analysis. These conditions generally led to a 2–5-fold increase in microtubule acetylation, approximating the approximately 3-fold increase observed with ethanol.3
Previously, we observed that ethanol promoted the formation of microtubules that appeared thicker and more gnarled, features of stable microtubules.3 However, when TSA-treated cells were stained for total α-tubulin, the change in morphology was much less pronounced. The tubules appeared somewhat thicker but not shorter or more gnarled (Fig. 1B[a,b]). This may likely be explained by the short TSA treatment (30 minutes versus 72 hours of ethanol treatment). In contrast, the acetylated α-tubulin staining pattern in TSA-treated cells was similar to that in ethanol-treated cells (Fig. 1B[c,d]), where increased labeling was detected in the microtubule organizing centers just adjacent to the apical membrane (the bile canalicular-like structures are indicated by asterisks).
Because microtubule acetylation is a marker for stable microtubules7 and because increased acetylation in ethanol-treated cells correlated with increased microtubule stability, we examined the effect of TSA on microtubule stability. We assayed intact cells for soluble versus polymeric acetylated tubulin after treatment with the microtubule poison nocodazole. Cells were treated with 33 μM nocodazole for the indicated times (Fig. 2) and lysed in a microtubule stabilizing buffer containing 0.15% Triton X-100. The soluble (S) tubulin (that is deacetylated) was released into the supernatant, whereas the stable polymeric (P) tubulin (that is acetylated) remained cell associated. In control cells, acetylated microtubule polymers were significantly depolymerized after 15 minutes of nocodazole; less than 50% remained polymeric (Fig. 2). In contrast, the polymers in TSA-treated cells were nocodazole resistant. Even after 60 minutes, 72% ± 11% of the acetylated polymeric tubulin remained compared with the 16% ± 8% observed in the controls. The levels of polymeric tubulin in TSA-treated cells were remarkably similar to values from ethanol-treated cells (72% versus ~75% polymeric tubulin, respectively, after 60 minutes; see Kannarkat et al.3), These data indicate that the TSA-induced increase in microtubule acetylation and stability was similar to that induced by ethanol, thus providing us with the appropriate agent to test our hypothesis. Also, these results suggest that increased acetylation induced microtubule stability, a topic currently being debated (see Discussion section).
We first examined the steady-state distributions of ASGP-R, a receptor whose itinerary is impaired by ethanol consumption.8–10 In control cells, the majority of ASGP-R was on intracellular, perinuclear structures (Fig. 3A). Although this intracellular staining was present in ethanol-treated cells, a striking increase in staining at or near the basolateral membrane was observed (Fig. 3B). Such a shift in distribution is consistent with impaired internalization of this recycling receptor from the basolateral surface. These results are also consistent with reports of impaired internalization in situ,8–10 further confirming that WIF-B cells are a good model for studying alcohol-induced hepatic injury.
To confirm our morphological observations, we quantitated ASGP-R distribution from micrographs using our previously published method.18 The ratio of intracellular to basolateral fluorescence for ASGP-R in control cells was 1.53, indicating that most receptors were intracellular (Table 1). In the presence of ethanol, this ratio was decreased to 0.88 (56% of control, P < 0.006), indicating more receptors were present at the basolateral surface, consistent with decreased internalization. The addition of the ADH inhibitor 4-MP prevented redistribution, and the ratio was similar to that of the control (1.28, P < 0.02), indicating that ethanol metabolism was required for the trafficking defect and that the impairment was likely mediated by acetaldehyde, results also consistent with previous studies.11 An increased basolateral population of ASGP-R was also observed in TSA-treated cells (Fig. 5A,B) but to a lesser extent than in ethanol treated cells (Table 1). Nonetheless, the 20% decrease in the ratio of intracellular to basolateral staining was found to be statistically significant (P<0.02). As described above, the less dramatic TSA effect was likely a result of the short treatment time.
To determine if the internalization defect was specific to recycling receptors, we examined the distribution of another class of proteins, apical residents. In hepatocytes, newly synthesized apical residents take an indirect route to the apical surface. They are transported from the TGN to the basolateral surface, where they are selectively retrieved by endocytosis and transcytosed to the apical surface.19 We tested whether ethanol impaired the basolateral internalization of transcytosing proteins by examining the distributions of APN (single-spanning protein) and 5′-NT (GPI-anchored protein). Importantly, these proteins are likely internalized via different mechanisms.19
In control cells, APN was found exclusively at the apical surface (Fig. 4A), whereas 5′-NT was found mainly at the apical surface, with a small population of the transcytosing population at the basolateral surface (Fig. 4C). Interestingly, increased basolateral staining was only observed for APN in the presence of ethanol (Fig. 4B). To ensure that the increased basolateral staining was not a result of the loss of tight junction integrity, which would allow lateral diffusion, we examined the distribution of HA321, a basolateral resident. In both control and treated cells, HA321 distribution was restricted to the basolateral surface (Fig. 4E,F; note the absence of staining at the canaliculi, indicated by asterisks), showing that lateral diffusion to the apical surface was blocked and confirming the integrity of the tight junctions.
Our morphological observations were confirmed when the ratio of apical to basolateral membrane staining was measured in control and treated cells (Table 1). The ratio for APN staining was drastically reduced, from 10.89 to 4.49 (41% of control values) in ethanol-treated cells, indicating increased basolateral levels, and as for ASGP-R, implying decreased basolateral internalization. In contrast, no significant change in the fluorescence intensity ratio was detected for 5′-NT (Table 1), suggesting no impairment of its internalization.
Remarkably, this selectivity was observed in TSA-treated cells. Morphologically, a slightly increased basolateral population of APN was observed in treated cells (Fig. 5C,D), but no change in the distributions of 5′-NT or HA321 was observed (Fig. 5E,F and G,H, respectively). These observations were confirmed quantitatively (Table 1). Despite the smaller difference for APN, the values were statistically significant when normalized to the percent inhibition for each experiment (P<0.05). As described earlier, this was likely a result of the short TSA incubation. However, it is interesting to note that the ratios for both ASGP-R and APN decreased to nearly the same levels in TSA-treated cells (19.5% and 21%, respectively), supporting the significance of these findings.
These results suggest that both agents impaired basolateral internalization. To confirm this directly, we monitored the trafficking of antibody-labeled ASGP-R. Live cells were labeled with ASGP-R antibodies specific to external epitopes and internalized for 15–30 minutes. Because tight junctions restrict antibody access to the apical surface, only antigens at the basolateral membrane were labeled. Cells were fixed and labeled with secondary antibodies to detect the trafficked antigen–antibody complexes. In control cells, a significant intracellular ASGP-R population was detected (Fig. 6A,C, arrows) with very little basolateral labeling, indicating that ASGP-R was rapidly internalized. In contrast, in both ethanol-treated cells (Fig. 6B) and TSA-treated cells (Fig. 6D), ASGP-R was mainly detected at the basolateral surface; very little intracellular labeling was observed. These results indicate that the increased ASGP-R steady-state basolateral staining was a result of impaired internalization.
Because ethanol is known to impair secretion in situ and in isolated hepatocytes, we examined whether ethanol and TSA impaired albumin secretion in WIF-B cells. For these experiments, cells were pretreated in the absence or the presence of 50 mM ethanol for 72 hours or of 50 nM TSA for 15 minutes. Cells were washed and reincubated in serum-free medium in the continued absence or presence of either agent. Either 0, 15, 30, or 60 minutes after reincubation, aliquots of the media were collected and analyzed for albumin secretion by immunoblotting. Values were plotted, and rates of albumin secretion were calculated. As in situ, ethanol impaired albumin secretion in WIF-B cells (Table 2). Although the rate decreased by only about 20%, this was significant (P < 0.03). TSA treatment also led to decreased secretion (Table 2), but in this case the impairment was more dramatic (>26% impaired) and was highly significant (P < 0.0001). We believe that this disparity might be explained by the ability of WIF-B cells to compensate for decreased secretion during the longer ethanol incubation. Nonetheless, microtubule acetylation and increased stability appear to correlate with impaired secretion.
We previously determined that ethanol induced increased acetylation and stability of microtubules in WIF-B cells and in livers from ethanol-fed rats.3 In the present study, we asked whether these microtubule alterations could explain ethanol-induced defects in protein trafficking. We compared the distribution and dynamics of selected proteins in ethanol- and TSA-treated cells. Importantly, TSA led to increased microtubule acetylation and stability to the same extent as that in ethanol. Both treatments led to the impaired internalization of ASGP-R and APN. Interestingly, the internalization of 5′-NT, a GPI-anchored protein, was not impaired, implying the effect was selective (that is, only certain internalization mechanisms were impaired) and specific (that is, the changes were due to altered microtubule dynamics). Furthermore, we determined that albumin secretion was impaired in ethanol- and TSA-treated cells. We concluded that increased microtubule acetylation and stability in part explain ethanol-induced defects in protein trafficking. These studies also further confirmed that WIF-B cells are a good model for studying alcohol-induced hepatotoxicity.
Microtubules exist as both dynamic and stable polymers. The latter population is characterized by a longer half-life, resistance to microtubule poisons (for example, cold and nocodazole), and specific posttranslational modifications.7 These modifications include the removal of a carboxy-terminal tyrosine, polyglutamylation, polyglycylation, and acetylation of lysine 40 on α-tubulin.7 The functions of these modifications and how they contribute to microtubule stability are not yet understood. In particular, the role of acetylation-induced microtubule stability has been the subject of debate. In some studies, acetylation appeared independent of microtubule stability,16,20,21 whereas in other studies including this one, increased acetylation induced increased microtubule stability.13–15 The reasons for these conflicting results are not known but may reflect cell type differences in microtubule dynamics. Also, different assays were used to determine stability that likely vary in sensitivity and approach (for example, measuring nocodazole resistance versus real-time tubulin dynamics) such that conflicting results were obtained.
There is evidence that different microtubule populations (and/or modification of these populations) support specific protein transport steps.22 Of particular interest are studies performed in WIF-B cells that used a novel microtubule depolymerizing drug, 201-F.23 This drug specifically depolymerized dynamic microtubules, leaving only stable acetylated polymers behind. In 201-F–treated cells, both secretion and transcytosis were impaired. Although the specific impaired transcytotic step was not identified, increased basolateral labeling of the apical proteins was observed. These results are remarkably consistent with our findings that in ethanol-treated or TSA-treated cells in which increased populations of stable microtubules were observed (presumably at the expense of dynamic microtubules), both albumin secretion and basolateral internalization were impaired.
Although lysine 40 is thought to reside in the lumen of the microtubule,24 it is possible that its acetylation may lead to altered tubulin conformation such that there are altered interactions with microtubule-associated proteins and motors. This hypothesis is supported by the findings that kinesin, dynein, and dynactin preferentially bound acetylated microtubules in neuronal cells.21,25,26 Enhanced kinesin binding to acetylated microtubules correlated with increased anterograde vesicle motility in these neuronal cells,21,26 a result inconsistent with the decreased secretion observed in our studies. However, secretion of interkleukin 1β was impaired in blood cells treated with specific HDAC6 inhibitors,27 consistent with our results suggesting that microtubule dynamics may differentially regulate vesicle trafficking in neuronal cells. This is further supported by the finding that vesicles recovered from livers of ethanolfed rats had decreased motility in vitro.28 Clearly, further studies are needed to fully understand these results.
Another open question is why the distribution of 5′-NT was not altered in treated cells. One possibility is that internalization mechanisms were differentially impaired. There are at least 3 major internalization routes in mammalian cells—clathrin-mediated, caveolae-mediated/raft-mediated, and non–clathrin-mediated/non–raft-mediated29—which are characterized by specific molecular players, cargoes, and regulators. In general, the receptors that displayed impaired endocytosis in ethanol-treated hepatic cells are internalized via clathrin-mediated pathways (ASGP-R, epidermal growth factor receptor, and insulin via its receptor).8–10,30,31 Although the basolateral internalization mechanism for 5′-NT has not been identified, its GPI anchor suggests it is internalized via a caveolae-mediated/raft-mediated pathway. Furthermore, from work performed in hepatocytes,32 5′-NT is known to have a longer apical arrival time than other residents (3.5 versus 1.5–2 hours), which may reflect differences in their internalization routes. Thus, we propose that the molecular machinery that drives clathrin-mediated endocytosis is more prone to adduction by acetaldehyde or other reactive metabolites such that it is selectively impaired by alcohol treatment. This is a hypothesis we are actively testing.
This study further confirms that WIF-B cells are a good model for studying alcohol-induced hepatotoxicity. Not only does this cell line exhibit a polarized phenotype in culture, it also maintains numerous liver-specific activities normally lost in isolated hepatocytes or missing from other hepatic cell lines. Importantly, these cells metabolize ethanol2 and display alcohol-induced defects in cellular processes as described in situ or in isolated hepatocytes. These impairments include decreased microtubule polymerization,3 defects in ASGP-R trafficking, and albumin secretion. Furthermore, we determined that microtubules in both WIF-B cells and in livers from alcohol-fed rats were hyperacetylated.3 Thus, we believe that WIF-B cells will enable meaningful mechanistic studies, allowing us to examine the molecular basis of alcohol-induced hepatic-injury.
Supported by the National Institute of Alcohol Abuse and Alcoholism (R21 AA015683 to P.L.T.).
Potential conflict of interest: Nothing to report.