An Autophagy-Enhancing Drug Promotes Degradation of Mutant α1-Antitrypsin Z and Reduces Hepatic Fibrosis

See allHide authors and affiliations

Science  09 Jul 2010:
Vol. 329, Issue 5988, pp. 229-232
DOI: 10.1126/science.1190354

Correcting a Liver Problem

The classical form of α1-antitrypsin (AT) deficiency is caused by a point mutation that alters the folding and causes intracellular aggregation of AT—an abundant liver-derived plasma glycoprotein. AT deficiency is the most common genetic cause of liver disease in childhood and can also lead to cirrhosis and/or hepatocellular carcinoma in adulthood. Carbamazepine is a drug known to be well tolerated in humans that enhances the intracellular degradation process known as autophagy. Now, Hidvegi et al. (p. 229, published online June 3; see the Perspective by Sifers) show that carbamazepine can reduce the severity of liver disease in a mouse model of AT deficiency by enhancing the degradation of misfolded accumulated AT.


In the classical form of α1-antitrypsin (AT) deficiency, a point mutation in AT alters the folding of a liver-derived secretory glycoprotein and renders it aggregation-prone. In addition to decreased serum concentrations of AT, the disorder is characterized by accumulation of the mutant α1-antitrypsin Z (ATZ) variant inside cells, causing hepatic fibrosis and/or carcinogenesis by a gain–of–toxic function mechanism. The proteasomal and autophagic pathways are known to mediate degradation of ATZ. Here we show that the autophagy-enhancing drug carbamazepine (CBZ) decreased the hepatic load of ATZ and hepatic fibrosis in a mouse model of AT deficiency–associated liver disease. These results provide a basis for testing CBZ, which has an extensive clinical safety profile, in patients with AT deficiency and also provide a proof of principle for therapeutic use of autophagy enhancers.

The classical form of α1-antitrypsin (AT) deficiency is caused by a point mutation (substitution of lysine for glutamate at residue 342) that alters the folding of an abundant liver-derived plasma glycoprotein during biogenesis and also renders it prone to polymerization (1). In addition to the formation of insoluble aggregates in the endoplasmic reticulum (ER) of liver cells, there is an 85 to 90% reduction in circulating concentrations of AT, the predominant physiologic inhibitor of neutrophil elastase. Liver fibrosis and carcinogenesis are caused by a gain–of–toxic function mechanism. Indeed, AT deficiency is the most common genetic cause of liver disease in childhood but can also present for the first time with cirrhosis and/or hepatocellular carcinoma in adulthood (1).

Genetic and/or environmental modifiers determine whether an affected homozygote is susceptible to liver disease (2). Two general explanations for the effects of such modifiers have been postulated: variation in the function of intracellular degradative mechanisms (3, 4) and/or variation in the signal transduction pathways that are activated to protect the cell from protein mislocalization and/or aggregation. As for degradation, the proteasome is responsible for degrading soluble forms of α1-antitrypsin Z (ATZ) (5), and macroautophagy is specialized for disposal of the insoluble polymers and aggregates (6, 7). However, disposal of ATZ may involve other degradative mechanisms, as yet not well defined (8, 9). In terms of cellular response pathways, accumulation of ATZ activates nuclear factor κB (NF-κB) and autophagy but not the unfolded protein response (6, 9, 10).

Because the autophagic response participates in both degradation of ATZ and in the cellular response to accumulation of ATZ in the ER, we examined whether a drug that enhances autophagy could ameliorate hepatotoxicity in this disorder. From a list of drugs that have been recently shown to enhance autophagic degradation of aggregation-prone proteins with polyglutamine repeats (1113), we selected carbamazepine (CBZ) for detailed studies of its effect on ATZ because it has the most extensive safety profile in humans.

First, we found that CBZ mediated a marked decrease in steady-state levels of ATZ in both the insoluble and soluble fractions in the HeLa inducible cell line HTO/Z (Fig. 1). The effect of CBZ was also specific because rapamycin, a drug that activates autophagy by inhibiting target of rapamycin (TOR) kinase, had no effect on ATZ levels (fig. S1). CBZ was dose dependent in the range of 1 to 60 μM (fig. S2) and did not affect wild-type AT levels in the HTO/M cell line or BiP levels in the HTO/Z line (fig. S3).

Fig. 1

Effect of CBZ on steady-state levels of ATZ in the HTO/Z cell line. Immunoblot analysis of HTO/Z cells treated with various concentrations of CBZ, separated into soluble and insoluble fractions, and then probed with antibodies to AT (top) and to GAPDH (bottom).

To further characterize the effect of CBZ on ATZ, we carried out pulse labeling and pulse-chase labeling experiments in the HTO/Z line. CBZ did not affect synthesis of ATZ (Fig. 2A), and disappearance of ATZ from the intracellular compartment was more rapid in cells treated with CBZ than in the untreated cells (Fig. 2, B and C). A statistically significant increase in disappearance of ATZ from the intracellular compartment was mediated by CBZ (P = 0.0007 by two-way analysis of variance with Bonferroni adjustment), with a half-time of 130 min compared to 200 min in untreated cells. The increase in intracellular disappearance of ATZ mediated by CBZ could not be attributed to enhanced secretion (Fig. 2B and fig. S4). Thus, CBZ appears exclusively to change the rate of intracellular degradation.

Fig. 2

Effect of CBZ on synthesis (A) and kinetics of secretion (B and C) of ATZ in the HTO/Z cell line. (A) After pulse labeling, cell lysates were immunoprecipitated with anti-AT. (B) Cell lysates (IC) and extracellular fluid (EC) were immunoprecipitated with anti-AT after pulse-chase labeling. (C) Kinetics of disappearance from IC was determined by densitometric scanning of fluorograms from five separate experiments. Data are shown as the mean ± SE. Dashed lines show the half-time for disappearance. Raw densitometric values for each time point are shown below the figure.

To determine whether CBZ enhances autophagy in the HTO/Z line, we examined its effect on isoform conversion of autophagosomal membrane-specific protein LC3, an indicator of autophagosome formation (Fig. 3A). The LC3-II to LC3-I ratio increased in a dose-dependent manner and was greater in the presence of lysosomal enzyme inhibitors, indicating that CBZ elicits increased autophagic flux. This effect of CBZ on autophagic flux exceeded the increase that results from intracellular accumulation of ATZ (fig. S5). Thus, CBZ stimulates autophagy in cells that have already activated the autophagic pathway in response to ER accumulation of ATZ.

Fig. 3

(A) Effect of CBZ on LC3 conversion in the HTO/Z cell line by immunoblot. Densitometric values are shown at the bottom. (B and C) Effect of CBZ on ATZ in autophagy-deficient (Atg5−/−) (B) versus wild-type (Atg5+/+) (C) cell lines. (D) Effect of CBZ on levels of the AT Saar variant in the HTO/Saar cell line compared to ATZ in the HTO/Z line. (E) Effect of CBZ on ATZ levels in the presence of proteasomal inhibitors. For the last 6 hours of incubation with CBZ (30 μM) or control, proteasomal inhibitors were added to some of the monolayers. The experiments were done as in Fig. 1. For loading control, immunoblots for GAPDH are shown in the lower panels. Similar results were obtained in three separate experiments.

To determine whether the effect of CBZ on ATZ degradation involved enhanced autophagy, we examined its effect on ATZ levels in an autophagy (Atg5)–deficient cell line (Fig. 3, B and C). CBZ mediated a decrease in levels of insoluble ATZ in the wild-type mouse embryonic fibroblast (MEF) cell line but not in the Atg5-deficient cell line. CBZ also mediated a decrease in levels of soluble ATZ in both wild-type and Atg5-deficient cells. Thus, CBZ enhances the disposal of insoluble ATZ by autophagy and has an independent action on the disposal of soluble ATZ by mechanism(s) that do not involve the conventional autophagic pathway.

To determine whether the effects of CBZ were specific for the Z variant of AT, we investigated its effect on disposal of AT Saar, a variant of AT that accumulates in the ER but does not aggregate and is predominantly degraded by a proteasomal mechanism (9). AT Saar was present only in the soluble fraction, but it was degraded by CBZ in a manner almost identical to that of ATZ (Fig. 3D), suggesting an effect of CBZ also on the proteasome.

Thus, we examined the effect of CBZ on steady-state levels of ATZ in the presence of proteasomal inhibitors (Fig. 3E). Although they had no effect on levels of insoluble ATZ, lactacystin and MG132 partially reversed the effect of CBZ on levels of soluble ATZ [lactacystin: reversal of 23.1 ± 14.0% (mean ± SD), n = 3 experiments; MG132: reversal of 12.3, average of n = two experiments ]. Increased levels of ATZ in the presence of lactacystin and MG132 alone provided validation for proteasome inhibitory activity under the conditions of these experiments. Thus, CBZ mildly enhances proteasomal degradation of ATZ and has an independent action on nonproteasomal mechanisms for disposal of soluble ATZ.

Next, we examined the effect of CBZ on hepatic load of ATZ in vivo using PiZ × GFP-LC3 mice. The PiZ mouse was created with the human ATZ gene as transgene. Although it differs from the human disorder in having normal circulating levels of the endogenous murine ortholog of AT, the PiZ mouse is a robust model of liver disease associated with AT deficiency, as characterized by intrahepatocytic ATZ-containing globules, inflammation, and increased regenerative activity, dysplasia, and fibrosis (14, 15). It has been bred onto the GFP-LC3 background to monitor autophagy (6). When administered at 250 mg kg−1 day−1 for 2 weeks by gavage, CBZ mediated a marked decrease in total, insoluble, and soluble ATZ in the liver (Fig. 4A). The treatment was also associated with a marked decrease in intrahepatocytic ATZ-containing globules (Fig. 4, B and C). Quantitative morphometry showed a decrease in globule-containing hepatocytes by a factor of 3.36 (P < 0.001 by Mann-Whitney rank sum test). Serum concentrations of human AT were not significantly affected by CBZ treatment (fig. S6), arguing against any effect on secretion of ATZ in vivo.

Fig. 4

In vivo effect of CBZ on (A) hepatic AT load, (B and C) globules, (D) autophagosomes, and (E) hepatic fibrosis in PiZ × GFP-LC3 mice. Male mice at 5 months of age were treated for 2 weeks with CBZ (250 mg kg−1) or solvent (dimethyl sulfoxide) by gavage. Samples from two control and two CBZ-treated mice are shown. (A) Immunoblot; (B) histochemical staining with periodic acid-Schiff and diastase; (C) immunostaining with anti-AT; (D) immunostaining with anti-GFP; (E) histochemical staining with Sirius red. Globules are purple in (B). Globules are red and nuclei blue in (C). Autophagosomes are green in (D). Scale bars, 100 μm.

Using indirect immunofluorescence, an increase in number of hepatic green fluorescent autophagosomes was detected in areas of liver that lacked AT-stained globules after CBZ treatment (Fig. 4D), and this was confirmed by quantitative morphometry (mean ± SD: 565.7 ± 185.7 μm2 in control versus 1055.3 ± 139.7 μm2 in CBZ; P = 0.049 by t test). The increase in autophagosomes mediated by CBZ superseded the increase that occurs predominantly in globule-containing hepatocytes from ATZ expression alone (6) (figs. S7 and S8). The effect of CBZ in vivo was specific in that rapamycin had no effect on hepatic ATZ levels (fig. S9). Next, we examined the effect of CBZ on hepatic fibrosis because it is a key feature of the liver disease associated with AT deficiency (14). CBZ mediated a marked decrease in fibrosis (Fig. 4E). Furthermore, there was a marked and statistically significant reduction in hepatic hydroxyproline concentration in PiZ mice treated with CBZ (mean ± SD: 1.21 ± 0.7 in CBZ versus 2.27 ± 1.02 μg per milligram of dry weight in control, P = 0.0074 by t test with Welch modification). Hepatic hydroxyproline content was decreased 46.7% by CBZ, reaching a level that was indistinguishable from that of the background FVB/N strain (fig. S10). CBZ also mediated a decrease in hepatic hydroxyproline concentration in the PiZ × IKKβΔhep mouse model (fig. S11). On this hepatocyte-specific NFκB-deficient background, there is more severe liver damage as reflected by hydroxyproline concentrations that are >150% of the levels in the PiZ mouse on the FVB/N background (fig. S10). CBZ treatment decreased levels of stellate cell activation markers, including smooth muscle actin, collagen 1A, and transforming growth factor β, but only the decrease in actin reached statistical significance (table S1).

To determine whether lower doses of CBZ for more prolonged time intervals could reduce hepatic fibrosis, we examined the effect of CBZ at lower doses for 6 weeks. Hepatic hydroxyproline concentrations decreased at the dose of 200 mg kg−1 day−1 but not at doses of 50 and 100 mg kg−1 day−1 (fig. S10). Although the lowest effective dose of CBZ (200 mg kg−1 day−1) was considerably higher than the doses used in humans (10 to 20 mg kg−1 day−1), effective doses of drugs can be 10 to 20 times as high in mice because of the higher ratio of surface area to body weight when compared to humans.

Thus, CBZ reduces the hepatic load of mutant ATZ and hepatic fibrosis in the PiZ mouse. Mechanistic studies indicate that CBZ increases both autophagic and proteasomal degradation of ATZ. That rapamycin does not enhance autophagic disposal of ATZ may mean that a TOR-independent pathway is involved in the effect of CBZ. The effect of CBZ on ATZ disposal cannot be fully accounted for by the proteasomal and conventional macroautophagic pathways. The capacity to enhance disposal of both insoluble and soluble ATZ could represent an important characteristic of CBZ as a potential therapeutic in that it might provide for elimination of the putative hepatotoxic form of ATZ, whether it is soluble monomeric, soluble oligomeric, and/or insoluble polymeric ATZ species.

Because it is theorized that clinically significant liver damage occurs only in AT-deficient patients who also have a “second” defect in quality control and that these second defects are heterogeneous among the affected population, one might conclude that CBZ would be effective only in individuals in whom the “second” defect is related to the specific mechanism of CBZ action. However, our results suggest that CBZ can enhance autophagy beyond the extent to which it has already been activated by the pathological state. CBZ also appears to affect several mechanisms of intracellular disposal and therefore may not require mechanistic specificity for a beneficial effect. It is also encouraging that CBZ reduced hepatic fibrosis in the PiZ × IKKβΔhep mouse model, which could be viewed as a mouse with a type of “second” defect—in this case, reduced functioning of the hepatocyte NF-κB signaling pathway.

In addition to its potential for the treatment of liver disease due to AT deficiency, CBZ should be considered for its ability to enhance intracellular disposal pathways for the treatment of other diseases in which tissue damage involves gain–of–toxic function mechanisms caused by misfolded or aggregation-prone proteins (13). Our results also provide further evidence for the concept that the endogenous protein homeostasis machinery can be used to prevent tissue damage from mutant proteins (16).

Supporting Online Material

Materials and Methods

Figs. S1 to S10

Table S1


References and Notes

  1. Further description of the mouse model is available as supporting material on Science Online.
  2. We are indebted to J. Englert for the hydroxyproline assay and to G. Silverman and J. Brodsky for critical review of the manuscript. The work was partially supported by NIH grants HL037784, DK076918, and RR022241.
View Abstract

Stay Connected to Science

Navigate This Article