To investigate the efficacy of TFEB-mediated enhancement of lysosomal degradation and autophagy on ATZ clearance, we have co-transfected mouse embryonic fibroblasts (MEF) with a plasmid that expresses TFEB under the control of the CMV promoter and with a plasmid expressing the ATZ under the CMV promoter. Transfected cells were subjected to a pulse-chase radiolabelling with 35
S-labelled Cys and Met and the resulting cell lysates and corresponding media were analysed by immunoprecipitation followed by SDS–PAGE analysis. This study showed that intracellular 52 kDa ATZ decreased more rapidly in TFEB-transfected cells compared to control cells transfected with a plasmid expressing the green fluorescent protein (GFP; ). The reduction of intracellular ATZ was associated with decreased mature 55 kDa ATZ in media of TFEB transfected cells compared to control cells (). TFEB-mediated increase of ATZ clearance was not observed in Atg7−/−
MEFs (Komatsu et al, 2005
), thus showing that functioning autophagy is needed for TFEB-mediated increase of ATZ clearance (). Treatment of MEF with proteasome inhibitor MG132 resulted in similar increase of steady state levels of intracellular ATZ in both GFP- and TFEB-transfected cells; thus TFEB does not appear to increase proteasomal degradation of ATZ (). ATZ protein was reduced in HeLa cells stably overexpressing TFEB (HeLa-CF7 cell line; Sardiello et al, 2009
) compared to control HeLa cells ().
TFEB induced autophagy dependent ATZ clearance in vitro
We next generated an HDAd vector that expresses the human TFEB cDNA under the control of a liver-specific phosphoenolpyruvate carboxykinase (PEPCK) promoter and a liver-specific enhancer (Brunetti-Pierri et al, 2005b
; HDAd-TFEB; Supporting Information Fig S1
) to investigate in vivo
the therapeutic potential of TFEB gene transfer for treatment of the liver disease of AAT deficiency. We evaluated the efficacy of TFEB hepatic gene transfer in the PiZ mouse model, a transgenic mouse that expresses the human ATZ gene under the control of its endogenous regulatory regions (Carlson et al, 1988
). PiZ mice recapitulate the features of liver disease observed in humans, i.e.
intrahepatocytic ATZ-containing globules, inflammation/regenerative activity and fibrosis (Hidvegi et al, 2010
). We injected 3-month-old PiZ mice (at least n
= 5 for each group) intravenously with the HDAd-TFEB vector at the dose of 1 × 1013
vp/kg. Control PiZ mice were injected with either saline or with 1 × 1013
vp/kg of a HDAd vector that expresses the unrelated, non-immunogenic, non-toxic alpha-fetoprotein (AFP) reporter gene under the control of the same expression cassette and within the same vector backbone (Brunetti-Pierri et al, 2006
; Supporting Information Fig S1
) of the HDAd-TFEB vector (HDAd-AFP). No changes in appearance, behavior and body weight (Supporting Information Fig S2
) were noted in PiZ mice injected with HDAd-TFEB compared to HDAd-AFP or saline injected mice for up to 6 months post-injection. The livers of PiZ mice injected with HDAd-TFEB showed at 4 weeks post-injection a dramatic reduction of both ATZ accumulation and ATZ-containing globules by periodic acid-Schiff (PAS) staining and immunofluorescence, respectively, compared to saline or HDAd-AFP injected mice (). The reduction of ATZ protein levels was confirmed by ELISA () and by Western blot (Supporting Information Fig S3
A) on hepatic protein extracts. The effect of HDAd-TFEB was specific and HDAd-AFP had no effect on hepatic ATZ levels (). PiZ mice were injected at 3 months of age, when a significant hepatic accumulation of ATZ is already established, as previously shown by PAS staining and ATZ immunostaining (Lindblad et al, 2007
; Teckman et al, 2002
) and confirmed by us (Supporting Information Fig S3
B). Given the significant reduction in hepatotoxic ATZ achieved in HDAd-TFEB injected mice (), these results suggest that liver expression of TFEB was effective in the removal of preformed ATZ inclusions. Although slightly increased, at 6 months post-injection livers of HDAd-TFEB injected mice continued to show a clear reduction in PAS staining and ATZ protein compared to control mice (). Sustained correction of ATZ accumulation is consistent with long-term expression of TFEB, which was still detected at 6 months post-injection by real time PCR on hepatic mRNA ().
TFEB gene transfer in PiZ mice promoted hepatic ATZ clearance
We observed a gradual decline in serum ATZ in PiZ mice injected with HDAd-TFEB and, at 16 weeks post-injection, the reduction of serum ATZ in these animals was approximately 43% of pre-injection levels. From 16 to 24 weeks post-injection, the amount of serum ATZ in HDAd-TFEB injected mice was significantly lower than HDAd-AFP and saline-injected mice (p < 0.05; ).
Serum ATZ is reduced in HDAd-TFEB injected mice
To rule out toxicity from TFEB hepatic gene transfer, we determined transaminases and levels of hepatocyte-specific mRNA in livers from mice injected with either HDAd-TFEB or control HDAd-AFP vector. In both experimental groups, alanine aminotransferase (ALT) and aspartate aminotransferase (AST) were found to be within the normal range or slightly above it at baseline and at various time points post-injection (Supporting Information Fig S4
). No statistical significant differences were found between HDAd-TFEB and HDAd-AFP injected mice for the mRNA levels of selected hepatocyte-specific genes, including mouse transferrin receptor (mTFRC), mouse ornithine carbamoyltransferase (mOTC) and mouse factor IX (mFIX) (Supporting Information Fig S5
). In summary, no clear evidence of liver toxicity was detected in HDAd-TFEB injected PiZ mice up to 6 months post-injection.
Consistent with TFEB-mediated activation of lysosome biogenesis (Medina et al, 2011
; Sardiello et al, 2009
), high levels of LAMP-1 were observed in livers of HDAd-TFEB injected animals (). Interestingly, a negative correlation was noted between ATZ and LAMP-1 immunostaining signals and the few areas positive for ATZ signals showed less increase in LAMP-1 expression (). SQSTM1/p62 is incorporated into completed autophagosome and are degraded in autolysosomes, thus serving as a read-out of autophagic degradation (Klionsky et al, 2012
; Pankiv et al, 2007
). The SQSTM1/p62 was reduced in the livers of mice injected with HDAd-TFEB compared to saline or HDAd-AFP injected mice (). An increase in LC3-I levels was also observed in livers of mice injected with HDAd-TFEB, compared to saline or HDAd-AFP injected mice (). Despite evidence of autophagy increase (enhanced SQSTM1/p62 degradation and transformation of ATZ inclusions into autophagosomes, see and ), the overall number of autophagosomes counted in thin sections decreased upon TFEB treatment (). This observation suggests TFEB gene transfer stimulates fusion of autophagosomes with lysosomes (which are also increased in number, see ), and thus it accelerates the rate of autophagy flux. Increased autophagy flux may also explain the apparently unchanged levels of LC3-II that may reflect increased LC3-II consumption in the autophagosome–lysosome fusion process (). Taken together, these results showed that hepatic gene transfer of TFEB enhances autophagy and lysosome biogenesis in the liver and reduces accumulation of ATZ in livers of PiZ mice.
Increased liver autophagy following hepatic TFEB gene transfer in PiZ mice
Increased ATZ in autophagolysosomes of HDAd-TFEB injected PiZ mice
ATZ content was increased in autophagolysosomes of HDAd-TFEB injected PiZ mice
Monomeric ATZ molecules bind together forming long, polymeric chains that reside in the ER of the cells in a conformation with a very long half-life. To determine the effect of HDAd-mediated gene transfer of TFEB on monomer and polymer ATZ pools, we analysed liver samples from HDAd-TFEB injected mice and corresponding controls using a previously published assay (An et al, 2005
). First, ATZ polymers were isolated from monomers in liver lysates under non-denaturing conditions and then separated polymer and monomer fractions were denatured and compared by quantitative immunoblot. The denaturation step reduces polymers to monomers and the resulting bands can be compared at the same molecular weight (An et al, 2005
). A significant decrease in both ATZ monomer and polymer was observed in livers of HDAd-TFEB injected mice compared to either saline or HDAd-AFP injected control mice ().
TFEB gene transfer reduced hepatic ATZ monomer and polymer in PiZ mice
Electron microscopy (EM) analysis of thin sections from livers of animals injected with saline or control HDAd-AFP vector revealed numerous membrane bound inclusions in hepatocyte cytoplasm (). These inclusions ranged from smaller (0.3–1 µm; see ) to larger sizes which were comparable to nuclei (up to 10 µm; see ), and exhibited membrane continuity with cisternae of rough endoplasmic reticulum (RER) decorated by ribosomes (). These features suggested that such structures are the sites of newly synthesized ATZ which accumulates and aggregates in the RER. Immunolabelling of ATZ in thin cryosections indicated a strong concentration of ATZ in the inclusions of control HDAd-AFP injected PiZ mice (). In contrast, most hepatocytes in HDAd-TFEB injected animals lack the inclusions observed in control PiZ mice (), with exception of few cells that still contained large ATZ aggregates (). Notably, these remaining aggregates were frequently surrounded by double membrane (), that indicates their transformation into an autophagic vacuole.
Hepatic TFEB gene transfer induced clearance of ATZ inclusions of PiZ mice
Immuno-EM revealed ATZ to be diffusely distributed along the RER profiles in the hepatocytes of HDAd-TFEB injected mice (). In control HDAd-AFP injected mice, several multi-vesicular body (MVB)-like structures, that correspond to ‘lysosomes’ or ‘autolysosomes’ based on their ultrastructural features (Saftig & Klumperman, 2009
), exhibited little or no ATZ (). In sharp contrast to control mice, significant amounts of ATZ protein were detected within MVB-like structures in HDAd-TFEB injected mice (). The elevated ATZ signal in MVB-like structures indicates the activation of ATZ degradation by the lysosomal pathway upon TFEB gene transfer. Indeed, gold particles in lysosome-like organelles were frequently associated with intraluminal vesicles that are actively involved in lysosome degradation (Saftig & Klumperman, 2009
), as shown by morphometric quantitative analysis (). Taken together these data, showed that TFEB hepatic expression enhances degradation of insoluble ATZ in autolysosomes.
Previous studies have shown mitochondrial autophagy and injury in the liver of PiZ mice and affected humans. However, it remains unclear whether ATZ retention in the ER is itself responsible for mitochondrial dysfunction or the autophagic response results in non-specific removal of normally functioning mitochondria in ATZ injured livers (Teckman et al, 2004
). The predominant ultrastructural feature of mitochondria in PiZ livers is the degeneration, which did not appear to occur within autophagolysosomes. Affected mitochondria appear as multilamellar structures within their limiting membrane and condensation of matrix and cristae (Teckman et al, 2004
). Therefore, we analyzed livers of HDAd-TFEB and control injected mice for evidence of mitochondrial damage at the ultrastructural level. Mitochondria of PiZ mouse hepatocytes exhibited frequently swollen cristae (Supporting Information Fig S6
A, arrows) that suggest mitochondrial dysfunction. In contrast, the mitochondrial structure of HDAd-TFEB injected mice was normal with thinner cristae (Supporting Information Fig S6
B, arrows). Moreover, HDAd-TFEB injected mice did not show a reduction in the number of mitochondria in livers (Supporting Information Fig S6
C), indicating that their consumption through mitophagy was not accelerated. In addition, increased levels of mitochondrial citrate synthase and CoxIV proteins were detected in HDAd-TFEB injected mice compared to HDAd-AFP or saline injected mice (Supporting Information Fig S6
D and E). Taken together, these data show that enhancement of autophagy mediated by TFEB gene transfer does not result in mitochondrial depletion but conversely, it may have a positive effect on mitochondria. Mitochondria are physically in close vicinity to the cisternae of the ER with which they interact (Arnaudeau et al, 2001
; Perkins et al, 1997
) and based on previous studies (Teckman et al, 2004
) and our results, mitochondrial dysfunction in AAT deficiency appears to be likely secondary to ATZ retention in the ER rather than due to abnormal non-specific increase in mitochondrial autophagy. Indeed this observation is consistent with phenotypic improvements by autophagy enhancers, such as rapamycin or carbamazepine, reported in previous studies on AAT deficiency (Hidvegi et al, 2010
; Kaushal et al, 2010
) and neurodegenerative disorders (Khandelwal et al, 2011
; Liu & Lu, 2010
; Thomas et al, 2011
While reduction of accumulated ATZ polymer is dependent upon increased autophagolysosome degradation, a different mechanism has to be involved in reduction of ATZ monomer detected by monomer analysis (). We found that ATZ mRNA levels were reduced in livers of HDAd-TFEB injected mice compared to either saline or HDAd-AFP injected control mice (). Because the cis-acting elements regulating ATZ expression are retained in the transgenic PiZ mice (Carlson et al, 1988
), we next determined whether the expression of endogenous mouse AAT (mAAT) was also affected by TFEB gene transfer. Indeed, we found that mAAT mRNA was reduced in HDAd-TFEB injected mice (), thus suggesting that ATZ monomer reduction occurred through a transcriptional effect.
TFEB gene transfer interrupted ATZ pathogenic vicious cycle
Previous studies have shown that ATZ accumulation induces ER overload response (EOR) that activates the NFκB pathway (Hidvegi et al, 2005
; Lawless et al, 2004
; Ron, 2002
) through IκB-α phosphorylation (Lawless et al, 2004
). NFκB activation further aggravates ATZ intracellular accumulation because it increases IL-6 transcription (Lawless et al, 2004
), which in turn enhances ATZ transcription through binding to an IL-6 responsive element of the AAT promoter (; Alam et al, 2012
; Morgan et al, 1997
). TFEB gene transfer interrupts this deleterious positive feedback loop and results in reduction of calnexin, a molecular chaperone of AAT and a sensor of EOR (Alam et al, 2012
; Lawless et al, 2004
; ), decreased NFκB synthesis and activation (), and normalizes mIL-6 mRNA levels (). NFκB is sequestered in the cytoplasm by the IκB inhibitory proteins (Brockman et al, 1995
; Whiteside et al, 1997
) and thus reduction of phosphorylated-IκB-α reflects a reduction in the activation of NFκB (). In summary, TFEB-mediated augmentation of ATZ polymer degradation interrupts the vicious cycle that increases the burden of intracellular ATZ resulting in ER stress, NFκB activation and IL-6 dependent activation of ATZ transcription (). In contrast to PiZ mice, TFEB gene transfer in C56BL/6 wild-type mice did not result neither in changes of NFκB protein levels (Supporting Information Fig S7
A) nor in changes of IL-6 mRNA (Supporting Information Fig S7
B), thus suggesting that reduced NFκB activation in PiZ mice is not a direct effect of TFEB gene transfer.
Hepatic fibrosis is a key feature of the hepatic disease in AAT deficiency and is secondary to hepatocyte apoptosis. Therefore, we next investigated whether TFEB gene transfer reduces ATZ-induced liver fibrosis. Collagen deposition was determined by Sirius red staining and by measurement of hepatic hydroxyproline content. HDAd-TFEB injection resulted in a long-term reduction of Sirius red staining () and hydroxyproline content in the livers of HDAd-TFEB injected mice compared to saline and HDAd-AFP injected animals ().
Correction of liver disease following TFEB gene transfer
Caspase-12 is involved in ER stress-induced apoptosis in both ATZ expressing cells and livers (Hidvegi et al, 2005
). Therefore, we investigated whether the reduction of ATZ load in HDAd-TFEB injected mice resulted in decreased activation of caspase-12. Western blot analysis showed that the ~42 kDa cleavage product that corresponds to activated caspase-12 was significantly diminished in HDAd-TFEB injected mouse livers compared to control livers (). We also observed a reduction in caspase-cleaved 89 and 24 kDa fragments of poly(ADP-ribose) polymerase-1 (PARP-1), which are generated during the execution of apoptotic program (). In summary, TFEB hepatic gene transfer decreased detrimental activation of liver apoptosis and fibrosis which underlines the pathogenesis of neonatal hepatitis, cirrhosis and hepatocellular carcinoma in AAT deficiency (Rudnick & Perlmutter, 2005