autophagy enhanced by microtubule- and mitochondrion ... · autophagy enhanced by microtubule- and...

Molecular and Cellular Pathobiology

Autophagy Enhanced by Microtubule- and Mitochondrion-Associated MAP1S Suppresses Genome Instability andHepatocarcinogenesis

Rui Xie, Fen Wang, Wallace L. McKeehan, and Leyuan Liu

AbstractDysfunctional autophagy is associated with tumorigenesis; however, the relationship between the two

processes remains unclear. In the present study, we showed that MAP1S levels immediately become elevatedin response to diethylnitrosamine-induced or genome instability-driven metabolic stress in a murine model ofhepatocarcinoma. Upregulation of MAP1S enhanced autophagy to remove aggresomes and dysfunctionalorganelles that trigger DNA double-strand breaks and genome instability. The early accumulation of an unstablegenome before signs of tumorigenesis indicated that genome instability caused tumorigenesis. After tumori-genesis, tumor development triggered the activation of autophagy to reduce genome instability in tumor foci.We,therefore, conclude that an increase in MAP1S levels triggers autophagy to suppress genome instability such thatboth the incidence of diethylnitrosamine-induced hepatocarcinogenesis and malignant progression are sup-pressed. Taken together, the data establish a link between MAP1S-enhanced autophagy and suppression ofgenomic instability and tumorigenesis. Cancer Res; 71(24); 1–10. �2011 AACR.

Introduction

Autophagy, or self-digestion, is a process that begins withthe formation of isolation membranes that engulf substratessuch as aggregated proteins or damaged organelles to formautophagosomes. During different stages of cell cycle, cellsnormally retain robust autophagic activity to remove aggre-gated proteins or damaged organelles such as mitochondria tomaintain cellular homeostasis (1). Under nutritive stresses,autophagy is activated to digest cellular organelles or proteinaggregates and recycle the basic components for survival (2).An overactivated autophagy causes organelle depletion andtype II programmed cell death (3). If the autophagic process isblocked before the stage of autophagosomal formation, thefragmented mitochondria will release cytochrome c and othersmall molecules to induce conventional apoptosis (4). If autop-hagosomes are not degraded efficiently, accumulated mito-chondria may become damaged by their own production ofsuperoxide and start to leak electrons and will lose theirmembrane potentials. Thus, diverse forms of aggregation andperinuclear clustering of mitochondria are formed and further

induce robust oxidative stress thatmight also trigger apoptoticcell death (5). Therefore, autophagy not only promotes survivalbut also controls cell death. Although a link between autop-hagic malfunction and cancer was established (6-11), themechanism by which autophagy suppresses tumorigenesisremains largely unknown (2).

MAP1S was originally described and named the chromo-some 19 open reading frame 5 (C19ORF5; refs. 12, 13). It is awidely distributed homologue of neuronal-specific MAP1Aand MAP1B (13, 14). Similar to MAP1A/B, full-length MAP1S(MAP1SFL) gives rise to multiple posttranslationally mod-ified isoforms, including heavy chain (HC) and light chain(LC) isoforms. Prolonged mitotic arrest or inhibition of the26S proteasome causes accumulation of an otherwise highlylabile short chain (SC) of MAP1S (15, 16). In collaborationwith leucine-rich PPR motif-containing protein (LRPPRC)that associates with mitochondria and interacts with Par-kinson disease–related mitophagy initiator Parkin, a tumorsuppressor and microtubular stabilizer RASSF1A, and LC3that associates with isolation membrane (12, 13, 16, 17),MAP1S bridges autophagic components with microtubulesand mitochondria to affect autophagosomal biogenesis anddegradation (16). Ablation of the Map1s gene in mice causesimpairment in both basal autophagy for clearance of abnor-mal mitochondria and nutritive stress–induced autophagyfor nutrient recycling (16). The MAP1SSC associates withmitochondria in addition to microtubules and causes irre-versible aggregation of dysfunctional mitochondria resultingin mitotic cell death (15). Therefore, the MAP1S-depletedmice serve as a good model to test the relationship betweenautophagy and tumorigenesis.

Authors' Affiliation: Center for Cancer and Stem Cell Biology, Institute ofBiosciences andTechnology, TexasA&MHealth ScienceCenter, Houston,Texas

Corresponding Author: Leyuan Liu, Center for Cancer and Stem CellBiology, Institute of Biosciences and Technology, Texas A&M HealthScience Center, Houston, TX 77030. Phone: 713-677-7518; Fax: 713-677-7512; E-mail: [email protected]

doi: 10.1158/0008-5472.CAN-11-2170

�2011 American Association for Cancer Research.

CancerResearch

www.aacrjournals.org OF1

Research. on January 13, 2018. © 2011 American Association for Cancercancerres.aacrjournals.org Downloaded from

Published OnlineFirst October 28, 2011; DOI: 10.1158/0008-5472.CAN-11-2170

http://cancerres.aacrjournals.org/

In this study, we show that in response to the acute increaseof oxidative stress imposed by exposure to a chemical carcin-ogen diethylnitrosamine (DEN), MAP1S levels in mouselivers are dramatically elevated to a peak within 1 day andthen decrease to an undetectable level 2 days after exposure.The acute elevation of MAP1S levels in mouse livers on DENexposure leads to activation of autophagy. Activated autop-hagy leads to removal of p62-associated aggresomes anddysfunctional mitochondria and reduction of DNA double-strand breaks (DSB) and genome instability. Because of theineffective autophagy machinery in the absence of MAP1S,the MAP1S-depleted mice accumulate higher levels of thep62 and g-H2AX–marked genome instability even after theacute effect of DEN exposure diminishes 2 days later. Thehigher levels of genome instability than in wild-type miceremain in the livers of DEN-treated adult MAP1S-depletedmice. After tumor foci are detectable in wild-type mice,MAP1S levels again shift to high levels in tumor foci butremain undetectable in their adjacent nontumor tissues.Elevated levels of MAP1S in tumor foci lead to reactivationof autophagy, further leading to reduction of genomic insta-bility within the tumor foci. Because of higher levels ofgenome instability in liver tissues before tumorigenesis andtumor foci after formation, the MAP1S-depleted mice devel-op more tumor foci and those benign adenomas are faster tobecome more malignant hepatocarcinomas. This is the firstreport from in vivo studies to show a relationship betweenthe autophagy-suppressed oxidative stress and genomeinstability–driven tumorigenesis from the earliest origin tothe latest stage of tumorigenesis.

Materials and Methods

Reagents and antibodiesAntibody against MAP1S (4G1) recognizing an epitope

between humanMAP1S sequence D667 and S767was obtainedfrom A&G Pharmaceutical, Inc. (14, 16). Primary antibodiesagainst b-actin, phosphorylated-H3, FITC-, and rhodamine-conjugated secondary antibodies were purchased from SantaCruz Biotechnology, Inc. Antibody against p62 (SQSTM1;rabbit polyclonal, BML-PW9860) was obtained from Enzo LifeSciences International, Inc. Antibody against g-H2AX (rabbitpolyclonal, A300-081A) was from Bethyl Laboratories, Inc. Thehorseradish peroxidase (HRP)-conjugated secondary antibo-dies were from Bio-Rad. DEN was purchased from Sigma-Aldrich. The ECL Plus Western Blotting Detection System wasprocured from GE Healthcare Biosciences.

Analysis of DEN-induced hepatocarcinogenesis in miceWild-type and MAP1S-depleted mice were created as

described (16) and housed in a pathogen-free animal facilityunder a standard 12-hour light/12-hour dark cycle with adlibitum water and chow. Animal protocols were approved bythe Institutional Animal Care and Use Committee, Institute ofBiosciences and Technology, Texas A&M Health Science Cen-ter. Hepatocarcinogenesis were induced with the hepatocar-cinogen DEN as described previously (18). Briefly, cohorts ofmale 15-day-old wild-type and MAP1S-depleted littermates

were subjected to a single intraperitoneal administration of10 mg/g body weight DEN dissolved in saline. Mice weresacrificed with euthanasia techniques at 4, 6, 8, 10, and 12months after birth. Immediately after euthanasia, whole-bodyweights were recorded and the livers were excised, weighed,and photographed. Grossly visible surface liver tumors werescored from the top and bottom of each liver. Large macro-scopic lesions and their adjacent normal tissueswere dissectedout. Tissues were either fixed or frozen in liquid nitrogenimmediately and stored at �80�C following standardprotocols.

Immunoblot, immunostaining, andfluorescent confocalmicroscopy

Lysates from normal liver tissues or tumor samples wereprepared and immunoblotting was carried out as described(16). Immunostaining was carried out on 5-mm-thick sec-tions mounted on Superfrost Plus slides (Fisher Scientific).Sections were then deparaffinized in xylenes and rehydratedthrough a graded series of ethanol/water solutions. Theantigens were retrieved by boiling in citrate buffer (10mmol/L sodium citrate sodium buffer, pH 6.0) for 20 min-utes at 100�C or according to instructions by manufacturersof the antibodies. The specifically bound antibodies weredetected with fluorescent- or HRP-conjugated secondaryantibody and visualized using the TSA Plus FluorescenceSystem. The sections were counterstained with TO-PRO-3iodide to label nuclei and observed under a Zeiss LSM 510confocal microscope.

Toobserve autophagosomes in vivo,mouse liver tissueswereharvested, frozen, and cryosectioned. As previously reported(16), the GFP-LC3 distribution in frozen sections was directlyobserved and recorded by fluorescent confocal microscopy.The acquired images were converted to 8-bit binary files, andthe number, size, and total occupied area of GFP-LC3 punctatefoci greater than 4 pixels in diameter on each image werecalculated by ImageJ software.

Partial hepatectomy and liver regenerationLiver regeneration induced by partial hepatectomy and

subsequent analysis were done as described previously (19).Briefly, two thirds of the liver was surgically removed from 1.5-month-old male mice (20). Mice were sacrificed 48 hours later.Samples were collected from the regenerated liver tissues andprocessed for analyses.

Isolation of liver nuclei and flow cytometric analysisMale mice of indicated age were sacrificed by cervical

dislocation. Their livers were rapidly removed and frozen inthe liquid nitrogen. The frozen livers were thawed and used toisolate nuclei as described previously (21). For flow cytometricanalysis, the nuclei pelletwas resuspended in propidium iodide(PI) buffer (50 mg/mL PI and 10 mg/mL RNAse in PBS). Theresuspended nucleiwere set for 30minutes at 4�Cand analyzedby flow cytometry. The individual profiles of each type of micewere superimposed on each other and DNA contents at 2Npeaksweremeasured. Thenumber of nuclei under each peak of2N to 8N was counted.

Xie et al.

Cancer Res; 71(24) December 15, 2011 Cancer ResearchOF2




Results

Depletion of MAP1S gene accelerates DEN-inducedhepatocarcinogenesis in miceTo investigate the tumor suppression function of the

MAP1S gene, we set up pairs of wild-type and MAP1S-depleted mice under identical condition without carcinogenand observed for tumorigenesis in all of organs. No abnor-mality associated with the onset of neoplasia was detected

up to the age of 12 months. For example, liver tissues ofMAP1S-depleted mice were as healthy as those of wild-typemice (Fig. 1A and B). We reasoned that although depletion ofMAP1S may not be sufficient to initiate tumorigenesis, thepresence of MAP1S may suppress tumorigenesis inducedunder certain types of stresses.

We started to induce hepatocarcinogenesis with DEN, aspecific initiator of hepatocyte carcinogenesis (18). Littermatesof wild-type (þ/þ) and MAP1S-depleted mice (�/�) were

Figure 1. Depletion of MAP1S promotes DEN-induced hepatocarcinogenesis in mice. A, a summary of tumorigenesis in the liver tissues of wild-type (WT) andMAP1S-depleted mice at different times after DEN treatment. Data are the mean � SD of each group, and difference was statistically determined by theStudent t test; a, P� 0.05. B, the morphology of livers of DEN-treated WT andMAP1S-depleted mice at different ages. (12), 12-month-old mice without DENtreatment. Bar, 10 mm. C, a comparative hematoxylin and eosin (H&E) staining of livers of DEN-treated 6-month-old WT and MAP1S-depleted mice.A normal liver of theWTmouse is comparedwith a liver with adenomas ofMAP1S-depletedmouse. Bar, 100 mm.D andE, a plot of tumor number (D) or ratio ofliver weight to body weight (E) to ages of DEN-treated WT and MAP1S-depleted mice as listed in A. F, a comparative H&E staining of livers of DEN-treated10-month-old WT and MAP1S-depleted mice. Bar, 100 mm. G, a plot of hepatocarcinoma (HCC) incidence to ages of WT and MAP1S-depleted miceafter DEN treatments as listed in A. �, P � 0.05.

MAP1S and Hepatocarcinogenesis

www.aacrjournals.org Cancer Res; 71(24) December 15, 2011 OF3




subjected to a single dose of DEN injection at 15 days post-natally when their livers underwent active proliferation. Theirlivers reached the full sizes of mature livers approximately 2weeks later. We checked for hepatocarcinogenesis and foundtumor foci in MAP1S-depleted liver tissues at 6 months and inwild-type liver tissues at 8 months after birth (Fig. 1A and B).Histologically, the tumor foci were primarily hepatocellularadenomas that caused compression but no disruption of thesurrounding normal tissue (Fig. 1C). Whenmice became older,their body weights were not significantly changed, but tumorincidence, average number of surface tumors per mouse, andratio of liver weight to body weight (LW/BW) were graduallyincreased. The MAP1S-depleted mice had significantly highertumor incidence, average number of surface tumors permouse, and ratio of liver weight to body weight (LW/BW)than wild-type mice at the same ages (Fig. 1A, B, D, and E).Therefore, MAP1S is a tumor suppression gene.

Further detailed examination revealed that hepatocellularcarcinoma (HCC) was detected 10 months after birth (Fig. 1Aand B). The typical trabecular hepatocarcinomas coexistedwith the lipid droplet-containing adenomas in aged mice (Fig.1F), but the MAP1S-depleted mice displayed higher incidenceof HCC than the wild-type (Fig. 1A, B, and G). Depletion ofMAP1S gene not only promoted the initiation of tumor foci butalso accelerated the conversion of hepatocellular adenomas tomalignant hepatocarcinomas.

Acute elevation of MAP1S levels immediately uponexposure to DEN leads to activation of autophagy andreduction of DNA DSB

Although the mechanism is not clearly deciphered, autop-hagy has been hypothesized to reduce chromosomal insta-bility and suppress tumorigenesis through elimination ofp62-associated aggregated proteins (7, 8). We attempted todeduce a role of MAP1S-regulated autophagy in tumorsuppression. The expression levels of MAP1S in normalhepatocytes were too low to be detected by immunoblotting,but were dramatically elevated immediately on exposure toDEN, and then reduced to an undetectable level 2 days afterexposure to DEN (Fig. 2A). Our previous results have sug-gested that MAP1S levels increase on autophagic activation(16). The wild-type accumulated fewer autophagosomeslabeled with GFP-LC3 punctate foci in hepatocytes thanMAP1S-depleted mice (Fig. 2B and C), indicating that theacute elevation of MAP1S caused activation of both autop-hagy initiation and autophagosomal degradation. Examina-tion of LC3-II levels representing a balance of autophago-somal generation and degradation at each sampling point indetail revealed that acute elevation of MAP1S on DENexposure enhanced both events but still did not efficientlyprocess all autophagosomes such that a certain amount ofautophagosomes accumulated and reached a peak at 12hours. The MAP1S-deficient mice may generate autophago-somes slowly such that LC3-II levels reached a peak at least12 hours later than the wild-type mice; in addition, they lessefficiently remove autophagosomes through lysosomes suchthat LC3-II levels dropped to base levels 5 days later than inthe wild-type mice (Fig. 2A and D).

MAP1S and GFP-LC3 signals distributed in cells in amutually exclusively manner, indicating that autophago-somes had been degraded efficiently in cells with elevatedlevels of MAP1S. Approximately 40% of cells had enhancedlevels of MAP1S and activated autophagy to degrade autop-hagosomes whereas the other 60% cells did not haveincreased levels of MAP1S to degrade autophagosomes.More than 80% of hepatocytes accumulated autophago-somes in the MAP1S-depleted mice (Fig. 2E and F), furtherconfirming the role of MAP1S in autophagy. Because ofactive autophagy, the wild-type hepatocytes were able toefficiently eliminate the p62-associated aggresomes anddysfunctional organelles. Two weeks after exposure, levelsof p62 reduced to levels close to but still higher than thosebefore exposure. In contrast, the MAP1S-deficient hepato-cytes maintained higher levels of p62 than the wild-typemice (Fig. 2A and D). Therefore, elevated levels of MAP1Slead to activation of autophagy to reduce the DEN-inducedaggresomes and dysfunctional organelles.

Protein aggresomes and dysfunctional organelles such asmitochondria and endoplasm reticulum induce oxidativestress that causes DNA DSB and genome instability(7, 8, 22). Intensity of DNA DSB is represented by the levelsof phosphorylatedH2AX (g-H2AX; ref. 23). NoDNADSB labeledwith g-H2X was found in untreated hepatocytes of both wild-type and MAP1S-depleted mice (Fig. 2G). Together with theacute elevation of MAP1S levels, DNA DSB was also dramati-cally intensified immediately on exposure to DEN (Fig. 2A, D,and H). Similar to the mutually exclusive distribution of GFP-LC3 punctate foci andMAP1S signals,MAP1S levels were lowerin the cells with DNADSB (Fig. 2G). The MAP1S-depleted miceexhibited higher levels of DNADSB in hepatocytes immediatelyon DEN treatment and maintained higher levels than thewild-type mice until livers became fully mature 2 weeks afterDEN treatment (Fig. 2A, D, and H). Therefore, depletion of theMAP1S gene impaired normal autophagy function, leading toaccumulation of higher levels of genome instability thatincrease the chance to form tumor foci in the future.

The DEN-induced DNA DSB and genome instability thatpersist in the adult liver tissues are suppressed byMAP1S

It is well known that chromosomal fragments with DSB tendto randomly fuse either correctly with the original brokenchromosomal fragment or erroneously with a fragment brokenfrom another chromosome. The erroneously ligated chromo-some undergoes breakage–fusion–bridge cycles in successivemitoses that eventually lead to chromosome instability (24).Majority of hepatocytes with DNA DSB on DEN treatment hadtheir broken ends fused such that no g-H2X signal could readilybe detected 1 month after exposure to DEN. We hypothesizedthat those hepatocytes could still generate daughter cells withDNA DSB because of the breakage–fusion–bridge mechanism.However, nearly no mitotic hepatocyte could be identified inthe adult livers (Fig. 3A); therefore, that it was technicallydifficult to test the hypothesis. To increase the mitotic index,we conducted partial hepatectomy to induce liver regenerationin 2-month-old mice, which led to a dramatic increase in thenumber of mitotic hepatocytes (Fig. 3A).

Xie et al.





Figure 2. Acute elevation of MAP1S levels immediately on DEN exposure leads to activation of autophagy and reduction of DNA DSB. A, an immunoblotanalysis of liver lysates fromWT (þ/þ) andMAP1S-depletedmice (�/�) injected with DEN (25 mg/g body weight). Liver tissues were collected immediately oron different days after DEN injection. FL, full-length MAP1S; HC, heavy chain of MAP1S. B, representative images showing GFP-LC3 punctatefoci that appeared in liver tissuesat the indicated times afterDEN treatment. TheGFP-LC3þ/�MAP1S�/�mice (�/�) and theirGFP-LC3þ/�MAP1Sþ/þ controls(þ/þ) were similarly treated with DEN as in A. Bar, 100 mm. C, quantitation of the total area of GFP-LC3 punctate foci with size larger than 100 pixel2. Data arethe average and SD of 10 randomly selected images in a field of 512 � 512 pixel2. The significance of differences was determined by the Studentt test; �, P < 0.05. D, a plot of relative intensities of p62 and g-H2AX bands shown in A with the value in WT mice at day 0 as the standard 1. Data are arepresentative set of 3 repeats, all ofwhichexhibit the same largedifferencesbetween theWTandknockout (KO) but varied relative intensities due to the variedintensity of the standards. E, relative distribution of GFP-LC3 and MAP1S in liver tissues 12 hours after DEN treatment. Tissue sections prepared as inB were fixed and subjected to immunostaining. Cell boundaries were outlined with white dot lines. Bar, 20 mm. F, percentage of cells with MAP1S-positivesignals or GFP-LC3 punctate foci 12 hours after DEN treatment. Samples were prepared and observed as described in E. Data are the mean of10 randomly selected images containing approximately 2,000 cells in total. G, relative distribution of MAP1S and g-H2AX in liver tissues 12 hours after DENtreatment. Pairs of MAP1Sþ/þ (þ/þ) and MAP1S�/� mice (�/�) were similarly treated with DEN as in A, and liver tissues were fixed and subjected toimmunostaining. Bar, 10 mm. H, representative images showing g-H2AX signals in liver tissues ofMAP1Sþ/þ (þ/þ) andMAP1S�/�mice (�/�) at the indicatedtimes after DEN treatment. Nuclei were counterstained. Bar, 100 mm.






Both normal and DEN-treated mitotic hepatocytes wereinduced into mitosis during liver regeneration. The untreatedmitotic hepatocytes exhibited no DNA DSB, whereas mitotichepatocytes that had been exposed to DEN 1.5 monthspreviously contained a large number of DNA DSB (Fig. 3B).The DNA DSB signals were seen to clearly overlap with thecondensed chromosomes in the mitotic hepatocytes (Fig. 3C).Relatively, fewer mitotic hepatocytes carried DNA DSB in thewild-type than in the MAP1S-depleted mice (Fig. 3B and D).Although partial hepatectomy created a highly artificial andunexpected biologic condition, the results indicated that hepa-tocytes in DEN-treated MAP1S-depleted mice had a higherchance than the wild-type to generate daughter hepatocyteswith genome instability if they had a chance to re-entermitosis.

We further examined the livers of 4-month-old DEN-treatedmice and found no tumor foci in either wild-type or MAP1S-depleted mice (Fig. 1A). Constant evolution of the destabilizedgenomes during the 3.5 months after exposure to DEN mightlead to preservation of DNA DSB although no tumor focus wasdetectable. The MAP1S-depleted mice contained slightly butsignificantly higher levels of DNA DSB in livers that did notdevelop into any detectable tumor focus yet (Fig. 4A and B).Examination of DNA contents by flow cytometry revealed thathepatocytes from both wild-type and the MAP1S-depleted

mice contained amounts of genomic DNA that had deviatedfrom their normal 2N DNA content (Fig. 4C). The depletion ofMAP1S gene caused a wider range of variation of 2N DNAcontents (Fig. 4C and D) and had a higher percentage ofpolyploidy hepatocytes (Fig. 4C and E). Polyploidy cells oftenlead to generation of aneuploidy and genome instability duringproliferation because of weakened mitotic checkpoint (25).Thus, the knockout mice had higher levels of tumorigenesis-driving genome instability.

Elevation of MAP1S levels within tumor foci of DEN-treated mice leads to activation of autophagy influx andreduction of DNA DSB

The persistence of higher genomic instability in the MAP1S-depleted hepatocytes represented a higher chance of tumor-igenesis, which explains why the MAP1S-depleted mice hadhigher tumor incidence and more surface tumors than thewild-type mice. However, the reason for the MAP1S-depletedmice developing more malignant HCC was not explained. Wemeasured the MAP1S levels adenomas of similar sizes in DEN-treated 10-month-old mouse livers. The levels of MAP1S wereundetectable in non–tumor-adjacent liver tissues from wild-type mice, and were similar to the levels in liver tissues ofuntreated mice (Fig. 2A), but dramatically elevated in the

Figure 3. MAP1S suppresses the DEN-induced DNA DSB and instability that persist in the fully developed liver tissues 1 month after DEN treatment. A, animmunostaining analysis showing that partial hepatectomy (PH) reactivates mitosis for liver regeneration. Tissue sections were stained with TOPRO-3to label nuclei (red) and anti-phosphorylated histone H3 (p-H3) to label mitotic cells (green). Bar, 50 mm. B, representative images showing mitoticcells with DNA DSB in regenerated liver tissues. Liver hepatectomy was carried out on bothMAP1Sþ/þ (þ/þ) andMAP1S�/� mice (�/�) 1 month after DENtreatment to induce hepatocyte proliferation. Status of DNA DSB and cell proliferation were identified by immunostaining with antibody against g-H2AXor p-H3 in addition to DNA-stained chromosomes. The untreated mice serve as control. Bar, 100 mm. C, an enlarged image showing mitotic cells with DNADSB in regenerated liver tissues of MAP1S-depleted mice. Bar, 50 mm. D, the percentage of mitotic cells with DNA DSBs to the total mitotic cellsin the regenerated liver tissues after hepatectomy.

Xie et al.





majority of tumor foci (Fig. 5A–C). It was likely that suchincrease was caused by metabolic stresses that resulted fromgenome instability.Cells in the aged livers contained different genomes

among different individuals (Fig. 4C and D). Each diversifiedgenome evolves and adopts a different set of autophagymachinery. Under such conditions, the LC3-II levels mayrepresent balances of different generation and degradationrates. As expected, the LC3-II levels were not significantlydifferent between tumor foci and neighboring tissues, andbetween wild-type and MAP1S-deficient mice (Fig. 5C andD). However, high levels of MAP1S indicate activation ofautophagy influx leading to reduction of autophagy sub-strates (16). More p62-associated protein aggregates anddysfunctional organelles accumulated in MAP1S-deficienttumor foci (Fig. 5C, E, and F), indicating dramatic reduction

of autophagy influx. The increase of MAP1S levels leads toincrease of autophagy influx and reduction of p62 levels inthe DEN-induced tumor foci of wild-type livers. The defi-ciency in autophagic response in the tumor foci of theMAP1S-depleted livers reflected by such accumulation cor-respondingly led to increase of DNA DSB (Fig. 5C, G, and H).Higher degree of genomic instability in the tumor foci led todevelopment of more malignant HCC. Thus, an activeautophagy suppressed genome instability in tumor foci suchthat a less malignant tumor was developed.

Discussion

Chromosomal instability, as the most common type ofgenome instability, mainly results from either whole-chromo-some aneuploidy due to mitotic errors or segmental

Figure 4. MAP1S suppresses theDEN-inducedDNADSBand instability that persist in liver tissues from4-month-old DEN-treatedmice. A, immunoblot of liverlysates from DEN-treated 4-month-old WT (þ/þ) and MAP1S-depleted mice (�/�). Samples were collected from 6 pairs of littermates. B, a plot ofrelative intensities of g-H2AXbands shown inAwith the value inWTmouse number 1 as the standard. Data are the average andSDof 6mice. The significancesof differences betweenWT and KOmicewere determined by the Student t test; �,P < 0.01. C, flow cytometric analyses of hepatic nuclei isolated from livers ofDEN-treated 4-month-old mice that did not exhibit any tumor foci. For each genotype, 8 mice were subjected to analysis for their DNAcontents by propidium iodide (PI) staining, and their histograms in different colors were superimposed with each other. D, the distribution of relative DNAcontents at 2N peaks. The DNA contents of 2N nuclei were quantified from data of 8 pairs of mice shown in C and normalized with the average contentsof allWTmice. E, thepercentagesof nucleiwithDNAcontents aroundpeaksof 2N, 4N, 6N, 8N, and>8N to the total number of nuclei in each sample shown inCwere quantified. Data are the average and SD of 8 mice. The significances of differences between WT and KO mice were determined by the Studentt test; �, P < 0.05.






aneuploidy due to chromosomal rearrangements includingdeletions, amplifications, or translocation (26). Although aneu-ploidy is not synonymous with chromosomal instability (26),aneuploidy was hypothesized to be the origin of tumorigenesis(27). Since then, it has been strongly debated whether pointmutation or genomic instability is the cause or a consequenceof tumorigenesis (28). Our results indicate that, early beforetumorigenesis, significant levels of genomic instability exist inthe liver tissues exposed to the chemical carcinogen DEN. Thisindicates that genomic instability may be the cause of hepa-tocarcinogenesis. The positive correlation between genomicinstability and malignancy of tumor foci as revealed in thisstudy further indicates that genomic instability is also a drivingforce in cancer development. Thus, high levels of genomicinstability provide high chances of tumor initiation andwide genetic variation to evolve into highly malignant tumorfoci.

An inverse relationship between autophagic activity andmalignant transformation has been established since 1999 (29).A high incidence of spontaneous tumors in systemic mosaicdeletion of ATG5, ATG7�/�, Beclin-1þ/�, and Beclin-1–inter-

active protein Bif-1�/� mutant mice enforced a role of autop-hagy in tumor suppression (6, 10, 11, 30). The ATG4C-depletedmice show an increased susceptibility to develop fibrosarco-mas induced by chemical carcinogens (9). Similarly, theMAP1S-depeleted mice have shown a suppressive role inDEN-induced hepatocarcinomas. The contradiction indicatesthat ATG5, ATG7, Beclin-1, and Bif-1may play roles in additionto autophagic regulation.

To understand the relationship between autophagy andtumor suppression, a link through p62 and g-H2AX has beenwell established (7, 8). p62, or sequestosome 1, associates withubiquitin-positive Mallory–Denk bodies in hepatocytes ofpatients with alcoholic hepatitis and a variety of other liverdiseases (31). The p62 protein serves as receptor to bind withubiquitinated substrates such as aggregated proteins anddysfunctional organelles including mitochondria for autopha-gosomal biogenesis, and its levels represent the amount ofaggregated proteins anddysfunctional organelles accumulatedin cells (32). Balanced autophagy aids cells in clearance ofaggregated proteins and dysfunctional organelles and, thereby,reduces p62 levels that serve as accurate monitors of the

Figure 5. Elevation of MAP1S levels leads to activation of autophagy influx and reduction of DNA DSBwithin tumor foci in livers of DEN-treated 10-month-oldmice. A, representative images showing the immunostaining patterns of MAP1S in adjacent tumor tissues (T) and normal tissues (NT) from livers of DEN-treated 10-month-old WT (þ/þ) and MAP1S-depleted mice (�/�). All bars in this figure are equal to 100 mm. B, a summary of percentage of MAP1S-positivetumor foci in livers ofWT (þ/þ) andMAP1SKOmice (�/�). Approximately 10 hepatoadenomaswere examined for each group. C, immunoblot of lysates fromadjacent tumor tissues (T) and normal tissues (NT) isolated from livers of DEN-treated 10-month-old WT (þ/þ) and MAP1S-depleted mice (�/�). The sameamount of total proteins was loaded for each lane. D, a plot of relative intensities of LC3-II bands shown in Cwith the value in the first WT tumor sample as thestandard. Data are the mean and SD of 3 repeats. E, representative images showing immunostaining profiles of liver tissues carrying hepatoadenoma withantibody against p62. F, a plot of relative intensities of p62bands shown inCwith the value in the firstWT tumor sample as the standard. Data are themean andSD of 3 repeats. The significances of differences between tumor and adjacent normal tissues and betweenWT andKOmicewere determined by the Student ttest; �,P < 0.05. G, representative images showing immunostaining profiles of liver tissues carrying hepatoadenomawith antibody against g-H2AX. H, a plot ofrelative intensities of g-H2AX bands shown in C. Data were collected and analyzed similarly as in F.

Xie et al.





intensity of cellular oxidative stress. A defect in autophagyenhances oxidative stress resulting in cell death (7). However,oxidative stress leads to DNA DSB (22, 33, 34) because it cansimultaneously subvert mitotic checkpoints (35, 36).From a cytogenetic perspective, a chromosomal fragment

with a break tends to randomly fuse with another brokenchromosomal fragment, which potentially leads to the forma-tion of a new chromosome with 2 centromeres if ligationoccurs between 2 fragments with centromeres. The cell con-taining the chromosome with 2 centromeres forms a chromo-somal bridge during mitotic metaphase and the resultantchromosomal bridge will break and generate new broken endsafter telophase (24, 37). Modern molecular biology indicatesthat DNA DSB will be repaired through either homologousrecombination (HR) or nonhomologous end-joining (NHEJ),whereas NHEJ is believed to be the predominantmechanism ofDSB repair in higher eukaryotes (38). Cells with DSBs willexperience a mitotic defect that will activate mitotic check-point and temporally result in mitotic arrest and mitotic celldeath (15). If cells can escape the arresting through the so-calledmitotic slippage, the resultant aneuploidy daughter cellsare generally lethal. Therefore, genome instability is intrinsi-cally tumor suppressive (39). However, if an aneuploidy cell canescape the mitotic checkpoint and survive, it potentially initi-ates a cascade of autocatalytic karyotypic evolution throughcontinuous cycles of chromosomal breakage–fusion–bridgeand eventually destabilizes the genome. A destabilized genomeresults in either inactivation of tumor suppressor genes oractivation of oncogenes through point mutation, rearrange-ment of wild-type genes in new locations on genome to createmalfunctions, or creation of fusion gene between differentwild-type genes to fulfill a new function. Such genetic variationgenerates the opportunity for phenotypic selection, whichresults in continuous evolution and diversity of cancer phe-notypes during the progression (24, 28, 40). Thus, autophagicdefect-resultant oxidative stress could either suppress tumor-igenesis if it can trigger cell death or promote tumorigenesis ifthe cells can survive the insult of oxidative stress.It has been known that animals exposed to DEN exhibit

increased levels of oxidative stress (41). We believe that suchincrease in oxidative stress may be caused by accumulation ofp62-associated protein aggresomes and dysfunctional orga-nelles. We knew that MAP1S positively regulates autophagicactivity and its increase on DEN exposure indicates an autop-

hagic activation. Activated autophagy leads to removal of p62and reduction of oxidative stress while animals with defectiveautophagy accumulate high levels of p62 and oxidative stress.MAP1S-activated autophagy causes reduction of both oxida-tive stress and genomic instability immediately on DEN expo-sure. The MAP1S-depleted mice accumulate higher levels ofgenomic instability and have higher probability to generatetumor progenitor cells thatwill be developed into visible tumorfoci several months later. For the tumor progenitor cells toevolve into large tumor foci, cells continue multiple rounds ofmitoses that will further destabilize the genome and results inaccumulation of oxidative stress (42). Such stress will reacti-vate MAP1S-regulated autophagic machinery and, in turn,reduce the levels of p62 and oxidative stress. The inactiveautophagy in MAP1S-depleted mice causes accumulation ofp62, oxidative stress, and genomic instability in tumor foci.Mice with a destabilized genomewill develop highly malignanthepatocarcinomas. The high incidence of tumor foci andhepatocarcinomas in the MAP1S-depleted mice clearly attestthat MAP1S suppresses hepatocarcinogenesis through regu-lation of autophagy.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

The authors thank Dr. Le Sun and Joe Corvera (A&G Pharmaceuticals,Inc.) for providing the anti-MAP1S mouse monoclonal antibody 4G1, which isnow sold by Precision Antibody with the catalog number AG10006. LC3 cDNAand GFP-LC3 transgenic mice were gifts from Dr. N. Mizushima, Departmentof Physiology and Cell Biology, Tokyo Medical and Dental University Grad-uate School and Faculty of Medicine. S. Nguyen and K. McKeehan providedexcellent technical assistance.

Grant Support

This work was supported by the DOD New Investigator Award W81XWH (L.Liu), NCI grant 1R01CA142862 (L. Liu), NCI P50 CA140388 (W.L. McKeehan), theSusan Komen Foundation (W.L. McKeehan), and aid from the John S. DunnResearch Foundation (W.L. McKeehan).

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicate thisfact.

Received June 25, 2011; revised October 13, 2011; accepted October 23, 2011;published OnlineFirst October 28, 2011.

References1. Liu L, Xie R, Nguyen S, Ye M, McKeehan WL. Robust autophagy/

mitophagy persists during mitosis. Cell Cycle 2009;8:1616–20.2. Mizushima N, Levine B, Cuervo AM, Klionsky DJ. Autophagy fights

disease through cellular self-digestion. Nature 2008;451:1069–75.3. Kroemer G, Jaattela M. Lysosomes and autophagy in cell death

control. Nat Rev Cancer 2005;5:886–97.4. Desagher S, Martinou JC. Mitochondria as the central control point of

apoptosis. Trends Cell Biol 2000;10:369–77.5. Boya P, Gonzalez-Polo RA, Casares N, Perfettini JL, Dessen P,

Larochette N, et al. Inhibition of macroautophagy triggers apoptosis.Mol Cell Biol 2005;25:1025–40.

6. Qu X, Yu J, Bhagat G, Furuya N, Hibshoosh H, Troxel A, et al.Promotion of tumorigenesis by heterozygous disruption of the beclin1 autophagy gene. J Clin Invest 2003;112:1809–20.

7. Mathew R, Karp CM, Beaudoin B, Vuong N, Chen G, Chen HY, et al.Autophagy suppresses tumorigenesis through elimination of p62. Cell2009;137:1062–75.

8. Mathew R, Kongara S, Beaudoin B, Karp CM, Bray K, Degenhardt K,et al. Autophagy suppresses tumor progression by limiting chromo-somal instability. Genes Dev 2007;21:1367–81.

9. Marino G, Salvador-Montoliu N, Fueyo A, Knecht E, Mizushima N,Lopez-Otin C. Tissue-specific autophagy alterations and increased






tumorigenesis in mice deficient in Atg4C/autophagin-3. J Biol Chem2007;282:18573–83.

10. Yue Z, Jin S, Yang C, Levine AJ, Heintz N.Beclin 1, an autophagy geneessential for early embryonic development, is a haploinsufficient tumorsuppressor. Proc Natl Acad Sci U S A 2003;100:15077–82.

11. Takahashi Y, Coppola D, Matsushita N, Cualing HD, Sun M, Sato Y,et al. Bif-1 interacts with Beclin 1 through UVRAG and regulatesautophagy and tumorigenesis. Nat Cell Biol 2007;9:1142–51.

12. Liu L, Amy V, Liu G, McKeehan WL. Novel complex integratingmitochondria and the microtubular cytoskeleton with chromosomeremodeling and tumor suppressor RASSF1 deduced by in silicohomology analysis, interaction cloning in yeast, and colocalization incultured cells. In Vitro Cell Dev Biol Anim 2002;38:582–94.

13. Liu L, McKeehan WL. Sequence analysis of LRPPRC and its SEC1domain interaction partners suggest roles in cytoskeletal organization,vesicular trafficking, nucleocytosolic shuttling and chromosome activ-ity. Genomics 2002;79:124–36.

14. Liu L, Vo A, Liu G, McKeehan WL. Distinct structural domains withinC19ORF5 support association with stabilized microtubules and mito-chondrial aggregation and genome destruction. Cancer Res 2005;65:4191–201.

15. Liu L, Xie R, Yang C, McKeehan WL. Dual function microtubule- andmitochondria-associated proteins mediate mitotic cell death. CellOncol 2009;31:393–405.

16. Xie R, Nguyen S, McKeehan K, Wang F, McKeehan WL, Liu L.Microtubule-associatedProtein 1S (MAP1S) bridges autophagic com-ponents with microtubules and mitochondria to affect autophagoso-mal biogenesis and degradation. J Biol Chem 2011;286:10367–77.

17. Davison EJ, Pennington K, Hung CC, Peng J, Rafiq R, Ostareck-Lederer A, et al. Proteomic analysis of increased Parkin expressionand its interactants provides evidence for a role in modulation ofmitochondrial function. Proteomics 2009;9:4284–97.

18. Huang X, Yang C, Jin C, Luo Y, Wang F, McKeehan WL. Residenthepatocyte fibroblast growth factor receptor 4 limits hepatocarcino-genesis. Mol Carcinog 2009;48:553–62.

19. Huang X, Yu C, Jin C, Kobayashi M, Bowles CA,Wang F, et al. Ectopicactivity of fibroblast growth factor receptor 1 in hepatocytes accel-erates hepatocarcinogenesis by driving proliferation and vascularendothelial growth factor-induced angiogenesis. Cancer Res 2006;66:1481–90.

20. Mitchell C,WillenbringH. A reproducible andwell-toleratedmethod for2/3 partial hepatectomy in mice. Nat Protoc 2008;3:1167–70.

21. Ho YF, Guenthner TM. Isolation of liver nuclei that retain functionaltrans-membrane transport. J Pharmacol Toxicol Methods 1997;38:163–8.

22. LiuL,Trimarchi JR,SmithPJ,KeefeDL.Mitochondrial dysfunction leadsto telomere attrition and genomic instability. Aging Cell 2002;1:40–6.

23. Paull TT, Rogakou EP, Yamazaki V, Kirchgessner CU, Gellert M,Bonner WM. A critical role for histone H2AX in recruitment of repairfactors to nuclear foci after DNA damage. Curr Biol 2000;10:886–95.

24. McClintock B. The fusion of broken ends of chromosomes followingnuclear fusion. Proc Natl Acad Sci U S A 1942;28:458–63.

25. Storchova Z, Pellman D. From polyploidy to aneuploidy, genomeinstability and cancer. Nat Rev Mol Cell Biol 2004;5:45–54.

26. Geigl JB, Obenauf AC, Schwarzbraun T, Speicher MR. Defining `chro-mosomal instability'. Trends Genet 2008;24:64–9.

27. Boveri T. Zur frage der entstehung maligner tumoren. Jena, Germany:Gustav Fischer Verlag; 1914.

28. Duesberg P, Rasnick D. Aneuploidy, the somatic mutation that makescancer a species of its own. Cell Motil Cytoskeleton 2000;47:81–107.

29. Liang XH, Jackson S, SeamanM, Brown K, Kempkes B, Hibshoosh H,et al. Induction of autophagy and inhibition of tumorigenesis by beclin1. Nature 1999;402:672–6.

30. Takamura A, KomatsuM, Hara T, Sakamoto A, Kishi C, Waguri S, et al.Autophagy-deficient mice develop multiple liver tumors. Genes Dev2011;25:795–800.

31. Zatloukal K, French SW, Stumptner C, Strnad P, Harada M, ToivolaDM, et al. From Mallory to Mallory–Denk bodies: what, how and why?Exp Cell Res 2007;313:2033–49.

32. Komatsu M, Waguri S, Koike M, Sou YS, Ueno T, Hara T, et al.Homeostatic levels of p62 control cytoplasmic inclusion body forma-tion in autophagy-deficient mice. Cell 2007;131:1149–63.

33. Marietta C, Thompson LH, Lamerdin JE, Brooks PJ. Acetaldehydestimulates FANCD2 monoubiquitination, H2AX phosphorylation, andBRCA1 phosphorylation in human cells in vitro: implications for alco-hol-related carcinogenesis. Mutat Res 2009;664:77–83.

34. Mishra PK, Raghuram GV, Panwar H, Jain D, Pandey H, Maudar KK.Mitochondrial oxidative stress elicits chromosomal instability afterexposure to isocyanates in human kidney epithelial cells. Free RadicRes 2009;43:718–28.

35. D'Angiolella V, Santarpia C, Grieco D. Oxidative stress overrides thespindle checkpoint. Cell Cycle 2007;6:576–9.

36. Li M, Fang X, Baker DJ, Guo L, Gao X, Wei Z, et al. The ATM-p53pathway suppresses aneuploidy-induced tumorigenesis. Proc NatlAcad Sci U S A 2010;107:14188–93.

37. Gisselsson D, Pettersson L, Hoglund M, Heidenblad M, Gorunova L,Wiegant J, et al. Chromosomal breakage–fusion-bridge events causegenetic intratumor heterogeneity. Proc Natl Acad Sci U S A 2000;97:5357–62.

38. Burma S, Chen BP, Chen DJ. Role of non-homologous end joining(NHEJ) in maintaining genomic integrity. DNA Repair (Amst) 2006;5:1042–8.

39. Weaver BA, Cleveland DW. Decoding the links between mitosis,cancer, and chemotherapy: the mitotic checkpoint, adaptation, andcell death. Cancer Cell 2005;8:7–12.

40. Mitelman F, Johansson B, Mertens F. The impact of translocationsand gene fusions on cancer causation. Nat Rev Cancer 2007;7:233–45.

41. Yamada K, Yamamiya I, Utsumi H. In vivo detection of free radicalsinduced by diethylnitrosamine in rat liver tissue. Free Radic Biol Med2006;40:2040–6.

42. Limoli CL, Giedzinski E, Morgan WF, Swarts SG, Jones GD, Hyun W.Persistent oxidative stress in chromosomally unstable cells. CancerRes 2003;63:3107–11.

Xie et al.





Published OnlineFirst October 28, 2011.Cancer Res Rui Xie, Fen Wang, Wallace L. McKeehan, et al. Instability and HepatocarcinogenesisMitochondrion-Associated MAP1S Suppresses Genome Autophagy Enhanced by Microtubule- and

Updated version

10.1158/0008-5472.CAN-11-2170doi:

Access the most recent version of this article at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. (CCC)Click on "Request Permissions" which will take you to the Copyright Clearance Center's

.http://cancerres.aacrjournals.org/content/early/2011/12/07/0008-5472.CAN-11-2170To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/lookup/doi/10.1158/0008-5472.CAN-11-2170

http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://cancerres.aacrjournals.org/content/early/2011/12/07/0008-5472.CAN-11-2170


autophagy enhanced by microtubule- and mitochondrion ... · autophagy enhanced by microtubule- and...

Documents