AIR

CC

HfeiAcvoc

cgTH

0d

1

purinic Apyrimidinic Endonuclease/Redox Effector Factor 1mmunoreactivity and Grading in Hepatocellular Carcinoma Risk ofelapse After Liver Transplantation

. Avellini, M. Orsaria, U. Baccarani, G.L. Adani, D. Lorenzin, V. Bresadola, F. Bresadola, and

.A. Beltrami

ABSTRACT

Apurinic apyrimidinic endonuclease (APE1)/redox effector factor 1 (Ref-1), which is amultifunction protein involved in both transcriptional regulation of gene expression duringadaptive cellular responses to oxidative stress and in the base excision repair pathway ofDNA lesions generated as a consequence of oxidant-induced base damage, contributes tothe maintenance of genome stability. APE1/Ref-1 is normally localized in the nucleus;cytoplasmic localization observed in several tumors has been correlated with a poorprognosis. Hepatocellular carcinoma (HCC) grading is an essential tool to predict the riskof relapse and patient prognosis, particularly in patients undergoing liver transplantation(OLT). The aim of this study was to identify the role of APE1/Ref-1 in predicting aposttransplant HCC relapse. We studied 48 patients transplanted for HCC to definegrading as well as nuclear and cytoplasmic APE1/Ref-1 expression within neoplastic versusnonneoplastic parenchyma. We defined a cutoff of 60% of cytoplasmic APE1/Ref-1expression to identify positive cases. At a minimum of 1.5-year follow-up after transplan-tation, 32 patients are alive and 16 patients are deceased after HCC relapse. Amonglow-grade HCC (grades 1 and 2), 76% of cases are alive; only 34% showed cytoplasmicAPE1/Ref-1 immunoreactivity. Among the high-grade cases (grades 3 and 4), 50% werealive with 64% showing cytoplasmic immunoreactivity. Nuclear reactivity was generallysimilar either in neoplastic or in cirrhotic livers, irrespective of the grade. These dataseemed to support the hypothesis of a predictive role of APE1/Ref-1 for HCC risk ofrelapse, which together with tumor grade by analysis of a pretransplant needle biopsy

should aid decision making for OLT.

oiisodav

SU

B

EPATOCELLULAR CARCINOMA (HCC) is thefifth most frequent neoplasm worldwide, accounting

or 5.6% of all human cancers,1 and is the third cause ofstimated cancer-caused deaths. The development of HCCs generally the final event of long-standing liver disease.ny agent leading to chronic liver injury and cirrhosis

onstitutes a possible risk for HCC, the most relevant beingiral infection (hepatitis B or C virus), alcohol intake, andther diseases such as nonalcoholic steatohepatitis, iron oropper deposition, and primary biliary cirrhosis.2

HCC is a long process,3 showing several molecularhanges with various phenotypes that suggest differentenetic and epigenetic alterations during carcinogenesis.he vast majority of HCC, develop in a cirrhotic liver.4

epatocyte DNA damage based on oxidative ground is S

041-1345/10/$–see front matteroi:10.1016/j.transproceed.2010.03.045

204

bserved in chronic inflammatory processes.5 It reflects anmbalance of an adequate base repair system that cannduce genomic and mitochondrial DNA damage.6 Atrong body of evidence suggests a crucial role of reactivexygen species (ROS) in the development of chronic liverisease. Evidence of oxidative stress has been detected inlmost all clinical and experimental chronic liver diseases ofarious etiologies and progression rates to fibrosis.7,8 The

From the Departments of Pathology (C.A., M.O., C.A.B.) andurgery (U.B., G.L.A., D.L., V.B., F.B.), University Hospital Udine,dine, Italy.Address reprint requests to Umberto Baccarani, Dott Umberto

accarani. Dipartimento di Scienze Chirurgiche, Università degli

tudi di Udine, Udine, Italy. E-mail: umberto.baccarani@uniud.it

© 2010 by Elsevier Inc. All rights reserved.360 Park Avenue South, New York, NY 10010-1710

Transplantation Proceedings, 42, 1204–1208 (2010)

imdt�cItR

fovdbacstpo(

aRoftliidcbccpcRirpt

MS

Wlcpt

S

Sthl

Fsosaa

I

IicaGdsPp6iosGwfacthnewecR

R

TfdbGlsgc

lbbcAH

cttn

APE/REF-1 IMMUNOREACTIVITY AND HCC RELAPSE 1205

ncreased ROS, which is associated with disease develop-ent and parenchymal damage, promotes oxidative stress

amage to proteins, lipids, and DNA as well as activation ofranscription factors (for instance, nuclear factor-�B (NF-B), STAT3, and AP-1), which are either involved inell-survival pathways or specifically activated in HCC.nflammatory activity and subsinusoidal fibrosis with al-ered blood-hepatocyte exchanges are the key factors inOS-induced damage.9

Apurinic apyrimidinic endonuclease (APE1)/redox effectoractor 1 (Ref-1) is a master regulator of cellular responses toxidative stress. APE1/Ref-1 is a multifunction protein in-olved in both transcriptional regulation of gene expressionuring adaptive cellular response to oxidative stress and in thease excision repair (BER) pathway of DNA lesions gener-ted as a consequence of oxidant-induced base damage,ontributing to the maintenance of the genome stability.10 Ittimulates DNA binding activity of AP-1 (Fos, Jun) pro-eins, as well as NF-�B, polyoma virus enhancer-bindingrotein 2 (PEBP2), early growth response-1, Myb, membersf the ATF/CREB family, hypoxia inducible factor-1�HIF-1�), HIF-like factor, Pax-5, and Pax-8.11

APE1/Ref-1 has been shown to be closely linked topoptosis. Altered levels or cellular locations of APE1/ef-1 have been observed in some cancers, includingvarian, cervical, prostate, and germ cell tumors.12 There-ore, APE1/Ref-1 appears to form a unique link betweenhe DNA BER pathway, cancer, transcription factor regu-ation, oxidative signaling, and cell-cycle control. Althoughts subcellular distribution in various mammalian cell typess mainly nuclear, cytoplasmic localization has also beenescribed10 and associated with distinct tumorigenic pro-esses.10 In particular, among lung, ovarian, thyroid,13 andreast14 cancers, a cytoplasmic distribution has been asso-iated with greater tumor aggressiveness. The possibleausal role played by this particular distribution in tumorrogression is, at present, completely unknown. Genomicellular changes and oxidative stress may trigger APE1/ef-1 responses. The aim of the present study was to

dentify the predictive role of APE1/Ref-1 for risk of HCCelapse and in association with tumor grade in analysis of aretransplant needle biopsy, aid decision making for liverransplantation.

ATERIALS AND METHODSubjects

e retrospectively studied 48 consecutive patients who underwentiver transplantation (OLT) for the diagnosis of HCC, which wasonfirmed by histological analysis at the time of surgery. All theatients have been followed for a minimum of 1.5 years afterransplantation.

ample Collection, Analysis, and Histological Grading

amples of HCC and surrounding liver cirrhosis (SLC) obtained athe time of the OLT were fixed in formalin for a maximum of 24ours and embedded in paraffin. Each liver sample was sufficiently

arge to include both the tumor and the surrounding cirrhosis. a

rom each sample, 4-�m sections were cut for one slide to betained with hematoxylin-eosin for morphological diagnosis. Thether slides were used for immunohistochemical analysis. HCCamples were histologically classified according to the Edmondsonnd Steiner criteria.15 All histological specimens were analyzed bysingle pathologist (C.A.).

mmunohistochemical Analysis

mmunohistochemical detection of APE1/Ref-1 was performed bymmunohistochemistry using the anti-APE1/Ref-1 mouse mono-lonal antibody, prepared as previously described,16 as the primaryntibody and the Super Sensitive TN Polymer HRP IHC (Bioenex, San Ramon, Calif, USA) as the detection system. The

ewaxed sections were treated with H2O2 and phosphate-bufferedaline (PBS) solution (40 mL H2O2 [140 vol], added to 160 mLBS) for 10 minutes at room temperature to inhibit endogenouseroxidase, before being immersed in 10 mmol/L citrate buffer (pH.0) for 40 minutes at 98°C. Sections washed with PBS were thenncubated with the primary antibody (1:50 diluted in PBS) at 4°Cvernight. After washing twice with PBS and incubation withecondary antibody (Super Sensitive TN Polymer HRP IHC, Bioenex) for 30 minutes at room temperature, peroxidase activityas developed with 3,3=-diaminobenzidine (used as a chromogen)

or 10 minutes. Mayer’s hematoxylin counterstain was subsequentlypplied to nuclei. A positive reactivity for APE1/Ref-1 was indi-ated by the presence of a dense, homogeneous brown staining ofhe nucleus and a granular brown staining of the cytoplasm of theepatocyte. APE1/Ref-1 expression evaluated by counting theumber of positive cells among 1000 hepatocytes on each slide wasxpressed as a percentage. Subcellular localization of APE1/Ref-1as classified as nuclear versus cytoplasmic on the basis of the 60%xpression value, which was used as cutoff value to define positiveases, considering as positive a nucleus or cytoplasm with APE1/ef-1 expression greater than or equal to 60%.

ESULTS

hirty two patients are alive with a minimum of 1.5 years ofollow-up after transplantation; 16 patients were deceasedue to HCC relapse. APE1/Ref-1 expression was comparedetween patients with highly differentiated (grade [G]1 to2 versus poorly differentiated (G3–G4) HCC. Among

ow-grade HCC, 76% of cases were alive and only 34%howed cytoplasmic APE1/Ref-1 immunoreactivity. High-rade cases showed 50% of patients still alive, with 64% ofases with cytoplasmic immunoreactivity.

Positive staining has been observed only at the nuclearevel of normal hepatocytes as well as endothelial andiliary duct cells. In contrast, expression of APE1/Ref-1 haseen detected not only in the nucleus but also in theytoplasm of HCC cells (Fig 1). Nuclear and cytoplasmicPE1/Ref-1 expressions were significantly higher amongCC than SLC.These findings suggested that the distinct nuclear versus

ytoplasmic localization of APE1/Ref-1 may have relevanceo the clinical outcomes of patients, as reported for otherypes of cancer.13,14 Because cutoffs of clinical relevance areot available, we split the value of APE1/Ref-1 expression

t 60% of neoplastic hepatocytes, defining a positive cyto-

pt

wvppn

tcgwoPs

aGsod

evss(h

gt

r0

Rwcc

D

Ht

Fceetp

1206 AVELLINI, ORSARIA, BACCARANI ET AL

lasm staining as a value of APE1/Ref-1 expression greaterhan or equal to 60%.

For various grades of neoplasia (G1, G2, and G3 � G4),e analyzed the mean value, standard deviation, medianalues and 25th and 75th percentiles of rate of APE1/Ref-1ositivity in the neoplasia (N) versus the surroundingarenchyma (C1) and in the distant parenchyma (C2;uclear and cytoplasmic).The Kruskal Wallis nonparametric method was applied

o evaluate whether the proportion of immunohistochemi-al positivity was significantly different between the threeroups based on HCC grades (G1, G2, G3 � G4). Thereas a significant difference between cytoplasmic positivityf neoplasia (Ncyto) among the three grades. (chi-square� .0342). There was no significant difference at the other

ites.As shown in Table 1 at 1.5-year follow-up, 3/4 (75%) G1

nd 22/29 (75.9%) G2 patients, but only 6/14 (43%) G3 �4 patients were alive free of relapse. These findings

uggested that there was only a trend to worse survival andf a higher risk of relapse among patients with poorlyifferentiated HCC (Fisher exact test P � .08).The relative risk of death or relapse at 1.5 years was

valuated with odds ratios (ORs) and 95% confidence inter-als (CIs) with nonconditional logistic regression analy-is. Using G1 cases as the reference group, G3 � G4 caseshowed higher relative risks of death or relapse at 1.5 yearsOR � 4.0, 95% CI 0.3–48.6), while G2 cases did not showigher relative risk.Negative immunohistochemical tests were the reference

roup, showing that cases with cytoplasmic reactivity within

ig 1. High-grade hepatocellulararcinoma (grade 3); hematoxylin-osin 10�. Apurinic apyrimidinicndonuclease/redox effector fac-or 1 (APE1/Ref-1) cytoplasmic ex-ression; APE1/Ref-1 20�.

he neoplastic tissue showed a relative risk of death or

elapse at 1.5 years of 50% greater (OR � 1.5; 95% CI.4–5.11).These findings suggested that only evaluation of APE1/ef-1 cytoplasmic positivity, but not the nuclear expression,as crucial to predict patient clinical outcomes. All theonfidence intervals were wide due to the small number ofases.

ISCUSSION

CC is one of the most common neoplasms worldwide andhe fifth most common cause of mortality from cancer. This

Table 1. Grade and Survival

Grade 1Survival

Total

Frequency PercentRow PctCol Pct

Death or Relapseat 1.5 y

Alive at1.5 y

1 1 3 42.13 6.38 8.51

25.00 75.006.25 9.68

2 7 22 2914.89 46.81 61.7024.14 75.8643.75 70.97

3–4 8 6 1417.02 12.77 29.7957.14 42.8650.00 19.35

Total 16 31 4734.04 65.96 100.00

Frequency missing � 1.

naas

oHpotcH

wtiltoc

ocEspmtEsbpt

ftptcntTgldi�pImamti

nr

finnmw

spn

chrtDrhhc

w1RWRd

aiHbd

mwat

pioechlsrtAo

R

t

APE/REF-1 IMMUNOREACTIVITY AND HCC RELAPSE 1207

eoplasm is almost always associated with cirrhosis and isssociated with the main risk factors of chronic hepatitis Cnd B infections, alcoholic cirrhosis, or hemochromato-is.2,4,7

Liver resection and transplantation are considered thenly two potentially curative treatments for this cancer.17

owever, the long-term prognosis remains poor with he-atic resection. For this reason, OLT may be viewed as theptimal treatment for HCC, because of the widest possibleumor resection margins and the removal of the underlyingirrhotic parenchyma, which is at risk for recurrence ofCC.18

However, the chronic shortage of donor organs togetherith the increased number of patients awaiting transplan-

ation makes this therapeutic option less available forndividual patients. Because of the limited organ supply,iver transplant centers usually tend to reserve transplanta-ion for patients with impaired liver function and smallligonodular HCC, considered within the currently ac-epted Milan criteria for transplantation.19

The grading of neoplasia represents the level of similarityf a tumor to nonneoplastic tissue, with a higher or lowerellular atypia. The grading for HCC is based on thedmondson and Steiner system described in 1954 with

ubsequent modifications proposed by Ishak: the mainarameters considered are nuclear features (polymorphism,embrane irregularity, nucleoli) as well as cellular archi-

ectural organization, cellular polymorphism, and necrosis.ven if controversial, grading together with vascular inva-

ion has been useful as a predictor for HCC biologicalehavior. To define strict criteria for the inclusion ofatients on the transplantation list, efforts have been madeo utilize pretransplantation HCC staging criteria.

Mazzaferro et al in 1996 established the Milan criteria20:or patients with a single HCC to be eligible for transplan-ation, the tumor could not exceed 5 cm in diameter; amongatients with multiple tumours, there could be no morehan three masses, none exceeding 3 cm in diameter. Theseriteria exclude patients in whom blood vessel or lymphode invasion are evident or suspected preoperatively. Arend toward expanding the Milan criteria is reflected byhe University of California, San Francisco criteria21: sin-le tumor nodule up to 6.5 cm; or three or fewer tumors, theargest being 4.5 cm with the sum of the total tumoriameters equal to 8 cm. Barcelona expanded criteria22

nclude one tumor �7 cm, three tumors �5 cm, five tumors3 cm, or downstaging to conventional Milan criteria with

retransplant adjuvant therapies. The Cancer of the Livertalian Program score23 includes Child-Pugh stage, tumororphology and extent, presence of portal vein thrombosis,

nd serum level of alpha-fetoprotein. These are all funda-ental parameters for the inclusion of HCC patients on the

ransplantation list. It is appropriate and necessary tonclude among these criteria an evaluation of grading.

A first grading determination can be made on a fineeedle biopsy performed after evaluation of the hemor-

hagic risk, given the lower risk of seeding with OLT. The L

ne needle biopsy has several limits: compared with aeoplastic nodule of some centimeters in diameter, a biopsyeedle of 1 to 2 cm in length and a few millimetres diameteray not obtain the poorly differentiated part of the tumorith a risk of undergrading the neoplasia.Immunohistochemical evaluation of the cytoplasmic expres-

ion of APE1/Ref-1 may be useful. This multifunctionalrotein is involved in several steps in the development ofeoplasia.Transcriptional regulation and BER are crucial functions

onnected with neoplastic pathogenesis and biological be-avior. Cytoplasmic relocalization of the protein may beelated to a cleavage phenomenon from APE1/Ref-1 inumor tissue, with a possible relation to mitochondrialNA repair and/or to the response with endoplasmic

eticulum to cytoplasmic oxidative stress. Unfavorable be-aviors of various neoplasias—lung, breast, thyroid, ovary—ave been observed to be associated with APE1/Ref-1ytoplasmic reactivity, as was evident in our series of HCCs.

In fact, we observed that poorly differentiated HCCsere associated with a higher risk of relapse or death within.5 years, and also that cytoplasmic expression of APE1/ef-1 was associated with higher risk of relapse or death.e have also shown that cytoplasmic expression of APE1/ef-1 was more frequently associated with poor HCCifferentiation.Therefore, probably APE1/Ref-1 cytoplasmic reactivity is

n expression of severe oxidative stress and DNA changesn neoplastic cells, which show more aggressive behavior.

igher grading and APE1/Ref-1 cytoplasmic reactivity cane considered markers of the risk of HCC relapse or ofeath within the first 1.5 years after OLT.These data together with the stage of neoplastic diseaseay be useful for decision making about listing patientsith HCC for OLT. Additional data are needed to providen explanation for the molecular events that are the basis ofhese observations.

In conclusion, APE1/Ref-1, a multifunctional protein,lays a central role in cellular response to oxidative stress,

nvolving both control of gene expression and maintenancef genome stability. HCC is characterized by increasedxpression of APE1/Ref-1 compared with surrounding liverirrhosis, with cytoplasmic localization of the protein. Aigher cytoplasmic accumulation was associated with a

ower degree of differentiation and significantly shorterurvival time. Collectively, these data suggested a possibleole of APE1/Ref-1 overexpression in the development ofhe HCC, indicating that the subcellular localization ofPE1/Ref-1 in HCC tissue might be used as an index of riskf death or relapse after liver transplantation.24

EFERENCES

1. Bosch FX, Ribes J, Cleries R, et al: Epidemiology of hepa-ocellular carcinoma. Clin Liver Dis 9:191, 2005

2. Llovet JM, Burroughs A, Bruix J: Hepatocellular carcinoma.ancet 362:1907, 2003

h

cD

am

tcTP

h

R

in

lA

ta

oAAC

aE

AO

a

ir

o

3

t2

tp

Ifw

ce

spg

t

1208 AVELLINI, ORSARIA, BACCARANI ET AL

3. Thorgeirsson SS, Grisham JW: Molecular pathogenesis ofuman hepatocellular carcinoma. Nat Genet 31:339, 20024. Simonetti RG, Camma C, Fiorello F, et al: Hepatocellular

arcinoma: a worldwide problem and the major risk factors. Digis Sci 36:962, 19915. Kawanishi S, Hiraku Y: Oxidative and nitrative DNA damage

s biomarker for carcinogenesis with special reference to inflam-ation. Antioxid Redox Signal 8:1047, 20066. Grisham JW: Molecular genetic alterations in primary hepa-

ocellular neoplasms: hepatocellular adenoma, hepatocellular car-inoma and hepatoblastoma. In Coleman WB, Tsongalis GJ (eds):he Molecular Basis of Human Cancer. Totowa, NJ: Humanaress, 2001, p 2697. Loguercio C, Federico A: Oxidative stress in viral and alco-

olic hepatitis. Free Radic Biol Med 34:1, 20038. Poli G, Parola M: Oxidative damage and fibrogenesis. Freeadic Biol Med 22:287, 19979. Jüngst C, Cheng B, Gehrke R, et al: Oxidative damage is

ncreased in human liver tissue adjacent to hepatocellular carci-oma. Hepatology 39:1663, 200410. Tell G, Damante G, Caldwell D, et al: The intracellular

ocalization of APE1/Ref-1: more than a passive phenomenon?ntioxid Redox Signal 7:367, 200511. Xanthoudakis S, Curran T: Identification and characteriza-

ion of Ref-1, a nuclear protein that facilitates AP-1 DNA-bindingctivity, EMBO J 11:653, 1992

12. Kelley MR, Moore DH, Cheng L, et al: Altered expressionf the multifunctional DNA base excision repair and redox enzymePE/Ref-1 in ovarian and prostate cancers. In: Proceedings of theACR Special Conference on DNA Repair Defects, San Diego,alif, 200013. Russo D, Arturi F, Bulotta S, et al: ApeI/Ref-I expression

nd cellular localization in human thyroid carcinoma cell lines. J

ndocrinol Invest 24:RC10, 2001 p

14. Puglisi F, Barbone F, Tell G, et al: Prognostic role ofpe/Ref-1 subcellular expression in stage I–III breast carcinomas.ncol Rep 9:11, 200215. Edmondson HA, Steiner PE: Primary carcinoma of the liver:study of 100 cases among 48,900 necropsies. Cancer 7:462, 195416. Pines A, Perrone L, Bivi N, et al: Activation of APE1/Ref-1

s dependent on reactive oxygen species generated after purinergiceceptor stimulation by ATP. Nucleic Acids Res 33:4379, 2005

17. Forner A, Heissheimer AJ, Isabel Real M, et al: Treatmentf hepatocellular carcinoma. Crit Rev Oncol Hematol 60:89, 200618. Belghiti J: Transplantation for liver tumors. Semin Oncol

2(6 Suppl 8):29, 200519. Llovet JM, Schwartz M, Mazzaferro V: Resection and liver

ransplantation for hepatocellular carcinoma. Semin Liver Dis5:181, 200520. Mazzaferro V, Regalia E, Doci R, et al: Liver transplanta-

ion for the treatment of small hepatocellular carcinomas inatients with cirrhosis. N Engl J Med 334:693, 199621. Decaens T, Roudot-Thoraval F, Hadni-Bresson S, et al:

mpact of UCSF criteria according to pre- and post-OLT tumoreatures: analysis of 479 patients listed for HCC with a shortaiting time. Liver Transpl 12:1761, 200622. Duffy JP, Vardanian A, Benjamin E, et al: Liver transplantation

riteria for hepatocellular carcinoma should be expanded: a 22-yearxperience with 467 patients at UCLA. Ann Surg 246:502, 2007

23. Masatoshi K, Hobyung C, Yukio O: A new prognosticystem for hepatocellular carcinoma: a retrospective study of 435atients: the Cancer of the Liver Italian Program (CLIP) investi-ators. Hepatology 28:751, 1998

24. Di Maso V, Avellini C, Crocè LS, et al: Subcellular localiza-ion of APE1/Ref-1 in human hepatocellular carcinoma: possible

rognostic significance. Mol Med 13:89, 2007