European Journal of Cancer (2013) xxx, xxx– xxx

A v a i l a b l e a t w w w . s c i e n c e d i r e c t . c o m

ScienceDirect

journa l homepag e : www.e j cancer . com

SH2D4A is frequently downregulated inhepatocellular carcinoma and cirrhotic nodules

0959-8049/$ - see front matter � 2013 Elsevier Ltd. All rights reserved.

http://dx.doi.org/10.1016/j.ejca.2013.11.018

⇑ Corresponding author: Address: Department of Pathology, Molecular Pathology Division, University of Basel, Schonbeinstrasse 40, 400Switzerland.

E-mail address: lterracciano@uhbs.ch (L.M. Terracciano).1 These authors equally contributed to the work as senior authors.

Please cite this article in press as: Quagliata L. et al., SH2D4A is frequently downregulated in hepatocellular carcinoma and cirrhotic noduJ Cancer (2013), http://dx.doi.org/10.1016/j.ejca.2013.11.018

Luca Quagliata a, Mariacarla Andreozzi a, Michal Kovac a, Luigi Tornillo a,Zuzanna Makowska b, Francesca Moretti e, Markus H. Heim b, Karl Heinimann c,d,Salvatore Piscuoglio a,c,d,1, Luigi Maria Terracciano a,⇑,1

a Institute of Pathology, University Hospital of Basel, Basel, Switzerlandb Department of Biomedicine, Hepatology Laboratory, University of Basel, Basel, Switzerlandc Research Group Human Genetics, Department of Biomedicine, University of Basel, Basel, Switzerlandd Division of Medical Genetics, University Children’s Hospital, Basel, Switzerlande Department of Biomedicine, Institute of Biochemistry and Genetics, University of Basel, Basel, Switzerland

KEYWORDS

Hepatocellular carcinomaSH2D4A

Cirrhotic nodulesTumour suppressor gene

Abstract Hepatocellular carcinoma (HCC) is the third leading cause of cancer-relateddeath worldwide. The lack of effective therapeutic options for advanced stage HCCs com-bined with an increasing incidence rate calls for the identification of early stage HCC molec-ular markers. SH2 Domain Containing 4A (SH2D4A) gene maps to human chromosome8p21.3 and encodes for SH(2)A. The chromosomal region containing SH2D4A is frequentlylost in colorectal, lung and HCC cancers. Our study aimed to investigate SH2D4A involve-ment in HCC pathogenesis combining mRNA expression, protein and clinical data.Transcriptome analysis performed on 37 HCC needle biopsies (matched with their corre-sponding non-neoplastic parenchyma) and five normal liver donor samples revealed thatSH2D4A is downregulated in HCC. Results were confirmed by quantitative real-time-poly-merase chain reaction (qRT-PCR), 25 out of 37 (67.6%) fresh frozen samples showedSH2D4A downregulation (p = 0.026). Furthermore, combining qRT-PCR and immunohis-tochemistry data we demonstrated a direct correlation between SH2D4A mRNA andSH(2)A protein levels. The analysis of a tissue microarray (TMA) containing 336 specimensconfirmed that SH(2)A is frequently reduced in HCC (56.8%) as well as in cirrhotic nodules(50.5%) compared to normal liver samples (31.1%). To conclude, our study revealed thatSH2D4A is frequently downregulated in HCC samples thus corroborating its putative roleas a tumour suppressor gene. In addition, we provide new evidence for SH2D4A

3 Basel,

les, Eur

2 L. Quagliata et al. / European Journal of Cancer xxx (2013) xxx–xxx

Please cite this article in press as: QuaglJ Cancer (2013), http://dx.doi.org/10.10

involvement in HCC pathogenesis demonstrating for the first time its deregulation incirrhotic nodules.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Liver cancer is among the leading cause of cancer-related death worldwide with an estimated over500,000 new cases per year representing the fifth andninth cause of mortality among male and female cancerpatients, respectively [1,2]. Unlike most malignancies,mortality from liver cancer has increased significantlyover the past 20 years and epidemiologic evidence indi-cates that the medical and economic burden of liver can-cer will increase significantly in Western populationsduring the next decades [2]. Hepatocellular carcinoma(HCC) represents the vast majority of liver cancer cases(up to 85%) [3]. Hepatocarcinogenesis is a multi-stepprocess involving different genetic alterations that leadto malignant transformation of the hepatocyte [4]. MostHCCs arise in a cirrhotic liver [5] and worldwide around80% of them are related to a chronic infection witheither hepatitis B (HBV) or hepatitis C virus (HCV)[6]. Moreover, a rising proportion of HCCs is ascribedto alcohol abuse and metabolic disorders [7,8]. In thelast years, the extensive heterogeneity of genomic lesionsreported in HCCs has suggested that liver cancer is theresult of a combination of both, genetic and epigenetic,alterations which affect more than one regulatory path-way [9–14]. Among these pathways the best describedare Wnt/b-catenin, mitogen-activated protein kinase(MAPK), p14ARF/p53, p16INK4A/Rb, transforminggrowth factor-b (TGF-b) and phosphatase and tensinhomologue deleted on chromosome 10 (PTEN)/Akt[2]. Furthermore, comparative genomic hybridisation(CGH) has shown frequent DNA copy number gainsat chromosomes 1q, 8q and losses at 1p, 4q, 8p, 13q,16q and 17q in HCC specimens [4,15,16]. Interestingly,the chromosomal region 8p21–23 has been shown tobe frequently lost in colorectal cancer (CRC) and othersolid tumours [17] and thus has been generally linked tocarcinogenesis [18]. More recently, using an integrativeapproach of high-resolution array-based CGH and geneexpression profiling, Roessler et al. characterised someof the genes located at 8p in HCC samples [15]. Amongthese, SH2 Domain Containing 4A (SH2D4A) (8p21.3locus) encodes for SH(2)A. This protein contains a pro-line-rich region, tyrosine residues with potential sites forphosphorylation by LCK protein tyrosine kinase (PTK)and an SH2 domain with high levels of homology to theT cell specific adapter protein (TSAd) and to the Adap-ter protein of unknown function (ALX) [19]. Theseproteins have been shown to regulate T-cell receptor(TCR signal) transduction in T cells [19]. With regardto subcellular localisation, however, TSAd and ALX

iata L. et al., SH2D4A is frequently16/j.ejca.2013.11.018

are found in both cytoplasmic and nuclear compart-ments and are transported between compartmentsthrough active mechanisms, while SH(2)A residesalmost exclusively in the cell cytoplasm [19]. In addition,SH(2)A inhibits oestrogen-induced cell proliferation bycompeting with phospholipase C-gamma (PLC-c) forbinding to the oestrogen receptor-1 (ESR1), thusrepressing oestrogen-induced proliferation [20]. SH(2)Ahas been found to directly interact with ESR1 andthe cortactin/hematopoietic-specific protein 1 (HS1)complex [20]. SH2D4A knockout mice develop normallywith no obvious phenotype [21].

Concerning SH2D4A involvement in carcinogenesis,its overexpression inhibited HCC cell colony formationand cell migration in both Hep3B and HuH1 cellsin vitro [15]. Furthermore, in vivo mouse experimentsusing the same SH2D4A overexpressing HCC cell linesdemonstrated that SH2D4A significantly reducedtumour incidence and volume [15]. On the basis of theseresults SH2D4A has been suggested as a putative newtumour suppressor gene driving HCC tumourigenesis[15].

In this study, using a combination of global tran-scriptomic, specific mRNA expression analysis, immu-nohistochemistry (IHC) and tissue microarray (TMA)technology, we provide new evidence for SH2D4A

involvement in HCC pathogenesis and further establishits tumour suppressor role in HCC.

2. Materials and methods

2.1. Patient’s samples and RNA Isolation

Patient’s specimens and clinico-pathological datawere obtained from the Institute of Pathology, Univer-sity Hospital of Basel, Switzerland. The ethics commit-tee of the Institute of Pathology, University Hospitalof Basel, has approved the study. HCC diagnosis wasverified by pathological examination, no anti-cancertreatments were given before biopsy collection. Tumourdifferentiation was defined according to Edmondson’sgrading system. Only biopsies containing at least 50%of tumour cells and no necrotic area have been used inthis study. The clinico-pathologic features of the sam-ples used in this study are summarised in Table 1 andSupplementary Table 1. Total RNA was isolated from82 frozen needle biopsy (37 HCC samples matched withtheir corresponding non-neoplastic liver parenchymaand eight normal liver donors) using the TRI-reagent(Ambion) method according to the manufacturer’sguidelines.

downregulated in hepatocellular carcinoma and cirrhotic nodules, Eur

Table 1Clinico-pathological features of the 37 patients with hepatocellularcarcinoma (HCC) used in this study.

Factors

AgeMedian 72Range 26–89>65 12 (32%)<65 24 (68%)

GenderMale 35 (95%)Female 2 (5%)

FociSingular 20 (54%)Multifocal 17 (46%)

AFP (>200 ng/ml) (seven missing values)� 8 (26%)+ 23 (74%)

AetiologyNone 2 (5.4%)Alcohol 20 (54%)Hepatitis B virus (HBV) 4 (10.8%)HBV + alcohol 2 (5.4%)Hepatitis C virus (HCV) 3 (8.1%)HCV + alcohol 6 (16.2%)

AFP: a-fetoprotein.

L. Quagliata et al. / European Journal of Cancer xxx (2013) xxx–xxx 3

2.2. Transcriptomic profiling

Transcriptomic profiling of 37 HCC needle biopsiesmatched with their corresponding non-neoplastic liverparenchyma and five normal liver donors was per-formed with GeneChip�Human Gene 1.0ST arrays(Affymetrix, Santa Clara, CA) based on manufacturer’sinstructions. The arrays were scanned with an Affyme-trix GeneChip Scanner 3000. Resulting CEL files weresubsequently imported into Qlucore software andRobust Multichip Average (RMA) normalised. Princi-pal component analysis discriminating between HCCfrom their corresponding non-neoplastic liver paren-chyma and normal liver donor samples was performedat p values < 0.001 and false discovery rate of and lowerthan 0.1%.

2.3. Relative expression of SH2D4A by quantitativereal-time polymerase chain reaction

SH2D4A mRNA expression on 37 HCC needle biop-sies matched with their corresponding non-neoplasticliver parenchyma and eight normal liver donors freshfrozen tissues was assessed using the TaqMan� Probe-Based Gene Expression Analysis (Applied Biosystems,Foster City, CA), using the SH2D4A probeHs00930134_m1 (Applied Biosystems, Foster City,CA). The measurements were normalised using thebeta-actin probe Cf03023880_g1 (Applied Biosystems,

Please cite this article in press as: Quagliata L. et al., SH2D4A is frequentlyJ Cancer (2013), http://dx.doi.org/10.1016/j.ejca.2013.11.018

Foster City, CA) [22]. The fold-changes in gene expres-sion were calculated using the standard DDCt method[23].

2.4. Immunohistochemistry

For immunohistochemical analysis (IHC) of SH(2)A,we analysed formalin-fixed paraffin embedded (FFPE)tissue samples of 23 out of our set of 37 HCC needlebiopsies (for the remaining 14 samples no FFPE tissuewas available for this analysis). In addition, we screenedan HCC tissue microarray (TMA) previously describedby Baumhoer et al. [24]. This array contains 434 tissuespecimens from both HCC and non-neoplastic livertissue samples, 336 were suitable for evaluations.Roughly 60% of patients whom specimens have beenused for TMA construction, underwent surgical resec-tion without prior treatment for HCC, the remaininghas been collected from autopsy cases. The non-neoplas-tic liver tissue specimens included 93 samples of livercirrhosis (hepatitis B virus [HBV]+, n = 20; hepatitis Cvirus [HCV]+, n = 18; HBV+/HCV+, n = 3; HBV+/hepatitis D virus [HDV]+, n = 4; HBV+/HDV+/HCV+, n = 1; alcohol abuse, n = 4; Wilson disease,n = 1; cryptogenic, n = 5 and unknown cause, n = 37),and 69 samples of normal liver parenchyma. Tissuesamples were retrieved from the archives of the Instituteof Pathology, University Hospital Basel, Switzerland,and the Department of Biomorphological Sciences,University ‘Federico II’ Naples, Italy. Histologic grad-ing of the 174 HCCs and classification of pre-neoplasticnodular liver lesions and non-neoplastic liver tissue sam-ples was performed by two experienced pathologistsaccording to the World Health Organisation classifica-tion and Edmondson & Steiner grading system. Sectionsof 4 lm of paraffin embedded tissue were immuno-stained using a primary antibody against SH(2)A(Abcam clone 103166 dilution 1:200) (http://www.pro-teinatlas.org/ENSG00000104611). Staining was carriedout as previously described [25]. Immunoreactivity wasscored semi-quantitatively by evaluating the stainingintensity as described by Allred et al. [26].

2.5. Statistical analysis

For statistical analysis, the Chi-square test (v2 test)and the Fisher’s exact test for non-parametric variablesand Student’s t-test for parametric variables were used,with all probabilities reported as two-tailed. Differencesin patient survival were assessed using the Kaplan–Meier method and analysed using the log-rank test inunivariate analysis. All tests were two sided and p

values < 0.05 considered to be statistically significant.Cut-off scores were selected by evaluating the receiver-operating characteristic (ROC) curves. The point onthe curve with the shortest distance to the coordinate

downregulated in hepatocellular carcinoma and cirrhotic nodules, Eur

4 L. Quagliata et al. / European Journal of Cancer xxx (2013) xxx–xxx

(0,1) was selected as the threshold value to classify casesas ‘positive/overexpressing’ or ‘negative/downregulated’[27]. Analyses were performed using the SPSS software(IBM).

Fig. 1. SH2 Domain Containing 4A (SH2D4A) mRNA expressionlevels. (A) Quantitative real-time-polymerase chain reaction (qRT-PCR) data showing SH2D4A expression in tumour free livers andhepatocellular carcinoma (HCC). These data demonstrate thatSH2D4A is frequently downregulated in hepatocellular carcinoma.(B) SH2D4A expression data stratification according to the aetiolog-ical related agent.

3. Results

3.1. SH2D4A mRNA is frequently downregulated in

HCC

Global transcriptomic analysis of 37 HCC biopsiesand their matched non-neoplastic counterpart plus fivenormal liver donor samples revealed that SH2D4A isnot only downregulated in HCC samples but also inthe adjacent tumour free area (Supplementary Fig. 1).A detailed description of the analysed cohort is providedin Table 1 and Supplementary Table 1. Of note, none ofthe patient received any therapeutic treatments at thetime of biopsy collection. We validated the array databy quantitative real-time polymerase chain reaction(qRT-PCR) evaluating mRNA expression levels in thesame set of samples with three additional normal liverdonors. In accordance with its suggested tumour sup-pressor role, our analysis demonstrated that SH2D4A

mRNA was downregulated in 25 out of 37 (67.6%)HCC samples. Furthermore, SH2D4A was normallyexpressed in 6 out of 37 (16.2%) and upregulated in 6out of 37 (16.2%) of tested specimens (Fig. 1A). Interest-ingly, histological evaluation revealed the presence ofMallory bodies (MBs) and intracellular hyaline bodies(IHBs) exclusively in HCC samples with upregulatedSH2D4A mRNA expression (Supplementary Fig. 2).In contrast, neither MBs nor IHBs were found inHCC samples with downregulated and normal SH2D4A

mRNA levels (Supplementary Fig. 2). Furthermore, fol-lowing stratification of HCC analysed samples based ontheir aetiologically related agent, we observed a trendexhibiting lower SH2D4A levels in alcohol-relatedHCC (p < 0.05) compared to virus-related samples(Fig. 1B). This aetiologically related difference was notobserved when analysing the corresponding adjacenttumour free area samples (Supplementary Fig. 3).

3.2. SH2D4A mRNA levels correlate with SH(2)A

protein expression in HCC

It has been shown that the alteration of gene expres-sion levels in cancer cannot always be directly correlatedwith protein levels, mainly due to post-transcriptionalprotein modifications. To determine the consequencesof altered SH2D4A mRNA expression at the proteinlevel, we thus performed IHC staining for SH(2)A onall available FFPE tissues (23 out of our set of 37HCC needle biopsies). A direct correlation betweenSH2D4A mRNA and SH(2)A protein levels wasobserved: samples with downregulated SH2D4A mRNA

Please cite this article in press as: Quagliata L. et al., SH2D4A is frequentlyJ Cancer (2013), http://dx.doi.org/10.1016/j.ejca.2013.11.018

presented with low SH(2)A levels; conversely, sampleswith higher SH2D4A mRNA expression showed highSH(2)A protein levels (Fig. 2). The cytoplasmic cellularlocalisation of SH(2)A was not altered in the analysedsamples (Fig. 3).

3.3. SH(2)A levels are reduced in HCC and cirrhotic

nodules

In order to further corroborate the results obtainedby qRT-PCR and IHC on whole FFPE tissue sections,we analysed SH(2)A protein levels in a TMA containing336 informative samples. Representative images ofdifferent staining intensity for normal liver and HCCtissues are shown in Fig. 3. In accordance to our

downregulated in hepatocellular carcinoma and cirrhotic nodules, Eur

Fig. 2. Correlation between SH2 Domain Containing 4A (SH2D4A)mRNA expression and protein levels in hepatocellular carcinoma(HCC). With this analysis we combined SH2D4A quantitative real-time-polymerase chain reaction (qRT-PCR) expression data andSH(2)A immunostaining. These data suggest a direct correlationbetween mRNA expression and protein content in HCC testedspecimens.

L. Quagliata et al. / European Journal of Cancer xxx (2013) xxx–xxx 5

previously described results, we found SH(2)A proteinlevel to be significantly reduced (p = 0.002) in HCCs(56.8%) as well as in cirrhotic nodules (50.5%)

Fig. 3. SH(2)A tissue microarray (TMA) immunostaining of liver and hdifferent protein levels for SH(2)A (intensity score: 0–1–2–3). (A and B) NoNormal liver with intensity 1 at 200� and 400�, respectively. (E and F) NoNormal liver with intensity 3 at 200� and 400�, respectively. (I–L) HepatocHCC with intensity 1 at 200� and 400�. (O and P) HCC with intensity 2 aScale bars correspond to 50 and 20 lm in panels A and B, respectively.

Please cite this article in press as: Quagliata L. et al., SH2D4A is frequentlyJ Cancer (2013), http://dx.doi.org/10.1016/j.ejca.2013.11.018

(p = 0.041) compared to normal liver samples (31.1%).In addition, we observed SH(2)A reduction also cir-rhotic nodules when compared to normal liver tissue(p = 0.041, respectively) (Table 2). No significant corre-lation between SH(2)A protein levels and clinico-patho-logical data (besides age) (Table 3), or overall survival(either in alcohol and virus-related or only in virus-related HCC) was found (Supplementary Fig. 4).

4. Discussion

In this work we used a combination of transcriptomicanalyses, qRT-PCR-based expression data and IHCanalysis to determine the extent of SH2D4A deregula-tion in HCC and cirrhotic nodules.

HCC is typically diagnosed late in its course, with amedian survival of approximately 6–20 months [28],making HCC one of the most aggressive cancer typeworldwide. Surgical resection, liver transplantationand, recently, local ablation therapies have achievedhigh curative rates in selected patients [29]. Howeveronly a small number of patients with early stage diagno-sis are eligible for such therapeutic approaches. In addi-tion, recurrence of HCC remains a major problem [30].

epatocellular carcinoma (HCC) punches. Representative pictures ofrmal liver with intensity 0 at 200� and 400�, respectively. (C and D)rmal liver with intensity 2 at 200� and 400�, respectively. (G and H)

ellular carcinoma (HCC) with intensity 0 at 200� and 400�. (M and N)t 200� and 400�. (Q and R) HCC with intensity 3 at 200� and 400�.

downregulated in hepatocellular carcinoma and cirrhotic nodules, Eur

Table 2Summary of observed SH(2)A protein levels in normal liver (n = 69), cirrhotic nodules (n = 93) and hepatocellular carcinoma (HCC) (n = 174).Immunoreactivity was scored semi-quantitatively by evaluating the staining intensity as described by Allred et al. [26]. Results were analysed usingthe Chi-square test (v2 test) and the Student’s t-test (p value).

Tissue type SH(2)A intensity n (%) v2 p

0 + ++ +++

Normal liver 22 (31.1) 29 (42) 16 (23.1) 2 (3.8) 14.5 0.002

HCC 99 (56.8) 49 (28.1) 19 (10.9) 7 (4.2)

Normal liver 22 (31.1) 29 (42) 16 (23.1) 2 (3.8) 8.252 0.041

Cirrhotic nodules 47 (50.5) 33 (35.4) 13 (14.1) 0 (0)

Cirrhotic nodules 47 (50.5) 33 (35.4) 13 (14.1) 0 (0) 5.721 0.127HCC 99 (56.8) 49 (28.1) 19 (10.9) 7 (4.2)

Table 3SH(2)A protein levels obtained by analysing tissue microarray (TMA) contained hepatocellular carcinoma (HCC) specimens. Differences instaining intensity have been correlated with available clinico-pathological factors. Immunoreactivity was scored semi-quantitatively by evaluatingthe staining intensity as described by Allred et al. [26]. Results were analysed using the Chi-square test (v2 test) and the Student’s t-test (p value).

Factors SH(2)A intensity (n) v2 p

0 + ++ +++

Age 665 20 15 8 4 7.957 0.047

>65 78 33 11 3

Gender Female 19 8 3 3 2.95 0.399Male 80 41 16 4

Nstage (four missing values) 0 87 44 19 6 2.685 0.4431 9 3 0 0

Tstage (14 missing values) (1–2) 55 39 15 3 5.938 0.115(3–4) 30 9 4 3

Mstage (four missing values) 0 80 42 17 5 1.202 0.7521 16 5 2 1

Edmondson stage Moderate 40 26 11 3 5.759 0.764Poor 17 4 2 1Well 40 19 6 3

Edmondson number 1 39 16 6 3 3.905 0.692 41 26 11 33 17 4 2 1

Multifocality (three missing values) No 42 25 11 3 3.081 0.799Yes 53 23 8 4

Vascular invasion (16 missing values) No 51 33 8 5 8.791 0.186Yes 37 12 6 1

Aetiology (49 missing values) Alcohol 16 10 5 1 4.027 0.673Viralinfection

49 27 10 5

6 L. Quagliata et al. / European Journal of Cancer xxx (2013) xxx–xxx

For later stage HCC, recently, Llovet et al. demon-strated that an oral multikinase inhibitor (Sorafenib) iseffective in improving patient’s survival [31]. Neverthe-less, although such treatment holds a significanttherapeutic benefit, it is associated with a number ofside-effects [32]. Therefore the identification of specificmolecular markers able to identify early stage of HCCsis urgently needed.

Applying the CGH technique we and other researchgroups have previously described that 8p loss is one ofthe major cytogenetic alterations in liver cancer [23–28]. In different tumour types microsatellite studies have

Please cite this article in press as: Quagliata L. et al., SH2D4A is frequentlyJ Cancer (2013), http://dx.doi.org/10.1016/j.ejca.2013.11.018

identified at least three different regions of allelic loss on8p, suggesting the presence of several tumour suppressorgenes located on 8p21, 8p22 and 8p23 [33]. The 8p losshas also been shown to be a site of preferential loss ofheterozygosity in a series of hepatocellular nodules, withincreasing frequency from cirrhosis to HCC [34]. In par-ticular we detected deletions on chromosome 8p21–23 in50% of high-grade hepatic dysplastic nodules, suggestinga possible critical role for the genes located in this regionduring the early stages of hepatocarcinogenesis [35]. Inthe last years the importance of genomic deletions at8p21–23 has been further outlined in other solid

downregulated in hepatocellular carcinoma and cirrhotic nodules, Eur

L. Quagliata et al. / European Journal of Cancer xxx (2013) xxx–xxx 7

tumours [15,17]. Recently Roessler et al. performed asystematic analysis of the genes contained within thisfrequently deleted locus to investigate their involvementin hepatocarcinogenesis. In this study the authors pro-posed that SH2D4A acts as a tumour suppressor genein HCC [15].

In accordance with these findings, investigating twoindependent cohorts of HCC samples we observedSH2D4A downregulation both at the mRNA (liver nee-dle biopsies) and the protein level (liver-TMA). Alcohol-related HCC samples presented with lower SH2D4A

levels compared to virus-related specimens. Further-more, we provide intriguing new experimental evidencefor SH2D4A involvement during the early stages ofHCC pathogenesis observing SH2D4A reduced expres-sion also in cirrhotic nodule samples. Importantly, wewould like to stress the point that human data wereobtained from snap-frozen needle biopsies samples col-lected from patients that did not received any previousHCC-tailored therapeutic treatments neither, as in thevast majority of similar studies, from surgical resectedspecimens often resulting in ischaemic or therapy-induced tissue alterations.

In a previous study, SH(2)A has been shown to com-pete with phospholipase C-c (PLC-c) for binding toESR1 [20]. Oestrogen-induced proliferation has beenrecognised as an important pathway to sustain tumourgrowth in breast cancer [20]. Li et al. therefore demon-strated an inhibitory effect of SH(2)A on cell prolifera-tion due to suppression of a non-genomic action ofoestrogen receptor-alpha (ERa) via an ERa/PLC-c/pro-tein kinase C (PKC) signalling pathway [20]. Based onthese data, it is also tempting to speculate that dimin-ished SH(2)A protein levels, as observed in about 60%of HCC analysed samples, might facilitate PLC-c bind-ing to ESR1. This in turn could lead to an increased pro-liferation rate in SH(2)A deficient cells sustaining cancergrowth. Conversely, in accordance with this hypothesis,the overexpression of SH(2)A in HCC cells has beenreported to attenuate tumour growth [15]. Similar obser-vation have also been made by our research group incolorectal cancer specimens (manuscript in preparation)suggesting that downregulation of SH2D4A could playan important role in the development of different cancertypes.

In the HCC cohorts investigated in this study theSH2D4A/SH(2)A levels failed to predict patient’s sur-vival (either in alcohol and virus-related HCC or onlyvirus-related HCC). However, previously Roessleret al. demonstrated that using a ‘molecular signature’made of 10 genes often altered in HCC, among whichalso SH2D4A, they were able to predict HCC patients’survival. Although this result might appear to be in con-trast with our data, a few points should be considered.In their study, Roessler et al. used 256 specimensobtained from patients who underwent radical resection

Please cite this article in press as: Quagliata L. et al., SH2D4A is frequentlyJ Cancer (2013), http://dx.doi.org/10.1016/j.ejca.2013.11.018

and that were mostly characterised (96.31% of the entirecohort) by a history of HBV infection. Importantly,patients that are subjected to radical resection constitutea strictly selected subgroup among all the HCC patients;conversely our study has been performed using needlebiopsies thus representing the full range of HCC stages.Additionally, when analysing SH(2)A levels and sur-vival, we took advantage of a TMA made of an hetero-geneous cohort of HCC specimens with respect to theiraetiological related agent, having either an history ofalcohol consumption or HCV/HBV infection. Alto-gether, the populations analysed in our study highly dif-fer from the one used in Roessler et al., thus makingdifficult a direct comparison of the results. Furthermoreour result is not surprising in the context of HCC since itis in line with the lack of predictive value of other impor-tant tumour suppressor genes. For example, p53 andp16 also fail to serve as prognostic markers though theirloss of expression in HCC is correlated with increasedtumour aggressiveness [36].

Additional insights helping to uncover the molecularfunction of SH2D4A/SH(2)A come from the histopa-thological analysis of HCC whole section slides. Thisevaluation revealed the presence of MBs and IHBsalmost exclusively in HCC samples where SH2D4A

was upregulated. MBs and IHBs consist of intracellularaggregates of keratin, heath shock proteins, ubiquitinand p62 [37]. Although the process leading to the forma-tion of such structures remains to be determined, it hasbeen suggested that alterations of the ubiquitin/protea-some pathway might be a key element [37]. Thus, a pos-sible role of SH2D4A/SH(2)A in the regulation ofprotein ubiquitination deserves further investigation.

In conclusion, in this study we confirm the frequentdownregulation of SH2D4A in two independent, largecohorts of HCC samples further supporting its previ-ously proposed function as a tumour suppressor [15].Furthermore, as we reported for the first time SH2D4A

to be downregulated also in cirrhotic nodules, our datapave the way for further studies attempting to validateSH2D4A as a new diagnostic marker for the early detec-tion of HCC. Finally, the identification of defectiveSH2D4A mRNA expression levels in HCC patientscould offer an opportunity to develop novel specifictherapeutic strategies for patient treatment.

Conflict of interest statement

None declared.

Acknowledgements

The study was supported by grants from the Univer-sity of Basel, the Krebsliga Zentralschweiz (K.H.) andOncosuisse (KLS-2867-08-2011 to L.M.T.).

downregulated in hepatocellular carcinoma and cirrhotic nodules, Eur

8 L. Quagliata et al. / European Journal of Cancer xxx (2013) xxx–xxx

Appendix A. Supplementary data

Supplementary data associated with this article canbe found, in the online version, at http://dx.doi.org/10.1016/j.ejca.2013.11.018.

References

[1] Thorgeirsson SS, Grisham JW. Molecular pathogenesis of humanhepatocellular carcinoma. Nat Genet 2002;31:339–46.

[2] El-Serag HB, Rudolph KL. Hepatocellular carcinoma: epidemi-ology and molecular carcinogenesis. Gastroenterology2007;132:2557–76.

[3] Gomaa AI, Khan SA, Toledano MB, Waked I, Taylor-RobinsonSD. Hepatocellular carcinoma: epidemiology, risk factors andpathogenesis. World J Gastroenterol 2008;14:4300–8.

[4] Farazi PA, DePinho RA. Hepatocellular carcinoma pathogenesis:from genes to environment. Nat Rev Cancer 2006;6:674–87.

[5] Schutte K, Bornschein J, Malfertheiner P. Hepatocellular carci-noma – epidemiological trends and risk factors. Dig Dis2009;27:80–92.

[6] Bosch FX, Ribes J, Cleries R, Diaz M. Epidemiology ofhepatocellular carcinoma. Clin Liver Dis 2005;9:191–211, v.

[7] Grewal P, Viswanathen VA. Liver cancer and alcohol. Clin LiverDis 2012;16:839–50.

[8] Mair RD, Valenzuela A, Ha NB, Ayoub WS, Daugherty T,Lutchman GA, et al. Incidence of hepatocellular carcinomaamong US patients with cirrhosis of viral or nonviral etiologies.Clin Gastroenterol Hepatol 2012;10(12):1412–7.

[9] Guichard C, Amaddeo G, Imbeaud S, Ladeiro Y, Pelletier L,Maad IB, et al. Integrated analysis of somatic mutations and focalcopy-number changes identifies key genes and pathways inhepatocellular carcinoma. Nat Genet 2012;44:694–8.

[10] Nault JC, Zucman-Rossi J. Genetics of hepatobiliary carcino-genesis. Semin Liver Dis 2011;31:173–87.

[11] Hoshida Y, Toffanin S, Lachenmayer A, Villanueva A, MinguezB, Llovet JM. Molecular classification and novel targets inhepatocellular carcinoma: recent advancements. Semin Liver Dis2010;30:35–51.

[12] Zender L, Villanueva A, Tovar V, Sia D, Chiang DY, Llovet JM.Cancer gene discovery in hepatocellular carcinoma. J Hepatol2010;52:921–9.

[13] Toffanin S, Hoshida Y, Lachenmayer A, Villanueva A, CabellosL, Minguez B, et al. MicroRNA-based classification of hepato-cellular carcinoma and oncogenic role of miR-517a. Gastroen-terology 2011;140(1618–1628):e1616.

[14] Longerich T, Mueller MM, Breuhahn K, Schirmacher P, BennerA, Heiss C. Oncogenetic tree modeling of human hepatocarcino-genesis. Int J Cancer 2012;130:575–83.

[15] Roessler S, Long EL, Budhu A, Chen Y, Zhao X, Ji J, et al.Integrative genomic identification of genes on 8p associated withhepatocellular carcinoma progression and patient survival. Gas-troenterology 2012;142(957–966):e912.

[16] Tornillo L, Carafa V, Richter J, Sauter G, Moch H, Minola E,et al. Marked genetic similarities between hepatitis B virus-positive and hepatitis C virus-positive hepatocellular carcinomas.J Pathol 2000;192:307–12.

[17] Emi M, Fujiwara Y, Nakajima T, Tsuchiya E, Tsuda H,Hirohashi S, et al. Frequent loss of heterozygosity for loci onchromosome 8p in hepatocellular carcinoma, colorectal cancer,and lung cancer. Cancer Res 1992;52:5368–72.

[18] Macartney-Coxson DP, Hood KA, Shi HJ, Ward T, Wiles A,O’Connor R, et al. Metastatic susceptibility locus, an 8p hot-spotfor tumour progression disrupted in colorectal liver metastases: 13candidate genes examined at the DNA, mRNA and protein level.BMC Cancer 2008;8:187.

Please cite this article in press as: Quagliata L. et al., SH2D4A is frequentlyJ Cancer (2013), http://dx.doi.org/10.1016/j.ejca.2013.11.018

[19] Lapinski PE, Oliver JA, Bodie JN, Marti F, King PD. The T-cell-specific adapter protein family: TSAd, ALX, and SH2D4A/SH2D4B. Immunol Rev 2009;232:240–54.

[20] Li T, Li W, Lu J, Liu H, Li Y, Zhao Y. SH2D4A regulates cellproliferation via the ERalpha/PLC-gamma/PKC pathway. BMBRep 2009;42:516–22.

[21] Lapinski PE, Oliver JA, Kamen LA, Hughes ED, Saunders TL,King PD. Genetic analysis of SH2D4A, a novel adapter proteinrelated to T cell-specific adapter and adapter protein in lympho-cytes of unknown function, reveals a redundant function in Tcells. J Immunol 2008;181:2019–27.

[22] Dydensborg AB, Herring E, Auclair J, Tremblay E, Beaulieu JF.Normalizing genes for quantitative RT-PCR in differentiatinghuman intestinal epithelial cells and adenocarcinomas of the colon.Am J Physiol Gastrointest Liver Physiol 2006;290:G1067–74.

[23] Livak KJ, Schmittgen TD. Analysis of relative gene expressiondata using real-time quantitative PCR and the 2(�Delta DeltaC(T)) method. Methods 2001;25:402–8.

[24] Baumhoer D, Tornillo L, Stadlmann S, Roncalli M, Diamantis EK,Terracciano LM. Glypican 3 expression in human nonneoplastic,preneoplastic, and neoplastic tissues: a tissue microarray analysis of4,387 tissue samples. Am J Clin Pathol 2008;129:899–906.

[25] Lugli A, Tzankov A, Zlobec I, Terracciano LM. Differentialdiagnostic and functional role of the multi-marker phenotypeCDX2/CK20/CK7 in colorectal cancer stratified by mismatchrepair status. Mod Pathol 2008;21:1403–12.

[26] Allred DC, Harvey JM, Berardo M, Clark GM. Prognostic andpredictive factors in breast cancer by immunohistochemicalanalysis. Mod Pathol 1998;11:155–68.

[27] Piscuoglio S, Lehmann FS, Zlobec I, Tornillo L, Dietmaier W,Hartmann A, et al. Effect of EpCAM, CD44, CD133 and CD166expression on patient survival in tumours of the ampulla of Vater.J Clin Pathol 2012;65:140–5.

[28] A new prognostic system for hepatocellular carcinoma: a retro-spective study of 435 patients: the Cancer of the Liver ItalianProgram (CLIP) investigators. Hepatology 1998;28:751–55.

[29] Ye SL, Takayama T, Geschwind J, Marrero JA, Bronowicki JP.Current approaches to the treatment of early hepatocellularcarcinoma. Oncologist 2010;15(Suppl. 4):34–41.

[30] McCormack L, Petrowsky H, Clavien PA. Surgical therapy ofhepatocellular carcinoma. Eur J Gastroenterol Hepatol2005;17:497–503.

[31] Llovet JM, Ricci S, Mazzaferro V, Hilgard P, Gane E, Blanc JF,et al. Sorafenib in advanced hepatocellular carcinoma. N Engl JMed 2008;359:378–90.

[32] Keating GM, Santoro A. Sorafenib: a review of its use inadvanced hepatocellular carcinoma. Drugs 2009;69:223–40.

[33] Pineau P, Nagai H, Prigent S, Wei Y, Gyapay G, Weissenbach J,et al. Identification of three distinct regions of allelic deletions onthe short arm of chromosome 8 in hepatocellular carcinoma.Oncogene 1999;18:3127–34.

[34] Maggioni M, Coggi G, Cassani B, Bianchi P, Romagnoli S,Mandelli A, et al. Molecular changes in hepatocellular dysplasticnodules on microdissected liver biopsies. Hepatology2000;32:942–6.

[35] Tornillo L, Carafa V, Sauter G, Moch H, Minola E, GambacortaM, et al. Chromosomal alterations in hepatocellular nodules bycomparative genomic hybridization: high-grade dysplastic nod-ules represent early stages of hepatocellular carcinoma. LabInvest 2002;82:547–53.

[36] Matsuda Y, Ichida T, Genda T, Yamagiwa S, Aoyagi Y, AsakuraH. Loss of p16 contributes to p27 sequestration by cyclin D(1)-cyclin-dependent kinase 4 complexes and poor prognosis inhepatocellular carcinoma. Clin Cancer Res 2003;9:3389–96.

[37] Denk H, Stumptner C, Fuchsbichler A, Muller T, Farr G, MullerW, et al. Are the Mallory bodies and intracellular hyaline bodiesin neoplastic and non-neoplastic hepatocytes related? J Pathol2006;208:653–61.

downregulated in hepatocellular carcinoma and cirrhotic nodules, Eur