Histopathology 2009, 54, 319–327. DOI: 10.1111/j.1365-2559.2009.03224.x

Down-regulated expression of the TSAP6 protein in liver is associated with a transition from cirrhosis to hepatocellular carcinoma Fre´de´rique Caillot, Romain Daveau, Maryvonne Daveau, Jean Lubrano,1 Gae¨lle Saint-Auret, Martine Hiron, Odile Goria,2 Michel Scotte,1 Arnaud Francois3 & Jean-Philippe Salier Inserm Unite´ 905 and Institut Fe´de´ratif de Recherches Multidisciplinaires sur les Peptides, Faculte´ de Me´decine-Pharmacie, 1 Service de Chirurgie Ge´ne´rale et Digestive, 2Service d’He´pato-Gastro-Ente´rologie, and 3De´partement de Pathologie, Centre Hospitalier Universitaire, Rouen, France Date of submission 31 March 2008 Accepted for publication 29 July 2008

Caillot F, Daveau R, Daveau M, Lubrano J, Saint-Auret G, Hiron M, Goria O, Scotte M, Francois A & Salier J-P (2009) Histopathology 54, 319–327

Down-regulated expression of the TSAP6 protein in liver is associated with a transition from cirrhosis to hepatocellular carcinoma Aims: Hepatocellular carcinoma (HCC) results from cirrhosis and, in Western Europe, hepatitis C virus and alcoholism are the predominant causes of this disease. We recently documented a global transcript repression in hepatocarcinoma nodules. The tumour suppressor activated pathway-6 (TSAP6) transcript codes for a transmembrane molecule that is an inducer of a caspase-3-dependent apoptotic pathway. The down-regulation of TSAP6 transcripts in HCC and perinodular cirrhosis, which contrasts with a sustained transcript level in HCC-free cirrhosis, has suggested that this hepatic protein level may provide a prognostic marker for HCC occurrence.

Methods and results: This protein was quantified by semiquantitative assessment of immunohistochemistry on samples from 42 cases HCC-free cirrhosis, 49 cases cirrhosis with HCC, 43 HCC associated with healthy liver and 31 controls. TSAP6 expression was linked to the liver state, healthy or cirrhotic without or with an HCC and to tumour grade. Conclusions: With biopsies periodically performed for surveillance purposes, the decreased expression of TSAP6 in cirrhotic tissue could reflect a decrease in the apoptotic process and could be interpreted as a warning sign. This evaluation of the TSAP6 level in cirrhotic liver conveys predictive information for the development of HCC.

Keywords: alcoholism, cancer liver, cirrhosis, hepatitis C, TSAP6) Abbreviations: HCC, hepatocellular carcinoma; HCV, hepatitis C virus; HL, healthy liver; IP, intensity and percentage of positive cells; PBS, phosphate-buffered saline; PTC, peritumoral cirrhosis; PTC1, peritumoral cirrhosis with tumour grade 1; PTC2, peritumoral cirrhosis with tumour grade 2; PTC3, peritumoral cirrhosis with tumour grade 3; PTHL, peritumoral healthy liver; TSAP6, tumour suppressor activated pathway-6

Introduction Hepatic parenchyma is a target for pathogenic molecules. For example, hepatitis B or C virus Address for correspondence: F Caillot, Inserm Unite´ 905, Faculte´ de Me´decine – Pharmacie, 22 Bvd Gambetta, 76183 Rouen cedex, France. e-mail: [email protected] ! 2009 The Authors. Journal compilation ! 2009 Blackwell Publishing Limited.

(HBV, HCV) as well as chronic alcohol abuse usually promote an acute inflammatory phase that remains unnoticed, but is followed by a chronic, pathological condition and cirrhosis1,2 and, in a limited number of cases, hepatocellular carcinoma (HCC).3 Indeed, in most instances, HCC results from cirrhosis4 and, in western Europe, HCV and alcoholism are the predominant causes of this disease.5,6 We have

320

F Caillot et al.

recently documented a global transcript repression in HCC nodules.7 Some repressed proteins are concerned with apoptosis, suggesting that reduction in apoptosis predominates over cell proliferation in the onset of HCC.7 Among these transcripts, the tumour suppressor activated pathway-6 (TSAP6) transcript codes for an ubiquitous, transmembrane molecule that participates in cell–cell communication by regulating the export and traffic of a so-called exosomal vesicule.8,9 Also, TSAP6 is an inducer of a caspase-3-dependent apoptotic pathway10,11 and a p53-dependent inhibitor of G2 ⁄ M transition.12 The involvement of TSAP6 in apoptosis arises from its interaction with p53, as well as its association with Nix, a proapoptotic Bcl2related protein, and Myt1 kinase, a negative regulator of G2 ⁄ M transition.12,13 Likewise, by virtue of apoptosis induction, pHyde, i.e. the rat orthologue of TSAP6, inhibits prostatic cancer cell proliferation.10,14,15 On such grounds, down-regulation of TSAP6 transcripts in HCC and perinodular cirrhosis, which both contrast with a sustained transcript level in HCC-free cirrhosis, has suggested that the TSAP6 protein level in liver may provide a prognostic marker for HCC occurrence.7 However, numerous discrepancies between transcript levels and matched protein levels have been previously noticed, which still makes inference of protein from transcript level a highly debatable matter.16,17

Patients and methods

taken from paraffin-embedded hepatic tissue. According to current French legislation and ethical guidelines, neither informed consent nor advice from an ethics committee were requested prior to histological analysis in resected tissues that would otherwise be disposed off. HCCs were graded according to the World Health Organization classification.18 Briefly, welldifferentiated HCC cells (grade 1) were smaller than adjacent hepatocytes and had nuclear cytoplasmic ratios above that of adjacent hepatocytes. Intermediate differentiated cells (grade 2) had elevated numbers of pseudoglandular structures. Poorly differentiated cells (grade 3) were found in livers with a limited number of residual trabecular, compact and vessel-poor structures. polyclonal anti-tsap 6 antibodies Rabbit polyclonal IgGs raised against the human TSAP6 residues 16-DSDSSLAKVPDEAPK-3012 were affinity purified on cyanogen bromide-activated Sepharose covalently covered with residues 16–30, and eluted with 100 mm glycine at pH 2.5. Antibody specificity was then verified by enzyme-linked immunosorbent assay with TSAP6 residues 16–30 as coating antigen, and affinity purified anti-TSAP6 IgGs or control IgGs isolated prior to immunization. IgGs were stored as frozen solutions [1300 lg ⁄ ml in phosphatebuffered saline (PBS)]. All the above steps were carried out by Eurogentec (Seraing, Belgium).

human livers a nd hcc h istology

immunohistochemistry

Neoplastic and peritumoral cirrhotic liver samples were derived from surgical resections. Patients with HCC-free cirrhosis were defined as those with a span of ‡ 3 years without any visible HCC after a first liver biopsy. For these patients liver samples used in the present study were provided either from fragments from surgery performed in order to confirm histologically the macroscopic diagnosis of cirrhosis (8 ⁄ 17 for alcoholic cirrhosis; 6 ⁄ 25 for HCV cirrhosis) during the course of a cholecystectomy, from a porta-caval anastomosis or from needle liver biopsy specimens performed during follow-up of cirrhosis (alcoholic 9 ⁄ 17, or post HCV 19 ⁄ 25) and retrieved from the Department of Pathology at the Charles-Nicolle Hospital in Rouen (France). Patients with HCC appearing in the context of a healthy liver (HL) were also included in the present study. Controls were livers operated on for benign tumours (i.e. adenoma, focal nodular hyperplasia). For each patient, sections were

Formaldehyde-fixed, paraffin-embedded, 5 mm-thick liver sections were deparaffinized in xylene and rehydrated on a Techmate 500 apparatus (Dako, Trappes, France). Heat-induced retrieval of formaldehyde-hidden epitopes was performed by microwaving (60"C) for 40 min in citrate buffer (1 mm, pH 6.0). After preliminary tests for optimal IgG concentration, the sections were incubated with rabbit anti-TSAP6 IgGs at 1.3 lg ⁄ ml for 30 min at room temperature. The secondary antibodies (biotinylated goat antirabbit IgGs; Dako Labelled Streptavidin-Biotin universal kit ⁄ Horseradish peroxidase) were incubated for 25 min at room temperature. After a further incubation with a strepta vidin–peroxidase complex, binding of antiTSAP6 or control IgGs was visualized using diaminobenzidene solution (75 mg 3.3¢-diaminobenzidine and 30 ll 50% H2O2 in 150 ml PBS) for 10 min. Counterstaining was achieved with haemalun, which provided blue staining of the nuclei.

! 2009 The Authors. Journal compilation ! 2009 Blackwell Publishing Ltd, Histopathology, 54, 319–327.

TSAP6 and hepatocarcinoma occurrence

se m iq u a n t i t a t iv e a s se ss m en t o f immunohistochemistry The percentage of positive cells was evaluated on the whole surface of the histological section and the intensity of immunoreactivity was estimated at the level of the cell membranes. Two independent visual scores from 0 to 4 were recorded by a trained pathologist using a Leica DMR microscope equipped with a camera. The number of positive cells (P score): 0, no positivity; 1, 50–75%; and 4, >75–100%. Determination of intensity of immunoreactivity (I score): 0, no reactivity; 1, very weak membranous reactivity only seen at magnification ·10; 2, membranous reactivity seen at magnification ·10; 3, moderate reactivity seen at magnification ·2.5; and 4, strong reactivity seen at magnification ·2.5. The IP score was obtained from the additional combination of the two parameters: intensity and percentage of positive cells. This score (from 0 to 8) for each studied sample therefore reflects the average of the pattern of reactivity observed throughout the entire surface of the histological section (surgical resection or biopsy) and was further statistically analysed in order to compare it among the different pathological liver samples.

321

HCC (alcoholism 13; HCV 36), 43 HCC in HL and 31 controls. Also, nine more patients (with a total of 12 liver resections), whose liver ranged from HCC-free cirrhosis to detectable HCC over a period of ‡ 3 years before the emergence of HCC, were studied. Affinity-purified anti-TSAP6 antibodies or control IgGs were tested in liver. With anti-TSAP6 antibodies, a membranous brown signal was found in most hepatocytes from HL (Figure 1A) and HCC-free cirrhosis (with a stronger signal, Figure 1B), whereas these signals were undetectable with control IgGs (Figure 1C,D). Hence, it was considered a TSAP6-specific signal which could be further analysed as two semiquantitative signals, i.e. intensity (I) and percentage of positive cells (P). Initially, considering the values obtained from the calculation of I or P scores, these two parameters were strongly linked (P < 2.10)16, Pearson’s v2 test). Moreover, the IP score obtained from the additional combination of the two parameters I + P was strongly linked to I and P, respectively (Cramer’s value = 0.86 for IP versus I and 0.83 for IP versus P). Therefore, for all comparative studies of TSAP6 expression carried out on hepatic sections showing different pathological conditions, only the IP score was used.

sta ti st ic al to o l s Pearson’s v2 was used to test for independence. If expected frequencies are too low, a better approximation can be obtained by applying the Yates’ correction for continuity. Based on the v2 statistic, the Cramer’s value is an indication of the extent of the relationship between the two variables. Fisher’s exact test is another way to examine the significance of the association between two variables in a 2 · 2 contingency table, and from the latter, specificity and sensitivity can be defined. Based on specificity (sensitivity), the positive (negative) predictive value is the proportion of patients with positive (negative) test results who are correctly diagnosed. All statistics and graphical outputs were carried out using the R language for statistical computing.19 For all tests, P < 0.05 was considered to be statistically significant.

Results semiquantitative i mmunodetection o f tsap 6 in l i ve r s a mp le s Liver samples were obtained from 42 HCC-free cirrhosis (alcoholism 17; HCV 25), 49 cirrhosis with

tsap 6 e x p r es s io n in p er i t u m o ra l a n d h c c - f re e c i r rh o s i s : a e t i o l o g y a n d tu m ou r - g r a d e e ff ec t s TSAP6 expression in HCC-free cirrhosis was not related to the aetiology (post HCV or post alcohol) of the cirrhosis (P = 0.15). So, for all further comparative studies, score values concerning HCC-free cirrhosis were analysed independently of the aetiological factor. In peritumoral cirrhosis (PTC), the TSAP6 IP score was not related to the aetiology of the cirrhosis (P = 0.31), but analysis according to the associated tumour grade showed an association very close to the level of significance (P = 0.07, Fisher’s exact test). In PTC, TSAP6 IP score and tumour grade were not linked (P = 0.69) when comparing PTC with tumour grade 1 (PTC1) versus PTC with tumour grade 2 (PTC2), but in contrast, for all other comparisons an association was observed between TSAP6 score and the associated tumour grade. Moreover, when considering the median score value of TSAP6, this latter was significantly smaller in PTC1 versus PTC with tumour grade 3 (PTC3), in PTC2 versus PTC3 and in PTC (1 + 2) versus PTC3. All statistical values are shown in Table 1.

! 2009 The Authors. Journal compilation ! 2009 Blackwell Publishing Ltd, Histopathology, 54, 319–327.

322

F Caillot et al.

HL

CIR

B

C

D

Control IgGs

Anti-TSAP6 IgGs

A

Figure 1. Specific TSAP6 immunohistochemistry. A, In healthy liver (HL). B, In cirrhotic tissue (CIR). C, D, Same hepatic sections incubated with the negative control (control purified IgGs).

tsap 6 e x p r e s s i o n i n a tu m o ur o c c u r ri n g i n c i r rh o t i c or he a l t h y l iv e r TSAP6 expression in tumour occurring in a cirrhotic liver does not depend on the aetiology of the latter (P = 0.99) but, in contrast, is linked to tumour grade. This expression strongly decreases with increasing tumour grade (P = 0.02). Statistically, TSAP6 expression in grade 1 was not different from that observed in grade 2 (P = 0.60), but became significant when comparing grade 1 with grade 3, grade 2 with grade 3 or grade (1 + 2) with grade 3. In the same manner, in tumour occurring in healthy liver, TSAP6 expression differed significantly according to the associated tumour grade (1 or 2), and the IP median score was lower in grade 2 tumours compared with grade 1. All detailed data are included in Table 1.

t s a p 6 ex p r e ss i on i n h ea lt h y li v er v er s us hcc-free cirrhosis, versus p tc and v ersus p e r i t u mo r a l he a l t h y l i v e r TSAP6 expression in HCC-free cirrhosis was significantly different to that observed in healthy liver (P = 1.87 · 10)9, Pearson’s v2 test), and the median score for TSAP6 was significantly higher in cirrhotic tissue without HCC (P = 1.15 · 10)12, Fisher’s exact test). Likewise, for PTC, TSAP6 expression was significantly different to that observed in HL, whatever the associated tumour grade. Moreover, the median score was higher compared with that observed in HL. All statistical values are shown in Table 2. In contrast, no significant difference was observed in the median score in PTHL1 (peritumoral healthy liver with associated tumour of grade 1) versus HL

! 2009 The Authors. Journal compilation ! 2009 Blackwell Publishing Ltd, Histopathology, 54, 319–327.

TSAP6 and hepatocarcinoma occurrence

Table 1. TSAP6 expression in peritumoral tissues and tumour nodes

323

Tissue*

n†

TSAP6 median score

PTC1 versus PTC2

18 versus 19

7 versus 7

0.69

NA

PTC1 versus PTC3

18 versus 12

7 versus 8

0.03

0.02

PTC2 versus PTC3

19 versus 12

7 versus 8

0.05

3.38 · 10)3

PTC1+2 versus PTC3

37 versus 12

7 versus 8

0.05

2.43 · 10)3

TC1 versus TC2

18 versus 19

5 versus 3

0.60

NA

TC1 versus TC3

18 versus 12

5 versus 1

0.02

0.13

TC2 versus TC3

19 versus 12

3 versus 1

0.04

Pearson’s v2 test‡

Median comparison§

0.08 )3

TC1+2 versus TC3

37 versus 12

3 versus 1

3.95 · 10

PTHL1 versus PTHL2

19 versus 20

6 versus 7

0.49

NA

THL1 versus THL2

19 versus 20

6 versus 2

0.05

0.02

0.05

*Peritumoral cirrhosis (PTC), tumour nodes occurred in cirrhotic liver (TC), peritumoral healthy liver (PTHL) and tumour nodes occurred in healthy liver (THL). Number indicates the associated tumour grade. †Number of samples for each group. ‡Independence between IP score and tumour grade using Pearson’s v2 test for count data with Yates’ continuity correction. §Fisher’ exact test for count data. This was done when the v2 test was significant (NA, not available).

(P = 0.63), among PTHL2 (peritumoral healthy liver with associated tumour grade 2) versus HL (P = 0.12) or among PTHL (1 + 2) versus HL (P = 0.23). For associated tumours of grade 3, no analysis was made due to the limited number of samples studied (n = 4). So, when a tumour occurs in HL, regardless of the associated tumour grade, the TSAP6 score did not differ from that observed in HL without HCC. TSAP6 expression in PTC was also significantly different to that observed in HCC-free cirrhosis, but this observation applies only to PTC1 (P = 4.30 · 10)3) and PTC2 (P = 4.23 · 10)3), with a lower median score in such PTC compared with that observed in HCC-free cirrhosis: P = 4.18 · 10)3 for PTC1 and P = 6.60 · 10)4 for PTC2. In contrast, no difference was observed for PTC3 (P = 0.73). Thus, TSAP6 expression strongly increased in HCCfree cirrhosis compared with its expression in HL, then decreased in PTC1 or PTC2 (but with a level higher compared with HL) and then again increased among PTC3. Moreover, if we compare TSAP6 expression in PTC versus PTHL, no significant difference was observed whatever the performed comparison: P = 0.13 for PTC1 versus PTHL1, P = 0.89 for PTC2

versus PTHL2 and P = 0.25 for PTC (1 + 2) versus PTHL (1 + 2). All these results are detailed and shown in Figure 2 and Table 2. tsap 6 expression a mong cirrhosis –t u m o u r p a ir e d a n d he a lth y l iv e r–tumour p aired cases Pairs with a tumour of grade 1, 2 or 3 arising in a cirrhotic liver will be respectively named PTC ⁄ T1, PTC ⁄ T2 or PTC ⁄ T3 respectively. Compared with its score in PTC, TSAP6 expression was greatly reduced in tumour whatever the grade: P = 7 · 10)3 for PTC ⁄ T1, 1 · 10)3 for PTC ⁄ T2 and 2 · 10)4 for PTC ⁄ T3 (Figure 3A). Specificity and sensitivity of the test increased from grade 1 to grade 3 to reach 100% for grade 3. Typical patterns of reactivity obtained according to the different physiopathological conditions are shown in Figure 3b, with strong membranous reactivity for the cirrhotic peritumoral liver (see Figure 3a,c,e), which becomes almost undetectable in the tumour (see Figure 3b,d,f), with an increasingly significant decrease as the grade increases.

! 2009 The Authors. Journal compilation ! 2009 Blackwell Publishing Ltd, Histopathology, 54, 319–327.

324

F Caillot et al.

Tissue*

n†

TSAP6 median score

HL versus CIR

31 versus 42

6 versus 8

1.87 · 10)9

1.15 ·10

HL versus PTC1

31 versus 18

6 versus 7

5.96 · 10)3

8.08 · 10)4

HL versus PTC2

31 versus 19

6 versus 7

2.26 · 10)2

1.33 · 10)3

HL versus PTC3

31 versus 12

6 versus 8

6.45 · 10)6

8.21 · 10)6

HL versus PTC1 + 2

31 versus 37

6 versus 7

1.50 · 10)3

3.34 · 10)4

HL versus PTHL1

31 versus 19

6 versus 6

0.63

NA

HL versus PTHL2

31 versus 20

6 versus 7

0.12

Pearson’s v2 test‡

Median comparison§ )12

NA )3

4.18 · 10)3

CIR versus PTC1

42 versus 18

8 versus 7

4.30 · 10

CIR versus PTC2

42 versus 19

8 versus 7

4.23 · 10)3

6.60 · 10)4

CIR versus PTC3

42 versus 12

8 versus 8

0.73

NA

CIR versus PTC1 + 2

42 versus 37

2.09 · 10

8 versus 7

Table 2. TSAP6 expression in healthy liver, hepatocellular carcinoma (HCC)free cirrhosis and peritumoral tissue (cirrhotic or healthy)

)3

1.38 · 10)4

*Healthy liver (HL), HCC-free cirrhosis (CIR) and cirrhotic peritumoral tissue (PTC) or healthy peritumoral tissue (PTHL). If necessary, number indicates the associated tumour grade. †Number of samples for each group. ‡Independence between liver tissue origin and IP score using Pearson’s v2 test for count data wit Yates’ continuity correction. §Fisher’s exact test for count data. This was done when the v2 test was significant (NA, not available).

NS

*** 8

IP

HL

**

*

CIR

NS

***

PTC1

PTC2

PTC3

PTHL1

PTHL2

n = 19

n = 12

n =19

n = 20

7 6 5 4 3

**

2 n = 31

n = 42

** n = 18

NS

Figure 2. TSAP6 expression in healthy liver (HL), hepatocellular carcinoma-free cirrhosis (CIR), peritumoral cirrhosis (PTC) and peritumoral healthy liver (PTHL). PTC1 ⁄ PTC2 ⁄ PTC3 were defined as peritumoral cirrhosis with, respectively, a grade 1 ⁄ 2 ⁄ 3 associated tumour. PTHL1 ⁄ PTHL2 were defined as peritumoral healthy liver with a grade 1 ⁄ 2 associated tumour. P-values, which were computed using v2 test for count data with the Yates’ continuity correction, were: *P < 5 · 10)2; **P < 1 · 10)2; ***P < 1 · 10)3. Box plots depict groups of numerical data through their five-number summaries: the smallest observation (the extreme of the lower whisker), the lower quartile (the lower hinge), the median (the horizontal thick line), the upper quartile (the upper hinge) and the largest observation (the extreme of the upper whisker). Box plots also indicate which observations, if any, might be considered as outliers (open circles). Note that the vertical lines (the whiskers) extend to at most 1.5 times the box height (the interquartile range) from either or both ends of the box. ! 2009 The Authors. Journal compilation ! 2009 Blackwell Publishing Ltd, Histopathology, 54, 319–327.

TSAP6 and hepatocarcinoma occurrence

A 8

* IP

PTC1

** PTC2

T1

325

*** T2

PTC3

T3

7 6 5 4 3 2 1 0 B

n = 18

n = 12

n = 19 PTC3

PTC2

PTC1

T1

e

c

a T2

b

T3

d

f

Figure 3. TSAP6 expression among cirrhosis–tumour paired PTC ⁄ T1, PTC ⁄ T2, PTC ⁄ T3. A, A box plot graphic representation is used (see legend to Figure 2) where significant differences between median scores in these three conditions were computed using Fisher’s exact test and were: *P < 1 · 10)2; **P < 1 · 10)4; ***P < 1 · 10)6. B, TSAP6 immunohistochemistry of hepatic sections corresponding to the abovementioned pathological conditions.

In the HL–tumour paired cases, no significant difference was observed in TSAP6 score between the HL and the tumour of grade 1, contrasting with a significant difference (P = 6.62 · 10)4) for a tumour of grade 2. tsap 6 , a t ra n sit ion mar ke r f r om c ir r ho sis to hepatocellular c arcinoma? For nine patients, from 12 biopsies performed in the follow-up of their cirrhosis, we evaluated the

TSAP6 score in the cirrhotic tissue at least 3 years before the emergence of HCC. In this context, we observed that this score was significantly different (P = 6.69 · 10)8) from that obtained in cirrhotic tissues that, over the same period, had not developed HCC. This score was significantly reduced (P = 1.47 · 10)7) in cirrhosis which 3 years later progressed to HCC. The negative predictive value of this test was 100% with a positive predictive value of 63%. Results are detailed in Table 3.

! 2009 The Authors. Journal compilation ! 2009 Blackwell Publishing Ltd, Histopathology, 54, 319–327.

326

F Caillot et al.

Table 3. TSAP6 expression in cirrhotic liver 3 years before the emergence or not of a hepatocellular carcinoma CIR before HCC (12)* TSAP6 median score

HCC-free CIR (42)*

5.5

8

IP < 8

12

7

IP = 8

0

35

v2 test†

6.69 · 10)8

Median comparison‡

1.47 · 10)7

Sensitivity

1.00

Specificity

0.83

PPV§

0.63

NPV§

1.00

*Cirrhosis before emergence of an HCC (CIR before HCC) or not (HCC-free cirrhosis). The number in parentheses is the number of samples for each group. †Independence between IP score and HCC outcome using Pearson’s v2 test for count data with Yates’ continuity correction. ‡Fisher’s exact test for count data. §Positive predictive value (PPV) or negative predictive value (NPV).

Discussion We have recently observed that TSAP6 mRNA in significantly decreased in hepatic tumour tissue compared with that observed in cirrhotic peritumoral tissue, the latter itself being significantly decreased compared with that in healthy liver.7 In determining the level of TSAP6 mRNA in cirrhotic livers without HCC, our results suggest a progressive decrease in the level of this gene from healthy liver to HCC, but with overexpression in HCC-free cirrhosis. In order to detect a correlation between protein and mRNA levels, which often leads to contradictions,16,17 we have quantified the TSAP6 protein level by immunohistochemistry. Our results have shown a strong and significant decrease in TSAP6 protein expression in tumour according to its level of differentiation, with the lowest values observed when the tumour is moderately or poorly differentiated. In contrast, in HCC-free cirrhosis and independently of aetiological factors, TSAP6 protein expression is significantly increased compared to its expression in healthy liver. Thus, such expression

is linked to the condition of the liver, healthy or cirrhotic, with or without the occurrence of an HCC and in the latter case according to tumour grade. These new results at the protein level concur with those previously described at the mRNA. The fact that TSAP6, both at the mRNA and protein level, was strongly down-regulated in tumours should be taken into account when using it as a new marker of hepatic carcinogenesis. Moreover, follow-up of postHCV cirrhosis carried out in nine patients before the occurrence of an HCC has shown modifications in TSAP6 expression that was already apparent in cirrhotic tissue at least 3 years before neoplastic transformation. Because there is an urgent need for better identification of tissue markers to characterize precancerous lesions, this observation should be thoroughly tested in a large population of patients with HCV cirrhosis in order to confirm this modification of the TSAP6 score from HCC-free cirrhosis toward a cirrhotic liver with an HCC. In this context, we believe that a patient with such a modified TSAP6 level should be followed by regular imaging studies. Since HCC is the third leading cause of cancer-related death in the world with an increasing incidence in Europe, there is a real challenge to identify HCC and its precursor lesions earlier. With this objective, a specific molecular signature to discriminate dysplastic nodules from early HCC in HCV cirrhosis has recently been identified and described.20 In the cirrhotic tissue of patients, with biopsies periodically performed for surveillance purposes, decreased expression of TSAP6 membranous protein could reflect a decrease in the apoptotic process and could be interpreted as a warning signal of the emergence of a cancerous lesion. The role of pHyde (rat orthologue of TSAP6) as a tumour suppressor gene product involved in the apoptotic process has been shown in rat during prostatic cancer.10,15 Experimental studies in vivo and in vitro models have shown that pHyde expression is significantly decreased in poorly differentiated prostatic carcinomas compared with that in well and intermediate ones.14 In contrast, no difference was observed prostatic benign benign hyperplasia, non-treated primary carcinomas and hormoneresistant carcinomas. Taken together, these studies suggest that normal or overexpression of pHyde might induce inhibition of the proliferative process involved in prostatic cancer, which could to some extent result in the induction of the apoptotic process. The present study has shown, for the first time in human liver, an important role for TSAP6 in controlling regulation of these above-mentioned processes, both of which are involved in hepatic carcinogenesis

! 2009 The Authors. Journal compilation ! 2009 Blackwell Publishing Ltd, Histopathology, 54, 319–327.

TSAP6 and hepatocarcinoma occurrence

and provides the basis for further studies. For example, in an experimental diethyl-nitrosamine-induced HCC,21 the induction of TSAP6 expression at various times during the course of HCC development could allow one to evaluate the capacity of TSAP6 to slow down or stop the carcinogenesis process and could allow the development of new therapeutic approaches to HCC.

Competing interests F.C. is a recipient of a fellowship from ARC. This work was supported in part by grants from ARC, IREB and Ligue contre le Cancer to J.-P.S.

Acknowledgement We thank Duclos Ce´lia for her excellent technical assistance in immunohistochemistry procedures.

References 1. Kyrlagkitsis I, Portmann B, Smith H, O’Grady J, Cramp ME. Liver histology and progression of fibrosis in individuals with chronic hepatitis C and persistently normal ALT. Am. J. Gastroenterol. 2003; 98; 1588–1593. 2. Ryder SD, Irving WL, Jones DA, Neal KR, Underwood JC. Progression of hepatic fibrosis in patients with hepatitis C: a prospective repeat liver biopsy study. Gut 2004; 53; 451–455. 3. Block TM, Mehta AS, Fimmel CJ, Jordan R. Molecular viral oncology of hepatocellular carcinoma. Oncogene 2003; 22; 5093–5107. 4. El-Serag HB. Hepatocellular carcinoma: recent trends in the United States. Gastroenterology 2004; 1; S27–S34. 5. Morgan TR, Mandayam S, Jamal MM. Alcohol and hepatocellular carcinoma. Gastroenterology 2004; 1; S87–S96. 6. Stickel F, Schuppan D, Hahn EG, Seitz HK. Cocarcinogenic effects of alcohol in hepatocarcinogenesis. Gut 2002; 51; 132–139. 7. Coulouarn C, Derambure C, Lefebvre G et al. Global gene repression in hepatocellular carcinoma and fetal liver, and suppression of dudulin-2 mRNA as a possible marker for the cirrhosis-to-tumor transition. J. Hepatol. 2005; 42; 860–869.

327

8. Yu X, Harris SL, Levine AJ. The regulation of exosome secretion: a novel function of the p53 protein. Cancer Res. 2006; 66; 4795– 4801. 9. Amzallag N, Passer BJ, Allanic D et al. TSAP6 facilitates the secretion of translationally controlled tumor protein ⁄ histaminereleasing factor via a nonclassical pathway. J. Biol. Chem. 2004; 279; 46104–46112. 10. Zhang X, Steiner MS, Rinaldy A, Lu Y. Apoptosis induction in prostate cancer cells by a novel gene product, pHyde, involves caspase-3. Oncogene 2001; 20; 5982–5990. 11. Amson RB, Nemani M, Roperch JP et al. Isolation of 10 differentially expressed cDNAs in p53-induced apoptosis: activation of the vertebrate homologue of the drosophila seven in absentia gene. Proc. Natl Acad. Sci. USA 1996; 93; 3953–3957. 12. Passer BJ, Nancy-Portebois V, Amzallag N et al. The p53inducible TSAP6 gene product regulates apoptosis and the cell cycle and interacts with Nix and the Myt1 kinase. Proc. Natl Acad. Sci. USA 2003; 100; 2284–2289. 13. Roger L, Gadea G, Roux P. Control of cell migration: a tumour suppressor function for p53? Biol. Cell 2006; 98; 141–152. 14. Porkka KP, Nupponen NN, Tammela TL, Vessella RL, Visakorpi T. Human pHyde is not a classical tumor suppressor gene in prostate cancer. Int. J. Cancer 2003; 106; 729–735. 15. Steiner MS, Zhang X, Wang Y, Lu Y. Growth inhibition of prostate cancer by an adenovirus expressing a novel tumor suppressor gene, pHyde. Cancer Res. 2000; 60; 4419–4425. 16. Unwin RD, Whetton AD. Systematic proteome and transcriptome analysis of stem cell populations. Cell Cycle 2006; 5; 1587–1591. 17. Greenbaum D, Colangelo C, Williams K, Gerstein M. Comparing protein abundance and mRNA expression levels on a genomic scale. Genome Biol. 2003; 4; 117. 18. Hamilton SR, Aaltonene LA. Classification of Tumor Pathology and Genetics of Tumors of the Digestive System. World Health Organization Classification of Tumors, 2000. 19. R Development Core Team. R: A Language and Environment for Statistical Computing. In R Fondation for Statistical Computing. Vienna, Austria: URL http://www.R-project.org, 2005. 20. Llovet JM, Chen Y, Wurmbach E et al. A molecular signature to discriminate dysplastic nodules from early hepatocellular carcinoma in HCV cirrhosis. Gastroenterology 2006; 131; 1758–1767. 21. Shiota G, Oyama K, Noguchi N, Takano T, Ito H, Kawasaki H. Oral administration of cholic acid promotes growth of liver tumors initiated by diethylnitrosamine in rats. Int J Oncol 1999; 15; 259–265.

! 2009 The Authors. Journal compilation ! 2009 Blackwell Publishing Ltd, Histopathology, 54, 319–327.