Inhibitor of differentiation/DNA binding protein 1 (Id-1) plays a pivotal role in the regulation of cell proliferation and carcinogenesis via inhibiting basic helix-loop-helix (HLH) transcription factors. Recently, Id-1 was found to repress p16 in tumorous tissue specimens including hepatocellular carcinoma (HCC), but its relevance in precancerous liver tissues is unknown.
Id-1 expression in the liver tissue specimens of 112 patients with cirrhosis without HCC was studied by immunohistochemical analysis. Correlations were investigated between Id-1 expression and clinicopathologic features, the status of p16, and the risk of HCC occurrence.
A high expression of Id-1 was observed in 42 patients (38%). The level of Id-1 expression was not associated with clinical standard parameters or the status of p16 in cirrhotic tissue specimens. The cumulative incidence of HCC development was significantly higher in a group of patients with high Id-1 expression (P = 0.0008). Multivariate analysis revealed that increased Id-1 expression is an independent significant factor for the risk of HCC development in patients with cirrhosis (relative risk = 2.75, P = 0.003).
Hepatocellular carcinoma (HCC) is one of the most frequently found malignancies worldwide, which mainly arises from cirrhosis caused by chronic hepatitis B virus/hepatitis C virus infection (HBV/HCV), alcohol abuse, and hemochromatosis. Many clinicians have suggested that cell proliferation status may be closely involved in the mechanistic nature of hepatocarcinogenesis. Follow-up studies using a bromodeoxyuridine assay, silver staining of nucleolar organizer region proteins (AgNOR), and immunostaining for Ki-67 and proliferating cell nuclear antigen (PCNA) have shown that a high rate of hepatocellular proliferation was closely associated with the increased risk of HCC development in patients with cirrhosis.1–8 This suggests that assessment of cell proliferation in liver biopsy samples might help to determine patients with cirrhosis who require careful surveillance of HCC.
Because DNA is prone to exposure of mutagens during cell replication, the relation between increased hepatocellular proliferation and carcinogenesis is plausible. Nevertheless, the regulatory mechanism of cell proliferation in hepatocarcinogenesis is unknown. Several clinical studies have reported no significant relation between serum levels of alanine aminotransferase (ALT) and aspartate aminotransferase and PCNA labeling index (LI) in liver tissue specimens from patients with cirrhosis,4, 7, 8 suggesting that inflammatory stimuli in the diseased liver cannot be fully attributable to the increased proliferation of hepatocytes.
Recent studies have shown that inhibitor of differentiation/DNA binding (Id) proteins, which act as dominant negative inhibitors of basic helix-loop-helix (HLH) transcription factors,9 play a pivotal role in the regulation of cellular proliferation.10 Among the members of the Id family, Id-1 is overexpressed in many types of tumor tissue specimens and may play a crucial role in carcinogenesis.11, 12 Because Id-1 is overexpressed in tissue specimens from dysplastic lesions in the pancreas,13 it is highly likely that Id-1 is implicated in the early step of carcinogenesis. More recently, Id-1 has been found to oppose an Ets-mediated activation of tumor suppressor p16,14 which is inactivated in approximately one-half the patients with HCC.15–17 The biologic role of Id-1 in hepatocarcinogenesis is intriguing, but no studies have examined the correlation between the status of Id-1 and HCC development in patients with cirrhosis.
To investigate the role of Id-1 in an early step of hepatocarcinogenesis, we examined the status of Id-1 expression in patients with cirrhosis by immunohistochemical staining and compared expression with clinical variables. Moreover, to determine whether overexpressed Id-1 represses p16 in cirrhosis, the correlation of Id-1 and p16 expression was investigated by immunohistochemical staining and methylation-specific polymerase chain reaction (PCR). Finally, to investigate the clinical significance of Id-1 in cirrhosis with regard to hepatocarcinogenesis, the relation between the status of Id-1 and the risk of HCC occurrence was analyzed by a retrospective follow-up study.
MATERIALS AND METHODS
The records of 112 patients who were histologically diagnosed with cirrhosis were retrieved from our files of liver biopsy performed over the past 20 years at Niigata University Graduate School of Medical and Dental Science (Niigata City, Japan). The cohort comprised 36 patients with HBV-associated cirrhosis, 54 patients with HCV-associated cirrhosis, and 22 patients with alcohol-induced cirrhosis. There were 91 men and 21 women with a mean age of 63 ± 9 years (range, 43–87 years). At the time of biopsy, none of the patients showed evidence of dysplastic nodules or HCC in the livers as assessed by ultrasonography (US), computed tomography (CT) scan, or magnetic resonance imaging (MRI) scan. Patients who had any episodes of HCC treatment before the time of biopsy were excluded from analysis, as were patients who received interferon therapy after the biopsy was performed.
As a rule, patients with cirrhosis had been regularly followed up at 1–3-month intervals with monitoring of serum α-fetoprotein (AFP) levels, and every 3–6 months with US, CT, or MRI scans. When the serum levels of AFP or characteristics of the US/CT scan/MRI scan imaging pattern changed during the follow-up, the occurrence of HCC was evaluated by US-guided biopsy or CT scan during arterioportography. All the patients in the study were followed for ≥ 6 years or until the development of HCC, and the follow-up period was set as the time from the biopsy to HCC occurrence or to the last observation. Informed consent was obtained from all patients to participate in the study, which was approved by the institutional guidelines of Niigata University Graduate School of Medicine and Dental Science.
All tissue samples were immediately fixed in neutral-buffered formalin for 3 days after the biopsy procedure was performed, and embedded in paraffin. For immunohistochemical staining of Id-1, deparaffinized thin-sliced sections were microwaved in 10 mM citrate buffer (pH 6.0) for antigen retrieval. They were then treated with 1% hydrogen peroxidase for 30 minutes to block endogenous peroxidase, followed by a 10% normal goat serum block. Tissue sections were incubated overnight at 4 °C with anti–Id-1 rabbit polyclonal antibody (1:50) (C-20; Santa Cruz Biotechnology, Santa Cruz, CA), which is known to successfully detect Id-1 in paraffin-embedded human tissue sections and not cross-react with Id-2, Id-3, or Id-4.13 Color development was carried out using the vector Elite ABC kit (Vector Laboratories, Burlingame, CA) with 3,3′-diaminobenzidine tetrahydrochloride (Sigma, St. Louis, MO) and the reacted sections were counterstained with hematoxylin. Preabsorption of the primary antibody with specific blocking peptide of Id-1 (Santa Cruz) was verified to abolish the immunoreactivity. Substitution of the primary antibody with normal immunoglobulin of the same species was used for negative controls. Tissue samples from four normal livers served as normal controls. Smooth muscle cells of vessels known to express Id-118 were regarded as an internal positive control. Cytoplasmic staining of the cells was considered positive because Id-1 lacks a nuclear localization signal.18–20 To confirm the reproduction of the staining intensity, each tissue sample was stained in duplicate, and immunostaining was assessed by counting 200 cells in the region of interest by 2 observers without knowledge of the clinical characteristics of the samples. Scoring of immunostaining for Id-1 expression was assessed by the percentage of positive cells and the staining intensity, which was based on experience from previous studies.19, 20 The percentage of positive cells was divided into 4 groups: 2 points, 11–50 % of the positive cells; 3 points, 51–80% of the positive cells; and 4 points, > 80% of the positive cells. The intensity of the immunostaining signals was categorized and adjusted to the internal positive control as follows: 1 point, weak immunoreactivity; 2 points, moderate immunoreactivity; and 3 points, strong immunoreactivity. The sum of the points for the percentage score and the intensity score were calculated, and specimens were categorized into 4 groups according to the total score: negative, < 10% of the immunoreactive cells regardless of the intensity; weak expressor, 2–3 points; moderate expressor, 4–5 points; and strong expressor, 6–7 points. We found that this scoring method minimized subjective grading of the immunostaining.
For evaluating the cell proliferation status of the tissue samples, the LI of PCNA was assessed by using antihuman PCNA monoclonal antibody (1:50) (PC10; Dako A/S, Carpinteria, CA). The percentage of immunoreactive cells for PCNA was calculated by counting ≥ 200 cells at high magnification (× 400). For immunostaining of p16, deparaffinized tissue sections were reacted with a rabbit polyclonal anti-p16 antibody (1:400) directed against the entire region of the human p16 protein (PharMingen, San Diego, CA) overnight at 4 °C. The following procedure of p16 immunostaining was the same as that for Id-1. According to previous reports,15–17 nuclear staining is considered to be a positive reaction. The degree of staining was graded as follows: weak expression, negative staining or < 50% of the immunoreactive cells; strong expression, positive staining in > 51% of the cells.
Methylation-Specific Polymerase Chain Reaction
Methylation-specific PCR (MSP) was performed to investigate the methylation status of the p16 gene in cirrhotic livers. Genomic DNA was extracted from deparaffinized tissue sections from the biopsy liver sample.16 DNA was modified with sodium bisulfite, and aliquots (50 ng) were amplified by PCR using primers specific for the unmethylated (5′-TTATTAGAGGGTGGGGTGGATTGT-3′, 5′-CAACCCCAAACCACAACCATAA-3′) or methylated (5′-TTATTAGAGGGTGGGGCGGATCGC-3′, 5′-GACCCCGAACCGCGACCGTAA-3′) p16 gene as previously described.21 The PCR products (10 μL) were resolved by electrophoresis on agarose gels containing ethidium bromide and visualized with ultraviolet illumination.
Because the number of the patients in each group was disproportionate, the patients with negative and weak Id-1 expressors were assembled into a group of low Id-1 expressors, and moderate and strong Id-1 expressors were put into a group of high Id-1 expressors for statistical analysis. The association between the status of Id-1 and clinicopathologic parameters was examined using chi-square analysis. When appropriate, a Mann–Whitney test was used to test for statistical differences between the groups. Prospective curves of the HCC-free period in individuals were calculated by the Kaplan–Meier method22 and the statistical significance between groups was determined by a log-rank test. The Cox proportional hazards model on Stat View 5.0 software (SAS Institute, Cary, NC) was used to evaluate the possible correlation of Id-1 status with the risk of HCC development in each of the patient groups. All reported P values are 2 sided and the data were considered statistically significant when P < 0.05.
Immunohistochemical Analysis of Id-1
It had previously been verified that the immunostaining pattern of the Id-1 antibody used in the current study corresponds to the gene expression in human tissue specimens.23 In normal control liver tissue specimens, immunostaining for Id-1 was undetectable or weakly expressed in a few hepatocytes (Fig. 1A). However, many patients with cirrhosis showed positive expression of Id-1 with diverse staining intensity. In our study sample, Id-1 staining was negative in 5 patients (4%), weak in 65 patients (58%), moderate in 32 patients (29%), and strong in 10 patients (9%) (Fig. 1B–F). When the patients were divided into 2 groups of low Id-1 expressor (negative or weak Id-1 expressors; n = 70) and high Id-1 expressor (moderate or strong Id-1 expressors; n = 42), there was no significant difference in clinical variables such as age, gender, blood platelet count, and serum ALT and AFP levels (Table 1). The mean PCNA LI tended to be higher in the group of high Id-1 expressors than in low Id-1 expressors (4.0 ± 2.2% vs. 3.1 ± 1.4%, P = 0.071; Mann–Whitney test) (Table 1).
Table 1. Comparison of Clinicopathologic Findings of Cirrhosis between High and Low Id-1 Expression
Id-1: inhibitor of differentiation/DNA binding protein 1; yrs: years; HBsAG: hepatitis B surface antigen; HCV: hepatitis C virus; ALT: alanine aminotransferase; AFP: α-fetoprotein; PCNA: proliferating cell nuclear antigen; LI: labeling index; HCC: hepatocellular carcinoma.
aP value: Id-1 low vs. high expressors.
No. of patients
Mean age (yrs)
63 ± 10
63 ± 9
63 ± 9
63 ± 9
61 ± 9
62 ± 9
ALT level (IU/L)
71 ± 21
84 ± 47
83 ± 46
76 ± 34
84 ± 31
78 ± 33
Platelet count (× 104/mL)
10.1 ± 4.5
9.3 ± 3.3
9.4 ± 3.4
8.4 ± 3.8
8.4 ± 5.0
8.4 ± 4.1
AFP level (ng/mL)
35 ± 20
26 ± 27
27 ± 27
18 ± 11
23 ± 25
19 ± 16
PCNA LI (%)
2.0 ± 0.7
3.2 ± 1.4
3.1 ± 1.4
3.8 ± 2.2
4.8 ± 2.3
4.0 ± 2.2
HCC occurrence during 5 yrs
Relation between the Status of p16 and Id-1
The relation between p16 expression and Id-1 status was evaluated by immunohistochemical analysis using serial sections of the same tissue sample. Three (3%) of the 112 patients showed weak immunohistochemical expression for p16, but all remaining patients showed strong p16 expression irrespective of the status of Id-1 (Fig. 2A). The Mann–Whitney U test showed no relation between the level of p16 and Id-1 expression in patients with cirrhosis (P = 0.650). MSP detected hypermethylation of the p16 gene in 4 of the 112 patients with cirrhosis (Fig. 2B). Of these, immunostaining for p16 was weak in 3 patients and strong in 1 patient, and the methylation status was strongly correlated with the immunostaining intensity for p16 protein (chi-square analysis: P < 0.0001).
Correlation of Id-1 Status with Hepatocellular Carcinoma Development in Patients with Cirrhosis
The median follow-up period of patients with cirrhosis was 80 months (range, 18–121 months). In 112 patients with cirrhosis, 21 patients (19%) developed HCC within 5 years. Of these, 7 were weak, 10 were moderate, and 4 were strong expressors (Table 1). The 5-year HCC-free rates as assessed by the Kaplan–Meier method were 87% and 63% in low and high Id-1 expressors, respectively, and the cumulative incidence of HCC development was significantly higher in high Id-1 expressors than in low Id-1 expressors (P = 0.0008) (Fig. 3). Multivariate analysis of HCC-free survival using the Cox proportional hazard model indicated that high Id-1 expressors were recognized as an independent significant factor for the risk of HCC development in patients with cirrhosis (relative risk [RR] = 2.75, P = 0.003). The RRs of most of the clinical parameters were > 1.00, but were not statistically significant (Table 2).
Table 2. Multivariate Analysis of Id-1 Status and Clinical Variables for HCC Development
In the current study, we examined Id-1 expression in 112 patients with cirrhosis and demonstrated that Id-1 expression is preferentially increased in patients with a high risk of HCC development. Immunohistochemical analysis showed that immunostaining for Id-1 was negative in 5 patients (4%) and positive in 107 patients (96%). The staining intensity was diverse among the 107 patients, with weak expression in 65 patients, moderate expression in 32 patients, and strong expression in 10 patients, possibly because the time of liver biopsy was heterogeneous among the patients in terms of the process of HCC development. However, we detected no significant differences in standard clinical parameters (age, serum ALT level, platelet count, and AFP levels) between the low and high Id-1 expressor groups.
We found that the mean PCNA LI of hepatocytes tended to be higher in the high Id-1 expressor group than in low Id-1 expressors, although statistical significance was not detected. Previous studies of AgNOR, Ki-67, and PCNA immunostaining have reported that increased hepatocellular proliferation was associated with the malignant property of cirrhotic livers,1–8 supporting the idea that cell cycle progression may be closely involved in the early step of hepatocarcinogenesis. Unfortunately, however, there has been little consensus regarding the clinical significance of cell cycle regulators in precancerous liver tissue specimens. Kang et al.24 reported that weak p53 expression was detected by immunohistochemical staining in 4 of 26 dysplastic nodules in patients with cirrhosis, whereas Choi et al.25 reported that p53, cyclin D1, and cyclin E proteins were not expressed in dysplastic nodules in either cirrhotic or normal livers. Wagayama et al.26 recently reported that the cumulative incidence of HCC was significantly higher in patients with cirrhosis with increased p21 expression. However, in that study, only 25 patients were evaluated. Therefore, another study comprising more patients may be required. Because cirrhosis is in a condition of perpetual inflammation, which consequently affects the turnover from cell death to renewal, obtaining accurate evidence of the relation between cell cycle regulators and carcinogenesis may be difficult.
Recently, several studies have shown that Id-1, a dominant negative regulator of basic HLH transcription factors, plays a crucial role in the mechanistic nature of increased cell proliferation during carcinogenesis.9–12 Currently, Id-1 has been found to initiate DNA synthesis by inducing cell cycle G1-S transition27 and extending the cell lifespan through inactivation of the RB/p16 pathway.28–30 Because Id-1 antagonizes the expression of differentiation-associated genes (e.g., p21, p15, and p16) via inhibiting DNA binding of bHLH or other activated proteins at the promoter regions,31, 32 overexpressed Id-1 may readily affect cell cycle machinery. Most importantly, the expression level of Id-1 has been indicated as a prognostic marker in several types of early-stage cancers. Schindl et al.33 reported that increased levels of Id-1 expression significantly influence prognosis in patients with cervical carcinoma Stage 1b. In addition, Schoppmann et al.34 reported that overexpression of Id-1 represented a strong independent prognostic marker in patients with lymph node-negative breast carcinoma. The biologic significance of Id-1 in the early step of carcinogenesis is supported by the findings of Maruyama et al.,13 who showed that Id-1 expression is significantly elevated in dysplastic and atypical papillary ducts in the pancreas as well as in cancer cells. In the current study, we found a significant increase in the cumulative incidence of HCC in the group of high Id-1 expressors during the long follow-up period (P = 0.0008). We also showed that Id-1 is an independent significant factor for the risk of HCC (RR = 2.75, P = 0.003), suggesting that Id-1 may become a useful marker for ascertaining cirrhosis patients who are more likely to develop HCC in the near term.
Investigating the signaling pathway elicited by Id-1 is an intriguing possibility to gain insight into the mechanism of early hepatocarcinogenesis. One of the most plausible candidates is p16, which is repressed by Id-128–30 and is reduced in approximately one-half of the patients with HCC.15–17 To investigate whether p16 is repressed by Id-1 in precancerous liver tissue specimens, we evaluated the expression of p16 and compared it with the status of Id-1 in each of the patients with cirrhosis. Immunohistochemical analysis showed that 109 of the 112 patients in the current study had expression of p16 in > 50% of the hepatocytes, irrespective of the levels of Id-1 expression. There was no significant relation between the level of p16 and Id-1 expression, indicating that transcriptional repression of p16 by Id-1 in cirrhotic livers is unlikely. Kaneto et al.35 reported that methylation of the p16 gene promoter was detected in 5 of 17 patients with cirrhosis and in 4 of 17 patients with chronic hepatitis with HBV/HCV infections, suggesting that the p16 gene is methylated in some patients with chronic liver injury as well as in some patients with HCC. In our study, although the methylated p16 gene was detected in only 4 of 112 patients with cirrhosis, the methylation status of p16 showed a strong correlation with reduced immunohistochemical staining (P < 0.0001). This evidence strongly indicates that functional loss of p16 in cirrhotic livers is mainly caused by DNA methylation, not by Id-1–mediated signaling. However, our finding is distinct from the recent study by Lee et al.,36 who showed a close inverse relation between the levels of Id-1 and p16 mRNA in human HCC tissue specimens. We surmise that Id-1 plays a critical role in early hepatocarcinogenesis, independent from the p16/RB signaling pathway, and in the later stage it promotes cancer aggressiveness via repression of p16 expression in HCC.
Cirrhosis is a terminal state of chronic liver injury, and many researchers have cautioned that the long-term inflammation per se may be a risk factor for the development of HCC. Recently, NF-κB has been found to link the mechanistic nature between inflammation and tumorigenesis in murine colitis-associated cancer and cholangitis-associated hepatoma.37, 38 This finding seems intriguing, because Id-1 is 1 of the main upstream regulators of NF-κB activity.39 To address whether the Id-1/NF- κB signaling pathway plays a crucial role in the molecular link between chronic hepatic injury and HCC, we are now investigating the status of NF-κB in cirrhotic livers with overexpressed Id-1.
To date, the prognosis of patients with HCC is still poor, and identifying useful molecular markers for predicting HCC occurrence in cirrhosis is needed. However, there have been no established indicators responsible for future HCC occurrence in cirrhotic livers. Our study suggests that Id-1 may be a significant molecular marker for the risk of HCC development in precancerous liver tissue specimens. For more precisely determining individuals with a high risk of hepatocarcinogenesis, repeat liver biopsy to determine changes in Id-1 expression as well as to investigate the biologic role of Id-1 in cirrhosis appears important.