High frequency of p16INK4A gene alterations in ... - Nature

Oncogene (1999) 18, 789 ± 795 ã 1999 Stockton Press All rights reserved 0950 ± 9232/99 $12.00 http://www.stockton-press.co.uk/onc

High frequency of p16INK4A gene alterations in hepatocellular carcinoma Choong Tsek Liew1, Hiu-Ming Li1, Kwok-Wai Lo1, Chon Kar Leow2, John YH Chan3, Lin Yee Hin4, Wan Yee Lau2, Paul Bo San Lai2, Boon Kian Lim1, Jin Huang1, Wai Tong Leung3, Shan Wu1 and Joseph Chuen Kwun Lee1 Departments of 1Anatomical and Cellular Pathology, 2Surgery, 3Clinical Oncology and 4Obstetric and Gynecology, Faculty of Medicine, The Chinese University of Hong Kong, Prince of Wales Hospital, Shatin, Hong Kong

The tumor suppressor gene p16 (CDKN2/MTS-1/ INK4A) is an important component of the cell cycle and inactivation of the gene has been found in a variety of human cancers. In order to investigate the role of p16 gene in the tumorigenesis of hepatocellular carcinoma (HCC), 48 cases of HCC were analysed for p16 alterations by: methylation-speci®c PCR (MSP) to determine the methylation status of the p16 promoter region; comparative multiplex PCR to detect homozygous deletion; PCR ± SSCP and DNA sequencing analysis to identify mutation of the p16 gene. We found high frequency of hypermethylation of the 5' CpG island of the p16 gene in 30 of 48 cases (62.5%) of HCC tumors. Moreover, homozygous deletion at p16 region were present in ®ve of 48 cases (10.4%); and missense mutation were detected in three of 48 cases (6.3%). The overall frequency of p16 alterations, including homozygous deletion, mutation and hypermethylation, in HCC tumors was 70.8% (34 of 48 cases). These ®ndings suggest that: (a) the inactivation of the p16 is a frequent event in HCC; (b) the p16 gene is inactivated by multiple mechanisms including homozygous deletion, promoter hypermethylation and point mutation; (c) the most common somatic alteration of the p16 gene in HCC is de novo hypermethylation of the 5' CpG island; and (d) in contrast to other studies, high frequency of genomic alterations are not uncommon in the 9p21 of the p16 gene. Our results strongly suggest that the p16 gene plays an important role in the pathogenesis of HCC. Keywords: hepatocellular carcinoma; p16; hypermethylation; homozygous deletion; mutation

Introduction The p16 tumor suppressor gene (TSG) is located on chromosome 9p21 and it is one of the most frequently altered genes observed in various human neoplasms (Kamb et al., 1994; Nobori et al., 1994; Okamoto et al., 1994). The p16 gene is believed to encode a negative regulatory protein which prevents the cell cycle progression from G1 to S phase by inhibiting the catalytic activity of the CDK4 or CDK6/cyclin D complex and subsequently Rb protein phosphorylation (Serrano et al., 1993; Kamb et al., 1994). Recent

Correspondence: CT Liew Received 17 February 1998; revised 11 August 1998; accepted 12 August 1998

investigations on human tumors have suggested that the p16 gene might be even more commonly mutated than the p53 gene (Marx, 1994). The alterations of the p16 TSG could lead to its inactivation and result in the deregulation of cell proliferation and tumorigenesis (Serrano et al., 1995). Multiple mechanisms of p16 inactivation, such as point mutations, homozygous deletion, loss of heterozygosity (LOH) and hypermethylation, have been reported in di€erent kinds of human tumors, including hepatocellular carcinoma (HCC) and HCC is a common and fatal cancer frequently seen on the Southern coast of China, Taiwan, Japan and Hong Kong. However, previous reports by Kita et al. (1996) and Biden et al. (1997) showed that homozygous deletions or mutations of the p16 gene were infrequent in a total of 85 cases of HCC samples from Japan and Australia. Furthermore, Kita et al. (1996) showed 17.9% of the HCC from Japan had LOH, while Biden and coworkers showed 44% of the HCC from Australia had LOH on the p16 loci. Moreover, a recent report by Chaubert et al. (1997) showed promoter hypermethylation of the p16 gene in 12 of 25 cases (48%) of HCC from Switzerland. Again, somatic mutations or homozygous deletions were not observed in all 25 cases of HCC studied. Homozygous deletion of the p16 gene has been reported in several types of primary tumors but the frequency of such occurrence in HCC is still unclear. (Kamb et al., 1994; Nobori et al., 1994; Caldas et al., 1994; Okamoto et al., 1995; Reed et al., 1996; Cairns et al., 1995). In this study, the status of the p16 tumor suppressor gene on chromosome 9 in 48 cases of HCC from Hong Kong were investigated for: (a) methylation on p16 promoter by methylationspeci®c PCR (MSP); (b) the loss of heterozygosity by microsatellite analysis and homozygous deletion by comparative multiplex PCR; and (c) mutation by PCR-SSCP and DNA sequencing.

Results Methylation of p16 promoter in HCC By using methylation-speci®c PCR (MSP), 30 of 48 (62.5%) cases of HCC tumors showed aberrant methylation at the 5' CpG island of the p16 gene (Figure 1 and Table 1). All 30 cases of HCC exhibiting MSP ampli®cation of methylated templates also showed ampli®cation of unmethylated templates. However, six of 30 (20.0%) of the corresponding non-tumor liver tissues showed weak MSP amplification of methylated templates.

p16 alterations in HCC CT Liew et al

790

High frequency of LOH on chromosome 9p21 of HCC In addition, high frequency of LOH region was observed on the chromosome 9p21 region of HCC. The number of informative cases and the frequency of LOH with nine microsatellite markers are listed in Figure 2. The highest frequency of LOH was observed with marker D9S1747 (63.0%) and the lowest with marker D9S171 (35.7%). In the order of decreasing frequency, LOH was observed as follows: D9S1747 (63%), D9S1751 (60%), D9S1752 (57.7%), D9S1748 (52.5%), D9S1749 (51.3%), IFNA (48.1%), D9S161 (44.0%) D9S1126 (41.4%) and D9S171 (35.7%)

Figure 1 MSP analysis of the p16 gene in HCC. The PCR products of p16 gene were analysed by agarose gel electrophoresis. f, Molecular weight marker fX174DNA/Hae II; N, nontumor liver tissue; T, tumor; U and M, MSP analysis speci®c for unmethylated and methylated 5'-CpG islands of p16, respectively

Table 1 Clinicopathological data and p16/9p21 status in 48 patients with HCC Case Tumor size no HBV (cm) Di€erentiation

Cirrhosis

AFP (mg/L)

Alb ALP 9p21 (g/L) (IU/L) allelic loss

p16 methylation Non-tumor Tumor U M U M

p16 HD

p16 mutation

HD HD HD HD HD

WT WT WT WT WT

+ + + + +

+ 7 7 7 7

+ + + + +

+ + + 7 7

5 30 37 26 47

+ + + + +

2.562.3 7.567 6.566 7.567 766.5

W-MD Clear cell WD MD MD

Yes No Yes Yes Yes

3650 309 26 971 45660

4 7 8 13 15 18 19 27 28 34 35 39 41 43

+ + + + + + + + + + + + + +

864.5 8.567.8 3.563 564 7.563.5 363 362.5 4.564.5 662.5 2.261.7 1.561.5 867 362.5 5.564.5

W-MD MD MD MD W-MD WD W-MD W-MD WD MD MD MD W-MD MD

No Yes Yes Yes No Yes Yes Yes No Yes Yes Yes Yes No

7326 26990 1321 76 46230 4788 329 35 11 ND 5.5 ND 610 4

43 43 33 33 34 30 32 39 33 ND 38 34 40 31

105 144 45 92 57 64 57 156 117 ND 113 133 95 70

LOH LOH LOH LOH LOH LOH LOH LOH LOH LOH LOH LOH LOH LOH

No No No No No No No No No No No No No No

HD HD HD HD HD HD HD HD HD HD HD HD HD HD

WT WT WT WT WT WT WT WT WT WT WT WT WT WT

+ + + + + + + + + + + + + +

7 7 7 + + 7 7 7 + + + 7 7 7

+ + + + + + + + + + + + + +

+ + + + + + + + + + + + + +

20 24 29 32 36 38 40 44 45 46 48

+ + + + + + + + + + +

5.564.5 362 2.562.5 2.561.5 565 665.5 4.363.8 764.5 2.562 1667 563.3

WD MD W-MD W-MD W-MD MD MD M-PD Clear cell W-MD WD

No Yes Yes No No Yes Yes No Yes No Yes

10 2727 439 115 49890 511 555 29 626 83 ND

40 41 28 40 32 41 40 34 33.6 39 ND

72 97 117 77 83 85 112 95 111 108 ND

LOH LOH LOH LOH LOH LOH LOH LOH LOH LOH LOH

No No No No No No No No No No No

HD HD HD HD HD HD HD HD HD HD HD

WT Exon 2 Mutation WT WT WT WT WT WT WT WT Exon 2 Mutation

+ + + + + + + + + + +

7 7 7 7 7 7 7 7 7 7 7

+ + + + + + + + + + +

7 7 7 7 7 7 7 7 7 7 7

1 3 6 9 10 12

7 + 7 + + +

2.462 4.564 262 2.862.2 4.764 863.7

MD M-PD MD W-MD WD W-MD

Yes Yes Yes Yes Yes Yes

6310 4706 4957 10 10 4800

39 47 42 30 27 32.5

308 82 231 156 207 169

Ret Ret Ret Ret Ret Ret

No No No No No No

HD HD HD HD HD HD

WT WT WT WT Exon 3 Mutation WT

+ + + + + +

7 7 7 7 7 7

+ + + + + +

+ + + + + +

17 21 23 25 31 33 42

+ + + + + + +

1467 2.562.5 3.563 766 665.5 261.4 7.567

M-PD WD Clear cell WD Clear cell W-MD WD

No No Yes No Yes Yes Yes

14715 694 4 12 45 4 11

30 42 43 40 29 40 28

122 52 65 94 226 92 116

Ret Ret Ret Ret Ret Ret Ret

No No No No No No No

HD HD HD HD HD HD HD

WT WT WT WT WT WT WT

+ + + + + + +

7 7 7 7 7 7 7

+ + + + + + +

+ + + + + + +

2 11 14 16 22

+ + 7 + +

462.3 1169 1167 2.562.5 563.6

M-PD WD MD WD MD

Yes Yes No Yes No

33

38 30 37 20 34

48 156 89 191 120

Ret Ret Ret Ret Ret

No No No No No

HD HD HD HD HD

WT WT WT WT WT

+ + + + +

7 7 7 7 7

+ + + + +

7 7 7 7 7

36 312 76

35 218 36 58 28 70 34.3 155 37 133

LOH LOH Ret LOH LOH

HBV, hepatitis B virus infection; AFP, a-Feto-protein; Alb, Albumin; ALP, Alkaline phosphatase; WD, well di€erentiated; MD, moderately di€erentiated; PD, poorly di€erentiated; LOH, loss of heterozygosity; Ret, retention of heterozygosity; HD, homozygous deletion; WT, wild type; U, unmethylated; M, methylated; +, positive; 7, negative; ND, not done

p16 alterations in HCC CT Liew et al

791

Figure 3 Homozygous deletion on exon 2 or p16 gene were con®rmed in ®ve HCC samples by comparative multiplex PCR. In these tumors, the signal for allele on exon 2 of p16 was highly reduced, whereas the band for control marker D9S126 showed similar intensity in both the non-tumor liver tissue (N) and tumor (T) lanes. Tumor 47 showed LOH on locus D9S126. Arrows indicate homozygous deletion on exon 2 of p16 gene Figure 2 Cases that allow the de®nition of the smallest common deleted region (SCDR), homozygous deletion (HD) on exon 2 of the p16 gene, and hypermethylation of the p16 in HCC. Four cases (cases 7, 20, 30 and 46) de®ne the SCDR. The SCDR was located between loci D9S1747 and D9S1748. This is a region of approximately 200 kilobase (kB) which includes the p16 gene. Percentages of loss of heterozygosity (LOH) in informative cases are listed on the far right column. Cases 5, 30 and 37 showed both HD on exon 2 and hypermethylation of the p16 gene. Solid circles, LOH; dotted circles, not informative; open circles, retention of heterozygosity. +, p16 homozygous deletion (HD) or hypermethylation; 7, absence of p16 HD or hypermethylation

respectively. A minimal common deletion region of approximately 200 kb, inclusive of the p16 gene which is located between loci D9S1747 and D9S1748, was observed. Twenty-nine of 48 (60.4%) tumors showed LOH on at least one locus on 9p21 region. The LOH was detected on 9p21 in the three HBV-negative examined. Deletion of p16 in HCC Moreover, homozygous deletion of the p16 gene were found in ®ve of 48 (10.4%) HCC samples by comparative multiplex PCR (duplex quantitative PCR) as shown in Figure 3. In these tumors, the signal for the alleles at p16 was highly reduced, whereas the band for the control marker D9S126 showed similar intensity in both the normal and tumor lanes. The evaluation of the intensity of the signal between di€erent alleles was determined by both visual examination and quantitated by densitometer. Homozygous deletion was scored if the signal intensity of the p16 in tumor tissues was at least tenfold less than the signal from the non-tumor tissues whereas the intensity of the D9S126 control allele was approximately equal to both tumor and non tumor tissues. This result clearly demonstrates the presence of homozygous deletion at p16 in HCC samples. Homozygous deletion at the p16 in these ®ve HCC cases were highly reproducible by multiplex PCR for at least three times. SSCP analysis of p16 in HCC In order to determine if mutations may account for p16 alterations in the rest of the HCCs, SSCP and sequencing were performed. Five PCR products covering the entire p16 coding region were used to screen for mutations by SSCP assays. Mobility shift was detected in three of 48 (6.3%) HCC tumors.

Figure 4 PCR ± SSCP analysis of p16 gene in HCC. The autoradiographs showed somatic mutations of p16 gene in cases 24, 48 and 10. DNA was isolated and ampli®ed by PCR from tumor (T) and corresponding non-tumor liver tissue (N). Arrows indicate the novel SSCP bands

Figure 4 shows the SSCP band shifts in HCC which are indicative of base changes. Mutations were identi®ed by direct sequencing of the three positive cases of HCC tumors as shown in Table 2. Mutations of p16 were observed in codons 74, 75 and 159 in HCC cases 48, 24 and 10 respectively. Nucleotide changes from GTG to ATG (Val?Met), CAC to TAC (His?Tyr) and CGG to CCG (Arg?Pro) were detected in these three cases. A representative example of sequencing data showing mutations at codon 75 (case 24) is shown in Figure 5. Overall, the frequency of p16 alterations with homozygous deletion, mutation or hypermethylation, were detected in 70.8% (34 of 48) of HCC tumors (Table 1). Among the 34 cases, two cases had both 9p21 LOH and p16 mutation; 16 cases had both 9p21

p16 alterations in HCC CT Liew et al

792

Table 2 p16 mutations in HCC Case No. 48T 24T 10T

Exon

Nucleotide number

Nucleotide change

Codon

Codon change

2 2 3

238 241 494

GTG to ATG CAC to TAC CGG to CCG

74 75 159

Val to Met His to Tyr Arg To Pro

The nucleotide positions of the base changes (location) derive from the numbering scheme used in Serrano et al. (1993, Nature, 366, 704)

Figure 5 Sequencing analysis of p16 gene in HCC. Sequencing data showing one base change in codon 75 (CAC to TAC) in case 24 of HCC

LOH and p16 hypermethylation; three cases (cases 5, 30 and 37) had both homozygous deletion and hypermethylation of the p16 gene; one case (case 10) was without LOH at 9p21 but had both p16 mutation and hypermethylation; 13 cases were without LOH, homozygous deletion and mutation at p16 locus but were hypermethylated at the p16 promoter region (Table 1). There was no statistical signi®cance observed in the correlation of hypermethylation, homozygous deletion and mutation with tumor size, tumor cell differentiation, cirrhosis and liver function tests including the levels of alpha-feto-protein, albumin and alkaline phosphatase. Discussion HCC is a very common cancer in the coastal areas of South China, Hong Kong, Taiwan, Japan and the subSaharan countries in Africa. It is currently the second commonest cancer death in Hong Kong (Hong Kong Hospital Authority, 1998). The tumor is highly associated with chronic hepatitis B infection and cirrhosis of the liver in these regions. Approximately 80% of the HCC in Hong Kong is associated with chronic hepatitis B virus infection and cirrhosis of the liver. This study demonstrated that the p16 tumor suppressor gene in HCC could be inactivated by di€erent mechanisms including hypermethylation, homozygous deletion and mutation. In this investigation, we detected hypermethylation in 30 of 48 (62.5%) cases of HCC. Hypermethylations usually occur in the 5'CpG islands, in or near the promoter and ®rst exon region of many genes (Baylin et al., 1998). The p16 tumor suppressor gene has a typical 5'CpG island in the promoter region that is unmethylated in normal tissues (Merlo et al., 1995; Herman et al., 1995).

Interestingly, three cases of HCC had both homozygous deletion on the second exon (2C), and hypermethylation on the 5'CpG island of the p16 gene while 16 cases had LOH and hypermethylation of p16 and two cases had mutation and LOH. This ®nding could represent multiple alterations in di€erent regions of the p16 gene and thus implicating multiple mechanisms for the inactivation of p16 in HCC. The redundancy of such alterations in the p16 region could be the result of gross deletion of chromosomal region which include partial deletion or inactivation of the p16 gene. Apparently, two independent mechanisms could be operating in the same region in which hypermethylation of the p16 loci may have occurred as an earlier event followed by deletion in the same region as a later event. The homozygous deletion of the p16 locus will also lead to the loss of p16b/p19ARF gene which share the same exons 2 and 3 of the p16 gene. Recent studies demonstrated that p19ARF inhibits the oncogenic actions of MDM2, blocks MDM2-induced degradation of p53 and enhance p53-dependent transactivation. The loss of p19ARF/p16b may lead to the inactivation of the p53 functions including G1 arrest and apoptosis (Pomerantz et al., 1998; Zhang et al., 1998). The homozygous deletion of the p16 locus may cause the inactivation of both of the RB and p53 pathways and may play an important role in the progression in some HCCs. It is interesting to note that weak ampli®cation of methylated DNA were observed in six of 48 (12.5%) cases of the non-tumor liver tissues in which strong signals of methylated sequence were observed in the corresponding tumor tissues. This weak ampli®cation could represent the presence of some tumor cells in the non-tumor liver tissues although frozen sections did not show any tumor cells. Alternatively, the weak signals of methylated sequence in the non-tumor liver tissues could implicate that such de novo methylation may occur in premalignant or early subcellular malignant changes in these cases. Our results are in keeping with an earlier study of hypermethylation in HCC, (Chaubert et al., 1997) which indicated that 12 of 25 (48%) cases of HCC, from a patient population of 24 white Swiss, had hypermethylation while no hypermethylation was detected in all of the corresponding non-tumor liver tissues. Although LOH on the p16 region is common in various types of tumors, yet the reported LOH in HCC showed varying frequencies (Kita et al., 1996; Biden et al., 1997; Chaubert et al., 1997). In addition, LOH on chromosome region 9p21 in HCC has not been thoroughly investigated. By using two polymorphic markers ¯anking the 9p21 region (IFNA and D9S126), Kita et al. (1996) reported 17.9% of the HCC from Japan had LOH on 9p21. On the other hand, Biden et al. (1997), used eight polymorphic markers ¯anking the same region, detected LOH in 44% of the HCC samples from Australia. Using a larger number of HCC cases and more microsatellite markers, our study showed the highest frequency of LOH (60.4%) on 9p21 region in HCC which is in accordance with results from other types of cancers. Kita et al. (1996) recently documented that the mutations or homozygous deletions of the p16/INK4A gene were infrequent in HCC from Japan. In addition, mutations of the p16 gene were detected only in three

p16 alterations in HCC CT Liew et al

of 62 (4.8%) HCC cases and no homozygous deletion was present in all 62 cases of HCC and six HCC cell lines (Kita et al., 1996). On the other hand, Biden et al. (1997) reported similar ®ndings and found no homozygous deletion at the p16 region in all 23 cases of HCC and four HCC cell lines while Hui et al. (1996) also reported the absence of mutation and homozygous deletion at the p16 region. However, they reported that the p16 protein was not expressed in three of six (50%) HCC cell lines and 11 of 32 (34%) HCC tumors and suggested that p16 is deregulated post-transcriptionally. More recently, Chaubert et al. (1997) also demonstrated that the lack of somatic mutation and homozygous deletion of the p16 gene in all 26 cases of HCC studied. Homozygous deletion of the p16 gene was reported in HCC cell lines from Japan (Okamoto et al., 1994). They detected homozygous deletion in one of eight (12.5%) HCC cell lines by Southern analysis. Furthermore, Qin et al. (1996) detected homozygous deletion on exon 2 in two of 24 (8.3%) HCC tumors from China. In this study, homozygous deletion was detected in ®ve of 48 cases (10.4%) of HCC from Hong Kong which is similar to those studies. The discrepancy of the frequencies in p16 alterations as reported is likely due to the lack of the examination of the methylation status of 5'CpG island of the p16 gene in these studies. On the other hand, it may be due to the di€erences in risk factors, such as HBV, HCV and alcohol, associated with the development of HCC in di€erent geographic regions. Chronic hepatitis B virus infection is a common risk factor in the population from China, Taiwan, and Hong Kong. While a¯atoxin B, a major etiological factor from speci®c region in China, is usually not considered as a risk factor in Hong Kong and Taiwan. On the contrary, chronic hepatitis C infection is more commonly observed in Japan, United States and Europe and recently, the association of chronic HCV infection and HCC has been reported from these regions (Tomimatsu et al., 1993; Gerber et al., 1992). Furthermore, HCC associated with chronic alcoholic liver disease and hemochromatosis are more commonly reported from Western populations such as Australia, Europe and United States. The prevalence of HCC associated with these risk factors in di€erent geographic areas is recognized and the speci®c genomic alterations may thus be di€erent. In the studies by Kita et al. (1996) and Hui et al. (1996), both of the studies involved population from Japan where the prevalence of hepatitis C related HCC is more common than HBV associated HCC. Furthermore, Chaubert et al. (1997) and Biden et al. (1997) studies presented HCC cases from the predominantly Caucasian populations from Switzerland and Australia respectively. In both studies, the cases presented showed varied associated risk factors and many were alcoholic and non-viral associated HCC. Almost two thirds of the HCC cases used in Biden et al. (1996) study were associated with alcoholic liver disease and hemochromatosis. It is interesting to note that the identi®cation of homozygous deletions in cell lines DNA is comparatively easier than human tumor tissues because the cell population is homogenous and the loss of the p16 gene sequence can be seen by either Southern blotting or PCR methods. In human tumor tissue samples, contamination by some normal cells will generate

some signals for retention of the gene. Therefore, to detect homozygous deletions in tissue samples, PCR reactions must be carefully controlled. An internal control must be used to con®rm that each template can be ampli®ed eciently and for quantitative comparison with the test sequence. This study demonstrated not only high frequency of p16 hypermethylation, it has also detected high frequency of LOH at the p16 gene on 9p21. The result is in accordance with reports for other kinds of tumors. More importantly, our results have clearly demonstrated the presence of homozygous deletions at the p16 gene by duplex quantitative PCR in our HCC samples. Thus, the overall frequency of p16 alterations in our HCC cases, including homozygous deletion, mutation and hypermethylation, detected was 70.8%. The three cases of HBV-negative HCC cases included in this study are too small to be representative and meaningful. Further study and comparison to elucidate the possible di€erence between HBV-positive and HBVnegative HCC should be conducted. This is the ®rst observation of high frequency of alterations of the p16 TSG with hypermethylation, homozygous deletion and mutation in mostly HBV associated HCC. In conclusion, the inactivation of the p16 gene at 9p21 by hypermethylation, homozygous deletion and mutation might play a critical role in the pathogenesis of HBV associated HCC. Materials and methods Specimens and DNA extraction Forty-eight cases of HCC and their corresponding nontumor liver tissues with con®rmed histopathological diagnosis were obtained immediately after hepatectomies at the Prince of Wales Hospital in Hong Kong from 1993 ± 1996. Forty-®ve cases of these samples were tested positive for hepatitis B surface antigen and HBV-DNA. Both tumor and non-tumor liver tissues were immediately snapped frozen in liquid nitrogen and stored at 7808C in a freezer before DNA extraction. The tumor tissue consisted of more than 90% tumor cells and the nontumor liver tissues had no tumor thrombi on frozen section examination. Prior to DNA extraction, the tissues were digested with RNase A (200 mg/ml) and proteinase K (1 mg/ml) after the tissues had been homogenized. DNA was then extracted with the QIAGEN Genomic DNA Kit (QIAGEN, Hilden, Germany) (Mischiati et al., 1993; Weising et al., 1991). Methylation speci®c PCR (MSP) of the p16 gene The CpG WIZTM p16 Methylation Assay Kit was obtained commercially (Oncor, Inc., Gaithersburg, MD, USA). An initial bisul®te reaction to modify the DNA was performed followed by PCR ampli®cation with speci®c primers designed to distinguish the methylated DNA from the unmethylated DNA. DNA (1 mg/100 ml) was denatured by NaOH (®nal concentration, 0.2 M) for 10 min at 378C. DNA Modification Reagent I was added (550 ml), vortexed and incubated at 508C for 20 h. The modi®ed DNA was puri®ed by using DNA Modi®cation Reagent II and III and eluted with 50 ml of water. The bisulfate modi®cation of DNA was completed by NaOH (®nal concentration, 0.3 M) treatment for 5 min at room temperature, followed by ethanol precipitation. DNA was resuspended in water and used immediately or stored at 7208C.

793

p16 alterations in HCC CT Liew et al

794

Hot start PCR was used for this assay. The PCR mixture contained 16 Universal PCR bu€ers, 4dNTPs (each at 1.25 mM), U or M primers (300 ng each per reaction). Annealing temperature was 658C for a total of 35 cycles. Each PCR product was directly loaded onto 3% agarose gels, stained with ethidium bromide and visualized under UV illumination. Microsatellite analysis of the Loss of heterozygosity (LOH) A panel of nine PCR based microsatellite markers ¯anking the 9p21 region were obtained from Research Genetics (Huntsville, AL, USA) or synthesized from Oligos Inc. (Wilsonville, OR, USA). The markers used were IFNA, D9S1751, D9S1749, D9S1747, D9S1748, D9S1752, D9S171, D9S126 and D9S161. Primers were labeled with [g-32P]ATP using T4 polynucleotide kinase (Armershan Corporation, UK). PCR ampli®cations were performed in 5 ml reaction volumes, including 15 ± 30 ng genomic DNA, 10 mM TrisHCl at pH 9.0, 50 mM KCl, 1.5 mM MgCl2, 62.5 mM of deoxynucleotide triphosphate, 0.1 unit of Taq DNA polymerase (Pharmacia, Uppsala, Sweden) and 0.5 pmol of each primer. Annealing temperatures were between 53 and 558C for a total of 40 cycles. The ampli®ed products were separated by 6 ± 7.5% denaturing polyacrylamide gels and ran at 60 W for 2 ± 3 h, dried and then processed for autoradiography using Kodak BioMax MR ®lm. Detection of the p16 gene homozygous deletion Homozygous deletion of the p16 gene was con®rmed by comparative multiplex PCR involving the ampli®cation of two di€erent sets of primer pairs in the same reaction mixture. D9S126 primer pair was used as a control to test for homozygous deletion at the p16 exon 2C locus. Two primer sets (p16 exon 2C and D9S126) were used in the same reaction to ensure the ampli®cation of the DNA and normalize the amount of products from the tumor DNA relative to that of the normal DNA for both of the test and control primers (Cairns et al., 1994). The PCR reaction was performed as described above in LOH assays with the total number of cycles decreased from 40 to 28. Annealing temperature was set at 558C. Results were analysed by the comparison of allele intensities in matched normal/tumor DNA by scanning densitometer with a computerized quantitative imaging system (Fluor-S MultiImager, BIORAD). Quanti®cation was performed using the MultiAnalyst/PC software. The relative ratio of both tumor and non-tumor alleles was determined, normalized, and then compared. A more than tenfold di€erence in the p16 allele intensity ratios between tumor and normal DNA was assigned as homozygous deletion. All of the comparative

multiplex PCR for the tumors with homozygous deletions were reproduced at least three times. PCR-Single strand conformation polymorphism (SSCP) All 48 HCC samples were screened for mutations on p16 exons 1, 2 (2A, 2B, and 2C) and 3 by SSCP analysis using primers as described by Hussussian et al. (1994). The primers were end-labeled as described above in LOH assays. For PCR ampli®cation, each sample was denatured at 948C for 5 min and subjected to 40 ampli®cation cycles (each cycle consisted of 30 s denaturing at 948C, 45 s annealing at 55 ± 628C, and 30 s extension at 728C), and a ®nal extension at 728C for 7 min. Following ampli®cation, 15 ml of denaturing loading dye was added to the 5 ml PCR reaction, denatured at 948C for 10 min, and then cooled on ice. In order to ensure the detection of mutations by SSCP analysis was performed in the absence of positive controls, the samples were separated by electrophoresis on two gels of di€ering composition under two di€erent sets of electrophoresis conditions. The two conditions were as follows: (a) a 60% neutral polyacrylamide gel containing 10% glycerol, ran at 3 W for 12 h at room temperature, and (b) a 6% nondenaturing polyacrylamide gel, ran at 30 W for 3 h at 48C. After electrophoresis, the gels were dried and exposed to Kodak BioMax ®lm for 4 ± 12 h. Variant SSCP bands were cut out from the gels and the DNA were eluted in 100 ml of 0.256 TE bu€er at 378C overnight. For sequencing, the variant SSCP bands and 2 ml of the eluted DNA were used as template for secondary unlabeled PCR reactions and carried out by using the conditions described above. DNA Sequencing The PCR product was puri®ed using QIAquick Gel Extraction kit (QIAGEN Hilden, Germany) and sequenced by the dideoxy chain termination method with a Thermo Sequenase cycle sequencing kit (Amersham LIFE SCIENCE, Inc. OH, USA). All of the mutations were further con®rmed by direct sequencing of the ampli®ed tumor and non-tumor DNA to identify germline mutations and polymorphisms. Acknowledgements This research is supported by the Research Grant Committee, Hong Kong (Project ID: 2140122). We are grateful to Professor CC Liew, Director of the Molecular Diagnostic Center of the University of Toronto, Canada for his advice and positive critical review of the manuscript. We also thank Pui Chung Wong for proof reading and typing our manuscript.

References Baylin SB, Herman JG, Gra€ JR, Vertino PM and Issa JP. (1998). Adv. Cancer Res., 72, 141 ± 196. Biden K, Young J, Buttenshaw R, Searle J, Cooksley G, Xu D-B and Leggett B. (1997). Hepatology, 25, 593 ± 597. Cairns P, Polascik TJ, Eby Y, Tokino K, Califano J, Merlo A, Mao L, Herath J, Jenkins R, Westra W, Rutter JL, Buckler A, Gabrielson E, Tockman M, Cho KR, Hedrick L, Bova GS, Isaacs W, Koch W, Schwab D and Sidransky D. (1995). Nature Genet., 11, 210 ± 212. Cairns P, Tokino K, Eby Y and Sidransky D. (1994). Cancer Res., 54, 1422 ± 1424. Caldas C, Hahn SA, da Costa LT, Redston MS, Schutte M, Seymour AB, Weinstein CL, Hruban RH, Yeo CJ and Kern SE. (1994). Nature Genet., 8, 27 ± 32.

Chaubert P, Gayer R, Zimmermann A, Fontolliet C, Stamm B, Bosman F and Shaw P. (1997). Hepatology, 25, 1376 ± 1381. Gerber MA, Shieh YSC, Shim KS, Thung SN, Demetris AJ, Schwartz M, Akyol G and Dash S. (1992). Am. J. Pathol., 141, 1271 ± 1277. Herman JG, Merlo A, Mao L, Lapidus RG, Issa JP, Davidson NE, Sidransky D and Baylin SB. (1995). Cancer Res., 55, 4525 ± 4530. Hong Kong Hospital Authority. (1998). Cancer incidence and mortality in Hong Kong 1993 ± 94. Hong Kong Cancer Report. Hui A-M, Sakamoto M, Kanai Y, Ino Y, Gotoh M, Yokota J and Hirohashi S. (1996). Hepatology, 24, 575 ± 579.

p16 alterations in HCC CT Liew et al

Hussussian CJ, Struewing JP, Golstein AM, Higgins PAT, Ally DS, Sheahan MD, Clark Jr WH, Tucker MA and Dracopoli NC. (1994). Nature Genet., 8, 15 ± 21. Kamb A, Gruis NA, Weaver-Feldhaus J, Liu Q, Harshman K, Tavtigian SV, Stochert E, Day RS, Johnson BE and Skolnick MH. (1994). Science, 264, 436 ± 440. Kita R, Nishida N, Fukuda Y, Azechi H, Matsuoka Y, Komeda T, Sando T, Nakao K and Ishizaki K. (1996). Int. J. Cancer, 67, 176 ± 180. Marx JA. (1994). Science, 264, 1846. Merlo A, Herman JG, Mao L, Lee DJ, Gabreilson E, Burger PC, Baylin SB and Sidransky D. (1995). Nature Med., 1, 686 ± 692. Mischiati C, Fiorentino D and Gambari R. (1993). Bioteshuiqere, 15, 146 ± 151. Nobori T, Miura K, Wu DJ, Lois A, Takabayashi K and Carson DA. (1994). Nature, 368, 753 ± 756. Okamoto A, Demetrick DJ, Spillare EA, Hagiwara K, Hussain SP, Bennett WP, Forrester K, Gerwin B, Serrano M, Beach DH and Harris CC. (1994). Proc. Natl. Acad. Sci. USA, 91, 11045 ± 11049. Okamoto A, Hussain SP, Hagiwara K, Spillare EA, Rusin MR, Demetrick DJ, Serrano M, Hannon GJ, Shiseki M, Zariwara M, Yue X, Beach DH, Yokota J and Harris C. (1995). Cancer Res., 55, 1448 ± 1451.

Pomerantz J, Schreiber-Agus N, Liegeois JN, Silverman A, Alland L, Chin L, Potes J, Chen K, Orlow I, Lee HW, Cordon-Cardo C and DePinho RA. (1998). Cell, 92, 713 ± 723. Qin L, Tang Z, Liu K, Ye S, He B and Zhou G. (1996). Oncol. Reports, 3, 405 ± 408. Reed AL, Califano J, Cairns P, Westra WH, Jones RM, Koch W, Ahrendt S, Eby Y, Sewell D, Nawroz H, Bartek J and Sidransky D. (1996). Cancer Res., 56, 3630 ± 3633. Serrano M, Gomz-Lahoz E, DePinho RA, Berch D and BarSagi D. (1995). Science, 267, 249 ± 252. Serrano M, Hannon GJ and Beach D. (1993). Nature, 366, 704 ± 707. Tomimatsu M, Ishiguro N, Taniai M, Okuda H, Saito A, Obata H, Yamamoto M, Takasaki K and Nakano M. (1993). Cancer, 72, 683 ± 688. Weising K, Beyermann B, Ramser J and Kahl G. (1991). Electrophoresis, 12, 159 ± 169. Zhang Y, Xiong Y and Yarbrough WG. (1998). Cell, 92, 725 ± 734.

795