(2002) 21, 6689 – 6693 ª 2002 Nature Publishing Group All rights reserved 0950 – 9232/02 $25.00 www.nature.com/onc SHORT REPORT High incidence of protein-truncating of the in liver metastases of colorectal carcinomas

Michiko Miyaki*,1,4, Takeru Iijima1,4, Masamichi Yasuno2, Yumi Kita1, Tsunekazu Hishima3, Toshio Kuroki4 and Takeo Mori2

1Hereditary Tumor Research Project, Tokyo Metropolitan Komagome Hospital, Tokyo 113-8677, Japan; 2Department of Surgery, Tokyo Metropolitan Komagome Hospital, Tokyo 113-8677, Japan; 3Department of Pathology, Tokyo Metropolitan Komagome Hospital, Tokyo 113-8677, Japan; 4Institute of Molecular Oncology, Showa University, Tokyo 142-8666, Japan

To clarify the significance of p53 mutations in liver protein and genetic mutations have been used as of colorectal , the character- prognostic indicators in various types of carcinomas. istics of p53 mutations from 51 liver metastases and 76 In breast, gastric and lung carcinomas, p53 mutations primary invasive carcinomas without liver metastasis appear to correlate with worse survival (Saitoh et al., (Dukes’ A, B and C) were compared. The frequency of 1994; Lim et al., 1996). In colorectal carcinoma, tumors with p53 mutations was 61% (31 out of 51) in various studies have reported that: p53 mutations are the liver metastases, and 51% (39 out of 76) in the associated with short survival (Hamelin et al., 1994; primary carcinomas without liver metastasis. Approxi- Goh et al., 1995), over expression of the p53 protein is mately 90% of the informative cases having p53 a useful prognostic indicator (Auvinen et al., 1994), showed 17pLOH. Mutations detected within and that positive p53 immunostaining is associated exons 4 – 10 of the p53 gene included missense, nonsense, with poor overall survival (Lanza et al., 1996). Where frameshift, inframe deletion, and inframe insertion these observations have been made in some studies, mutations. Out of the tumors with p53 mutations, we such correlations have not been observed in other found that the percentage of tumors with protein- studies (Bell et al., 1993; Morrin et al., 1994; Mulder et truncating mutations (nonsense and frameshift muta- al., 1995; Smith et al., 1996). Accordingly, the tions) was extremely higher in liver metastases (16 out of prognostic value of p53 mutations in colorectal 31, 52%) than in primary carcinomas without liver carcinogenesis is controversial. These previous studies metastasis (5 out of 39, 13%) (P=0.0005). The present have noted only the occurrence pf p53 nuclear staining results suggest that protein-truncating mutations of the or mutations, and have not taken into considerations p53 gene are more relevant than missense mutations as the type of p53 mutation. It is also still not clear one of the prognostic factors in liver metastasis of whether p53 mutations are involved in liver metastasis colorectal carcinomas. of colorectal carcinomas. To clarify the relationship Oncogene (2002) 21, 6689 – 6693. doi:10.1038/sj.onc. between p53 mutations and liver metastasis of color- 1205887 ectal carcinoma, we compared the types of p53 mutations in liver metastases and primary carcinomas Keywords: p53 mutation; protein-truncating mutation; without liver metastasis, and found a significantly liver metastasis; colorectal carcinoma higher incidence of protein-truncating mutations of the p53 gene in liver metastases of colorectal carcinomas. Fifty-one liver metastases of colorectal carcinomas The p53 gene is the most widely altered tumor and 76 primary colorectal invasive carcinomas without suppressor gene in human carcinomas, including liver metastasis (Dukes’ A, B and C) from sporadic colorectal carcinomas (Nigro et al., 1989; Harris and cases were analysed from exons 4 – 10 of the p53 gene Hollstein, 1993). Inactivation of p53 function by gene by means of PCR – SSCP and sequencing methods. mutation and allele loss, which has been observed to Mutations of the p53 gene were detected in 31 tumors occur at the carcinoma developmental stage (Kikuchi- (61%) of the 51 liver metastases, and in 39 tumors Yanoshita et al., 1992), results in increased genetic (51%) of the 76 primary carcinomas without liver instability (Fukasawa et al., 1996), and is assumed to metastasis, as listed in Table 1. Approximately 90% of lead to increased malignancy. Accumulation of the p53 the tumors having p53 mutations showed 17pLOH. Two of the primary carcinomas without liver metas- tasis had two mutations each, therefore the total *Correspondence: M Miyaki, Hereditary Tumor Research Project, number of p53 mutations detected in primary Tokyo Metropolitan Komagome Hospital, 3-18-22 Honkomagome, carcinomas was 41. Bunkyo-ku, Tokyo 113-8677, Japan; E-mail: [email protected] Out of the total mutations, single-base substitutions Received 24 March 2002; revised 5 July 2002; accepted 18 July accounted for 27 out of 31 (87%) in liver metastases, 2002 and 35 out of 41 (85%) in primary carcinomas without p53 mutation in liver metastasis of colorectal carcinoma M Miyaki et al 6690 Table 1 p53 mutations in liver metastases of colorectal carcinomas and primary colorectal carcinomas without liver metastasis Metastasis or Amino acid Tumor Dukes’ stage Exon Codon Mutation change 17pLOH

PLK327CaLim Liver meta 4 113 TTC?TCC Phe?Ser No loss PLK317CaLim Liver meta 5 133 ATG?AAG Met?Lys Loss PLK306CaLim Liver meta 5 158 CGC?CAC Arg?His Loss PLK266CaLim Liver meta 5 173 CTG?ATG Val?Met Loss PLK347CaLim Liver meta 5 175 CGC?CAC Arg?His Loss PLK333CaLim Liver meta 6 192 CAG?TAG Gln?Stop Loss PLK345CaLim Liver meta 6 196 CGA?TGA Arg?Stop Loss PLK206CaLim Liver meta 6 213 CGA?TGA Arg?Stop Loss PLK235CaLim Liver meta 6 213 CGA?TGA Arg?Stop No loss PLK247CaLim Liver meta 6 213 CGA?TGA Arg?Stop Loss PLK256CaLim Liver meta 6 213 CGA?TGA Arg?Stop Loss PLK288CaLim Liver meta 6 213 CGA?TGA Arg?Stop Loss MY79 Liver meta 6 213 CGA?TGA Arg?Stop NI PLK267CaLim Liver meta 7 241 C deletion Codon 246-7 Stop Loss PLK248CaLim Liver meta 7 245 GGC?GAC Gly?Asp Loss PLK229CaLim Liver meta 7 248 CGG?CAG Arg?Gln NI PLK208CaLim Liver meta 7 248 CGG?TGG Arg?Trp Loss MY53 Liver meta 7 249 AGG?GGG Arg?Gly NI PLK268CaLim Liver meta 8 270 TTT?TTG Phe?Leu Loss PLK226CaLim Liver meta 8 273 CGT?TGT Arg?Cys Loss PLK287CaLim Liver meta 8 273 CGT?TGT Arg?Cys No loss PLK292CaLim Liver meta 8 271-6 16-bp deletion Codon 344-5 Stop Loss PLK227CaLim Liver meta 8 282 CGG?TGG Arg?Trp Loss PLK238CaLim Liver meta 8 282 CGG?TGG Arg?Trp NI PLK231CaLim Liver meta 8 285 GAG?GTG Glu?Val Loss PLK283CaLim Liver meta 8 292 AAA?TAA Lys?Stop Loss PLK234CaLim Liver meta 9 327-8 T insertion Codon 335-6 Stop NI PLK291CaLim Liver meta 10 341 C deletion Codon 344-5 Stop Loss PLK215CaLim Liver meta 10 342 CGA?TGA Arg?Stop Loss PLK355CaLim Liver meta 10 342 CGA?TGA Arg?Stop Loss H19CaLim Liver meta 10 342 CGA?TGA Arg?Stop Loss MY245 C 4 111-2 G insertion Codon 122-3 Stop No loss MY20 B 5 151 CCC?GCC Pro?Ala Loss MY115 B 5 156-9 4-bp del, 1-bp ins 4 aa change Loss MY72 B 5 157 GTC?TTC Val?Phe Loss H8Ca B 5 161 GCC?ACC Ala?Thr Loss COK80Ca C 5 170 ACG?TCG Thr?Ser Loss MY15 C 5 175 CGC?CAC Arg?His Loss MY65 B 5 175 CGC?CAC Arg?His Loss MY148 C 5 175 CGC?CAC Arg?His Loss MY302 C 5 176 TGC?CGC Cys?Arg NI 8 282 CGG?TGG Arg?Trp MY221 A 5 177 CCC?GCC Pro?Ala Loss MY112 B 5 174-80 18-bp deletion 6 aa deletion Loss PLK289 C 5 175-82 23-bp deletion Codon 184-5 Stop No loss MY223 C 6 196 CGA?TGA Arg?Stop No loss 10 342 CGA?TGA Arg?Stop MY204 A 6 204 GAG?TAG Glu?Stop Loss MY69 C 6 211 ACT?ATT Thr?Ile Loss COK32 B 6 211 ACT?ATT Thr?Ile NI MY22 C 6 213 CGA?TGA Arg?Stop Loss MY206 C 7 237 ATG?ATA Met?Ile NI COK30Ca C 7 238 TGT?TAT Cys?Tyr Loss MY162 B 7 245 GGC?AGC Gly?Ser Loss MY155 B 7 245 GGC?GTC Gly?Val NI H21Ca C 7 245 GGC?GTC Gly?Val Loss COK44Ca C 7 248 CGG?CAG Arg?Gln Loss COK81Ca B 7 248 CGG?CAG Arg?Gln Loss MY92 C 7 248 CGG?CAG Arg?Gln Loss COK49Ca C 7 248 CGG?TGG Arg?Trp Loss MY39 B 7 248 CGG?TGG Arg?Trp Loss MY62 B 7 248 CGG?TGG Arg?Trp Loss MY176 C 7 248 CGG?TGG Arg?Trp Loss MY279 A 7 248 CGG?TGG Arg?Trp Loss MY231 C 7 253-4 6-bp deletion 2 aa deletion NI MY160 C 8 273 CGT?CAT Arg?His Loss MY209 C 8 273 CGT?CAT Arg?His NI MY250 B 8 278 CCT?GCT Pro?Ala Loss Continued

Oncogene p53 mutation in liver metastasis of colorectal carcinoma M Miyaki et al 6691 Table 1 (Continued ) Metastasis or Amino acid Tumor Dukes’ stage Exon Codon Mutation change 17pLOH

COK23Ca A 8 282 CGG?TGG Arg?Trp Loss MY44 C 8 282 CGG?TGG Arg?Trp Loss MY243 C 8 282 CGG?TGG Arg?Trp NI MY248 B 8 264-5 51-bp insertion 17 aa insertion Loss

Fifty-one liver metastases of colorectal carcinomas and 76 primary colorectal invasive carcinomas without liver metastasis (Dukes’ A, B and C) were obtained from 127 Japanese sporadic cases after obtaining informed consent. Surgically removed tumor tissues were histopathologically diagnosed, and appropropriate areas of tumor tissues were frozen at 7808C until they were used for DNA analyses. A portion of the sample of colorectal carcinomas without liver metastasis are the same as those included in the previous report (Kikuchi-Yanoshita et al., 1992). DNA was extracted from frozen tissue samples and corresponding normal tissues using proteinase K, SDS, and phenolchloroform. Exons 4 – 10 of the p53 gene were analysed by the PCR – SSCP method. DNA fragments in abnormal bands were amplified by asymmetrical PCR and then sequenced by dideoxytermination method as previously described (Miyaki et al., 1994). PCR primers for exons 5 – 8 were the same as those previously reported (Kikuchi-Yanoshita et al., 1992), and primers for exon 4 were sense TGATGAGAATTCCCATCTACAGTCCCCCTT, and antisense TGATGAGAATTCGCAACTGACCGTGCAAGT, for exon 9 were sense TGATGAGAATTCGCCTCTTTCCTAGCACTG, and antisense TGATGAGAATTCCCAAGACTTAGTACCTGA, and for exon 10 were sense TGATGAGAATTCCTCTGTTGCTGCAGATCC, and antisense TGATGAGAATTCGCTGAGGTCACTCACCT. 17pLOH was analysed in acrylamide gel electrophoresis after amplification of DNA by PCR using microsatellite primers, TP53 and D17S250. NI: Not informative

liver metastasis, and the most frequent direction of Table 2 Types of p53 mutations in liver metastases colorectal single-base substitutions was G : C to A : T, which was carcinomas and primary colorectal carcinomas without liver metastasis 21 out of 27 (78%) in liver metastases and 26 out of 35 (74%) in primary carcinomas without liver metastasis. Number of mutation (%) Primary carcinoma without With respect to the type of mutation (Table 2), liver liver metastasis metastases showed 15 out of 31 (48%) missense Type of mutation Liver metastasis (Dukes’ A, B and C) mutations, 12 out of 31 (39%) nonsense mutations, and 4 out of 31 (13%) frameshift mutations leading to Missense 15 (48) 31 (75) Inframe deletion, insertion 0 (0) 4 (10) stop codons downstream. In contrast, primary carci- Nonsense 12 (39) 4 (10) nomas without liver metastasis showed 31 out of 41 Frameshift 4 (13) 2 (5) (75%) missense mutations, 4 out of 41 (10%) inframe Total mutation detected 31 (100) 41 (100) deletion and insertion mutations, 4 out of 41 (10%) Protein-truncating mutations (nonsense and frameshift) are 16 out of nonsense mutations, and 2 out of 41 (5%) frameshift 31 (52%) in liver metastasis, and 6 out of 41 (15%) in primary mutations. In total, protein-truncating mutations of the carcinoma without liver metastasis (P=0.0009) p53 gene (nonsense and frameshift mutations) were 16 out of 31 (52%) in liver metastases, and 6 out of 41 (15%) in primary carcinomas without liver metastasis mutations) at both the inside and outside of exons 5 – (P=0.0009). In other words, out of the tumors with 8, as well as missense mutations inside of exons 5 – 8. p53 mutations, the percentage of tumors with protein- There were several hot spots of p53 mutations, such as truncating mutations was 16 out of 31 (52%) in liver codons 213 and 342 (nonsense mutations in exon 6 and metastases, and 5 out of 39 (13%) in primary 10, respectively) in liver metastases, and codons 175, carcinomas without liver metastasis (P=0.0005). 245, 248 and 282 (missense mutations in exons 5 – 8) in To examine whether p53 mutations in liver metas- primary carcinomas without liver metastasis (Figure 1). tases also existed in primary tumors, 10 matched liver Although the majority of mutations in primary metastases and primary carcinomas, which were carcinomas without liver metastasis existed in the resected simultaneously from the same patients, were DNA binding domain (exons 5 – 8), some nonsense compared by SSCP and sequencing. In nine out of 10 and frameshift mutations in liver metastases existed cases, both tumors showed identical mutations. outside of this domain (exon 9 and 10). All nonsense The database of previous reports contains more than and frameshift mutations existing both inside and 1300 p53 mutations in colorectal carcinomas outside of the DNA binding domain resulted in the including hot spots of mutations at codons 175, 245, loss of the COOH-terminal region, and some of these 248, 273 and 282 (in exons 5 – 8), all of which are protein-truncating mutations may be insensitive to p53 missense mutations. In the case of breast , the nuclear staining by immunohistochemisty, because of pattern of p53 mutations has been reported to differ loss of the nuclear localization signals or instability of between those inside and those outside of exons 5 – 8 truncated proteins. (Hartmann et al., 1995). No missense mutations have The present study found that the percentage of been detected outside of exons 5 – 8, while the majority tumors with protein-truncating mutations was extre- of mutations inside exons 5 – 8 have been missense mely high in liver metastasis compared to that in mutations. primary carcinoma without liver metastasis, in spite of Our analysis of exons 4 – 10 of the p53 gene detected a similar frequency of p53 mutations in both groups, as protein-truncating mutations (nonsense and frameshift well as a similar frequency of single-base changes

Oncogene p53 mutation in liver metastasis of colorectal carcinoma M Miyaki et al 6692 understood, and multiple genetic changes may be involved in metastasis. With respect to p53 mutations, one explanation may be a functional difference between missense and protein-truncating mutations in the p53 protein (May and May, 1999). Missense mutations in exons 5 – 8 disturb the DNA binding activity of the p53 protein, but other domains remain unchanged. Whereas, protein-truncating mutations cause the loss of the COOH-terminal region, which includes the teramerization domain (codons 323 – 356) (Pavletich et al., 1993; Clore et al., 1994), the apoptotic domain (codons 363 – 393) (Wang et al., 1996) and the nuclear localization signals (codons 316 – 325, 369 – 375 and 379 – 384) (Shaulsky et al., 1990). Since tetramerization is essential for DNA binding of the p53 protein, loss of this domain results in complete nullification of p53 protein activity. Null activity caused by the loss of the COOH-terminal region may result in more serious damage to the p53 protein than just the distortion of DNA binding by missense mutations in the DNA Figure 1 Positions and types of p53 mutations detected in liver metastases of colorectal carcinomas and those in primary carcino- binding domain. It has been reported that the absence mas without liver metastasis of the p53 protein in p53-null mice did not augment the frequency of initiation or the rate of promotion, but greatly enhanced malignant progression of chemi- cally induced skin tumors, concluding that the loss of (including G : C to A : T) in both. Our data suggests p53 activity is a causal step in malignant progression that protein-truncating mutations will more likely (Kemp et al., 1993). It is also possible that null- contribute to worsened conditions, such as leading to mutations, which result from protein-truncating muta- liver metastasis, and also that the prediction of liver tions and loss of the other allele of the p53 gene, has a metastasis is possible at the primary carcinoma stage, larger contribution to the progression of human since almost all p53 mutations in liver metastases were colorectal carcinomas. already present in the matched primary carcinomas. The present observation on the characteristics of p53 Moreover, in our cases of liver metastases, survival mutations in liver metastasis suggests that protein- time after hepatic rejection tended to be shorter for truncating mutations are more relevant than missense patients with protein-truncating p53 mutations than for mutations as one of the prognostic factors in liver those with missense p53 mutations (data not shown). metastasis of colorectal carcinomas. To clarify the role The reason for a larger contribution of protein- of p53 mutations in the development and progression truncating mutations in liver metastasis is difficult to of colorectal and other carcinomas, more detailed understand, because the molecular mechanism of liver examination of the different effects of various types of metastasis of colorectal carcinoma is not yet fully mutations may be important.

References

Auvinen A, Isola J, Visakorpi T, Koivula T, Virtanen S and Hartmann A, Blaszyk H, McGovern RM, Schroeder JJ, Hakama M. (1994). Br. J. Cancer, 70, 293 – 296. Cunningham J, De Vries EMG, Kovach JS and Sommer Bell SM, Scott N, Cross D, Sagar P, Lewis FA, Blair GE, SS. (1995). Oncogene, 10, 681 – 688. Taylor GR, Dixon MF and Quirke P. (1993). Gastro- Kemp CL, Donehower LA, Bradley A and Balmain A. enterology, 104, 57 – 64. (1993). Cell, 74, 813 – 822. Clore GM, Omichinski JG, Sakaguchi K, Zambrano N, Kikuchi-Yanoshita R, Konishi M, Ito S, Seki M, Tanaka K, Sakamoto H, Appella E and Gronenborn AM. (1994). Maeda Y, Iino H, Fukayama M, Koike M, Mori T, Science, 265, 386 – 391. Sakuraba H, Fukunari H, Iwama T and Miyaki M. (1992). Fukasawa K, Choi T, Kuriyama R, Roulong A and Cancer Res., 52, 3965 – 3971. VandeWoude GF. (1996). Science, 271, 1744 – 1747. Lanza Jr G, Maestri I, Dubini A, Gafa R, Santini A, Ferretti Goh H-S, Yao J and Smith DR. (1995). Cancer Res., 55, S and Cavazzini L. (1996). Am. J. Clin. Pathol, 105, 604 – 5217 – 5221. 612. Hamelin R, Laurent-Puig P, Olschwang S, Jego N, Asselain Lim BHG, Soong R, Grieu F, Robbins PD, House A and B, Remvikos Y, Girodet J, Salmon RJ and Thomas G. Iacopetta BJ. (1996). Int. J. Cancer, 69, 200 – 204. (1994). Gastroenterology, 106, 42 – 48. May P and May E. (1999). Oncogene, 18, 7621 – 7636. Harris CC and Hollstein M. (1993). NewEngl.J.Med.,329, 1318 – 1327.

Oncogene p53 mutation in liver metastasis of colorectal carcinoma M Miyaki et al 6693 Miyaki M, Konishi M, Kikuchi-Yanoshita R, Enomoto M, Saitoh S, Cunningham J, De Vries EMG, McGovern RM, IgariT,TanakaK,MuraokaM,TakahashiH,AmadaY, Schroeder JJ, Hartmann A, Blaszyk H, Wold LE, Schaid Fukayama M, Maeda y, Iwama T, Mishima Y, Mori T and D, Sommer SS and Kovach JS. (1994). Oncogene, 9, 2869 – Koike M. (1994). Cancer Res., 54, 3011 – 3020. 2875. Morrin M, Kelly M, Barrett N and Delaney P. (1994). Gut, Shaulsky G, Goldfinger N, Ben-Ze’ev A and Rotter V. 35, 1627 – 1631. (1990). Mol. Cell Biol., 10, 6565 – 6577. Mulder JWR, Baas IO, Polak MM, Goodman SN and Smith DR, Ji C-Y and Goh H-S. (1996). Br.J.Cancer,74, Offerhaus GJA. (1995). Br. J. Cancer, 71, 1257 – 1262. 216 – 223. Nigro JM, Baker SJ, Preisinger AC, Jessup JM, Hostetter R, Wang XW, Vermeulen W, Coursen JD, Gibson M, Lupold Cleary K, Bigner SH, Davidson N, Baylin S, Devilee P, SE, Forrester K, Xu G, Elmore L, Yeh H, Hoeijmakers Glover T, Collins FC, Weston A, Modali R, Harris CC and JHJ and Harris CC. (1996). Dev., 10, 1219 – 1232. Vogelstein B. (1989). Nature, 342, 705 – 708. Pavletich NP, Chambers KA and Pabo CO. (1993). Genes Dev., 7, 2556 – 2564.

Oncogene