MOLECULAR CA RClNOGENES/S3:68-74 (1990)

Increased Levels of Phorbin, c-myc, and Ornithine Decarboxylase RNAs in Human Colon Cancer Jose G. Guillem,’ Miriam F. Levy, Ling Ling Hsieh, Mark D. Johnson, Paul LoGerfo, Kenneth A. Forde, and 1. Bernard Weinstein Comprehensive Cancer Center (JGG, MFL, LLH, MDJ, PL, KAE IBW) and Departments of Surgery (JGG, PL, KAF) and Medicine (IBW), Columbia Universiu New York, New York

Our previous work on protein kinase C (PKC) and colon cancer has shown altered levels of PKC activity in human colon tumors, as well as activation of PKC by colon tumor promoters such as bile acids. To understand further the role o f PKC in colon carcinogenesis, we analyzed the expression of phorbin, a gene induced by PKC activation, in a series of different stages o f human colon tumors. As shown by northern blot analyses of poly (A)+ RNA, higher levels o f phorbin RNA were seen in 26 colon tumor samples than in their adjacent normal colonic mucosa. There also appeared t o be a correlation between the abundance of phorbin RNA in t h e tumors and the extent of invasion (tumor-to-normal tissue phorbin RNA ratio = 4.2,8.0, and 11.9 for Dukes’ A, B, and C, respectively). Phorbin RNA was also abundant in a human colon cancer line (HT29). We also examined the expression of other mitogen-responsive genes (c-myc, ODC, and p-actin) in a set o f 19 colon tumor samples. All tumors displayed significant (mean 3.8-fold) increases in the level of c-myc RNA compared with their adjacent normal colonic mucosa. About 47% and 16% of these tumor samples also showed increased levels of ODC (mean 3.1-fold) and p-actin (mean 1.6-fold) RNA, respectively. The increased levels of c-myc, ODC, and p-actin RNA did not correlate with the extent o f tumor invasion. Taken together, these results demonstrate that human colon tumors usually display increased levels o f both phorbin and c-myc RNAs. The marked increases in phorbin RNAsuggest that this could serve as a useful biomarker in studies on human colon cancer. Key words: Protein kinase C, colon neoplasm, phorbin, myc, orthinine decarboxylase INTRODUCTION Colon cancer is the most common cancer in the United States, accounting for about 60,000 deaths per year [ l l . Despite the magnitude of the disease, survival figures differ little from those of 30 years ago [ l l . In addition, the present staging systems are prognosticallysignificant only for advanced stages of colorectal cancer. In light of reports showing remarkable differences in survival within similarly staged tumors, the traditional staging criteria may be inadequate [21. These facts have highlighted the need to develop a more fundamental approach to this disease. A better understanding of the etiological factors and the molecular events involved in colon tumor growth and metastasis could lead to the identification of tumor cell markers that would be useful in detecting disease early, accurately predicting a tumor’s metastatic potential, and designing new strategies that achieve selective tumor cell destruction. Extensive data suggest that colon carcinogenesis proceeds through distinguishable stages of initiation, promotion, and progression. To begin to elucidate the molecular mechanisms governing this process, we have studied the role of protein kinase C (PKC) in colon carcinogenesis. PKC is a Ca2+-and phospholipid-dependent protein kinase that plays a key role in signal transduction by mediating the actions of certain growth factors, hormones, and tumor promoters [3]. In cell culture systems the activation of PKC 0 1990 WILEY-LiSS, INC.

by phorbol ester tumor promoters is associated with increased expression of several genes includingc-f& c-myc, ornithine decarboxylase (ODC), and phorbin [3,4]. The latter gene was identified in our laboratory in a cDNA subtraction library and is a particularly useful marker of PKC activation [4; unpublished data]. Phorbin is an acronym for PHORBol-ester-INduced;the phorbin gene’s expression correlates strongly with PKC activation. In previous studies we have obtained several lines of evidence that PKC may play a role in colon carcinogenesis. We have shown that bile acids, putative colon tumor promoters, can activate PKC in vitro and induce ODC expression in intact cells in a manner analogous to phorbol ester tumor promoters [5,6]. In addition, we have demonstrated decreased levels of PKC in human colon carcinomas compared with adjacent normal colon mucosa, suggesting downregulation of this enzyme in the tumors [7]. To understand further the possible role of PKC in colon carcinogenesis, as well as identify biomarkers of colon tumor growth and invasion, we measured in the present study the levels of phorbin RNA in a series of 26 primary human

’Corresponding author: Departmentof Surgery, Box 177, ColumbiaPresbyterian Medical Center, 622 W. 168th St., New York, NY 10032. Abbreviations: FCS, fetal calf serum; GTC, guanidiurn thiocyanate; ODC, ornithine decarboxylase; PKC, protein kinase C; SSC, standard saline citrate; TIMP, tissue inhibitor of metalloprotease.

lNCREASED RNA LEVELS IN COLON CANCER

colon tumors that varied in their degree of invasiveness. Since tumor invasion is thought to be caused, in part, by increased breakdown of the extracellular matrix by proteases produced by the tumor and possibly caused by abnormalities in growth factor function [8],the expression of phorbin was of particular interest. Phorbin shows strong sequence homology with a tissue inhibitor of metalloprotease (TIMP) and a growth hormone termed erythroidpotentiating factor [9,10]. In this study we also examined a set of 19 colon tumors for levels of three additional mitogen-responsive genes: c-myc, ODC, and p-actin. MATERIALS AND METHODS Chemicals [32P]NTPsand Hybond-N hybridization transfer membranes were purchasedfrom Amersham (Arlington Heights, IL). Restriction enzymes, DNA polymerases, and modifying enzymes were purchased from New England Biolabs (Beverly, MA) and Boehringer Mannheim (Indianapolis, IN). Tissue Samples All specimens were obtained from 26 patients during surgery. After the colectomy specimen was opened, particulate matter was rinsed off with cold saline. A rectangular, near full-thickness slice of the.tumor representative of the peripheraland central zones was carefully obtained in order to avoid any necrotic tissue. At the same time, normal mucosa samples were obtained at least 12 cm from the tumor by sharply dissecting the mucosalsubmucosa layer away from the muscle layers. Representativesamples (0.5-1 g) of the tumor and correspondingnormal mucosa were then quick-frozen in liquid nitrogen and stored at - 70°C. until analyzed later. Detailed histologicalanalyses with hematoxylin-phloxine-saffron-stained samples were performed on all tumor and normal mucosa samples. The adenocarcinomas were classified according to Dukes' original classification 11 11. Dukes' A, carcinoma limited to the bowel wall; Dukes' B, carcinoma through the wall but not involving lymph nodes; Dukes' C, carcinoma through the bowel wall with involved lymph nodes; Dukes' D, distant metastases. Cell Cultures The HT29 human colon cancer cell line was provided by the American Type Culture Collection (Rockville, MD). This cell line was used because of its well-documented metastatic and tumorigenic potential [ 121. Cultures were propagated in Dulbecco's modified Eagle's medium plus 10% heat-inactivated fetal calf serum (FCS) (GIBCO, Grand Island, NY) and incubated at 37°C in a humidified atmosphere of 5% C02, 95% air. Poly(A) RNA was extracted from postconfluent plates.

+

RNA Preparation Portions of the frozen tissue samples (approximately0.5-1 g) were homogenized in 9 mL of 4 M guanidium thiocyanate (GTC), using a polytron homogenizer (Brinkman, Westbury, NY). Total RNA was isolated by centrifugation through a 5.7 M CsCl gradient. The poly(A)+ fraction was

69

isolated following two rounds of selection through an oligo(dT) cellulose column. In the cell culture studies, cells were scraped from culture plates into cold phosphatebuffered saline, pelleted by centrifugation, and then processed as described above. (See references 4'13 for further details.) Northern Blot Analysis

+

Twenty micrograms of total RNA (or 5 p,g of poly(A) RNA) was separated by electrophoresison 1% agarosed% formaldehyde gels and transferred to Hybond-N transfer membranesin.10x SSC.Themembraneswerethenhybridized to nick-translated 32P-labeledprobes, washed at 68°C to 1 x and 0.1 x SSC stringency, and autoradiographed.The c-myc, ODC, and p-actin transcripts were detected with the appropriate 32P-labelednick-translated probes on the same total RNA blot by washing and rehybridization, as previously described [13]. The phorbin transcript was detected on a poly(A)+ RNA blot. Membranes were exposed to Kodak x-ray films using intensifying screens a t - 70°C. Autoradiographic band intensitywasquantifiedusing a densitorneter. The fold increase of a particular RNA species in the tumor samples was obtained by dividing the level of that RNA species detected in the tumor by that detected in the corresponding sample of normal colonic mucosa, based on densitornetric scanning of the northern blots. All lanes had equivalent amounts of total or poly(A)+ RNA by ethidium bromide staining. DNA Probes The following probes were used: c-myc, a 1.5-kb fragment obtained by Clal-EcoRI digestion of a pBR322 recombinant 1141; ODC, a 2.4-kb fragment obtained by EcoRI-BamHI digestion of an SP65 recombinant 11 51; pactin, a 2.1-kb Pstl fragment obtained by digestion from a pBR322 recombinant [161; and phorbin, a 0.8-kb EcoRl fragment of an SP6 recombinant [4]. 32P-labelednicktranslated probes were prepared from these fragments, as previously described [131. RESULTS Distribution of Tumor Cases and Yields of RNA The clinical and pathologic parameters for the 26 cases analyzed for phorbin RNA and a separate set of 19 cases (plus one case of Crohn's disease) analyzed for c-myc, ODC, and p-actin RNAs are given in Tables 1 and 2, respectively. The latter set of cases also included one case of Crohn's disease. The tissues examined included samples from all regions of the colon and tumors that varied considerably in their Dukes' stage and tumor grade. RNAs recovered from the colon tumors and their corresponding adjacent normal colonic mucosa were initially assessed for yield and possible degradation. In general, about 1 mg of total RNA was obtained from each gram of wet weight of tumor tissue. The yield of total RNA from normal colonic mucosa was somewhat less, i.e., about 0.50 mg per gram wet weight of tissue. The yields of poly(A)+ RNA from the tumor and normal mucosa samples ranged from 2-5% of the total RNA. In the studies on phorbin RNA we used

70

GUILLEM ETAL.

Table 1. Clinicopathologyof Colon Samples Whose RNAs Were Studied by Northern Blot Analysis Patient number Dukes' Stage A 1 2 3 4 5 6 7 Dukes' Stage B 8

9 10 11 12 13 14 15 16 17 18 Dukes' Stage C 19 20 21 22 23 24 25 26

Age

Sex

Site of tumor

Grade

75 79 73 79 82 68 73

M M M

Ascending Colon Descending Colon Splenic Flexure Sigmoid Sigmoid Transverse Cecum

Mod-Well Mod (Focal) Mod Mod Mod-Well (Focal)

1.9 3.3 2.7 5.0 8.5 3.5 4.8

56 70 76 77 64 78 66 67 66 51 93

M

Descending Colon Splenic Flexure Ascending Colon Sigmoid

6.0 6.5 18.0 1.5 9.4 3.5 5.8 6.2 12.5 10.2 8.4 20.0 10.0 34.5 6.2 4.7 3.5 9.7 6.3

78 75 61 46 65 76 71 81

F

M M M

Lymph nodes

M M

Ascending Colon Cecum Ascending Colon

F F F

Right Colon Sigmoid

Mod Well-Mod Mod-Poor Well Mod Mod Well Mod-Well Mod Mod-Poor Mod

Sigmoid Sigmoid Right Colon Sigmoid Sigmoid Right Colon Transverse Colon Sigmoid

Mod Mod Mod-Poor Poor Mod Poor Mod Mod-We1I

F F

M F F

M F F

M

M F F F

Phorbin RNA TIN*

*T, tumor; N, normal. This ratio indicates the fold increase in levels of phorbin RNA when a given tumor sample was compared with the normal mucosa sample from the same patient, based on densitometric scanning analysis of the northern blots. tMetastases to liver.

Table 2. Clinicopathologyof Colon Samples Whose RNAs Were Studied by Northern Blot Analysis Relative levels of tumor RNAs* Patient number 1 2 3 4 5 6 7 8 9 lot 11 12 13 14 15 16 17 18 19 20

Age

61 80 43 82 85 68 56 62 75 54 70 94 77 72 65 67 70 66 65 87

Sex

Site of tumor

F Rectum M Rectum F Sigmoid F Transverse Colon F Cecum M Ascending Colon F Rectosigmoid M Cecum F Sigmoid F Crohn'sDisease M Sigmoid M Ascending Colon M Rectum M Sigmoid F Descending Colon F Sigmoid F Colon F Cecum M Sigmoid F Ascending Colon

Dukes' stage Grade Lymph nodes C

A A B C C

C B B C B C B B C

B A C C

Mod Mod Well Mod Poor Mod Well Mod Well Poor Well Mod Well Poor Mod Mod Mod Well Mod

+ -

+ +

+ -

-

+ +

+ -

+

-

+

+

c-myc

3 3

ODC

p-actin

1.6 6

1.6 1

-

-

6 6 4 10 4 100 2 4 4 3 2 3 3 4

2 5 5 2 2 10 1 5 4 1

1 1 1.9

1

2

2 3 3 3

3 2

1

1

2 3 4 1 2.5 1

1.3 1 1 1 1.4 1.5 1.2 1.5

*The ratio of the level of the indicated RNA in the tumor sample compared with that of the same RNA species in the adjacent normal colonic mucosa of the same case. For additional details see Figures 2 and 3 and text. tSample number 10 is Crohn's disease of the ileum. The normal rnucosa in this case was obtained from the ascending colon.

71

INCREASED RNA LEVELS IN COLON CANCER

+

poly(A) RNA samples (Figure 1 and Table 1) because of the lower abundance of this RNA, whereas in the northern blot analyses using the c-myc, ODC, and p-actin probes, we used the total RNAsamples (Figures 2 and 3). Ethidium bromide staining of the gels used in the northern blot analyses of all of these samples indicated that all of the lanes contained equal amounts of nondegraded RNA. Levels of Phorbin RNA in Colon Tumors A striking finding was that the abundance of phorbin RNA transcripts, 0.8 kb in size, was increased in all of the 26 colon tumor samples examined with the phorbin probe, when each tumor sample was compared with the corresponding adjacent normal mucosa (Figure 1 and Table 1). Densitometry studies indicated that amongst the 26 samples, the tumor-to-normal ratio for phorbin RNA ranged from 1.5 to 34.5. When taken together, the data indicate that this ratio correlated wiith the extent of tumor invasion, using the Dukes' staging system (Table 3). Thus, the mean ratios of tumor-to-normal mucosa for levels of phorbin RNA were 4.2,8.0, and 11.9 for Dukes' A, B, and C tumors, respectively. The data on the 26 cases shown in Table 1 revealed a significant linear correlation with the Dukes' stage (r = 0.422; P