Lack of cyclin D1 overexpression in gastric carcinogenesis

P Blok,1 M E Craanen,2 P J van Diest,3 W Dekker4 & G N J Tytgat5

1Department of Pathology, Westeinde Ziekenhuis, Lijnbaan, The Hague; 2,3Departments of Gastroenterology and Pathology,University Hospital ‘Vrije Universiteit’, Amsterdam; 4Department of Internal Medicine, Kennemer Gasthuis, Haarlem; and5Department of Gastroenterology, Academic Medical Centre, Amsterdam, The Netherlands

Date of submission 28 January 1999Accepted for publication 19 May 1999

Blok P, Craanen M E, van Diest P J, Dekker W & Tytgat G N J

(2000) Histopathology 36, 151–155

Lack of cyclin D1 overexpression in gastric carcinogenesis

Aims: Cyclin D1 overexpression was examined in earlygastric carcinomas and precursor lesions with thefollowing aims; (1) to assess the chronology of cyclinD1 overexpression in various stages of gastric carcino-genesis, (2) to correlate cyclin D1 overexpression withthe Lauren type, the grade of differentiation and the typeof growth pattern of the tumours and (3) to correlatecyclin D1 overexpression with clinical parameters, inparticular lymph node metastasis and overall prognosis.Methods and results: Forty-five paraffin-embedded gastrect-omy specimens from early carcinomas were examined forthe presence of various precursor lesions. The Lauren type,the grade of differentiation and the type of growth patternwere reassessed for all early carcinomas. Cyclin D1overexpression was examined using the monoclonalantibody DCS-6. Cyclin D1 overexpression was absentfrom all precursor lesions. Ten early carcinomas (22%)

were cyclin D1 positive without significant differenceswhen stratified according to Lauren type, grade ofdifferentiation, type of growth pattern or lymph nodestatus. Univariate analysis failed to show a significantdifference in 5-year surival rate between cyclin D1positive and negative early carcinomas (90% vs. 94%).Conclusions: Cyclin D1 protein overexpression does notplay a role in the progression from normal to neoplasticgastric mucosa and does not discriminate betweenintestinal and diffuse type early gastric carcinomas ofCaucasian origin. Moreover, mechanisms other thancyclin D1 protein overexpression underlie the reporteddifference in biological behaviour of early gastriccarcinomas with different types of growth pattern.Finally, although it appears that cyclin D1 does not haveprognostic significance, studies on larger numbers,including advanced carcinomas, are warranted.

Keywords: cyclin D1, early gastric cancer, growth pattern, intestinal metaplasia, Lauren classification, precursor/precancer

Introduction

Transitions between different phases of the cell cycle areregulated at checkpoints. One of the most importantcheckpoints of the cell cycle is the so-called restrictionpoint in late G1, at which the cell commits itself to DNAreplication and at which both positive and negativeexternal signals are integrated into the cell cycle.Deregulation at this point may lead to cell growth andcell division, both insensitive to external cues. As such,it may also result in the uncontrolled replication ofdamaged DNA in the cells, leading to a progressive

accumulation of genetic aberrations that may contri-bute to the tumour phenotype.

In mammalian cells, the regulation of the restrictionpoint in late G1 is controlled by D-type cyclins and theirpartner cyclin dependent kinases (CDKs), primarilyCDK4.1 It is thought that these cyclins act as integratorsof growth factor-mediated signals as can be appreciatedfrom the fact that cyclin D1 synthesis immediatelyceases after withdrawal of colony-stimulating factor 1from macrophages.2,3 Although there are three D-typecyclins (D1, D2 and D3), evidence has accumulated thatparticularly implicates cyclin D1 in human tumorigen-esis. First, cyclin D1 is overexpressed as a result of achromosomal inversion, inv(11)(p15;q13) in para-thyroid adenomas.4 Second, the cyclin D1/PRAD1

Histopathology 2000, 36, 151–155

q 2000 Blackwell Science Limited.

Address for correspondence: Dr P Blok, Department of Pathology,Leyenburg Ziekenhuis, Leyweg 275, 2545CH The Hague, TheNetherlands.

gene is identical with the bcl-1 proto-oncogene and istranslocated in a high proportion of mantle-celllymphomas.5 Third, cyclin D1 is located on chromo-some 11q13, a region frequently amplified in a varietyof human carcinomas, e.g. lung, breast, bladder andsquamous head and neck carcinomas.6–9 Finally, theD-type/CDK4 complexes can act as protein kinases ofpRb (retinobastoma tumour suppressor gene protein)in G1 phase, further strengthening the link betweenD-type cyclins and tumourigenesis.10,11 With regard togastric carcinoma, data on cyclin D1 overexpressionare almost nonexistent. Moreover, the few studies thathave been published are limited in that no precursorlesions and almost no early carcinomas were includedand in that the gastric carcinomas included were Asianin origin.12,13 Since early carcinomas are thought toevolve into advanced carcinomas, it is questionablewhether data from advanced carcinomas can be readilytranslated to early carcinomas. Moreover, analogous tonon small cell lung carcinomas, it may well be thatAsian and Caucasian (early) gastric carcinomas arequite dissimilar in molecular profile due to differencesin racial background and/or geographical exposure tovarious (environmental) carcinogens.14–16 Finally, inview of the reported association of cyclin D1 over-expression with poor prognosis, e.g. in squamous headand neck carcinomas,17 studies on early gastriccarcinomas should include data on the types ofgrowth pattern. Although the prognosis of surgicallytreated early gastric carcinoma is generally excellent,18

its metastatic potential and postoperative 5-year survi-val rate appear to be correlated with the type of growthpattern. Especially the penetrating A type, in contrast tothe other growth types, has been reported to have ahigh propensity for blood vessel invasion and lymphnode metastasis and to have the worst prognosis aftersurgery due to early postoperative hepatic metastasis.19

In view of the above, we studied cyclin D1 over-expression in a series of Caucasian early gastriccarcinomas with the following aims: (1) to assess overallcyclin D1 overexpression in early carcinomas andprecursor lesions; (2) to correlate cyclin D1 overexpres-sion, if any, with the Lauren type, the grade ofdifferentiation and the type of growth pattern; and (3) tocorrelate cyclin D1 overexpression with clinical para-meters, in particular lymph node metastasis and overallprognosis.

Material and methods

T I S S U E S

Forty-five gastrectomy specimens from patients with

early gastric carcinoma (EGC), previously classified aseither intestinal or diffuse type according to Lauren20

were retrieved from the files. The tissues had been fixedin 10% buffered formalin and processed to paraffin waxby routine methods. Chronic active gastritis and chronicatrophic gastritis were diagnosed on haematoxylin andeosin (H & E) stained sections using the criteria of theSydney classification.21 Whenever intestinal metaplasia(IM) was present in the mucosa, additional 5 mmsections were cut and stained with alcian blue pH 2.5/periodic acid–Schiff (AB pH 2.5/PAS) and high-iron-diamine/alcian blue (HID/AB) for subtyping of IMaccording to Filipe.22 Dysplasia and the grade ofdifferentiation of the tumours were scored accordingto the criteria of the World Health Organization.23 Earlygastric carcinoma was defined according to theJapanese Society of Gastrointestinal Endoscopy.24 Thetumour histology according to Lauren and the type ofgrowth pattern according to Kodama et al.19 werereassessed using the original paraffin-embedded H & Estained sections.

I M M U N O H I S T O C H E M I S T RY

The antibody used was the mouse monoclonal antibodycyclin D1/PRAD-1 Ab-1 raised against human recom-binant full length cyclin D1 protein (clone DCS-6,Neomarkers, Freemont, CA), diluted 1:100 in combina-tion with the avidin–biotin–peroxidase method, asdescribed by Hsu et al.25 For each case, sections wereprepared with and without antigen retrieval. Briefly,5 mm sections were cut and de-waxed in xylene. Afterrehydration through graded alcohol, the sections wereimmersed in 0.3% hydrogen peroxide in methanol for30 min to block endogenous peroxidase activity. Afterwashing with phosphate-buffered saline (PBS), sectionswere treated with 10% normal goat serum to reducenonspecific antibody binding. All steps were carried outat room temperature. To improve immunoreactivity,sections were treated with an antigen retrieval solution(10 mM citric acid monohydrate pH 6.0), heated in amicrowave. After washing with PBS, the primaryantibody was applied. After further washing with PBS,sections were incubated with a biotinylated rabbit anti-mouse second-stage antibody (Dako, Glostrup,Denmark). Following further washing with PBS,peroxidase-conjugated streptavidin was applied. Peroxi-dase activity was demonstrated by adding 3,30-diaminobenzidine as a chromogen. The sections werecounterstained with haematoxylin. Negative controlsections were processed without the primary antibody.Positive control sections were processed from a breastadenocarcinoma, previously shown to express high

152 P Blok et al.

q 2000 Blackwell Science Ltd, Histopathology, 36, 151–155.

levels of cyclin D1 protein. Four sections from eachtumour were scored to correct for tumour heterogene-ity. Analogous to previous studies, cases were regardedas negative when <5% tumour cell nuclei showedstaining.26,27

S TAT I S T I C S

The Fisher’s exact test was used in the statisticalanalysis of the data. Survival data were analysedaccording to the Kaplan–Meier method. The log-ranktest was used to compare survival data between groups.A P-value <0.05 was considered significant.

Results

H I S T O L O G Y

Chronic active gastritis was observed in all gastrectomyspecimens. Chronic atrophic gastritis was present in 35gastrectomy specimens. Since IM was multifocallypresent, 230 foci of type I IM, 147 foci of type II IMand 34 foci of type III IM were identified. Four areas ofgastric dysplasia were found in the mucosa adjacent tothe tumours and were graded as: mild (n ¼ 1), moderate(n ¼ 2) and severe (n ¼ 1). There were 20 intestinal type(four highly differentiated, 11 moderately differentiated,five poorly differentiated) and 25 diffuse-type earlycarcinomas. Five early carcinomas had lymph nodemetastasis, none (0%) of the mucosal carcinomas andfive (28%) of the carcinomas extending into thesubmucosa (P ¼ 0.007).

G RO W T H PAT T E R N

Twenty-seven early carcinomas (60%) were confined tothe mucosa and 18 (40%) extended into the submu-cosa. According to growth pattern, there were 24 smallmucosal-type EGC (all limited to the mucosa), fivesuperficially spreading type EGC (three mucosal, twosubmucosal), and 16 penetrating (Pen) type EGC (fourPen A type, 12 Pen B type).

C YC L I N D 1 I M M U N O R E AC T I V I T Y

Without antigen retrieval, none of the lesions stainedpositively. After pretreatment, cyclin D1 overexpressioncould not be demonstrated in chronic active gastritis,chronic atrophic gastritis, subtypes of intestinalmetaplasia and dysplasia. However, in accordancewith published data,28 a weak staining reaction wasnoted in scattered cell nuclei in the proliferative zoneof the gastric foveolar epithelium and in the lower

third of intestinalized gastric tubules. Overall, 10 earlycarcinomas (22%) were cyclin D1 positive (Figure 1)with a range of positive cell nuclei between 10% and50%. In each tumour, heterogeneity in nuclear stainingintensity was noted. No significant difference was foundbetween intestinal and diffuse type tumours, beingpositive in 20% and 24% of the cases, respectively. Incases of intestinal type tumours, stratification accordingto grade of differentiation did not result in a significantdifference in cyclin D1 overexpression (one out of fourwell differentiated, two out of 11 moderately differen-tiated and one out of five poorly differentiated). Whenstratified according to lymph node status, three of the fivetumours (60%) with lymph node metastasis and sevenout of 40 tumours (18%) without lymph node metastasiswere cyclin D1 positive (P ¼ 0.065). Finally, stratifiedaccording to growth pattern, no significant differenceswere found either, namely; Pen A type 0%, Pen B type17%, mucosal type 28%, superficial type 20%.

Cyclin D1 and early gastric cancer 153

q 2000 Blackwell Science Ltd, Histopathology, 36, 151–155.

Figure 1. Intestinal type early gastric carcinoma with markednuclear cyclin D1 positivity (× 20).

F O L L O W- U P A N D S U RV I VA L

Five-year follow-up was complete in all 45 patients.Three patients died within 2 months after operation dueto mesenteric artery thrombosis (one), progressive renalfailure (one) and myocardial infarction (one). At necropsy,no evidence of residual malignancy was found. Anotherpatient died nearly 2 years after initial treatment due toa myocardial infarction without evidence of tumourrecurrence at necropsy. During follow-up three patientsdied: one died of tumour recurrence in the gastricremnant and suspected liver metastasis, 11 monthsafter initial treatment. Two years after operation,another patient died of peritonitis carcinomatosa,which was not further evaluated. A third patient died28 months later from locoregional recurrence withretroperitoneal lymph node metastasis. Univariate ana-lysis failed to show a significant difference between cyclinD1 positive and negative tumours, with 5-year survivalrates of 90% and 94%, respectively (P ¼ 0.6).

Discussion

Although, in general, there is a good correlationbetween cyclin D1 gene (CCND1) amplification andcyclin D1 protein overexpression, it appears thatmechanisms other than amplification may also resultin overexpression of cyclin D1. Nearly 30% of immuno-histochemically positive invasive breast carcinomas didnot show gene amplification, whereas similar discre-pancies have been reported in in-situ ductal breastcarcinomas. Possible mechanisms that may underliethese observations are, e.g. increased transcription ofthe CCND1 gene, or enhanced protein stability.8,29–32 Inview of the above, it may well be that immuno-histochemical analysis is better suited to explore cyclinD1 deregulation than is DNA-based analysis. In thisimmunohistochemical study, we used antibody DCS-6since this antibody has been shown to be a highlyspecific reagent for cyclin D1 detection on paraffin-embedded sections, lacking the cross-reactivity with e.g.cyclin D2 as has been reported for polyclonal anti-bodies.28 None of the precursor lesions showed cyclinD1 overexpression, indicating that cyclin D1 over-expression apparently does not play a role at thepremalignant stages of gastric carcinogenesis. Withregard to the 10 (22%) cyclin D1 positive earlycarcinomas, the immunohistochemical expression ofcyclin D1 was heterogeneous with varying nuclearintensity from cell to cell per tumour. An obviousexplanation for such variations is cell cycle-dependentoscillation as has been shown to persist in lymphomas,breast carcinomas and several tumour cell lines.33,34 No

significant difference was found between intestinal anddiffuse type carcinomas, despite their alleged differencein pathogenesis.35 As such, this study adds to data fromother studies14,36–38 showing that—at least in Cauca-sian early carcinomas—both Lauren types do sharesimilarities in molecular profile, suggesting a pathway ofacquistion of various genetic alterations, irrespective ofhistological type. Unfortunately, whereas a negativecorrelation has been described in breast carcinomasbetween cyclin D1 overexpression and histologicalgrade,27 the small numbers in our study precludesvalid statements in this regard. Finally, from a clinicalpoint of view, no significant difference in cyclin D1overexpression was found when stratified according togrowth pattern nor was there a statistically significantassociation with lymph node metastasis and 5-yearsurvival rate, although a P-value of 0.06 (lymph nodemetastasis) suggests that analysis of larger numbersmight reach statistical significance. One should takeinto consideration, though, that analogous to breastcarcinomas, a potential confounding factor in theassessment of cyclin D1 overexpression on survival ingastric carcinoma patients might be the possibility thatdifferent causes of overexpression may confer a differentimpact on survival.27

In conclusion, cyclin D1 protein overexpression doesnot play a role in the progression from normal toneoplastic gastric mucosa and does not discriminatebetween intestinal and diffuse type early gastriccarcinoma of Caucasian origin. Moreover, mechanismsother than cyclin D1 protein overexpression underliethe reported differences in biological behaviour of earlygastric carcinomas with different types of growthpattern. Finally, although it appears that cyclin D1does not have prognostic significance, studies onlarger numbers, including advanced carcinomas, arewarranted.

References1. Hunter T, Pines J. Cyclins and cancer II. Cyclin D and CDK

inhibitors come of age. Cell 1994; 79; 573–582.2. Matsushime H, Roussel MF, Ashmun RA, Sherr CJ. Colony-

stimulating factor 1 regulates novel cyclins during the G1 phaseof the cell cycle. Cell 1991; 65; 701–713.

3. Sherr CJ. Mammalian G1 cyclins. Cell 1993; 73; 1059–1065.4. Motokura T, Bloom T, Kim HG et al. A novel cyclin encoded by a

bcl-linked candidate oncogene. Nature 1991; 350; 512–515.5. Withers DA, Harvey RC, Faust JB, Melnyk O, Carey K, Meeker TC.

Characterization of a candidate bcl-1 gene. Mol. Cell Biol. 1991;11; 4846–4853.

6. Berenson JR, Koga H, Yang J et al. Frequent amplification of thebcl-1 locus in poorly differentiated squamous cell carcinoma ofthe lung. Oncogene 1990; 5; 1343–1348.

154 P Blok et al.

q 2000 Blackwell Science Ltd, Histopathology, 36, 151–155.

7. Proctor AJ, Coombs LM, Cairns JP, Knowles MA. Amplification atchromosome 11q13 in transitional cell tumors of the bladder.Oncogene 1991; 6; 789–795.

8. Buckley MF, Sweeney KJ, Hamilton JA et al. Expression andamplification of cyclin genes in human breast cancer. Oncogene1993; 8; 2127–2133.

9. Berenson JR, Yang J, Mickel RA. Frequent amplification of the bcl-1locus in head and neck squamous cell carcinomas. Oncogene1989; 4; 1111–1116.

10. Dowdy SF, Hinds PW, Louie K, Reed SI, Arnold A, Weinberg RA.Physical interaction of the retinoblastoma protein with human Dcyclins. Cell 1993; 73; 499–511.

11. Kato J, Matsushime H, Hiebert SW, Ewen ME, Sherr CJ. Directbinding of cyclin D to the retinoblastoma gene product pRb andpRb phosphorylation by the cyclin D-dependent kinase CDK4.Genes Dev. 1993; 7; 331–342.

12. Nakamura M, Katano M, Fujimoto K, Morisaki T. A newprognostic strategy for gastric carcinoma: mRNA expression oftumor growth-related factors in endoscopic biopsy specimens.Ann. Surg. 1997; 226; 35–42.

13. Akama Y, Yasui W, Yokozaki H et al. Frequent amplification of thecyclin E gene in human gastric carcinomas. Jpn J. Cancer Res.1995; 86; 617–621.

14. Craanen ME, Blok P, Top B et al. Absence of ras gene mutations inearly gastric carcinomas. Gut 1995; 37; 758–762.

15. Rodenhuis S, Slebos RJC. Clinical significance of rasoncogene activation in human lung cancer. Cancer Res. 1992;52; 2665s–2669s.

16. Suzuki Y, Orita M, Shiraishi M, Hayashi K, Sekiya N. Detection ofras gene mutation in human lung cancers by single-strandconformation polymorphism analysis of polymerase chain reac-tion products. Oncogene 1990; 5; 1037–1043.

17. Michalides RJ, van Veelen NM, Kristel PM et al. Overexpression ofcyclin D1 indicates a poor prognosis in squamous cell carcinomaof the head and neck. Arch. Otolaryngol. 1997; 123; 497–502.

18. Craanen ME, Dekker W, Ferwerda J, Blok P, Tytgat GNJ. Earlygastric cancer; a clinicopathologic study. J. Clin. Gastroenterol.1991; 31; 274–283.

19. Kodama Y, Inokuchi K, Soejima K, Matsusaka T, Okamura T.Growth patterns and prognosis in early gastric carcinoma. Cancer1983; 51; 320–326.

20. Lauren P. The two histological main types of gastric carcinoma:diffuse and so-called intestinal-type carcinoma. Acta Pathol.Microbiol. Scand. 1965; 64; 31–49.

21. Price AB. The Sydney system: histological division. J. Gastro-enterol. Hepatol. 1991; 6; 209–222.

22. Filipe MI. Histochemistry of intestinal mucins. In Whitehead R ed.Gastrointestinal and Oesophageal Pathology, 1st edn. Edinburgh,London, Melbourne, New York: Churchill Livingstone, 1989; 71.

23. Watanabe H, Jass JR, Sobin LH, eds. WHO InternationalHistological Classification of Tumors. The Histological Typing of

Oesophageal and Gastric Tumors. Berlin, Heidelberg, New York,Paris, Tokyo: Springer Verlag, 1990.

24. Murakami T. Pathomorphological diagnosis. Definition and grossclassification of early gastric cancer. In Murakami T ed. EarlyGastric Cancer. Gann Monograph on Cancer Research, Vol. 11. Tokyo:University of Tokyo Press, 1971; 53–55.

25. Hsu SM, Raine L, Fanger H. The use of the avidin–biotin–peroxidase complex (ABC) in immunoperoxidase techniques: acomparison between ABC and unlabeled antibody procedures.J. Histochem. Cytochem. 1985; 19; 577–580.

26. Michalides R, Hageman P, van Tinteren H et al. A clinicopatho-logical study on overexpression of cyclin D1 and of p53 in a seriesof 248 patients with operable breast cancer. Br. J. Cancer 1996;73; 728–734.

27. van Diest PJ, Michalides RJ, Jannink L et al. Cyclin D1 expressionin invasive breast cancer. Correlations and prognostic value. Am.J. Pathol. 1997; 150; 705–711.

28. Bartkova J, Lukas J, Strauss Bartek J. Cell cycle-related variationand tissue-restricted expression of human cyclin D1 protein.J. Pathol. 1994; 172; 237–245.

29. Gillet CE, Fantl V, Smith R et al. Amplification and overexpressionof cyclin D1 in breast cancer detected by immunohistochemicalstaining. Cancer Res. 1994; 54; 1812–1817.

30. Simpson JF, Quan DE, O’Malley F, Odom-Maryon T, Clarke PE.Amplification of CCND1 and expression of its product, cyclin D1,in ductal carcinoma in situ of the breast. Am. J. Pathol. 1997; 151;161–168.

31. Schuuring E, Verhoeven E, Mooi WJ, Michalides RJAM. Identifica-tion and cloning of two overexpressed genes, U21B31/PRAD1and EMS1, within the amplified chromosome 11q13 region inhuman carcinomas. Oncogene 1992; 7; 355–361.

32. Welcker M, Lukas J, Strauss M, Bartek J. Enhanced proteinstability: a novel mechanism of D-type cyclin overabundanceidentified in human sarcoma cells. Oncogene 1996; 13; 419–425.

33. Lukas J, Pagano M, Staskowa G, Draetta G, Bartek J. Cyclin D1protein oscillates and is essential for cell cycle progression inhuman tumor cell lines. Oncogene 1994; 9; 707–718.

34. Bartkova J, Lukas J, Muller H, Lutzhoft Strauss M, Bartek J. CyclinD1 protein overexpression and function in human breast cancer.Int. J. Cancer 1994; 57; 353–361.

35. Correa P. A human model of gastric carcinogenesis. Cancer Res.1988; 48; 1319–26.

36. Craanen ME, Blok P, Dekker W, Offerhaus GJA, Tytgat GNJ.Chronology of p53 protein accumulation in gastric carcino-genesis. Gut 1995; 36; 848–852.

37. Blok P, Craanen ME, Dekker W, Offerhaus GJA, Tytgat GNJ.Absence of c-erbB-2 overexpression in gastric carcinogenesis.Histopathology 1998; 32; 335–339.

38. Blok P, Craanen ME, Dekker W, Offerhaus GJA, Tytgat GNJ. Noevidence for inactivation of wild-type p53 by MDM2 over-expression in gastric carcinogenesis. J. Pathol. 1998; 186; 36–40.

Cyclin D1 and early gastric cancer 155

q 2000 Blackwell Science Ltd, Histopathology, 36, 151–155.