Prognostic value of FHIT, CTNNB1, and MUC1 expression in non-–small cell lung cancer

Human Pathology (2008) 39, 126–136

www.elsevier.com/locate/humpath

Original contribution

Prognostic value of FHIT, CTNNB1, and MUC1 expression in non-small cell lung cancerB Matthias Woenckhaus MD a,⁎,1 , Johannes Merk MD b,1 , Robert Stoehr PhD c , Frank Schaeper MD d , Andreas Gaumann MD a , Karsten Wiebe MD e , Arndt Hartmann MD a , Ferdinand Hofstaedter MD a , Wolfgang Dietmaier PhD a a

Department of Pathology, University of Regensburg, D-93053 Regensburg, Germany Department of Chest Surgery, Hospital Berlin-Buch, D-13125 Berlin-Buch, Germany c Department of Urology, University of Regensburg, D-93053 Regensburg, Germany d Department of Pathology, Hospital Berlin-Buch, D-13125 Berlin-Buch, Germany e Department of Thoracic Surgery, University of Regensburg, D-93053 Regensburg, Germany b

Received 12 December 2006; revised 3 May 2007; accepted 17 May 2007

Keywords: Lung cancer; Prognosis; Immunohistochemistry

Summary Comprehensive expression analysis using microarrays has identified a number of differentially expressed genes in smoke-exposed bronchial epithelium and non–small cell lung cancers (NSCLCs). To evaluate the prognostic relevance of these proteins in NSCLCs, we used immunohistochemistry to investigate the expression of β-catenin (CTNNB1), dickkopf, Xenopus, homolog of 3 (DKK3 gene), fibroblast growth factor receptor 3 (FGFR3), fragile histidine triad (FHIT), tumor protein p53 (TP53), mucin1 (MUC1), topoisomerase II α (TOP2A), and glutathione Stransferase-Pi (GST) in a cohort of patients (n = 125). We correlated the expression data with clinicopathologic features and clinical outcome. In addition, SNaPshot multiplex assays (Applied Biosystems, Darmstadt, Germany) and restriction fragment length polymorphism analysis were used to screen for activating point mutations at the hot spots of FGFR3 in a cohort of 30 samples of NSCLC. Using Kaplan-Meier analysis, we observed significantly better overall survival in adenocarcinomas compared with squamous cell cancers (P = .049). Loss of FHIT expression showed a strong association with shorter overall survival in both histologic types of NSCLC (squamous cell cancers, P b .001; adenocarcinomas, P = .001). In adenocarcinomas, the cytoplasmic expression of β-catenin was associated with shorter survival (P = .012); MUC1 expression was associated with worse prognosis in patients with squamous cell cancers (P = .002). The nuclear staining of TP53 (P = .008) and TOP2A (P = .059) was associated with cancers without lymphonodal metastases. A correlation with positive staining of TOP2A (P = .03) and FGFR3 positivity (P = .057) was found in adenocarcinomas of male patients. Positive MUC1 stainings were associated with squamous cell cancers of male patients (P = .03). DKK3 expression did not show any significant association with clinical outcome or pathologic features. The screening of the FGFR3 sequence in lung cancers showed only wild-type sequences and did not detect mutations in the known hot spots for FGFR3 mutations. We conclude that

☆ The manuscript has been approved by all authors. The requirements for authorship have been met, and each author believes that the manuscript represents honest work. ⁎ Corresponding author. E-mail address: [email protected] (M. Woenckhaus). 1 These authors contributed equally to this article.

0046-8177/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.humpath.2007.05.027

Immunohistochemistry in NSCLC and prognosis

127

the immunohistochemical loss of FHIT expression and the positivity for β-catenin and MUC1 in NSCLC are useful prognostic markers, whereas the variable expression of TP53, TOP2A, and FGFR3 in relation to the different histologic types of NSCLC and sex of the patients is suggestive for different underlying molecular pathways. © 2008 Elsevier Inc. All rights reserved.

1. Introduction The main carcinogens such as polycyclic aromatic hydrocarbons, N-nitrosamines, and aromatic amines responsible for most lung cancers are contained in tobacco smoke [1]. However, the exact mechanisms leading to molecular alterations in bronchial epithelium in smokers and subsequent lung cancer development are still poorly understood. In previous studies, we were able to identify in smokers an overlapping signature of differentially expressed genes and allelic losses in tumor tissues and histologic inconspicuous lung tissues [2,3]. These new studies gave new insights into the molecular mechanisms of lung carcinogenesis and support the hypothesis that in smokers, the imbalance between oxidative stress and xenobiotic protective systems is responsible for the accumulation of DNA damages [4,5]. The 8 antibodies used detect the expression of gene products on the protein level, which are involved in the tobacco-induced carcinogenesis of lung cancers (β-catenin [CTNNB1]; dickkopf, Xenopus, homolog of 3 [DKK-3 gene]; fibroblast growth factor 3 [FGFR3]; fragile histidine triad [FHIT]; tumor protein p53 [TP53]; mucin1 [MUC1]; topoisomerase II α [TOP2A]; and glutathione S-transferase-Pi [GST]). GST plays an important role in detoxification by catalyzing the conjugation of hydrophobic and electrophilic compounds with reduced glutathione and was down-regulated in lung cancers. TOP2A catalyzes the relaxation of supercoiled DNA and is associated with DNA proliferation and DNA repair [6,7]. Both GST and TOP2A are involved in molecular protective systems of the cells [8]. During the smoking history, the imbalance between the smoke-induced oxidative stress and weakness of xenobiotic cellular protective systems leads to the accumulation of irreparable DNA damages [9]. The genomic locus of the FHIT gene is one of the most fragile sites of the human genome and is localized on the chromosomal region 3p14.2 [10]. Previous studies by us [11] and others found allelic loss on 3p14.2 in patients with long-term smoking history most frequently in bronchial epithelium and corresponding lung cancers [12]. Although the exact biological function of FHIT protein is still unknown, it is conceivable that FHIT plays a role in cell proliferation and is considered as an early molecular lesion in smokers. Various other studies have shown an association between the amount of cigarettes smoked and p53 mutations [13,14]. In the course of carcinogenesis, the p53 gene mutations appear early but subsequent to the genetic loss of 3p in histologically visible preneoplastic lesions and cancer tissues [15]. Gene expression profiles have also shown members of the Wnt pathway dysregulated. The Wnt pathway may occur at different levels,

including extracellular components of the signaling cascade [16]. Recently, we were able to find WIF I [17], a component of the Wnt pathway, down-regulated in lung cancers. Dickkopf 3 is another protein that also has the ability to bind Wnt proteins. DKK3 is localized at the chromosomal region 11p15. It is known as an inducer in the development of amphibian head structures and mediates its effect by antagonizing the Wnt signaling. In non–small cell lung cancer (NSCLC), allelic losses and abnormal promoter methylation were detected at the chromosomal region 11p15, where the DKK3 gene is localized [18-20]. β-catenin is a member of the Wnt signaling cascade and is related to cadherin-mediated cell-cell adhesion systems. In lung cancers, the immunohistologic loss of membranous staining and positive cytoplasmic or nuclear staining of β-catenin was described [21]. In addition, gene expression profiles found FGFR3 up-regulated in smoke-exposed lung tissue [2] and lung cancers; and recently, various studies have shown that oxidants in cigarette smoke mediate cell signaling in airways and pulmonary epithelium that regulate the expression pattern of mucins [22]. The immunohistochemical expression of 2 of those genes (DKK3, FGFR3) was completely unknown in lung cancers. The further used 6 genes showed conflicting results in the literature. Only one single immunohistochemical study found an increased expression of GST corresponding to increasing tumor volume and progressive cellular dedifferentiation [5]. In contrast to that finding, our previous study [2] has shown dysregulated RNA levels for GST in early lesions of carcinogenesis and smoke-exposed lung tissue. Our own [11] and other studies [23-26] support that molecular alterations on the FHIT region are a further early target for tobacco smoke–induced carcinogenesis. However, conflicting results still exist about the clinical and prognostic relevance of loss of FHIT in lung cancers. The loss or marked reduction in FHIT expression has shown variations between tumor types, but several studies did not show an association to worse survival from lung cancers [27,28]. In contrast to that, others reported about the negative prognostic value of loss of FHIT [29]. Conflicting results were also given for p53 [30-32]. MUC1 [33,34] and CTNNB1 [35-37] were described as worse prognostic markers. The aim of our study was to clarify the prognostic relevance of those genes in the lung cancers of our cohort.

1.1. Patients and tissue samples Our study was designed to clarify the clinical and prognostic relevance of differentially expressed genes in

128

M. Woenckhaus et al.

125 tissue samples of NSCLC by immunohistochemistry on the protein level (78 men, 47 women; mean age at diagnosis, 67 years; range, 43-87 years) of patients who underwent surgical therapy at the University Hospital Regensburg, Germany (1993-2003), or at the Department of Thoracic Surgery, Hospital Berlin-Buch, Berlin, Germany (20012003). Tumor stage, histologic type, and grade were assigned according to the International Union Against Cancer and World Health Organization grading. Sixtyeight percent (85/125) of the patients had a strongly positive smoking history (mean pack-years, 51). In 40 patients, exact information about smoking history was not available. No patient was clinically labeled as a nonsmoker. The recruitment period was from 1994 to 2005; the clinical follow-up data were documented by the regional clinical tumor registries in Regensburg and Berlin, and clinical follow-up controls were documented in the Department of Thoracic Surgery, Chest Hospital Berlin-Buch. The clinicohistopathologic data are listed in Table 1. In squamous cell cancers, the median overall survival was 28.6 months (men, 29.8 months; women, 17.1 months; median, 19 months; range, 12.923.0 months). In adenocarcinomas, the median overall survival was 35.2 months (men, 36.2 months; women, 32.2 months; median, 37 months; range, 2-74.7 months). With 60 squamous cell cancers, 44 adenocarcinomas, and 21 lung cancers with other differentiation, the distribution of the different histologic tumor types reflects approximately the epidemiological frequency of different histologic types of NSCLC. Because of the extensive immunohisto-

Table 1

Clinicopathologic data Variables

Sex

Male Female pT stage pT1 pT2 pT3 pT4 pTx pN stage pN0 pN1 pNX Grading G1 G2 G3 Histology SCC AC BAC LCLC Others Overall survival b36 mo N36 mo No exact information

Percentage Number 75 24 22 63 4 3 8 69 30 8 4 44 52 48 35 2 8 6 62 34 3

94/125 30/125 28/125 79/125 5/125 3/125 10/125 87/125 38/125 11/125 5/125 55/125 65/125 60/125 44/125 4/125 8/125 9/125 78/125 43/125 4/125

Abbreviations: SCC, squamous cell carcinoma; AC, adenocarcinoma; BAC, bronchoalveolar cancer; LCLC, large cell lung cancer.

chemical sections and limited tumor tissue for mutation analysis of FGFR3, we used paraffin-embedded tumor tissue blocks from 30 additional NSCLCs (14 squamous cell carcinomas, 16 adenocarcinomas; 23 men, 7 women; mean age at the diagnosis, 60 years; smoking history, mean 49 pack-years) collected from 2000 to 2003 at the University Hospital Regensburg. Independent histologic examination of the tumor tissue was performed according to the World Health Organization criteria to define the tumor-node-metastasis stage.

2. Materials and methods 2.1. Immunohistochemistry and DNA isolation For each case, one representative block was retrieved and serial 8-μm paraffin sections were prepared for immunohistochemistry. Immunohistochemical studies used an avidin-biotin peroxidase method with a diaminobenzidine chromatogen. Two-micrometer sections of formalin-fixed, paraffin-embedded tissue samples were incubated overnight; and after antigen retrieval (microwave oven for 35 minutes at 250 W), immunohistochemistry was carried out in a NEXES immunostainer (Ventana, Tucson, AZ) following manufacturer's instructions. The following primary antibodies were used: GST (1:50; Böhringer, Mannheim, Germany), TOP2A (1:100; Böhringer), TP53 (1: 1000; Oncogen Research Products, Cambridge, MA), FHIT (1:50; PAD: ZR44, Zymed Laboratories, San Francisco, CA), β-catenin (1: 50; Santa Cruz Biotechnology Inc, Santa Cruz, CA), DKK3 (1:50, Santa Cruz Biotechnology Inc), FGFR3 (1:50; Santa Cruz Biotechnology Inc), and MUC1 (1:4000; MA695; Novocastra, Newcastle, UK). Negative controls without primary antibody were included in each experiment. The immunohistochemical stainings did show fairly homogenous stainings in most cancers. A cutoff of 5% positive cells was used for all investigated proteins to consider a tumor positive because we also found heterogenous distribution especially in archived tissue blocks that were stored for a longer time. Deparaffinized tissue sections (5 μm) of 30 additional NSCLCs were stained with methylene blue for 15 seconds and subsequently microdissected either manually or by laser microdissection (PALM Microlaser Technology, Bernreid, Germany) by a pathologist (M. W.) to isolate pure tumor cells. DNA was extracted using the MagNa Pure LC DNA Isolation Kit II and the MagNa Pure LC with corresponding software 2.0 (Roche Diagnostics GmbH, Mannheim, Germany) following the manufacturer's instructions. These DNA probes were used for SNaPshot assays (Applied Biosystems, Darmstadt, Germany). Corresponding sections for immunohistochemical staining for FGFR3 were performed. As described above, 4-μm paraffin sections were prepared for immunohistochemistry of FGFR3.

Immunohistochemistry in NSCLC and prognosis

2.2. Mutation analysis of FGFR3 by SNaPshot assay The previously described SNaPshot multiplex assay [38] was used to screen for the most prominent activating FGFR3 point mutations (R248C, S249C, G372C, Y375C, G382R, A393E, K652E, K652M, K652Q, K652T; Fig. 3). The codons are numbered according to the open reading frame of the FGFR3IIIb isoform, which is predominantly found in epithelial cells. Two new antisense primers were added to the original assay to screen for mutations S373C (5′-T19GAGGATGCCTGCATACACAC-3′) and G382R (5′-T56GAACAGGAAGAAGCCCACCC-3′). Concentrations of the new primers used in the assay were 1.0 and 0.6 pmol/μL, respectively. The recently described G697C mutation [39] was determined using restriction fragment length polymorphism (RFLP) analysis. A polymerase chain reaction fragment (209 base pairs [bp]) containing codon 697 was amplified using sense primers: 5′-AGGTGTCTGTCCTGGGAGTCTC-3′ and 5′-GGCCAGGGATGCCACTCACAGG-3′ (annealing temperature, 64.8°C). The wild-type sequence contains one MspI site, and digestion with MspI (New England Biolabs, Frankfurt, Germany; 20 U per reaction, digestion overnight at 37°C) results in 2 bands (109 bp + 100 bp). In case of a mutation at codon 697, the fragment cannot be digested with MspI. After digestion, fragments were separated on 3% agarose gels and visualized under ultraviolet light by using 0.05% ethidium bromide.

2.3. Statistical analysis Statistical analyses were performed by using SPSS version 13.0 (SPSS, Chicago, IL). Fisher exact test was used to examine categorical data, whereas Wilcoxon rank sum test was used to examine continuous data. All tests of significance were 2-sided, and P values of less than .05 were considered as significant. Pearson χ2 test and 2-sided Fisher exact test were used to study the statistical association between clinicopathologic, immunohistochemical, and molecular data. Survival curves comparing patients with or without any of the factors were calculated using the Kaplan-Meier method, with significance evaluated by 2-sided log-rank statistics. For tumor-specific survival, patients were censored at the time of their last tumor-free clinical follow-up appointment or their date of death related or not related to the tumor.

3. Results 3.1. Expression of GST, TOP2A, TP53, FHIT, b-catenin, DKK3, MUC1, and FGFR3 and relation to clinicopathologic features The investigation gave informative results for GST in 72% (91/125; 87 cases positive, 4 cases negative), TOP2A in

129 78% (98/125; 70 cases positive, 28 cases negative), TP53 in 76% (95/125; 40 cases positive, 55 cases negative), FHIT in 66% (83/125; 70 cases positive, 13 cases negative), β-catenin in 70% (88/125; 79 cases positive, 9 cases negative), DKK3 in 64% (80/125; 56 cases positive, 24 cases negative), MUC1 in 76% (96/125; 73 cases positive, 23 cases negative), and FGFR3 in 65% (82/125; 39 cases positive, 43 cases negative) of the investigated tumors. Representative immunohistochemical stainings are shown in Fig. 1A to F. A significantly better overall survival was observed in adenocarcinomas (n = 44 cases, 24 deaths; average survival, 35 months; median, 37 months) compared with squamous cell cancers (n = 60 cases; 45 deaths; average survival, 28 months; median, 18 months) (P = .049, Fig. 2A). Statistical analysis demonstrated that positive MUC1 expression was significantly associated with worse prognosis in squamous cell carcinomas (P = .03, Fig. 2B). In both squamous cell carcinomas and adenocarcinomas, loss of FHIT expression was strongly associated with shorter overall survival (squamous cell cancers, P b .001; adenocarcinomas, P = .001; Fig. 2C, D). Positive cytoplasmic expression of β-catenin (CTNNB) in adenocarcinomas was associated with shorter survival (P = .01, Fig. 2F), whereas the β-catenin expression in squamous cell cancers did not correlate to the clinical outcome (Fig. 2E). The results of the Kaplan-Meier analyses are shown in Table 2 and Fig. 2A to F. Positive TP53 staining was more often found in adenocarcinomas without lymph node metastases than in metastasized patients (P = .008, Table 3). In addition, there was a strong tendency for positive TOP2A staining in patients with lymphonodal-negative adenocarcinomas compared with lymphonodal-positive patients (P = .059, Table 3). MUC1 expression was significantly more often detectable in adenocarcinomas than in squamous cell cancers (P = .002, Table 3). In relation to the sex of the patients, positivity for TOP2A in NSCLC was more frequent in men (P = .04, Table 3). In addition, positive DKK3 staining was predominantly seen in NSCLC of men compared with women (P = .057). The immunohistochemical results of FGFR3 showed no significant sex-related differences. The calculations were repeated separately only for patients with pT1/pT2/N0 status. Corresponding to our former results, we found worse prognosis associated with loss of FHIT expression (P b .001). The tendency was similarly directed; but perhaps mainly because of the reduced number of cases, the P values of the other stainings did not reach statistical significance (CTNNB1, P = .07; MUC1, P = .19; p53, P = .78; DKK3, P = .78; TOPA1, P = .73; GST, not available). In addition, further multivariable Cox regression models including pT status; pN status; and FHIT, CTNNB1, and MUC1 staining were performed. The calculations revealed that the expression of FHIT (P = .002; 95% confidence

130

M. Woenckhaus et al.

Fig. 1 Representative immunoreactivity for FHIT, β-catenin, DKK3, FGFR3, MUC1, GST, TOP2A, and TP53 in NSCLC. A, FHIT: strong cytoplasmic expression in squamous cell cancer and corresponding bronchial epithelia. B, β-Catenin: strong cytoplasmic expression in SCC. C, DKK3: strong cytoplasmic expression in SCC. D, FGFR3: diffuse cytoplasmic staining for FGFR3 in AC. Strong cytoplasmic positivity for MUC1 (E) and for GST (F). Strong nuclear positivity in AC for TOP2A (G) and TP53 (H). SCC indicates squamous cell carcinoma; AC, adenocarcinoma.

Immunohistochemistry in NSCLC and prognosis

131

Fig. 2 Distribution of time (months) to death as estimated by the Kaplan-Meier method. All patients with squamous cell cancer and AC (A). NSCLC with negative and positive immunohistochemistry of MUC1 (B), FHIT (C), and FGFR3 (D). The expression of β-catenin in SCC (E) and AC (F).

interval, 10.38-1.70; risk ratio, 4.211) and the pT stage (pT1 stage: P = .003; 95% confidence interval, 0.41-0.014; risk ratio, 0.75; pT2 stage: P = .009; 95% confidence

interval, 2.00-0.052; risk ratio, 0.32) were significantly associated with reduced survival. The other variables did not reach significance.

132 Table 2

M. Woenckhaus et al. Prognostic factors regarding overall survival Squamous cell cancer

Adenocarcinoma

Overall survival (mo)

Clinicohistologic data Stage Ia Stage Ib Stage IIa Stage IIb Stage IIIa Stage IIIb Stage IV Immunohistochemistry GST positive Negative TOP2A positive Negative TP53 positive Negative FHIT positive Negative CTNNB 1 positive Negative DKK3 positive Negative MUC1 positive Negative FGFR3 positive Negative

Overall survival (mo)

nanalyzable

Events

Average survival

4 28 0 16 7 2 0

1 17 0 15 7 2 0

51 33 0 15 20 5 0

32 15 0 13 13 1 0

45 3 32 15 16 28 28 7 33 5 27 11 28 18 18 20

31 3 23 12 13 21 13 7 23 4 17 7 25 7 11 16

31 17 27 29 24 28 39 10 31 28 30 29 21 39 30 25

20 16 18 22 9 22 32 10 19 20 22 19 ND 15 18 15

Median survival

P

nanalyzable

.001

NS NS NS b .001 NS NS .002 NS

Events

Average survival

Median survival

P

.008

14 14 2 8 2 1 3

3 8 1 8 2 1 1

47 33 31 14 32 26 41

34 32 12 9 15 26 ND

34 1 33 8 21 20 30 5 36 2 23 9 35 3 17 17

15 0 18 4 10 13 12 5 20 2 10 5 1 20 8 10

ND ND 33 42 37 30 40 16 35 8 37 37 35 37 33 34

ND ND 32 49 19 21 21 14 37 3 49 48 ND 37 32 32

NS NS .01 .01 NS NS NS

NOTE. Events refer to number of deaths. Abbreviations: NS, not significant; ND, not determined.

3.2. Mutation analysis of FGFR3 by SNaPshot assay All cases were successfully analyzed for FGFR3 mutations by SNaPshot assay and RFLP analysis. No mutations were detected; all cases showed wild-type sequence. The independently performed immunohistochemical staining for FGFR3 showed in the 21 informative cases 5 positive and 5 negative cases in adenocarcinomas and 8 positive and 3 negative cases in squamous cell cancers (Fig. 3A and B).

4. Discussion Recently performed microarray analyses [2,3,40] have identified a number of differentially expressed genes in smoke-exposed lung and lung cancer tissue. Our study aimed to investigate if some of these genes were related to clinical outcome in NSCLC. A well-characterized cohort of NSCLC patients with both squamous cell carcinomas and adenocarcinomas was investigated. Comparing the overall survival rates between both major histologic subgroups of NSCLC, we found adenocarcinomas associated with better prognosis than squamous cell cancers (P = .049). Most squamous cell

cancers commonly arise centrally with early connections to the branched lymphatic vessels of the lung radix, whereas the adenocarcinomas are more frequently observed as peripheral nodules that do not have such favorable conditions for tumor progress [41]. Recently, further clinical studies identified the smoking status as worse prognostic factor for patients with NSCLC. To evaluate the prognostic relevance of smokingrelated gene products in lung cancers, we compared the immunohistochemical expression of these genes with the overall survival rates of patients with NSCLC. The antibodies we used detect gene products on the protein level, which are involved in the tobacco-induced early steps of carcinogenesis. Corresponding to these, we [2,11,17] and others have found molecular alterations on the DNA and RNA level in lung cancer tissue and corresponding smokeexposed lung tissue [3,42,43]. The immunohistochemical stainings did show fairly homogenous stainings in most cancers. However, we also found heterogenous distribution especially in archived tissue blocks that were stored for a longer time. Several studies reported about a diminished immunoreactivity over time [44-46]. Little is known about the processes responsible for loss of antigenicity. Oxidations and drying are presumed mechanisms. To include in our

133

33

6 13 13

60 .057 15

8 7

49 NS 16

8 15

64 NS 7

2 16

54 31 43

9 20

50 NS 26

50 15

72 2

2 Female

Sex Male

4 2 0 LCLC

6

3

8

2 0 BAC

34

33 1

.03

12

3 2

0 3

20

21

NS

3

4 3 6 1 7 0

23

1 1

9 36

4 0

2 30

3 1

5

10

4

6 2

3

0

35

NS

30

1

3 4

2

17

17

NS

NS 18 .004 20 28 NS 27 NS 11 33 NS 5 28 7 NS 16 28 NS 32 NS 15 45

Histologic type of NSCLC SCC 3 AC 1

7

14 25

Lymph node status pN0 0 pN1, N2, N3 2

2

57

18

NS 15

10 27 27 4 37

NS 16 50

27 2

NS 6 50

18 5 8 25

.059 26 47 NS 13 60

0 pT4

1 3

0 pT3

20

1

1 2

2 2

31

.008

7

2

1 0

0 3

3 0

1 3

2 0

0

17

NS

17

4

2 1

0

44

NS

26

2 0

1

1

NS 22

13 8

32

NS 47

18 4

16

NS 37

14 6

17

NS 53

18 NS 1

6 43

20 1

11

NS 27

10 10

37

NS 47

16 5 0

23

NS

20 2 pT2

Tumor stage pT1

Negative Positive P Negative Positive P Negative Positive P Negative Positive P Negative Positive P Negative Positive P Negative Positive P Clinicopathologic Negative Positive P data

TOP2A GST Variables

Table 3

Correlation of immunohistochemical and clinicohistologic data

TP53

FHIT

CTNNB 1

DKK3

MUC1

FGFR3

Immunohistochemistry in NSCLC and prognosis

retrospective study such long time archived tissue blocks with reduced antigenicity, we used a low cutoff of 5% positive cells for all investigated proteins to consider a tumor positive. An additional scoring for immunohistochemical intensity did not seem to be valid because of the variable immunoreactivity according to the length of archived time of the used paraffin samples. Although we found the pT stage associated with worse prognosis [47], other studies based on x-ray and low-dose computed tomographic screening did not find a reduction in death from lung cancers in earlier tumor stages [48,49]. The discussion is ongoing [50-53], and several studies suggest that the early occult spread of tumor cells is most relevant in biological behavior of aggressive tumor types [47-49]. Therefore, we think it is useful to investigate the effect of immunohistologic stainings not only in small cohorts of the same pT status but also in cohorts of patients with extended range of cancer stages. However, we also repeated the calculations separately for patients with pT1/ pT2/N0 status. Corresponding to our former results, we found worse prognosis associated with loss of FHIT expression (P b .001). Perhaps mainly because of the reduced number of cases, the P values of the other stainings did not reach statistical significance; but the tendency was similarly directed. Besides the histologic type of NSCLC, the immunohistochemical loss of FHIT expression was associated with significantly shorter overall survival in squamous cell cancers and in adenocarcinomas. Identical results were reported by other groups [29,50], but there are also contradictory results [27,28]. These discrepancies may be explained by a smaller number of patients in those studies. These results were also confirmed by the multivariable Cox regression. The coding region for FHIT is located at 3p14.2 and is one of the most fragile sites of the human genome [54]. Various studies support the assumption that FHIT is an early molecular target of tobacco smoke carcinogens [10]; but recently, allelic loss at 3p14.2 and loss of FHIT staining were also described in patients with usual interstitial lung disease and no smoking history [55,56]. Although the loss of FHIT might be caused by different molecular events, the strong association of negative FHIT expression with poor prognosis shown in our study suggests that the molecular alteration of FHIT represents a relevant molecular pathway in the carcinogenesis of NSCLC. Recently, we found that allelic losses at 3p14.2 in NSCLC and adjacent bronchial epithelia are associated with an accumulation of chromosomal changes [11]. It was also shown that a synergism of the FHIT- and p53-mediated tumor suppression exists with FHIT-mediated stabilization of the p53 protein, which could be a possible explanation for the accumulation of additional genetic events after loss of FHIT during carcinogenesis in the lung [57]. The reduced membranous expression of β-catenin (CTNNB) and the abnormal accumulation in cytoplasm and/or nuclei of tumor cells in adenocarcinomas of the lung were associated with shorter survival (P = .012). Our result

134

M. Woenckhaus et al.

Fig. 3 A, Representative example of mutation analysis by SNaPshot assay. All 8 investigated mutation hot spots show wild-type sequence. Codons 249 and 652 are detected twice for internal verification. B, RFLP analysis for codon 697. Polymerase chain reaction fragments (209 bp) containing wild-type sequence in codon 697 could be digested with MspI (100 bp + 109 bp; both fragments appear as one band on the gel). In case of a mutation, the MspI site is destroyed. 1 to 14, DNA from tumor samples; 15, DNA from the urothelial cell line UROtsa (representing normal urothelium) without MspI digestion; 16, UROtsa DNA after MspI digestion; 17, negative control. All analyzed samples showed wild-type sequence for codon 697. Abbreviation: St, size standard.

corresponds well to the findings of others [58] who described abnormal β-catenin expression associated with a worse prognosis of NSCLC. Abnormal signal transduction of the Wnt pathway could be responsible for the pathologic accumulation and expression of β-catenin in NSCLC [17]. DKK3 is known as a secreted inhibitor of the Wnt signaling pathway [20]. Allelic losses and abnormal promoter methylation in the chromosomal region 11p15 where the DKK3 gene is localized were recently identified in NSCLC [59]. Corresponding to these findings, we found negative staining for DKK3 in a larger number of NSCLC. However, no statistically significant correlation was found to overall survival. Although our study includes only a couple of cases with MUC1-negative staining, the consecutive statistical analysis revealed MUC1 significantly associated with poor prognosis of squamous cell cancers (P = .002). The association of MUC1 expression with progressive cellular dedifferentiation in NSCLC was also described by other studies [33,34]. As seen on the RNA level, we found FGFR3 up-regulated on the immunohistochemical level [2] but did not find association with poor prognosis or other clinicopathologic features. The screening for the most frequent activating FGFR3 point mutations by SNaPshot assay and RFLP analysis in NSCLC did not show any changes, suggesting that causes other than point mutations are responsible for the up-regulation of FGFR3 in lung cancers [60]. Nuclear staining of TP53 was significantly more often detectable in more advanced T stages of NSCLC. The frequencies of TP53 positivity did not show significant differences between the various types of NSCLC. Although TP53 is known to be involved in various molecular mechanisms of carcinogenesis and tumor progressions corresponding to other studies, we could not demonstrate an association to overall survival [61]. Positive immunohistochemistry for TOP2A was found in adenocarcinomas without metastases, but TOP2A also did not show any correlation with overall survival of the patients. Although the expression of TOP2A was described as

strongly associated with cell proliferation [62], we did not find an association to worse prognosis. Recently, in a proteomic-based analysis, GST was described as up-regulated and as a potentially useful marker for detection of NSCLC [63]. Corresponding to this, we found a strong GST expression in all lung cancers on the protein level by immunohistochemistry but no significant correlation with clinicopathologic data. Different molecular pathways with activating epidermal growth factor receptor mutations were identified in female subjects with adenocarcinomas and negative smoking history [64]. In our patients with positive smoking history, we found positive staining for TOP2A, DKK3, FGFR3, and MUC1 more frequently in NSCLC of male than female patients. These results raise the question of whether the variable distribution of protein expression especially in male subjects with smoking history reflects different molecular pathways or is only the result of variable smoking-induced molecular damages. In summary, we identified loss of FHIT expression and the immunohistochemical cytoplasmic positivity for β-catenin and MUC1 to be associated with poor prognosis in NSCLC. The variable expression of TP53, TOP2A, FGFR3, and DKK3 in relation to the different histologic types of NSCLC and sex of the patients is suggestive for different underlying molecular pathways. Activating point mutations of FGFR3 in lung cancers, which were identified in cancers of other anatomical sites, are not present in NSCLC.

References [1] Hecht SS. Tobacco smoke carcinogens and lung cancer. J Natl Cancer Inst 1999;91:1194-210. [2] Woenckhaus M, Klein-Hitpass L, Grepmeier U, et al. Smoking and cancer-related gene expression in bronchial epithelium and non– small-cell lung cancers. J Pathol 2006;210:192-204. [3] Spira A, Beane J, Shah V, et al. Effects of cigarette smoke on the human airway epithelial cell transcriptome. Proc Natl Acad Sci U S A 2004;101:10143-8.

Immunohistochemistry in NSCLC and prognosis [4] Nakabeppu Y, Sakumi K, Sakamoto K, Tsuchimoto D, Tsuzuki T, Nakatsu Y. Mutagenesis and carcinogenesis caused by the oxidation of nucleic acids. Biol Chem 2006;387:373-9. [5] Lan Q, Mumford JL, Shen M, et al. Oxidative damage-related genes AKR1C3 and OGG1 modulate risks for lung cancer due to exposure to PAH-rich coal combustion emissions. Carcinogenesis 2004;25: 2177-81. [6] Lewis SJ, Cherry NM, Niven RM, Barber PV, Povey AC. Associations between smoking, GST genotypes and N7-methylguanine levels in DNA extracted from bronchial lavage cells. Mutat Res 2004;559:11-8. [7] Sorensen M, Autrup H, Tjonneland A, Overvad K, RaaschouNielsen O. Glutathione S-transferase T1 null-genotype is associated with an increased risk of lung cancer. Int J Cancer 2004;110:219-24. [8] Tews DS, Nissen A, Kulgen C, Gaumann AK. Drug resistance– associated factors in primary and secondary glioblastomas and their precursor tumors. J Neurooncol 2000;50:227-37. [9] Bagchi D, Bagchi M, Stohs SJ, et al. Free radicals and grape seed proanthocyanidin extract: importance in human health and disease prevention. Toxicology 2000;148:187-97. [10] Wistuba II, Behrens C, Virmani AK, et al. High resolution chromosome 3p allelotyping of human lung cancer and preneoplastic/preinvasive bronchial epithelium reveals multiple, discontinuous sites of 3p allele loss and three regions of frequent breakpoints. Cancer Res 2000;60:1949-60. [11] Grepmeier U, Dietmaier W, Merk J, et al. Deletions at chromosome 2q and 12p are early and frequent molecular alterations in bronchial epithelium and NSCLC of long-term smokers. Int J Oncol 2005;27: 481-8. [12] D'Agostini F, Izzotti A, Balansky R, Zanesi N, Croce CM, De Flora S. Early loss of Fhit in the respiratory tract of rodents exposed to environmental cigarette smoke. Cancer Res 2006;66:3936-41. [13] Slattery ML, Curtin K, Anderson K, et al. Associations between cigarette smoking, lifestyle factors, and microsatellite instability in colon tumors. J Natl Cancer Inst 2000;92:1831-6. [14] Kondo K, Tsuzuki H, Sasa M, Sumitomo M, Uyama T, Monden Y. A dose-response relationship between the frequency of p53 mutations and tobacco consumption in lung cancer patients. J Surg Oncol 1996; 61:20-6. [15] Breuer RH, Postmus PE, Smit EF. Molecular pathology of non–smallcell lung cancer. Respiration 2005;72:313-30. [16] Mazieres J, He B, You L, Xu Z, Jablons DM. Wnt signaling in lung cancer. Cancer Lett 2005;222:1-10. [17] Wissmann C, Wild PJ, Kaiser S, et al. WIF1, a component of the Wnt pathway, is down-regulated in prostate, breast, lung, and bladder cancer. J Pathol 2003;201:204-12. [18] Tsuji T, Nozaki I, Miyazaki M, et al. Antiproliferative activity of REIC/Dkk-3 and its significant down-regulation in non–small-cell lung carcinomas. Biochem Biophys Res Commun 2001;289:257-63. [19] Kobayashi K, Ouchida M, Tsuji T, et al. Reduced expression of the REIC/Dkk-3 gene by promoter-hypermethylation in human tumor cells. Gene 2002;282:151-8. [20] Hsieh SY, Hsieh PS, Chiu CT, Chen WY. Dickkopf-3/REIC functions as a suppressor gene of tumor growth. Oncogene 2004;23:9183-9. [21] Nozawa N, Hashimoto S, Nakashima Y, et al. Immunohistochemical alpha- and beta-catenin and E-cadherin expression and their clinicopathological significance in human lung adenocarcinoma. Pathol Res Pract 2006;202:639-50. [22] Poitiers DeW, Lord PW, Biles B, Whimster WF. Bronchial gland histochemistry in lungs removed for cancer. Thorax 1980;35:546-51. [23] Sozzi G, Sard L, De Gregorio L, et al. Association between cigarette smoking and FHIT gene alterations in lung cancer. Cancer Res 1997; 57:2121-3. [24] Kim JS, Kim H, Shim YM, Han J, Park J, Kim DH. Aberrant methylation of the FHIT gene in chronic smokers with early stage squamous cell carcinoma of the lung. Carcinogenesis 2004;25: 2165-71.

135 [25] Kim H, Kwon YM, Kim JS, et al. Tumor-specific methylation in bronchial lavage for the early detection of non–small-cell lung cancer. J Clin Oncol 2004;22:2363-70. [26] Zochbauer-Muller S, Wistuba II, Minna JD, Gazdar AF. Fragile histidine triad (FHIT) gene abnormalities in lung cancer. Clin Lung Cancer 2000;2:141-5. [27] Feng X, Li L, Gao Y, et al. Fhit protein expression in lung cancer studied by high-throughput tissue microarray. Bull Cancer 2007;94: E8-E11. [28] Zheng H, Tsuneyama K, Takahashi H, et al. Expression of PTEN and FHIT is involved in regulating the balance between apoptosis and proliferation in lung carcinomas. Anticancer Res 2007;27:575-81. [29] Toledo G, Sola JJ, Lozano MD, Soria E, Pardo J. Loss of FHIT protein expression is related to high proliferation, low apoptosis and worse prognosis in non–small-cell lung cancer. Mod Pathol 2004;17: 440-8. [30] Puglisi F, Minisini AM, Aprile G, et al. Balance between cell division and cell death as predictor of survival in patients with non–small-cell lung cancer. Oncology 2002;63:76-83. [31] Minami K, Saito Y, Imamura H, Okamura A. Prognostic significance of p53, Ki-67, VEGF and Glut-1 in resected stage I adenocarcinoma of the lung. Lung Cancer 2002;38:51-7. [32] Campling BG, el-Deiry WS. Clinical implications of p53 mutations in lung cancer. Methods Mol Med 2003;75:53-77. [33] Lau SK, Weiss LM, Chu PG. Differential expression of MUC1, MUC2, and MUC5AC in carcinomas of various sites: an immunohistochemical study. Am J Clin Pathol 2004;122:61-9. [34] Awaya H, Takeshima Y, Yamasaki M, Inai K. Expression of MUC1, MUC2, MUC5AC, and MUC6 in atypical adenomatous hyperplasia, bronchioloalveolar carcinoma, adenocarcinoma with mixed subtypes, and mucinous bronchioloalveolar carcinoma of the lung. Am J Clin Pathol 2004;121:644-53. [35] Retera JM, Leers MP, Sulzer MA, Theunissen PH. The expression of beta-catenin in non–small-cell lung cancer: a clinicopathological study. J Clin Pathol 1998;51:891. [36] Kase S, Sugio K, Yamazaki K, Okamoto T, Yano T, Sugimachi K. Expression of E-cadherin and beta-catenin in human non–small cell lung cancer and the clinical significance. Clin Cancer Res 2000;6: 4789-96. [37] Zhu CQ, Shih W, Ling CH, Tsao MS. Immunohistochemical markers of prognosis in non–small cell lung cancer: a review and proposal for a multiphase approach to marker evaluation. J Clin Pathol 2006;59: 790-800. [38] van Oers JM, Lurkin I, van Exsel AJ, et al. A simple and fast method for the simultaneous detection of nine fibroblast growth factor receptor 3 mutations in bladder cancer and voided urine. Clin Cancer Res 2005;11:7743-8. [39] Burger M, Denzinger S, Hammerschmied CG, et al. Elevated microsatellite alterations at selected tetranucleotides (EMAST) and mismatch repair gene expression in prostate cancer. J Mol Med 2006;84:833-41. [40] Powell CA, Spira A, Derti A, et al. Gene expression in lung adenocarcinomas of smokers and nonsmokers. Am J Respir Cell Mol Biol 2003;29:157-62. [41] Wilde J, Haenselt V, Luft D, et al. The relationship of histological type and tumor location to prognosis in 1000 patients with lung resection with special reference to adenocarcinoma. Pneumologie 1990;44:1287-93. [42] Shah V, Sridhar S, Beane J, Brody JS, Spira A. SIEGE: smoking induced epithelial gene expression database. Nucleic Acids Res 2005;33:D573-9. [43] Sanz-Ortega J, Roig F, Al-Mousa MM, et al. 17p13 (p53 locus), 5q21 (APC locus) and 9p21 (p16 locus) allelic deletions are frequently found in oral exfoliative cytology cells from smoker patients with non–small-cell lung cancer. Histol Histopathol 2007;22:541-5. [44] Olapade-Olaopa EO, Ogunbiyi JO, MacKay EH, et al. Further characterization of storage-related alterations in immunoreactivity of archival tissue sections and its implications for collaborative

136

[45]

[46]

[47]

[48]

[49] [50] [51]

[52]

[53]

[54]

M. Woenckhaus et al. multicenter immunohistochemical studies. Appl Immunohistochem Mol Morphol 2001;9:261-6. Magi-Galluzzi C, Montironi R, Prete E, Kwan PW, Delellis RA. Effect of microwave oven heating times on androgen receptor antigen retrieval from paraffin-embedded prostatic adenocarcinoma. Anticancer Res 1996;16:2931-6. Blind C, Koepenik A, Pacyna-Gengelbach M, et al. Antigenicity testing by immunohistochemistry after tissue oxidation. J Clin Pathol 2007 [Epub ahead of print]. Henschke CI, Yankelevitz DF, Libby DM, Pasmantier MW, Smith JP, Miettinen OS. Survival of patients with stage I lung cancer detected on CT screening. N Engl J Med 2006;355:1763-71. Bach PB, Jett JR, Pastorino U, Tockman MS, Swensen SJ, Begg CB. Computed tomography screening and lung cancer outcomes. JAMA 2007;297:953-61. Bach PB, Kelley MJ, Tate RC, McCrory DC. Screening for lung cancer: a review of the current literature. Chest 2003;123:72S-82S. Read C, Janes S, George J, Spiro S. Early lung cancer: screening and detection. Prim Care Respir J 2006;15:332-6. Maeda J, Inoue M, Okumura M, et al. Detection of occult tumor cells in lymph nodes from non–small cell lung cancer patients using reverse transcription–polymerase chain reaction for carcinoembryonic antigen mRNA with the evaluation of its sensitivity. Lung Cancer 2006;52:235-40. Minamiya Y, Ogawa J. The current status of sentinel lymph node mapping in non–small cell lung cancer. Ann Thorac Cardiovasc Surg 2005;11:67-72. Coello MC, Luketich JD, Litle VR, Godfrey TE. Prognostic significance of micrometastasis in non–small-cell lung cancer. Clin Lung Cancer 2004;5:214-25. Wang Y, Zhou Q, Zhang S, et al. Reduction of FHIT gene expression in primary lung cancer: relationship with the proliferation and apoptosis of tumor cell. Zhonghua Yi Xue Yi Chuan Xue Za Zhi 2001;18:178-82.

[55] Wong MP, Lam WK, Wang E, et al. Primary adenocarcinomas of the lung in nonsmokers show a distinct pattern of allelic imbalance. Cancer Res 2002;62:4464-8. [56] Uematsu K, Yoshimura A, Gemma A, et al. Aberrations in the fragile histidine triad (FHIT) gene in idiopathic pulmonary fibrosis. Cancer Res 2001;61:8527-33. [57] Nishizaki M, Sasaki J, Fang B, et al. Synergistic tumor suppression by coexpression of FHIT and p53 coincides with FHIT-mediated MDM2 inactivation and p53 stabilization in human non–small cell lung cancer cells. Cancer Res 2004;64:5745-52. [58] Choi YS, Shim YM, Kim SH, et al. Prognostic significance of Ecadherin and beta-catenin in resected stage I non–small cell lung cancer. Eur J Cardiothorac Surg 2003;24:441-9. [59] Kobayashi K, Ouchida M, Tsuji T, et al. Reduced expression of the REIC/Dkk-3 gene by promoter-hypermethylation in human tumor cells. Gene 2002;282:151-8. [60] Jebar AH, Hurstx AH, Tomlinson DC, et al. FGFR3 and Ras gene mutations are mutually exclusive genetic events in urothelial cell carcinoma. Oncogene 2005;24:5218-25. [61] Woenckhaus M, Bubendorf L, Dalquen P, et al. Nuclear and cytoplasmic Maspin expression in primary non–small cell lung cancer. J Clin Pathol 2007;60:483-6. [62] Keith WN, Tan KB, Brown R. Amplification of the topoisomerase II alpha gene in a non–small cell lung cancer cell line and characterisation of polymorphisms at the human topoisomerase II alpha and beta loci in normal tissue. Genes Chromosomes Cancer 1992;4:169-75. [63] Chen G, Gharib TG, Huang CC, et al. Proteomic analysis of lung adenocarcinoma: identification of a highly expressed set of proteins in tumors. Clin Cancer Res 2002;8:2298-305. [64] Shigematsu H, Lin L, Takahashi T, et al. Clinical and biological features associated with epidermal growth factor receptor gene mutations in lung cancers. J Natl Cancer Inst 2005;97:339-46.