Differential expression and biodistribution of cytokeratin 18 and desmoplakins in non-small cell lung carcinoma subtypes

Lung Cancer 36 (2002) 133– 141

www.elsevier.com/locate/lungcan

Differential expression and biodistribution of cytokeratin 18 and desmoplakins in non-small cell lung carcinoma subtypes Geoffrey D. Young a,b,e, Thomas S. Winokur b, Robert J. Cerfolio c, Brian A. Van Tine b,e, Louise T. Chow d, Victor Okoh a, Robert I. Garver Jr a,* a

Department of Medicine, Uni6ersity of Alabama at Birmingham, and Birmingham VAMC, 701 South 19th Street, LHRB 339, Birmingham, AL 35294, USA b Department of Pathology, Uni6ersity of Alabama at Birmingham, Birmingham, AL 35294, USA c Department of Surgery, Uni6ersity of Alabama at Birmingham, Birmingham, AL 35294, USA d Department of Biochemistry and Molecular Genetics, Uni6ersity of Alabama at Birmingham, Birmingham, AL 35294, USA e The Medical Scientist Training Program, Uni6ersity of Alabama at Birmingham, Birmingham, AL 35294, USA Received 23 July 2001; received in revised form 7 December 2001; accepted 11 December 2001

Abstract Adenocarcinoma (AC), squamous cell carcinoma (SCC) and adenosquamous carcinoma (ASC) of the lung are morphologically distinguished in part by cyto-architectural features. However, little is known about the relative expression and distribution of cyto-architectural proteins among AC, SCC and ASC. Initial microarray analysis revealed significant differences in expression of two cyto-architectural genes in AC, SCC and ASC. Desmoplakin (DP) 1 and 2, which link desmosomes to intermediate filaments, was strongly expressed in SCC relative to AC and ASC. Cytokeratin 18 (CK18), an intermediate filament that is commonly linked to desmoplakin, was strongly expressed in AC and ASC relative to SCC. Western blot analysis demonstrated that AC and ASC had abundant CK18 protein, whereas CK18 was weakly detected in SCC. DP 1 and 2 are strongly expressed in SCC and minimally expressed in AC and ASC. However, the ratio of one to the other is the same in SCC and AC, but DP2 is lost in ASC. Microscopic analysis with fluorescence-labeled antibodies for CK18 and DP 1 and 2 revealed abundant membrane localization of DP and minimal perinuclear localization of CK18 in SCC. In contrast, in both AC and ASC, the CK18 protein was diffusely distributed within the cytoplasm, and DP showed both membranous and cytoplasmic localization. In conclusion, the data here shows that AC, SCC and ASC each have specific patterns of DP 1 and 2 and CK18 gene expression, protein content and biodistribution. © 2002 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Lung neoplasms; Intermediate filaments; Desmosomes

1. Introduction Non-small cell lung cancer (NSCLC) encompasses several distinct subtypes, including adenocarcinoma (AC), squamous cell carcinoma (SCC) and adenosquamous carcinoma (ASC), which are in part identified by cyto-architectural features apparent on light microscopic evaluation [1,2]. One of the most prominent differences is intercellular bridging, which is Abbre6iations: AC, adenocarcinoma; ASC, adenosquamous carcinoma; CK, cytokeratin; DP, desmoplakin; NSCLC, non-small cell lung cancer; SCC, squamous cell carcinoma. * Corresponding author. Tel.: +1-205-934-7556; fax: + 1-205-9756969. E-mail address: [email protected] (R.I. Garver, Jr).

seen in SCC and in portions of ASC, but not observed in AC. A component of intercellular bridges are desmosomes, intercellular junctional structures involved in the maintenance of epithelial integrity. Ultrastructural analysis has shown that desmosomes appear as dense, extracellular plaques between two adjoining cells. These extracellular plaques are immediately contiguous with inner dense plaques within the cytoplasm of each of the adjoining cells (as reviewed in [3]). Desmosomes are formed by the assembly of several protein components, including transmembrane cadherins (desmogleins, desmocollins, plakoglobin) and desmoplakins [4]. The cadherins serve to link the extracellular plaques with the adjoining inner dense plaques beneath the cell membrane [3,5]. The desmoplakin

0169-5002/02/$ - see front matter © 2002 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 1 6 9 - 5 0 0 2 ( 0 1 ) 0 0 4 8 6 - X

134

G.D. Young et al. / Lung Cancer 36 (2002) 133–141

proteins are the most abundant desmosomal proteins located at the cytoplasmic portion of the desmosome [3]. Desmoplakin 1 and 2 (DP 1 and 2) are 250 and 215 kDa nonglycosylated proteins, respectively, representing splice variant protein products of one gene [6,7]. All normal desmosome bearing tissues contain DP1; the relative amount of DP2 is more variable among different tissues [8,9]. Importantly, a growing body of evidence strongly supports the notion that the intracellular cytokeratin network is linked with desmosomes via the carboxyl terminus of desmoplakin [10– 13]. Cytokeratins are intermediate filament proteins commonly present in both normal and transformed epithelial cells. More than 20 cytokeratins have been described in human tissues, and these types are broadly grouped into Type I and Type II on the basis of isoelectric points and antigenic determinants [14]. Cytokeratins (CK) form heterodimers, typically a type I and a type II, which polymerize to form mature intermediate filaments [15]. CK8, CK18 and CK19 are commonly expressed in simple and transformed epithelial cells, observations which have been exploited in both diagnostic and detection studies. Other studies have suggested that the expression of CKs are modulated by events relevant to neoplasia; such as the upregulation of CK8 synthesis by p53 [16] and downregulation of CK8 by c-Ha-ras transformation [17], as well as phosphorylation of CK8 and CK19 by protein kinase C— epsilon [18] and mediators in the epidermal growth factor pathway [19]. Studies of both cytokeratin and desmosomal proteins in NSCLC have been limited. Several reports have examined the potential utility of CKs as diagnostic (e.g. [20]) or prognostic (e.g. [21]) markers. Much less is known about the specifics of desmosomal gene expression, or desmosomal protein content and distribution in NSCLC. One study of primary NSCLC tissues found that the number of desmosomes per unit area inversely correlated with the presence of metastatic disease [22], and another found that reduced expression of plakoglobin was associated with a worse prognosis [23]. There have been no reports that have examined mRNA expression, protein content and distribution of desmosomal proteins in NSCLC, and whether these aspects of desmosomal biology differ among NSCLC subtypes. In the studies reported here, microarray analysis of primary NSCLC subtypes revealed significant differences in the gene expression patterns of CK18 and DP 1 and 2 in AC, SCC and ASC. These observations were extended by Western blot, immunohistochemistry, and immunofluorescence microscopy of DP and CK18. These studies demonstrated that these three NSCLC subtypes had distinct patterns of DP and CK18 gene expression, protein content and biodistribution.

2. Materials and methods

2.1. Cell lines and primary tissues Primary normal human bronchial epithelial (NHBE) cells were grown under conditions suggested by the supplier (Clonetics, Walkersville, MD). Primary lung cancer specimens were obtained from the operating room under the auspices of a protocol approved by the UAB IRB. Diagnosis of AC, SCC and ASC were made according to WHO/IALSC criteria [2]. A minimum of 10% of the minor subtype was required for the diagnosis of ASC. Only moderately to well differentiated specimens were used for the study, with the exception of one poorly differentiated ASC used in the Western blot analysis. The specimens were grossly dissected and flash frozen in ultra-cooled isopentane, and stored in liquid nitrogen vapors. Additional primary specimens used for immunohistochemical and immunoflourescent analyses were prepared from paraffin blocks of archival specimens. A-549, SCC-25, H520, H596 and NIH 3T3 cell lines were obtained from ATCC (Rockville, MD) and grown under conditions we have previously reported [24].

2.2. Tumor enrichment Tumor enrichment procedures were performed on all specimens as a means of minimizing the effects of non-neoplastic cellular elements on the analyses. Primary tissues were embedded in OCT medium (VWR, Atlanta, GA) at an optimum cutting temperature and 5–8 mm frozen sections applied to glass slides. Slides were stained using a standard hemotoxylin and eosin (H&E) protocol, and areas where the neoplastic cells comprised at least 80% of the cellular elements were identified as ‘tumor enriched’. The corresponding areas in the OCT blocks were excised and served as the source of ‘tumor-enriched primary tissue’.

2.3. Microarray analysis Microarray analysis identified genes that were differently expressed in AC, SCC, and ASC. Total cellular RNA was isolated from disrupted, NHBE cell line and tumor-enriched primary tissue by affinity column purification using the reagents and protocol of a kit (Mini Prep™, Qiagen, Valencia, CA). RNA was assayed for quantity and quality by ultraviolet spectroscopy (A260/ 280) and agarose gel electrophoresis. Four mg of the final total cellular RNA were used as a template to synthesize 32P-dATP labeled cDNA using array specific primers (Clontech, Palo Alto, CA) and enzymes, buffers, and instructions provided with a kit (Strip-EZ RT™, Ambion, Austin, TX). The labeled cDNAs were then hybridized to the array according to instructions

G.D. Young et al. / Lung Cancer 36 (2002) 133–141

provided by the manufacturer (Atlas Human Cancer 1.2 Microarray™, Clontech, Palo Alto, CA). Hybridization signals were the identified by autoradiography for 1 – 3 days. Analysis software provided by the manufacturer (Atlas Image v1.0a™, Clontech, Palo Alto, CA) was used to compare arrays after each autoradiogram was normalized to background. This software also averaged signals so that several different specimens of each cell type produced a composite array for comparisons between the different NSCLC subtypes.

2.4. Western blotting Western blot analysis was used as a means of qualitatively and semi-quantitatively assessing selected protein products of cyto-architectural genes shown to be differently expressed among the NSCLC subtypes by the microarray experiments. Cell lysates were made from NSCLC cell lines and tumor-enriched specimens by homogenization in a sodium dodecyl sulfate (SDS)lysis buffer containing protease inhibitors (Complete™, Roche, Indianapolis, IN). Lysates were clarified by centrifugation, and protein concentrations measured by the Bradford reaction using the reagents and protocol supplied with a kit (Protein Assay, BioRad, Hercules, CA). Equal amounts of each specimen (30 mg) were electrophoresed in 7.5% SDS-polyacrylamide gels under denaturing conditions. The size fractionated protein was electroblotted to nitrocellulose membranes. The membranes were treated with 5% non-fat dry milk in phosphate buffered saline (PBS) with 0.1% Tween-20 (PBS-T) for 1 h to block nonspecific antibody binding. The membranes were washed in PBS-T and then incubated with primary antibodies. The CK18 antibody (DC-10, Santa Cruz Biotech, Santa Cruz, CA, 1:200 dilution) is a mouse IgG1 monoclonal antibody to human CK18. The DP 1 and 2 antibody (AHP 320, Serotec, Raleigh, NC, 1:200 dilution) is a rabbit monoclonal antibody to human desmoplakin 1 and 2. Following the 1 h incubation with primary antibodies, membranes were washed and incubated with the appropriate horse radish peroxidase (HRP)-conjugated secondary antibody (Jackson ImmunoResearch, West Grove, PA, 1:2000 dilution) for 1 h and then washed with PBS-T. The bound secondary antibodies were revealed by chemilluminescent reagents (ECL™, Amersham/Pharmacia, UK) and autoradiography. Apparent molecular weights were estimated by comparison to markers (Bio-Rad, Hercules, CA). Positive controls included lysates from A-549 (human lung adenocarcinoma) cells for CK18 and SCC-25 (human oropharyngeal squamous cell carcinoma) cells for DP 1 and 2. The NIH 3T3 fibroblast cell line served as the negative control.

135

2.5. Immunohistochemistry To assess intra-tumoral protein expression and distribution, primary tumor sections were subjected to immunohistochemical analysis (IHC). IHC was performed on 5 –8 mm sections of frozen tissue, and on deparaffinized/rehydrated sections from formalin-fixed, paraffin-embedded tissue applied to polylysine-coated slides. The slides were first air-dried for 5 min and fixed with 100% methanol for 5 min. Endogenous peroxidase activity was quenched by a 25-min incubation in 100% methanol/0.3% hydrogen peroxide. Slides were stained using the Vectastain™ protocol (Vector Labs, Burlingame, CA), with a 1/200 of the same primary antibodies used for the Western blots. Slides were then stained with the peroxidase localizing chromogen diaminobenzadine (DAB) for 2 min, rinsed and counterstained in Mayer’s hematoxylin for 3 min (Sigma-Aldrich, St. Louis, MO). Controls in these experiments included incubations in DAB alone, or in PBS in lieu of primary antibody.

2.6. Fluorescent IHC and confocal microscopy Intracellular distribution of CK18 and DP 1 and 2 were simultaneously assessed using fluorochrome-labeled antibodies visualized by confocal microscopy as reported in [25], with the following modifications. About 5 –8 mm frozen sections fixed in paraformaldehyde as well as 5–8 mm deparafinized/rehydrated sections from formalin-fixed, parafin embedded sections were kept on polylysine-coated slides. Endogenous peroxidase activity was quenched by a 30-min incubation in PBS/3% hydrogen peroxide, and nonspecific antibody binding was blocked by incubation in 50% goat serum in PBS. After blocking, slides were incubated in rabbit anti-DP antibody (diluted 1:100 in PBS-T) in 50% goat serum for 1 h, washed in PBS-T and incubated for 1 h with HRP-labeled goat antirabbit secondary antibody (Jackson Immuno research, West Grove, PA) (diluted 1:100 in PBS) in 50% goat serum. After washing in PBS-T, desmoplakin was stained for 10 min by a fluorescein labeled tyramide using the reagents provided in a kit (TSA™, NEN, Boston, MA). Any remaining peroxidase activity after this step was quenched by a 15 min incubation in 3% hydrogen peroxide in PBS. These slides were subsequently stained for CK18 using the above DP protocol substituting an anti-CK18 primary antibody (diluted 1:100 in PBS) in 50% goat serum, an antimouse secondary antibody (Jackson Immuno research), and a cyanine-3 labeled tyramide (TSA™, NEN, Boston, MA). Slides were then counter-stained with DAPI, a blue fluorescent nuclear staining substrate, for 5 min, mounted with an anti-fade compound (Prolong™, Molecular Probes, Eugene, OR).

136

G.D. Young et al. / Lung Cancer 36 (2002) 133–141

All specimens had additional sections processed in parallel by the following combinations as controls for specific as well as nonspecific cross reactions between the individual primary antibodies, secondary antibodies, and tyramides: (1) primary mouse and secondary mouse/tyramide alone; (2) primary rabbit and secondary rabbit/ tyramide alone; (3) primary mouse and secondary rabbit/tyramide; (4) primary rabbit and secondary mouse/tyramide; (5) each secondary antibody alone plus tyramides; (6) both secondary antibodies together plus tyramides. Imaging was performed on a Leica DMIRBE inverted epifluorescence/Nomarski microscope outfitted with Leica TCS NT Laser Confocal optics.

3. Results Microarray analysis showed significant differences in the expression of specific cyto-architectural genes in primary NSCLC subtypes and normal human bronchial epithelial (NHBE) cells. Two genes with significant differences among AC, SCC and ASC were CK18 and DP (Fig. 1). The signals in Fig. 1 were generated from arrays of two unrelated tissues, representing each of the three NSCLC subtypes. NHBE cells were examined as an example of non-transformed respiratory epithelium. This analysis found CK18 to be strongly expressed in AC, and weakly expressed in ASC. In contrast, NHBE and SCC had negligible levels of CK18 expression. DP 1 and 2 were strongly expressed in SCCs, with minimal expression in AC and ASC. It is important to note that DP 1 and 2 protein products are derived from splice variant transcripts of a single gene, and this array probe bound both DP 1 and 2 transcripts. Thus, SCC appeared distinct from both AC and ASC in the expression of these two genes. Western blot analysis of CK18 protein content generally paralleled gene expression results (Fig. 2). In primary tissues (Fig. 2a), the dominant 48 kDa band identified by

the CK18 antibody was similar in size to previous, similar analyses of this protein [26]. Both AC and ASC had readily detectable amounts of CK18, although the intensity of the bands suggested a higher relative content of CK18 in AC. In contrast, SCC primary tissues had much lower amounts of CK18 protein. Similar findings for relative CK18 protein content were observed in lysates of cell lines derived from AC, SCC and ASC (Fig. 2b). Thus, the relative protein content based on this semiquantitative Western blot analysis of CK18 also showed that the AC and ASC were more similar to each other than to the SCC tissues and cell line. The Western blot analysis of DP 1 and 2 showed a striking difference between the NSCLC subtypes examined (Fig. 3). A dominant 250 kDa band corresponding to the predicted size of DP1 [7] was observed in all of the NSCLC primary tissues examined (Fig. 3a). The 215 kDa band corresponding to the predicted size of DP2 [7] was present in comparable amounts in AC and SCC with an intensity roughly equivalent to DP1. In marked contrast, the DP2 band was absent in two of three primary ASC, and present in much lower amounts relative to DP1 in the third primary ASC. Data from NSCLC cell lines (Fig. 3b) recapitulated what was observed in primary tissues, including a lack of DP2 in the ASC cell line. Thus, in the context of DP 1 and 2 content, AC and SCC were more similar to each other than to ASC. Immunohistochemical staining showed that CK18 and DP 1 and 2 were present within the neoplastic cells of the tissues, but their distribution and signal strength differed among subtypes (Fig. 4, one representative primary tissue of each subtype shown). Strong cytoplasmic staining of CK18 was observed within neoplastic cells of all (n= 5) the AC, similar to that shown in Fig. 4. A moderate cytoplasmic staining was detected in neoplastic cells in all (n = 5) of the ASC tumors in both squamous-like and adenocarcinoma-like areas. In con-

Fig. 1. Selected microarray signals from NHBE cell line and primary NSCLC tissue. Cell line and tissue subtype shown above, probe signal identified on left. Each column represents data from a separate primary NSCLC tumor.

G.D. Young et al. / Lung Cancer 36 (2002) 133–141

137

Fig. 2. Western blots of CK18. Shown is the dominant band identified by the CK18 primary antibody. Each lane represents protein lysate from separate tumor-enriched primary lung cancer specimens (A) or cell lines (B); the subtype is indicated by the text above.

Fig. 3. Western blots of DP 1 and 2. Shown are the dominant signals generated by the primary desmoplakin antibody binding. Each lane represents protein lysate from separate tumor-enriched primary lung cancer specimens (A) or cell lines (B); the subtype is indicated by the text above.

trast, only a sporadic perinuclear staining was detected within cells in two of five SCC, with no detectable staining in the remaining three cases (data not shown).

The DP antibody showed strong staining within the neoplastic cells of all SCCs, with moderate staining seen in all ACs and ASCs. The stromal cells did not have

G.D. Young et al. / Lung Cancer 36 (2002) 133–141

138

Fig. 4. Immunohistochemical analysis of primary NSCLC tissues. Shown are representative sections from one of each NSCLC subtype (200 × ). The ‘control’ (first left column) represents sections processed without the primary antibody to demonstrate that the secondary antibody was binding specifically. The center column shows the results of the CK18 antibody binding. The right column shows the results of the DP 1 and 2 antibody binding. Each row corresponds to roughly consecutive specimens from each primary NSCLC tumor. Antibody bound protein is identified by the dark brown staining.

detectable signals of either CK18 or DP in specimens shown in the figure, or in the additional specimens not shown here. Thus, the protein differences seen in the Western blot analysis appear to correspond to protein differences within the neoplastic cells themselves, with minimal contribution from the stromal elements present in these primary tissues. Microscopy of primary NSCLC subtypes stained with fluorochrome-tagged antibodies directed towards

CK18 and DP demonstrated differences in the intracellular distribution of these proteins. A total of seven primary tissues for each NSCLC subtype (AC, SCC, ASC) were examined by this analysis, and the results are shown in Table 1. Illumination of a flourochrometagged CK18 antibody demonstrated strong staining within the cytoplasm of all ACs and ASCs. In SCCs, however, only sporadic, perinuclear staining was observed in four of seven tissues. Illumination of a

Table 1 Summary of fluorescence results

AC SCC ASC

Strong cytoplasmic CK18 staining

Peri-nuclear CK18 staining

Membranous DP staining

Diffuse cytoplasmic DP staining

7/7 0/7 7/7

0/7 4/7 0/7

7/7 7/7 7/7

7/7 0/7 7/7

Shown is the relative number of each NSCLC tumor exhibiting characteristic staining patterns of CK18 and DP.

G.D. Young et al. / Lung Cancer 36 (2002) 133–141

139

Fig. 5. High power fluorescent labeling of a squamous-like portion of ASC and a SCC. (A) High power (1600 ×) color composite of a fluorescently labeled squamous-like region of an ASC; DAPI nuclei (blue), CK18 (red) and DP 1 and 2 (green) arrow indicates an area where DP and CK18 appear to be in proximity. (B) Corresponding hemotoxylin and eosin stained section of the region of ASC shown in (A) indicating this area of the tumor was primarily squamous-like (C) High power (1600 × ) color composite of a fluorescently labeled SCC; DAPI nuclei (blue), CK18 (red) and DP 1 and 2 (green). (D) Corresponding hemotoxalin and eosin stained section of the SCC.

fluorochrome-tagged desmoplakin antibody demonstrated abundant punctate staining along the cell membranes in SCCs. In contrast, ACs and ACSs showed both membranous and cytoplasmic DP staining. Thus, the intracellular distribution of CK18 and DP, were very similar in AC and ASC when compared with SCC. Importantly, the squamous-like portions of ASC had a DP and CK18 staining pattern different than that observed in SCC (Fig. 5). Fig. 5A shows a high power view of a squamous-like region of an ASC, showing the cytoplasmic distribution of CK18 (in red) with apparent proximity to the membranous DP (see arrow). Similar staining patterns were observed in both adenocarcinoma-like and squamous-like portions of all ASCs.

In striking contrast, Fig. 5C shows a SCC demonstrating perinuclear staining of CK18, with no apparent proximity to the membranous DP. Thus, in the context of CK18 and DP distribution, both the adeno-like and squamous-like portions of ASC were similar to AC, and distinctly different from SCC.

4. Discussion In the present study, it was found that SCC expressed the major desmosomal protein, desmoplakin, more strongly than in either AC or ASC. However, the intermediate filament gene encoding CK18 was most

140

G.D. Young et al. / Lung Cancer 36 (2002) 133–141

strongly expressed in AC, weakly expressed in ASC, with undetectable expression in SCC. The microarray used in this study included all members of the cytokeratin family including CK8, the usual paired partner of CK18, which had an expression pattern that paralleled what was seen for CK18 (data not shown). The array also contained probes for desmocollin, desmoglein and plakoglobin. Desmocollin and desmoglein were not differently expressed among the different subtypes evaluated by this microarray analysis. However, plakoglobin was found to be more strongly expressed in SCC when compared to ASC and AC. This study focused on the desmoplakin in part due to the fact that the difference in expression between these subtypes was greater than that observed for plakoglobin, as well as the fact that desmoplakin interacts with intermediate filament proteins, such as CK18, which was also found to have markedly different expression patterns among AC, SCC and ASC. The Western blot analyses did not provide precise quantitative analysis of either protein, but the apparent amounts of CK18 approximated the relative gene expression patterns observed in the microarrays: highest in the AC, less in the ASC, and least in the SCC. In several of the ACs, the CK18 band had a slightly higher molecular weight component not seen in the SCCs or ASCs, suggesting the possibility of differences in posttranslational modification of the protein in AC. The Western analysis of DP 1 and 2 showed that though both forms of desmoplakin were present in SCC and AC, DP2 was absent or present in lower amounts in ASC tissues studied in this report. This is consistent with previously reported variability in DP2 content between different tissues [8,9]. The differences in DP2 expression can be regarded as evidence that ASC is a unique subtype, rather than a mixture of squamous and adenocarcinoma cells. In this context, the microscopy data shown here demonstrated that both the squamouslike and adenocarcinoma-like regions of ASC had a similar biodistribution of CK18 and DP proteins, which was distinctly different from SCC, providing additional evidence that ASC is not simply a mixture of SCC and AC cells. Our findings are in broad agreement with recent, prior studies that have shown ASC is of clonal origin, and therefore unlikely to represent a form of multicentric lung cancer comprised of separate squamous cell and adenocarcinoma components [27,28]. The immunohistochemistry studies showed that the CK18 and DP proteins were primarily intratumoral, and little appeared to be present in surrounding stromal elements. The strong CK18 staining observed in the AC and ASC specimens compared with minimal staining in the SCC generally agrees with an earlier study that found CK 8 and 18 were present in AC, and more variably expressed in SCC [29]. Further study will be necessary in order to determine whether CK18 modu-

lates the distribution of desmoplakin in AC and ASC, or whether the expression of CK18 by itself is an important determinant of NSCLC differentiation into a specific subtype. Fluorescence microscopy demonstrated significant differences in the distribution of desmoplakin proteins among the three NSCLC subtypes examined here. In the SCC, desmoplakin was primarily localized to the vicinity of the cell membrane, consistent with the location of other desmosomal components. In contrast, the desmoplakin proteins in AC and ASC tumors had both a membranous and cytoplasmic distribution. It is unclear whether the difference in distribution of desmoplakin in AC and ASC is associated with any difference in function.

5. Conclusion It is concluded that AC, ASC and SCC can be distinguished by specific cyto-architectural features that include CK18 and DP content and biodistribution. Whether these differences confer specific clinical characteristics associated with each subtype will require further investigation.

Acknowledgements The authors would like to thank J. Michael Ruppert for assistance with tumor enrichment, Brian Streib for assistance in graphic production and the UAB Tissue Procurement Core. This study was supported in part by a VA Merit Review grant awarded to Robert I. Garver, Jr.

References [1] Travis WD, Travis LB, Devesa SS. Lung cancer. Cancer 1995;75:191 – 202. [2] Travis WD, Colby TV, Corrin B, Shimosato Y, Brambilla E, in collaboration with Sobin LH and pathologists from 14 countries. Histological typing of lung and pleural tumors, 2nd ed. In: World Health Organization. International Histological Classification of Tumors. Berlin, Heidelberg: Springer-Verlag, 1999. [3] Green KJ, Stappenbeck TS. The desmosomal plaque: role in attachment of intermediate filaments to the cell surface. In: Citi S, editor. Molecular mechanisms of epithelial cell junctions: from development to disease. R.G. Landes Company, 1994:157 – 71. [4] Kowalczyk AP, Bornslaeger EA, Norvell SM, Palka HL, Green KJ. Desmosomes: intercellular adhesive junctions specialized for attachment of intermediate filaments. Int Rev Cytol 1999;185:237 – 302. [5] North AJ, Bardsley WG, Hyam J, et al. Molecular map of the desmosomal plaque. J Cell Sci 1999;112:4325 – 36. [6] Mueller H, Franke WW. Biochemical and immunological characterization of desmoplakins I and II, the major polypeptides of the desmosomal plaque. J Mol Biol 1983;163:647 – 71.

G.D. Young et al. / Lung Cancer 36 (2002) 133–141 [7] Garrod DR. Desmosomes and cancer. Cancer Surv (Cell Adhesion and Cancer) 1995;24:97 –111. [8] Angst BD, Nilles LA, Green KJ. Desmoplakin II expression is not restricted to stratified epithelia. J Cell Sci 1990;97:247 – 57. [9] Green KJ, Jones JCR. Desmosomes and hemidesmosomes: structure and function of molecular components. FASEB J 1996;10:871 – 81. [10] Smith E, Fuchs E. Defining the interactions between intermediate filaments and desmosomes. J Cell Biol 1998;141:1229 – 41. [11] Stappenbeck TS, Lamb J, Corcoran CM, Green KJ. Phosphorylation of the desmoplakin COOH terminus negatively regulates its interaction with keratin intermediate filament networks. J Biol Chem 1994;269:29351 –4. [12] Gigi-Leitner O, Geiger B. Antigenic interrelationship between the 40-kilodalton cytokeratin polypeptide and desmoplakins. Cell Motility Cytoskeleton 1986;6:628 –39. [13] Bornslaeger EA, Corcoran CM, Stappenbeck TS, Green KJ. Breaking the connection: displacement of the desmosomal plaque protein desmoplakin from cell – cell interfaces disrupts anchorage of intermediate filament bundles and alters intercellular junction assembly. J Cell Biol 1996;134:985 –1001. [14] Schaafsma HE, Ramaekers FCS. Cytokeratin subtyping in normal and neoplastic epithelium: basic principles and diagnostic applications. Path Ann 1994;29(part 1):21 –62. [15] Cress AE, Dalton WS. Multiple drug resistance and intermediate filaments. Cancer Metast Rev 1996;15:499 –506. [16] Mukhopadhyay T, Roth JA. Functional inactivation of p53 by antisense RNA induces invasive ability of lung carcinoma cells and downregulates cytokeratin synthesis. Anticancer Res 1996;16:1683 – 90. [17] Paine TM, Fontanini G, Basolo F, Geronimo I, Elliott JW, Russo J. Mutated C-Ha-ras oncogene alters cytokeratin expression in the human breast epithelial cell line MCF-10A. Am J Pathol 1992;140:1483 –8. [18] Omary MB, Baxter GT, Chou C-F, Riopel CL, Lin WY, Strulovici B. PKC-e related kinase associates with and phosphorylates cytokeratin 8 and 18. J Cell Biol 1992;117(3):583 – 93. [19] Baribault H, Blouin R, Bourgon L, Marceau N. Epidermal growth factor-induced selective phosphorylation of cultured rat hepatocyte 55-kDa cytokeratin before filament reorganization

141

and DNA synthesis. J Cell Biol 1989;109:1665 – 75. [20] Ordonez NG. Value of cytokeratin 5/6 immunostaining in distinguishing epithelial mesothelioma of the pleura from lung adenocarcinoma. Am J Surg Pathol 1998;22(10):1215 – 21. [21] Bergqvist M, Brattstrom D, Hesselius P, et al. Cytokeratin 8 and 18 fragments measured in serum and their relation to survival in patients with non-small cell lung cancer. Anticancer Res 1999;19:1833 – 6. [22] McDonagh D, Vollmer RT, Shelburne JD. Intercellular junctions and tumor behavior in lung cancer. Modern Pathol 1991;4(4):436 – 40. [23] Pantel K, Passlick B, Vogt J, et al. Reduced expression of plakoglobin indicates an unfavorable prognosis in subsets of patients with non-small cell lung cancer. J Clin Oncol 1998;16:1407 – 13. [24] Harper E, Dang W, Lapidus RG, Garver RI Jr. Enhanced efficacy of a novel controlled release paclitaxel formulation (PACLIMER Delivery System) for local-regional therapy of lung cancer tumor nodules in mice. Clin Cancer Res 1999;5:4242 – 8. [25] Jian Y, Van Tine BA, Chein W, Shaw GM, Broker TB, Chow LT. Concordant induction of cyclin E and p21cip1 in differentiated keratinocytes by the human papillomavirus E7 protein inhibits cellular and viral DNA synthesis. Cell Growth Differen 1999;10:101 – 11. [26] Wells MJ, Hatton MWC, Hewlett B, Podor TJ, Sheffield WP, Blajchman MA. Cytokeratin 18 is expressed on the hepatocyte plasma membrane surface and interacts with thrombin-antithrombin complexes. J Biol Chem 1997;272:28574 – 81. [27] Niho S, Yokose T, Kodama T, Nishiwaki Y, Mukai K. Clonal analysis of adenosquamous carcinoma of the lung. Jpn J Cancer Res 1999;90:1244 – 7. [28] Kanazawa K, Ebina M, Ino-oka N, et al. Transition from squamous cell carcinoma to adenocarcinoma in adenosquamous carcinoma of the lung. Am J Pathol 2000;156:1289 – 98. [29] Broers JL, Ramaekers FC, Rot MK, Oostendorp T, Huysmans A, von Muijen GN. Cytokeratins in different types of human lung cancer as monitored by chain-specific antibodies. Cancer Res 1988;48:3221 – 9.