Expression of the focal adhesion protein paxillin in lung cancer and its relation to cell motility


Lung cancer can lead to abnormalities of the actin cytoskeleton structure which may be important in transformation. In this study, we have investigated the expression of the cytoskeletal associated protein paxillin in lung cancer. Paxillin is a 68 kDa focal adhesion protein, with four tandem LIM domains at the C-terminus, involved in growth factor receptor, integrin and oncogenic signaling such as v-src, BCR/ABL, and E6 of the papilloma virus. In non-small cell lung cancer (NSCLC) cell lines, paxillin localized to the focal adhesions. The possible role of paxillin in lung cancer cells was assessed by overexpressing green fluorescence protein (GFP)-paxillin construct in two separate NSCLC cell lines (Calu-1 and H661). Over the course of 48 h, GFP-paxillin consistently caused the cells to become round and to decrease cell motility as compared to normal controls, GFP-N-terminus paxillin, or GFP-LIM transfected cells. Because some lung cancers may be quite aggressive and metastasize quickly, which may be related to the cytoskeleton, we determined the expression of paxillin in NSCLC and small cell lung cancer (SCLC) cell lines and patient tumor tissues. Expression of paxillin in NSCLC and SCLC cell lines were determined by Northern blot and Western blot analysis. The expression of paxillin was consistently low in SCLC cell lines, whereas there was paxillin expression in NSCLC cell lines. There was a variability of expression of paxillin in NSCLC tumor tissue as compared to normal lung tissue. In contrast, by immunohistochemistry, we show that there was no detectable expression of paxillin in 5/5 SCLC patients. This data suggests that absence or low level of paxillin protein expression may cause certain lung cancers, such as SCLC, to be more motile and possibly more aggressive.


Lung cancer is the leading cause of death in the United States with projected 171 500 people to die from this disease in 1998 (Parker et al., 1997). Abnormalities in oncogenes, such as Ras (Slebos and Rodenhuis, 1992), Myc (Prins et al., 1993), Bcl-2 (Pezzella et al., 1993) and c-Erb-2 (Kern and Filderman, 1993) and tumor suppressor genes, such as Rb (Carbone and Kratzke, 1996), p53 (Sidransky and Hollstein, 1996), p16INK4A (Shapiro et al., 1995), may contribute to the pathogenesis of this devastating disease. The abnormalities in lung cancer may also reflect changes in the cytoskeleton and cytoskeleton associated proteins. Previously, it has been shown that transforming oncogenes such as v-src (Brown and Cooper, 1996), v-crk (Birge et al., 1993), BCR/ABL (Sattler and Salgia, 1997), and E6 (bovine and human papilloma virus) (Tong and Howley, 1997) interact with the cytoskeleton in oncogenesis. It is therefore possible that the various molecular abnormalities in lung cancer also contribute to transformation via the cytoskeleton.

The cytoskeleton is composed of multiple parts and it is the actin cytoskeleton which is involved in regulation of cell shape, cell motility and adhesion. The deregulation of cytoskeletal function contributes to transformation (Lo and Chen, 1994). The actin cytoskeleton interacts with the extracellular matrix and intracellular molecules via the focal adhesion. The focal adhesion contains proteins such as tensin (Lo et al., 1994) and vinculin (Coll et al., 1995) which are thought to also have tumor suppressor properties. Focal adhesion proteins are also key regulators of some signal transduction events (Weisberg et al., 1997).

Paxillin, a 68 kDa protein with four tandem LIM domains and multiple serine/tyrosine phosphorylation sites, is a unique focal adhesion protein with its ability to interact with certain oncogenes (Salgia et al., 1995b; Turner, 1994). Paxillin is tyrosine phosphorylated in response to oncogenes such as v-src (Schaller and Parsons, 1994), v-crk (Birge et al., 1993) and BCR/ABL (Salgia et al., 1995a). This focal adhesion protein is tyrosine phosphorylated in response to growth factors such as PDGF/EGF (Rankin et al., 1996), IL-3 and steel factor (Salgia et al., 1995b), and neuropeptides such as bombesin (Charlesworth et al., 1996). β1 and β2 integrin activation on the cell surface also lead to phosphorylation of paxillin (Charlesworth et al., 1996). Multiple kinases, including p125FAK and RAFTK, can directly interact with paxillin (Salgia et al., 1996a).

To better define the cytoskeletal abnormalities in lung cancer, we evaluated the expression of paxillin in non-small cell lung cancer (NSCLC) and small cell lung cancer (SCLC). Paxillin localized to focal adhesions with vinculin in lung cancer cell lines, and overexpression of full length paxillin in NSCLC cell lines causes cells to round up and become less motile. It is also shown here that there is paxillin expression in NSCLC cell lines and patient tumor tissue samples, with only a few tumors having abnormal paxillin expression. Whereas, there was a generalized relative decrease in expression of paxillin in SCLC – both cell lines and tumor specimens.


Localization of paxillin in focal adhesions in lung cancer cell lines

Previously, it had been shown that paxillin and vinculin localize to the focal adhesions of several adherent cell lines, such as mouse NIH3T3 and human FS-2 cells. We have performed cell staining for paxillin and vinculin in the human diploid lung line MRC-5 and various NSCLC cell lines (Calu-1 and SK-LU-1) (Figure 1). Using the anti-paxillin antibody 5H11, paxillin is localized to the focal adhesion in these cells (Salgia et al., 1996a). As a control, vinculin is in the focal adhesions as expected. This is also true for all NSCLC cell lines described in this paper (data not shown).

Figure 1

Paxillin and vinculin localize to the focal adhesions in lung cancer cell lines. Immunofluorescence of a diploid lung cell line (MRC-5) and NSCLC cell lines (Calu-1 and SK-LU-1) using anti-paxillin antibody or anti-vinculin antibody showing focal adhesion structures

In contrast to NSCLC, SCLC are much smaller cells and the actin and vinculin staining pattern is diffuse (data not shown). There was no staining seen with paxillin in SCLC cell lines.

Paxillin overexpression causes cell morphology and motility changes in NSCLC

Since paxillin appears to be expressed in adherent cell lines, we determined if there was any effect on cell morphology and cell motility by overexpression in NSCLC cell lines (calu-1 and H661). To determine the effects of overexpression of paxillin, we have constructed GFP-paxillin, GFP-PE7 (containing the N-terminus of paxillin), and GFP-LIM domains of paxillin. As seen in Figure 2, by overexpression, GFP-paxillin and GFP-PE7 during the first 24 h are localized to the cytoplasm; whereas GFP-LIM is localized both to the nucleus and cytoplasm. As a control, GFP alone is also both nuclear and cytoplasmically localized.

Figure 2

GFP-paxillin overexpression in NSCLC cell lines causes morphology and motility changes. (a) Immunofluorescence and phase-contrast microscopy of various GFP-transfected Calu-1 cells observed with time-lapse video microscopy. Calu-1 cells were either transfected with GFP-paxillin, GFP-PE7(N-terminus), GFP-LIM, or GFP vector as a control. Cells were processed and video-taped as described in Materials and methods for 48 h. (b) GFP-paxillin transfected Calu-1 cells (arrows) were followed for 48 h and each frame represents pictorial representation at 6 h intervals (b-i). The GFP-paxillin overexpressing cells rounded up, did not exhibit pseudopodia/filopodia projections, or ruffling as observed in the non-transfected cells surrounding the transfected cells. The non-transfected cells in the field exhibited normal cell motility as characterized by cell movement, membrane ruffling, formation of filopodia/lamellipodia

During the overexpression studies, when the transient transfections were observed for a 48 h or longer, the GFP-paxillin overexpressed cells became round and under time-lapse videomicroscopy had less membrane ruffling with less lamellipodia and filopodia and decreased migration. As controls, there was no change seen in cell morphology or cell motility in GFP-PE7, GFP-LIM or in the GFP overexpressed cells. These experiments have been performed at least five times each in Calu-1 and H661 NSCLC cell lines with the same results. Using another vector pRC/CMV and the various paxillin constructs, to study cell morphology and cell motility, the results were exactly the same as using the GFP vector (data not shown).

The same experiments were performed in SCLC cell lines (as shown in Figure 3). It is appreciated that SCLC cells are much smaller than NSCLC. Also, these cells tend to cluster much more, and have minimal ruffling or pseudopodia projections. Transfection of GFP-paxillin, GFP-PE7, GFP-LIM, or GFP alone does not cause any morphological changes. These results have been performed at least three times. With these imaging techniques, we are unable to appreciate gross differences between paxillin and control over-expression cells. It would be useful to make stable cell lines expressing various paxillin constructs and test homing/migration/invasiveness in an animal model.

Figure 3

GFP-paxillin overexpression in SCLC cell lines. Immunofluorescence and phase-contrast microscopy of GFP-transfected SCLC HTB-119 cells observed with time-lapse video microscopy. SCLC HTB-119 cells were either transfected with GFP-paxillin, GFP-LIM, or GFP-vector alone. Shown here are representative images taken over the course of 48 h of GFP-paxillin construct transfected cells. The transfection efficiencies are lower than that for NSCLC cell lines as shown. Arrows indicate transfected cells confirmed by GFP. GFP-paxillin overexpression effects on morphology are not appreciated due to the SCLC cells round morphology before transfection

Expression of Paxillin in NSCLC cell lines and decreased expression in SCLC cell lines

To determine the expression of paxillin in various lung cancer cell lines, we performed Northern blot analysis. Several non-small cell lung cancer (NSCLC) and small cell lung cancer (SCLC) lines were used to compare the levels of expression of paxillin and the expression of the SH2/SH3 domain containing adapter protein CRKL was utilized as a control (Figure 4). Inter-blot variability did not affect the observations. The membrane was processed for northern analysis using cDNAs of paxillin, CRKL and β-actin-specific probes sequentially. Expression of paxillin was found in the NSCLC cell lines; and in the SCLC cell lines, the expression of paxillin was lower as compared to NSCLC cell lines. As controls, CRKL (Salgia et al., 1995c) and β-actin were expressed in all cell lines at a consistent level.

Figure 4

Northern analysis of various NSCLC and SCLC cell lines using paxillin and CRKL as probes. Total RNA was isolated and northern analysis was performed as described under Materials and methods. Cell line RNA is identified above each lane. Top panel, Paxillin as the probe; middle panel, CRKL as the probe; bottom panel, β-actin as a control probe

Western blot analysis was also used to determine the expression of paxillin and several other focal adhesion proteins in NSCLC and SCLC cell lines (Figure 5). For both the NSCLC and SCLC cell lines, talin and vinculin expression was similar in all cell lines. CRKL was also expressed at consistent levels, and in the SCLC cell lines there was a doublet at 39 kDa for CRKL which may indicate phosphorylation of CRKL in these SCLC cell lines. The protein expression of paxillin was variable in the NSCLC cell lines with very low levels in 3/5 SCLC cell lines. Paxillin expression is lower, on average, in SCLC cell lines. The cell lines MRC-5, Calu-1, Calu-6, SK-LU-1, SW900 showed high levels of expression, while Calu-3 and H69 expressed an intermediate amount, and the H520, H661, H128 cell lines expressed very low levels of paxillin. In the NSCLC cell lines the p68 band of paxillin was the dominant band, with some expression of the lower bands of paxillin (p46 and p33). As shown previously, the p46 and p33 forms may be degradation forms of p68 paxillin (Salgia et al., 1997; Yamaguchi et al., 1994). In the SCLC cell lines, the dominant paxillin band was the p46 form. The p68 and p33 forms of paxillin in SCLC cell lines was expressed at very low levels.

Figure 5

Expression of paxillin is decreased in SCLC cell lines as assessed by immunoblotting. Cell lysates from the various NSCLC and SCLC cell lines were processed as described in Materials and methods. Lysates were applied to a 7.5% SDS – PAGE gel and transferred to an Immobilon-P membrane. The membrane was initially immunoblotted with anti-phosphotyrosine (4G10) antibody, and thereafter sequentially stripped and reprobed with anti-talin antibody, anti-vinculin antibody, anti-paxillin antibody, and finally with anti-CRKL antibody. There is decreased amount of p68 paxillin in most SCLC cell lines, with heterogeneity of expression in NSCLC cell lines. The p46 and p33 forms recognized by the anti-paxillin antibody may represent cleaved products of p68 paxillin. The expression of talin, vinculin, and CRKL appears to be similar in the various cell lines

Since the phosphorylation of paxillin may be important in its function, we determined the tyrosine-phosphorylation of paxillin in various cell lines. Using three NSCLC and three SCLC cell lines, there was no appreciable tyrosine phosphorylation of paxillin (Figure 6).

Figure 6

Tyrosine phosphorylation of paxillin in lung cancer cell lines. Lysates from three NSCLC and three SCLC cell lines were generated and processed as described. Paxillin immunoprecipitates were applied to a 7.5% SDS – PAGE gel and transferred to an Immobilon-P membrane. The membrane was initially immunoblotted with anti-phosphotyrosine antibody (4G10) (A) and thereafter stripped and probed with anti-paxillin antibody, clone 5H11 (B). No tyrosine phosphorylation of paxillin was observed

Expression of paxillin in NSCLC lung cancer patient samples and decreased expression in SCLC patient samples

Expression of paxillin was determined using Northern blot analysis and Western blot analysis in NSCLC patient samples. Both tumor samples and adjacent normal samples from the same lung were utilized for northern analysis and RNA probed using paxillin and β-actin-specific probe sequentially (Figure 7). Only a few adenocarcinoma, squamous, and large cell patient samples showed a variability in the expression of paxillin as compared with normal controls. The differences between lung cancer samples and adjacent normal lung samples varied as much as between different patients. As shown in Figure 7, one adenocarcinoma patient sample (#6) and one squamous cell lung cancer patient sample (#12) showed decreased paxillin expression as compared to normal control. In large cell cancer sample, four patient samples (#18, 19, 20, 21) had decreased paxillin expression and four samples (#17, 22, 23, 24) had increased paxillin expression in tumor tissue as compared to normal controls.

Figure 7

Northern analysis of paxillin expression in NSCLC tumor tissue samples. Using paired tumor (T) and adjacent normal (N) lung tissue samples from adenocarcinoma, squamous cell, and large cell lung cancer, Northern analysis was performed on total RNA. Each panel was probed using a paxillin probe and β-actin as a control probe

Western blot analysis was also used to determine protein expression of paxillin in patient samples with NSCLC (Figure 8). In the patient samples 1/4 adenocarcinoma; 4/4 squamous lung cancer; and 0/4 large cell lung cancer expressed paxillin at different levels as compared to normal controls. In the patient samples, the p68 paxillin form appears to be the dominant protein expressed. Even though in cell lines, there was some level of expression of p46 and p33 forms, in patient samples these forms are not as highly expressed. As a control, expression of CRKL was shown to be consistent in the various patient protein samples. The numbers represent individual patients. Northern blot analysis of the same tissue as Western was not performed secondary to limited availability of tissue.

Figure 8

Western blot analysis of paxillin expression in NSCLC tumor tissue samples. Using paired tumor (T) and adjacent normal (N) lung tissue samples from adenocarcinoma, squamous cell, and large cell lung cancer, immunoblotting was performed on extracted lysates. Lysates were prepared as described in Materials and methods. Lysates were applied to a SDS – PAGE 7.5% gel and transferred to an Immobilon-P membrane. The membrane was immunoblotted with anti-paxillin antibody, thereafter stripped and reprobed with anti-CRKL antibody

Finally, immunoperoxidase staining was used to determine expression of paxillin in patient tissue samples (especially SCLC). Both SCLC and NSCLC samples were used. The tissues were stained with anti-paxillin, anti-keratin and hematoxylin and eosin (H&E) (Figure 9). By immunohistochemistry, paxillin was shown to localize into cytoplasm. Lung cancer samples were positive for keratin staining, which is used as a positive control. The NSCLC tissue sample showed a localized cytoplasmic pattern for the paxillin staining, while the SCLC tissue sample showed no paxillin staining at all. We have determined immunoperoxidase staining in five separate NSCLC and SCLC patient samples each with the same results. We have seen positive reactivity with paxillin in NSCLC, whereas in the five SCLC patient samples, there is no reactivity.

Figure 9

Decreased expression of paxillin in SCLC cells by immunoperoxidase staining. Tissue samples from a patient with NSCLC or SCLC were prepared as described in Materials and methods. Slides were stained with anti-paxillin antibody, anti-keratin antibody, or hematoxylin and eosin. Appreciated is the detection of keratin in NSCLC and SCLC; however, there is no staining of paxillin in the SCLC sample


Lung cancer is a devastating illness with a plethora of abnormalities at the molecular level. With abnormalities of oncogenes and tumor suppressor genes, there are eventual abnormalities of the cytoskeleton. Changes in the cytoskeleton are very important in the carcinogenesis of lung cancer. Many cytoskeletal associated proteins have putative tumor suppressor functions. However, how cytoskeletal proteins in various tumors are expressed is just beginning to be characterized. A focal adhesion protein which may be a key regulator of oncogenesis is paxillin (Salgia et al., 1995b).

Previous groups have shown that both paxillin and vinculin localize to the focal adhesions of adherent cell lines (Turner, 1994). In NSCLC cell lines, we show that both paxillin and vinculin localize to focal adhesions. We show in this study that in GFP-paxillin overexpressed cells, localization was to the cytoplasm, while the GFP-LIM and GFP control localized to both the nucleus and the cytoplasm. After 48 h, the GFP-paxillin transfected cells were seen to round up, whereas the GFP-LIM and GFP control transfected cells expressed no change in morphology. Cell motility, as assessed by migration, membrane ruffling, and formation of lamellipodia and filopodia was inhibited by paxillin overexpression in lung cancer cell lines.

The cytoskeleton is very important in cell motility. The degradation and reformation of the actin cytoskeleton in lamellipodia and filopodia helps the cell move along the basement membrane (Lauffenburger, 1996). In normal cells, the cytoskeleton is very stable and there is little movement of cells. In cells which have become cancerous, the cytoskeleton is disrupted in such a way as to increase the motility of the cell (Sattler and Salgia, 1997). The normal cell will firmly attach to the membrane and stay in one place, while cancerous cells will increase their movement along the basement membrane due to a loss of a normal cytoskeleton. Previously it has been shown that metastatic lung cancers show changes in their cytoskeletal arrangement, such as a reduction in the amount of F-actin and a decrease in the integrity of the cytoskeleton itself (Ambros et al., 1975; Bernal et al., 1983). F-actin has been shown to collect at the periphery of the cell and in some outgrowths. Metastasizing cells tend not to form clusters and exhibit no cell-to-cell contacts, which is the opposite of normally functioning cells. The observation that overexpression of paxillin causes adherent epithelial NSCLC to change morphology with decreased motility may lead to further insights on the role of paxillin in normal and transformed cells.

Paxillin is one of few cytoskeletal proteins that has been recently shown to interact with growth factor receptors, integrins, and certain oncogenes. The paxillin protein was originally identified in Rous sarcoma virus-transformed chick embryo fibroblasts using a monoclonal antibody produced by Glenney and Zokas (1989). The protein has been cloned from human and chicken cells and proved to have an interesting structure; It has an N-terminus with many sites for interaction with proteins containing SH2 and SH3 domains, and a C-terminus that contains four tandem LIM domains (Salgia et al., 1995b; Turner and Miller, 1994). Paxillin localizes to focal adhesions where it interacts with several other proteins that also localize to this region of the cell. Tensin, talin, vinculin, and the unique tyrosine kinases p125FAK and RAFTK are but a few proteins that interact with paxillin at the focal adhesions (Bockholt and Burridge, 1993; Salgia et al., 1996a,b; Schaller et al., 1995). Paxillin can be tyrosine phosphorylated in response to many oncogenic factors, such as v-src, v-crk, BCR/ABL and HPV/BPV E6 (Bockholt and Burridge, 1995; Glenney and Zokas, 1989; Salgia et al., 1995b; Tong and Howley, 1997). Also, cytoplasmic kinases, growth factors, and β1 and β2 integrin cross-linking causes paxillin to be phosphorylated (Schaller et al., 1995). This phosphorylation facilitates the interactions between paxillin and several other proteins in the focal adhesions. In addition to the 68 kDa paxillin protein (designated α form), Mazaki et al., (1997) have reported two other isoforms of paxillin, β and γ. Isoforms β and γ have distinct amino acid insertions at the same site on paxillin. The β isoform was found to bind firmly to FAK but weakly to vinculin, while the γ isoform exhibited the reverse binding pattern. The mRNA of isoforms β and γ were not detected in normal tissue, however several cancer cell lines express both α and β isoforms simultaneously.

Paxillin expression in NSCLC and SCLC cell lines and patient samples was also determined for possible role towards carcinogenesis. In the different types of NSCLC – adenocarcinoma, squamous lung cancer, and large cell lung cancer – there was no consistent overexpression or underexpression of paxillin as compared to normal controls. The expression of paxillin was very low or non-existent in SCLC samples. Again, this may be important because SCLC tends to have different clinical and molecular biology characteristics as compared to NSCLC. SCLC are non-adherent cell lines of neuroendocrine origin, whereas NSCLC are adherent cell lines of epithelial origin. SCLC also metastasize more frequently and earlier than NSCLC.

Since paxillin expression is decreased in SCLC and there may be a few NSCLC tumor specimens exhibiting different levels of paxillin, it is conceivable that paxillin expression in certain subtypes of lung cancer could be used in prognostication. Prognostic factors are used as determinants to how widespread or aggressive a lung cancer could potentially be. Factors such as clinicopathologic (including clinical observations, tumor size, histological subtype, tumor differentiation, and lymphatic and blood vessel invasion), serum tumor markers (CA-125 and CEA) and molecular markers (oncogenes or tumor suppressors, metastatic propensity markers, differentiation markers and proliferation markers) are important prognosticators in lung cancer (Strauss, 1997). A mixture of these factors are used to determine how far along a lung cancer has progressed and can give a general idea as to how the patient may respond to different types of therapies. Theoretically, paxillin may be useful as a prognostic factor. The expression of paxillin in SCLC is very low as compared to NSCLC samples. Since SCLC is a more aggressive cancer, this data may point to the fact that paxillin is needed to keep cancerous cells from metastasizing. Future studies on paxillin expression in lung cancers should include its expression in lung cancer samples at different stages of the disease.

In conclusion, we have shown expression of paxillin in NSCLC and decreased expression in SCLC. By overexpression studies, paxillin may play an important role in cell morphology and cell motility. For future studies, it would be useful to identify if there are subsets of lung cancer which can be prognosticated based on paxillin expression. Also, it would be useful to identify which portion of paxillin may be important in cellular dynamics of motility.

Materials and methods

Cell lines, cell culture and tissue samples

The NSCLC and SCLC cell lines were obtained from the American Type Culture Collection (Rockville, MD, USA). NSCLC cells (Calu-1, Calu-3, Calu-6, SK-LU-1, SW900, H520, H661) were maintained in Dulbecco's modified Eagle's medium (DMEM) (Mediatech, Washington, DC, USA), 10% (v/v) fetal calf serum (PAA Laboratories Inc., Newport Beach, CA, USA). SCLC (H69, H128, H209, H82, H345) cells were maintained in RPMI 1640 medium (Mediatech, Washington, DC, USA), 10% (v/v) fetal calf serum. All cell lines were then incubated at 37°C with 10% CO2 (Shapiro et al., 1995). Tissue samples of normal and tumor were obtained by members of the Division of Thoracic Surgery, Brigham and Women's Hospital under IRB approved protocols.

Production of green fluorescence protein constructs

GFP-paxillin, GFP-PE7 (N-terminal paxillin without the LIM domains) and GFP-LIM expression constructs were prepared in the pEGFP-C1 vector (Invitrogen, Carlsbad, CA, USA). Full length paxillin (human, amino acids 1 – 557), PE7 (amino acids 1 – 324) and LIM domains (human, amino acids 325-557) cDNAs were released from the pGEX-3X constructs by digestion with EcoRI (Salgia et al., 1995b). They were ligated into EcoRI-digested pEGFP-C1 vectors to obtain the full length paxillin, PE7 and LIM domains GFP constructs.

Transfection and cell staining

Cells were transfected using LipoFECTAMINE (GIBCO – BRL, Gaithersburg, MD, USA) essentially as manufacturer's directions with minor modifications. Briefly, Calu-1 and H661 cells were plated at a density of 1.5×105 cells per well of a six-well tissue culture plate 24 h before transfection. Two μg of the respective pEGFP constructs and 5 μg of LipoFECTAMINE were diluted separately into 100 μl of OPTI-MEM Reduced Serum Medium (GIBCO BRL) followed by mixing of the DNA and lipid and incubation at room temperature for 60 min. During this incubation, the adherent cells were washed twice with prewarmed, OPTI-MEM medium. OPTI-MEM medium was then added to the DNA-lipid mixture to bring the volume to 1 ml, and the mixture was overlaid onto the cells for a 5 h incubate at 37°C. After 5 h, 1 ml of OPTI-MEM with 20% FCS and 1% penicillin-streptomycin was added, and the cells were incubated overnight. The cells were washed twice in complete medium, and then changed to complete DMEM with 1.0 mg/ml G418 (GIBCO – BRL) after 24 h. After an initial phase of selection, the G418 concentration was reduced to 800 μg/ml for maintenance of enriched populations of transfected cells.

For cell staining, cells were trypsinized and replated on glass coverslips, and processed for immunofluorescence analysis as has been described in detail previously (Salgia et al., 1995a).

Time-lapse video microscopy

Cells were cultured on uncoated plastic tissue culture plates (35-mm plates; Becton Dickinson Labware) in a temperature controlled chamber at 37°C in their standard media with air and CO2. The cells were examined by video microscopy utilizing a Olympus IX70 inverted microscope, Omega temperature control device, Optronics Engineering DEI-750 video camera, Olympus OEV142 TV and Panasonic AG6740 time-lapse S-VHS video recorder (Salgia et al., 1997). The digital video images were captured and printed with a Sony UP5600MD Color Video Printer.

Preparation of cell lysates and immunoblotting

Cell lines were lysed in lysis buffer (20 mM Tris, pH 8.0, 150 mM NaCl, 10% glycerol, 1% Nonidet P-40 and 0.42% NaF) containing inhibitors (10 μl of 100 mM phenylmethylsufonyl flouride, 10 μl of 100 mM Na3VO4, 5 μl of aprotinin (Sigma, St Louis, MO, USA) and 2 μl of 10 mg/ml leupeptin in final volume of 1 ml) (Salgia et al., 1995b). Protein from human tissues was prepared by extracting in TRISOLV according to the manufacturer's directions (GIBCO – BRL). Cell lysates were separated by 7.5% SDS – PAGE under reducing conditions, electrophoretically transferred to ImmobilonTM polyvinylidene diflouride (Millipore, Bedford MA, USA), and processed for immunoblotting as per established methods using the enhanced chemiluminescense technique (Amersham Corp.) (Sattler et al., 1997). Immunoblots with various antibodies are as shown under Results. Antibodies used were mouse monoclonal anti-paxillin (clone 5H11, UBI), mouse monoclonal anti-CRKL (clone 5-6), mouse monoclonal anti-phosphotyrosine (clone 4G10, a generous gift from Dr Brian Druker, Oregon Health Science University, Portland, OR, USA), mouse monoclonal anti-talin (clone 8d4, Sigma) and mouse monoclonal anti-vinculin (clone VIN-11-c, Sigma, St Louis, MO, USA).

RNA isolation and Northern analysis

Total RNA from various cell lines and human tissue was obtained by extracting in TRISOLV according to the manufacturer's directions (Gibco Life Sciences). Fifteen μg of each RNA sample was separated by electrophoresis in formaldehyde-agarose gels in 3-(N-morpholino)propane-sulfonic acid (MOPS) buffer and transferred to a Nylon membrane (GeneScreen, NEN Research Products, Boston, MA, USA). Hybridization probes used for Northern blotting were labeled with 32P by using the random primed DNA labeling kit (Boehringer Mannheim Biochemicals, Indianapolis, IN, USA). cDNAs for full length paxillin, CRKL (a generous gift of Dr John Groffen, Children's Hospital, Los Angeles, CA, USA), and β-actin were used as specific probes for northern analysis.

Immunoperoxidase staining

Immunostaining was performed on cryostat sections of tissue using a standard peroxidase-anti-peroxidase method (Sambrook et al., 1989). Cryostat sections were sequentially fixed in 2% paraformaldehyde for 5 min and methanol for 10 min. Thereafter, sections were preincubated with goat serum to block nonspecific binding and subsequently incubated with the appropriate primary antibody (mouse monoclonal anti-paxillin), rabbit anti-keratin, or control immunoglobulin. Endogenous peroxidase was subsequently blocked by immersing the cryostat sections in 1% H2O2 in methanol. Thereafter, secondary antibody was applied, followed by mouse or rabbit peroxidase-anti-peroxidase. All antibody incubations were carried out for 1 h and followed by washes with PBS/0.1% gelatin. Following antibody treatment, samples were subsequently developed with diaminobenzidine-H2O2, counterstained with hematoxylin, dehydrated in alcohols and xylene and mounted for microscopy.


  1. Ambros V, Chen L and Buchanan J. . 1975 Proc. Natl. Acad. Sci. USA 72: 3144–3148.

  2. Bernal S, Baylin S, Shaper J, Gazdar A and Chen L. . 1983 Cancer Res. 43: 1798–1808.

  3. Birge RB, Fajardo JE, Reichman C, Shoelson SE, Songyang Z, Cantley LC and Hanafusa H. . 1993 Mol. Cell. Biol. 13: 4648–4656.

    CAS  Article  Google Scholar 

  4. Bockholt SM and Burridge K. . 1993 J. Biol. Chem. 268: 14565–14567.

  5. Bockholt SM and Burridge K. . 1995 Cell Adhes. Commun. 3: 91–100.

  6. Brown M and Cooper J. . 1996 Biochim. Biophys. Acta - Rev. Ca 1287: 121–149.

  7. Carbone D and Kratzke R. . 1996 Lung cancer: Principles and practice. Mitchell J, Johnson D and Turrisi A. (eds).. Lippincott-Raven Publishers: Philadelphia. pp 107–121.

    Google Scholar 

  8. Charlesworth A, Broad S and Rozengurt E. . 1996 Oncogene 12: 1337–1345.

  9. Coll J, Benzeev A, Ezzell R, Fernandez J, Baribault H, Oshima R and Adamson E. . 1995 Proc. Natl. Acad. Sci. USA 92: 9161–9165.

  10. Glenney J and Zokas L. . 1989 J. Cell. Biol. 108: 2401–2408.

  11. Kern J and Filderman A. . 1993 Clinics Chest Med. 14: 31–41.

  12. Lauffenburger D. . 1996 Nature 383: 390–391.

    CAS  Article  Google Scholar 

  13. Lo S and Chen L. . 1994 Cancer Metas. Rev. 13: 9–24.

  14. Lo S, Weisberg E and Chen L. . 1994 Bioessays 16: 817–823.

    CAS  Article  Google Scholar 

  15. Mazaki Y, Hashimoto S and Sabe H. . 1997 J. Biol. Chem. 272: 7437–7444.

    CAS  Article  Google Scholar 

  16. Parker S, Tong T, Bolden S and Wingo P. . 1997 CA Cancer J. Clin. 47: 5–27.

  17. Pezzella F, Turley H, Kuzu I, Tungekar MF, Dunnill MS, Pierce CB, Harris A, Gatter KC and Mason DY. . 1997 New Engl. J. Med. 329: 690–694.

  18. Prins J, De Vries E and Mulder N. . 1993 Anticancer Res. 13: 1373.

  19. Rankin S, Hooshmandrad R, Claessonwelsh L and Rozengurt E. . 1996 J. Biol. Chem. 271: 7829–7834.

    CAS  Article  Google Scholar 

  20. Salgia R, Avraham S, Pisick E, Li JL, Raja S, Greenfield EA, Sattler M, Avraham H and Griffin JD. . 1996a J. Biol. Chem. 271: 31222–31226.

    CAS  Article  Google Scholar 

  21. Salgia R, Brunkhorst B, Pisick E, Li JL, Lo SH, Chen LB and Griffin JD. . 1995a Oncogene 11: 1149–1155.

  22. Salgia R, Li J-L, Ewaniuk D, Pear W, Pisick E, Burky S, Ernst T, Sattler M, Chen L and Griffin J. . 1997 J. Clin. Invest. 100: 46–57.

    CAS  Article  Google Scholar 

  23. Salgia R, Li JL, Lo SH, Brunkhorst B, Kansas GS, Sobhany ES, Sun YP, Pisick E, Hallek M, Ernst T, Tantravahi R, Chen LB and Griffin JD. . 1995b J. Biol. Chem. 270: 5039–5047.

    CAS  Article  Google Scholar 

  24. Salgia R, Pisick E, Sattler M, Li JL, Uemura N, Wong WK, Burky SA, Hirai H, Chen LB and Griffin JD. . 1996b J. Biol. Chem. 271: 25198–25203.

    CAS  Article  Google Scholar 

  25. Salgia R, Uemura N, Okuda K, Li JL, Pisick E, Satter M, de Jong R, Drucker B, Heisterkamp N, Chen LB. et al. 1995c J. Biol. Chem. 270: 29145–29150.

    CAS  Article  Google Scholar 

  26. Sambrook J, Fritsch EF and Maniatis T. . 1990 Molecular Cloning: A Laboratory Manual, Second edn. Cold Spring Harbor Laboratory Press, Cold Spring Harbor.

    Google Scholar 

  27. Sattler M and Salgia R. . 1997 Cytokine & Growth Factor Rev. 8: 63–79.

  28. Sattler M, Salgia R, Shrikhande G, Verma S, Uemura P, Law SF, Golemis EA and Griffin JD. . 1997 J. Biol. Chem. 272: 14320–14326.

    CAS  Article  Google Scholar 

  29. Schaller MD, Otey CA, Hildebrand JD and Parsons JT. . 1995 J. Cell. Biol. 130: 1181–1187.

  30. Schaller MD and Parsons JT. . 1994 Curr. Opin. Cell Biol. 6: 705–710.

  31. Shapiro G, Edwards C, Kobzik L, Godleski J, Richards W, Sugarbaker D and Rollins B. . 1995 Cancer Res. 55: 505–509.

  32. Sidransky D and Hollstein M. . 1996 Ann. Rev. Med. 47: 285–301.

  33. Slebos R and Rodenhuis S. . 1992 J. Natl. Cancer Inst. 13: 23–29.

  34. Strauss G. . 1997 Hem. Onc. Clinics North Am. 11: 409–434.

  35. Tong XA and Howley PM. . 1997 Proc. Natl. Acad. Sci. USA 94: 4412–4417.

  36. Turner CE. . 1994 Bioessays 16: 47–52.

    CAS  Article  Google Scholar 

  37. Weisberg E, Sattler M, Ewaniuk DS and Salgia R. . 1997 Crit. Rev. Oncogenesis 8: 343–358.

  38. Turner CE and Miller JT. . 1994 J. Cell. Sci. 107: 1583–1591.

  39. Yamaguchi R, Maki M, Hatanaka M and Sabe H. . 1994 FEBS Letts. 356: 114–116.

Download references


This work was partially supported by: Jose Carreras International Leukemia Foundation Fellowship FIJC-95/INT (to MS), NIH grants DK50654 (to JDG) and CA75348 (to RS).

Author information



Corresponding author

Correspondence to Ravi Salgia.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Salgia, R., Li, JL., Ewaniuk, D. et al. Expression of the focal adhesion protein paxillin in lung cancer and its relation to cell motility. Oncogene 18, 67–77 (1999).

Download citation


  • paxillin
  • lung cancer
  • cell motility

Further reading