Re-expression of p16INK4a in mesothelioma cells results in ... - Nature

22 downloads 89 Views 863KB Size Report
Absence of expression of the p16IKN4a gene product is commonly observed in mesothelioma tumors and cell lines, while wild-type pRB expression is ...
Oncogene (1998) 16, 3087 ± 3095  1998 Stockton Press All rights reserved 0950 ± 9232/98 $12.00 http://www.stockton-press.co.uk/onc

Re-expression of p16INK4a in mesothelioma cells results in cell cycle arrest, cell death, tumor suppression and tumor regression Sandra P Frizelle1, Jon Grim2, Joan Zhou1, Pankaj Gupta1, David T Curiel2, Joseph Geradts3 and Robert A Kratzke1 Department of 1Medicine, Section of Hematology/Oncology, Minneapolis Veterans A€airs Medical Center and the University of Minnesota Medical School, Minneapolis, Minnesota 55417, the 2Gene Therapy Program, University of Alabama at Birmingham, Birmingham, Alabama 35294 and the 3Department of Pathology and Laboratory Medicine, University of North Carolina School of Medicine, Chapel Hill, North Carolina 27599, USA

Absence of expression of the p16IKN4a gene product is commonly observed in mesothelioma tumors and cell lines, while wild-type pRB expression is maintained. We have examined the biologic and potential therapeutic role of re-expressing p16INK4a gene product in mesothelioma cells and tumors. Following transduction with a p16INK4a expressing adenovirus (Adp16), over-expression of p16INK4a in mesothelioma cells resulted in cell cycle arrest, inhibition of pRB phosphorylation, diminished cell growth, and eventual death of the transduced cells. Expression of p16INK4a protein was accompanied by decreased expression of pRB as detected by immunoblot and immunohistochemistry. Experiments in mesothelioma xenografts demonstrated inhibition of tumor formation, tumor growth arrest and diminished tumor size and spread. p16INK4a gene product expression was also demonstrated in intraperitoneal xenografts of human mesothelioma cells. These results demonstrate that p16INK4a gene transfer may play a therapeutic role in the treatment of mesothelioma. Keywords: p16INK4a; mesothelioma; tumor suppressor

Introduction Rb and p16INK4a gene expression are inversely correlated in a variety of tumors including lung (Kratzke et al., 1996; Otterson et al., 1994; Sakaguchi et al., 1996; Shapiro et al., 1995) and bladder carcinomas (Geradts et al., 1995), as well as astrocytomas (Ueki, 1996). This inverse correlation highlights the importance of releasing Rb-mediated suppression to entry into S-phase of the cell cycle in the development of these cancers. In a similar manner, the previously reported observation that all mesothelioma cell lines and tumors appear to express wildtype RB protein is also correlated with the absence of p16INK4a gene product expression in virtually all of the mesothelioma tumors and cell lines examined (Kratzke et al., 1995). However, unlike other adult onset tumors, where accumulation of acquired genetic defects are common, mesotheliomas are marked by a relative absence of abnormalities in the common molecular target genes such as p53, ras, DCC, APC, or WT1 (Metcalf et al., 1992; Pass and Mew, 1996),

Correspondence: RA Kratzke Received 9 September 1997; revised 19 January 1998; accepted 20 January 1998

although frequent mutations in the NF2 gene have recently been reported (Bianchi et al., 1995; Sekido et al, 1995). Thus, while lack of p16INK4a expression may be one the most common, and perhaps near universal, molecular defect in mesothelioma, it may also be accompanied by relatively few other abnormalities. Correction of this single molecular defect may provide an interesting and unique biologic approach to therapy in this disease. The initial clinical presentation of mesothelioma is usually con®ned within the pleural or peritoneal cavity (Pass and Mew, 1996). This limited initial disease manifestation makes mesothelioma an attractive disease to attempt to treat with gene replacement therapies. Clinical trials with gene based therapies in mesothelioma have already began using the novel and interesting concept of targeting tumors with gene products that confer selective toxicity on recipient tumor cells to an administered drug (Treat et al, 1996). Successful gene transfer in malignant pleural e€usions has already been demonstrated in this disease in animal models using adenoviral vectors (Smythe et al., 1994). In view of the near absolute lack of e€ective treatment for this disease, we have sought to develop reexpression of p16INK4a in mesothelioma as a potential gene based treatment of this disease. Results Mesothelioma cells are easily transduced by the Adp16 vector (Figures 1 and 2) at high eciency in vitro. High levels of p16INK4a expression were achieved after even 1 hour exposure to virus. Expression of p16INK4a in the previously p16INK4a negative mesothelioma cells resulted in marked hypophosphorylation of pRB (Figure 3), correlating with cell cycle arrest at the G1\S transition as expected (DeCaprio et al., 1992) (Table 1). Interestingly, the level of pRB expression was diminished in p16INK4a re-expressing cells consistent with previous observations in other cells types (Sandig et al., 1997). This cell cycle arrest was followed by evidence of cell death when analysed in two of the mesothelioma cell lines tested (Table 2). The percentage of dead cells as measured by propidium iodide staining was increased over twofold in H2373 cells, and over ®vefold in H2461 cells, while having minimal increase in the H2087 (p16 positive/pRB positive) lung cancer cells. Cell death was seen primarily in the mesothelioma cells and correlated with growth suppression (Figure 4). The phenomenon of cell death following re-expression of p16INK4a in malignant cells expressing

p16INK4a expression in mesothelioma SP Frizelle et al

3088

wild-type p53 has been recently documented in other cells (Sandig et al., 1997). Mutations of p53 in mesothelioma tumors and cell lines have rarely been detected, although some have been identi®ed (Carbone et al., 1997; Pass and Mew, 1996). Standard immunohistochemical analysis for p53 mutations in

Figure 1 Immunoblot of Adp16 transduced mesothelioma cells. The p16INK4a negative mesothelioma cell line H2373 (lane 1) was transduced with the Adp16 vector (lanes 3 and 4). Whole cell lysates were immunoblotted with a-p16INK4a antisera. Expression of the Adp16 virus results in a slightly truncated but fully functional p16INK4a protein that migrates faster on electrophoresis than the full length protein (described in Materials and methods). H2009, a p16INK4a positive lung cancer cell line (lane 2), demonstrates the full length p16INK4a gene product. H2009 transduced with Adp16 results in expression of both forms of the p16INK4a protein (lane 5)

these mesothelioma cell lines did not show, for the most part, evidence for expression of a mutant p53 gene product, although we have not performed sequence analysis in these cell lines (data not shown). Titration of the Adp16 virus in the H2373 cells demonstrated that at an M.O.I. of approximately 102 there is a 50% suppression of cell viability at 4 days (Figure 5). A time course of the results of the Adp16 treatment of the H2373 mesothelioma cells (p16INK4a negative/pRB positive) compared with the H630 colon cancer cell line (p16INK4a positive/pRB positive) further demonstrates the marked inhibition and cell death that follows a single treatment with the Adp16 vector in the mesothelioma cell line as compared to non-mesothelioma cells (Figure 6) In order to test whether this is speci®c to p16INK4a negative mesothelioma cells will require the generation of p16INK4a transductants since no naturally occurring p16INK4a mesothelioma cell lines have been identi®ed (Kratzke et al., 1995). After demonstrating the ecient transfer of p16INK4a gene into mesothelioma cells, we wished to test the e€ect p16INK4a expression would have on cell and tumor growth in vivo. Using four di€erent mesothelioma cell lines, we demonstrated that expression of p16INK4a

a

Figure 3 pRB expression in Adp16 transduced mesothelioma cells. The pRB positive cell line, H2373, was transduced with the Adlacz virus (lanes 2 and 3) or the Adp16 virus lanes 4 and 5), and whole cell lysates of the resulting cells were immunoblotted with a-pRB antisera. The pRB negative cell line (H2009) demonstrate absence of staining (lane 1). pRB hyperphosphorylation, as demonstrated by the migration of both the p105 and p110 species (lanes 2 and 3), is inhibited in the Adp16 transduced cells (lanes 4 and 5). Lanes 4 and 5 in Figure 3 are the same transduced cells shown in Figure 1, lanes 2 and 3

Table 1 S-phase suppression

b

Mesothelioma cell lines (p167/pRb+) S G2 G0/G1

Figure 2 H2373 cells transduced with Adp16. Immunohistochemistry using an a-p16INK4a antibody in Adp16 transduced (a) and Adlacz transduced (b) cells

H2373 (control) H2373 (Adlacz)

46.8% 48.6%

9.0% 6.2%

30.2% 30.4%

H2595 (control) H2595 (Adlacz) H2595 (Adp16)

62.0% 58.7% 69.7%

9.3% 10.8% 3.6%

18.4% 20.3% 16.4%

H2461 (control) H2461 (Adlacz) H2461 (Adp16)

51.9% 59.2% 68.7%

15.9% 15.4% 4.8%

16.9% 16.8% 13.1%

H513 (control) H513 (Adlacz) H513 (Adp16)

39.5% 37.7% 50.3%

9.1% 9.7% 5.4%

33.4% 33.6% 28.1%

H2087 (control) H2087 (Adlacz) H2087 (Adp16)

Lung cancer cell line (p16+/pRb+) 29.1% 47.6% 8.8% 32.7% 8.4% 39.7% 26.2% 46.2% 9.0%

p16INK4a expression in mesothelioma SP Frizelle et al

results in markedly fewer viable cells at 72 h and 7 days as compared to Adlacz treated cells (Figure 4). This evident suppression of growth was accompanied by a large number of mesothelioma cells dying in the p16INK4a treated cells as previously discussed (Table 2). A p16INK4a positive, pRB negative cell line, the H2009

Table 2 Cell death Mesothelioma cell lines (p167/pRb+) % dead H2373 (control) H2373 (Adlacz) H2373 (Adp16)

4.2% 5.1% 12.5%

H2461 (control) H2461 (Adlacz) H2461 (Adp16)

7.8% 4.4% 36.7%

H2087 (control) H2087 (Adlacz) H2087 (Adp16)

Lung cancer cell line (p16+/pRb+)

9.9% 11.8% 13.9%

lung cancer cell line (Otterson et al., 1994), failed to demonstrate any inhibition in cell growth following similar Adp16 treatment. A p16INK4a positive, pRB positive colon cancer cell line, H630 (Otterson et al., 1994), also failed to be suppressed by p16INK4a overexpression. One of the mesothelioma cell lines, H2595, also showed evidence of some growth suppression following transduction with the Adlacz vector, although the cell number was much more diminished in the Adp16 treated cells (Figure 4b). This e€ect of the Adlacz vector on H2595 cells is of interest and perhaps represents a non-speci®c e€ect seen only on this cell line. The remaining cell lines were largely una€ected by transduction with Adlacz. In order to further delineate the mechanism of cell death that we observed following Adp16 treatment of the mesothelioma cell lines, we performed a TUNEL assay on Adp16 treated cells (Figure 7). Both H2461 and H2373 mesothelioma cells treated with Adp16 demonstrated evidence of programmed cell death (Figure 7b and d). We next wished to test if the Adp16 vector could e€ect mesothelioma cells grown in an animal model. 16107 mesothelioma cells were mixed with Adp16 virus at a M.O.I. of 100 prior to inoculation. Treatment of the mesothelioma cells with the Adp16

a

b

Figure 4 Inhibition of mesothelioma cell growth following Adp16 transduction: Four mesothelioma cell lines, H512, H2373, H2461, H2595 (p16INK4a negative, pRB positive), one lung cancer cell line H2009 (p16INK4a positive, pRB negative), and one colon cancer cell line H630 (p16INK4a positive, pRB positive) were transduced with either Adp16, Adlacz, or not transduced as indicated. Viable cells were counted at 3 days (a) or 7 days (b). Cell viability was determiend by trypan blue exclusion of trypsinized cells. Results are expressed as a percent of the nontransduced cells

Figure 5 Dose response titration of Adp16: Following exposure to an increasing multiplicity of infections (M.O.I.) were used in the H2373 mesothelioma cell line. Cell viability was assessed 3 days later by trypan blue dye exclusion

3089

p16INK4a expression in mesothelioma SP Frizelle et al

3090

virus resulted in the presence of no detectable subcutaneous tumors (data not shown). At the time of injection with the Adp16 treated cells, the same mice were injected on the contralateral ¯ank (the left ¯ank) with identical cells treated for one hour prior to injection with the Adlacz construct. The left ¯ank (Adlacz treated cells) of the mice developed tumors while no tumors were observed on the right ¯anks (Adp16 treated cells) of the mice. Although subcutaneous xenografts in athymic nude mice is a model system that has been used in other studies, we wished to extend the ®ndings to an intraperitoneal xenograft which is more analogous to human mesothelioma. A similar trend, although less striking than in the subcutaneous tumors, was observed in intraperitoneal xenografts following mixing Adp16 with mesothelioma cells prior to inoculation (Table 3). The treatment of transformed cells prior to injection is obviously not re¯ective of the manner in which established human tumors are treated. In order to demonstrate a biologic e€ect on established tumors, 16107 H2373 cells were injected into the right and left ¯anks of athymic mice and allowed to reach approximately 1 cc3 in size (approximately 21 days). After measuring the individual size of the tumors, direct injections of either Adp16 (right ¯ank tumor) or Adlacz (left ¯ank tumor) were administered three

times per week for 2 weeks (six injections per tumor). The Adp16 injected tumors were approximately 50% smaller in volume at the end of the series of injections, and showed evidence for tumor regression during the second week of injections (Figure 8). Although the

Figure 7 TUNEL assay of Adp16 treated mesothelioma cells: The H2373 (a through c) and H2461 (d through f) mesothelioma cell lines were treated with the Adp16 vector and assayed with a TUNEL assay. Adp16 treated (b and e), untreated (a and d), and etoposide treated (c and f) cells seen. Nuclear fragmentation is noted in both Adp16 and etoposide treated cells

Table 3 In vivo e€ects of Adp16 on mesothelioma intraperitoneal xenografts Treatment H2373 cells alone

Animal Score 1 2 mean

H2373 cells plus Adp16 at time of injection

1 2 3 4 5 mean

H2373 cells with six Adp16 treatments Figure 6 Cell growth after exposure to Adp6: Following exposure to a M.O.I. of 100, cell growth and viability was evaluated over a 6 day period in (a) H2373 (p16 negative/pRB positve) and (b) H630 (p16 positive/pRB positive) cells

1 2 3 mean

SD

7 9 8.0

1.4

0 1 3 3 4 2.2

1.6

6 7 7 6.7

0.6

p16INK4a expression in mesothelioma SP Frizelle et al

reason for the diminished size after 1 week of treatments is not known, it is consistent with the time course of 4 ± 6 days necessary to begin to see a maximum inhibition in vitro (Figures 4 and 6). Following resection of the tumors (1 week after the ®nal injection), subsequent analysis by immunoblot using anti-human p16INK4a antisera revealed expression of the p16INK4a gene product in the injected tumors (Figure 9). Immunohistochemistry of these same tumors also showed localized expression of human p16INK4a in the tumors (Figure 10a). Expression of p16INK4a protein was localized to approximately less than 10% of the tumor. Interestingly, the relatively focal expression of p16INK4a that might be expected from a direct, and therefore localized, injection into the body of the tumor was accompanied by an associated decrease or loss of detectable pRB expression in the newly p16INK4a positive portions of

the tumor (Figure 10b). After demonstrating that p16INK4a re-expression could result in diminished size in established tumors, we investigated if similar results could be obtained in the intraperitoneal model. Four weeks following the intraperitoneal injection of 16107 H2373 mesothelioma cells, treatment was started with intraperitoneal injection of Adp16 three times a week for 2 weeks (six intraperitoneal treatments). These mice were contrasted to control mice which received cancer cells but no treatment at all. Animals were sacri®ced 72 h after the last injection and examined by autopsy. Mice receiving the Adp16 injections had a trend towards both smaller and less di€use intraperitoneal tumors after only six intraperitoneal treatments when compared to untreated mice, but the biologic e€ect over this short time period was small, as was the number of mice tested (Table 3). Tumors from the Adp16 treated mice also demonstrated evidence of human p16INK4a expression on immunohistochemistry (Figure 11) consistent with successful intraperitoneal delivery of the p16INK4a gene.

a

Figure 8 Tumor growth following Adp16 treatment. Established subcutaneous mesothelioma tumors (n=10) were injected three times a week for 2 weeks. Results are expressed as a percentage of the tumor size at the time treatments were started. Mice received injections of either Adp16 (right ¯ank) or Adlacz (left ¯ank)

Figure 9 p16INK4a expression in tumors treated with Adp16 or Adlacz. Tumors treated with either Adp16 or Adlacz were resected and subjected to immunoblot analysis with a-human p16INK4a antibody. The Adp16 vector expresses a slightly smaller protein than the seen in the p16INK4a positive lung cancer cell line, H2009, as previously described. Three sets of tumors (right (Adp16) and left (Adlacz) ¯anks) are shown

b

Figure 10 Expression of p16INK4a and pRB gene products in tumors treated with Adp16 or Adlacz. Xenografts using the H2373 cell line (p16INK4a negative, pRB positive) were treated with either Adp16 or Adlacz for 2 weeks, resected and subjected to immunohistochemistry with either a-human p16INK4a antibody (a) or a-human pRB antibody (b). Focally positive immunostaining (arrow) of p16INK4a is demonstrated (a). The same area of tumor demonstrates lack of pRB (arrow) staining while the corresponding p16INK4a negative area retains pRB positively

3091

p16INK4a expression in mesothelioma SP Frizelle et al

3092

a

b

Figure 11 Expression of p16INK4a and pRB gene products in intraperitoneal tumors treated with Adp16 or Adlacz. Intraperitoneal H2373 xenografts (p16INK4a negative, pRB positive) were treated with either Adp16 (a) or Adlacz (b) for 2 weeks, resected and subjected to immunohistochemistry with a-human p16INK4a antibody

Discussion Mesothelioma is newly diagnosed in approximately 5000 patients in the United States each year (Antman et al., 1989). The disease is infrequently cured by aggressive surgery and radiation, but largely remains an untreatable form of cancer. Despite this grim prognosis, early metastases are rare and the tumor generally is con®ned to the pleural cavity initially, followed by local invasion of surrounding tissues. There is no standard chemotherapy for the disease, and response rates to currently available agents is very low. In view of the high frequency of absence of p16INK4a expression in the tumors, we have investigated if there is a potential role for p16INK4a gene replacement therapy in this mesothelioma. In the current study, we have demonstrated that a p16INK4a adenovirus can absolutely inhibit tumor implantation in murine xenografts, as well as lead to smaller tumor size in established mesothelioma xenografts. In addition, treatment of mesothelioma cell lines with the Adp16 vector leads to direct cell death in the cell lines tested. The possibility of treating human disease with the Adp16 vector is appealing given the limited anatomic location of mesothelioma and the initial lack of metastatic disease, and

adenoviral transduction in human subjects in this disease is already being tested (Treat et al., 1996). Mesothelioma appears to be a disease that can be approached by gene transfer treatments. Although over-expression of p16INK4a will lead to cell cycle arrest in the presence of wild-type RB protein, this would not necessarily lead to cell death and tumor shrinkage. Recently it has been reported that p16INK4amediated cell cycle arrest leads to apoptosis of transformed cells if those cells also express wild-type p53 (Sandig et al., 1997). Although a strategy employing gene transfer of both p53 and p16INK4a could be used in the setting of mesothelioma, that may prove to be unnecessary given that low rate of p53 mutations detected in mesothelioma cell lines and tumors (Carbone et al., 1997; Pass and Mew, 1996). We detected evidence of only a single cell line with a p53 mutations by immunohistochemistry, but this method of detection is subject to false negatives and false positives (Bodner et al., 1992). Nonetheless, given the current published data, it is likely that only a small portion of mesothelioma tumors will express mutant p53 protein, or fail to lack p16INK4a protein expression. Thus single gene (p16INK4a) replacement may lead to apoptosis of mesothelioma cells. The utility of p16INK4a expression in mediating cell cycle arrest and possible tumor cell apoptosis is also limited by the need for the presence of functional wildtype pRB. No mutations or rearrangements of the Rb gene have been noted in mesothelioma (De Luca et al., 1997; Kratzke et al., 1995). However, mutations in other proteins involved in the G1/S transition could also render the replacement of p16INK4a meaningless. Rearrangements or over-expression of cyclin D family members have been reported in a wide variety of cancers, but not in mesothelioma (Sherr, 1996). In a similar manner, mutations in the p16INK4a associated cyclin-dependent-kinase (cdk4) have been reported in certain kindreds of familial melanoma (Wolfel et al., 1995; Zuo et al., 1996). These reported cdk4 mutations render the kinase insensitive to p16INK4a-mediated inhibition and thus allow continued G-phase cyclin mediated phosphorylation of pRB and lack of cell cycle arrest. If similar mutations are found to occur in mesotheliomas, this would likely abrogate any strategy designed to re-express p16INK4a in the p16INK4a negative cell. No such cdk4 mutations have been reported in mesotheliomas, but currently we are working to identify if such mutations are present. Another potential mechanism of subverting the Rbmediated block to cell cycle entry involves the transforming DNA viruses and some of their associated antigens. An intriguing report documented the presence of SV40-like DNA sequences in mesothelioma (Carbone et al., 1994). It is possible if SV40 or a related virus is present in these tumors and expressing the associated large T antigen (which can bind to pRB and `inactivate' it (DeCaprio et al., 1988)), then e€orts to manipulate the G1/S transition at the molecular level may prove futile. We did not attempt to identify any such sequences in the current study. Nonetheless, expression of p16INK4a using the Adp16 vector did result in cell cycle arrest and pRB hypophosphorylation indicating that if such SV40 sequences exist in these cells, their physiologic e€ect on cdk4/cyclin/pRB binding are of limited signi®cance in these cell lines.

p16INK4a expression in mesothelioma SP Frizelle et al

However, other potential molecular targets for these viral antigens, such as p53, exist within the cell and may present alternative pathways for molecular carcinogenesis (Carbone et al., 1997). Although mesothelioma presents a disease that may potentially be approached by gene therapy, it seems unlikely that this strategy would be capable of reaching every cancer cell. Re-expression of other tumor suppressor genes have been associated with the `bystander' e€ect wherein neighboring tumor cells are killed by mechanisms dependent on their proximity to tumor cells which have been recipients of a suppressing gene vector (Fujiwara et al., 1994). It is possible that such a bystander e€ect may play a role in p16INK4amediated tumor suppression, although we have not evaluated this in the current investigation. The fact that a relatively limited number of localized injections could results in detectable tumor shrinkage (Figure 8) may be evidence of such a mechanism, but veri®cation of a true bystander e€ect awaits further study. The ultimate utility of re-expression of p16INK4a in mesothelioma will need to be established in clinical trials. We have demonstrated, however, that a p16INK4a expressing vector that can be delivered in a anatomically localized disease site has the potential of not only inhibiting cell growth and tumor implantation, but also of leading to cell death and apoptosis. This vector can successfully be used in vivo to re-express p16INK4a in mesothelioma xenografts as well. Currently, we are testing xenografts of a spectrum of human mesothelioma cell lines in the intraperitoneal model to determine if repeated injections of the p16INK4a vector can result in a survival advantage in the animals treated with p16INK4a. If these experiments prove successful, potential use of p16INK4a based gene therapies in this disease will be limited by current vector eciency. The development of more ecient or less immunogenic vector systems may allow this type of therapy to be delivered in the near future to patients with this largely untreatable and highly lethal form of cancer.

Materials and methods Cell lines and cell culture The cell lines used in this study were derived, characterized and propagated as previously published (Kratzke et al., 1995; Pass and Mew, 1996; Pass et al., 1995). All the studies used RPMI 1640 media supplemented with 10% fetal bovine serum. The mesothelioma cell lines H513, H2373, H2595, H2461, the lung cancer (NSCLC) cell lines H2009 and H2087, and the colon carcinoma cell line H630, grow as adherent cultures in standard tissue culture dishes. Immunoblotting Immunoblotting for p16INK4a and pRB was carried out as previously published (Kratzke et al., 1995). Antisera for both p16INK4a (rabbit polyclonal) and pRB (murine monoclonal) were obtained from PharMingen (San Diego, CA). Whole cell lysates from mesothelioma cell lines were prepared from 56106 cells in 1 ml of lysis bu€er (50 mM Tris-HCl pH 7.5, 250 mM NaCl, 5 mM EDTA, 0.1% nonidet P-40, 50 mM NaF, 1 mM phenylmethylsulfonyl ¯uoride), and approximately 100 mg total protein was electroblotted onto nitrocellulose following separation on a

7.5% SDS ± PAGE gel for pRB detection and 15% SDS ± PAGE gel for detection of p16INK4a. Nitrocellulose ®lters were blocked for 1 h using 5% dry milk/1% bovine serum albumin (BSA) in 16phosphate bu€ered saline (PBS) and incubated overnight at 48C in a sealed bag with either a 1 : 200 dilution of the a-RB monoclonal antibody (PharMingen, San Diego, CA) or with a 1 : 1000 dilution of an a-p16INK4a antiserum (PharMingen) in 15 ml of 5% dry milk/1% BSA in 16PBS. Detection was performed using enhanced chemiluminescence as recommended by the manufacturer (ECL, Amersham, Arlington Heights, IL) for the monoclonal RB antibody, or using 26106 c.p.m. of [125I]protein A (Amersham) for the polyclonal p16INK4a antiserum followed by exposure to XAR-5 ®lm. Immunohistochemistry 5 mm paran sections were mounted on coated glass slides, baked for 20 min at 608C, and dewaxed. Following antigen retrieval (Geradts et al., 1996), the slides were incubated for 2 h with a murine monoclonal a-human RB antibody (clone 3C8, QED, San Diego, CA) at a concentration of 1 mg/ml. For p16INK4a immunohistochemistry, sections were immunostained with either a rabbit polyclonal a-human p16INK4a antiserum (PharMingen) or a murine monoclonal a-human p16INK4a antibody (clone 175 ± 405, PharMingen). The polyclonal antisera was used at a dilution of 1 : 400 at 48C as previously described (Geradts et al., 1995; Kratzke et al., 1995). The monoclonal p16INK4a antibody was used at a concentration of 2 mg/ml at room temperature following an antigen retrieval step in hot citrate bu€er. Detection for all immunohistochemistry utilized the Vector Elite kit (Burlingame, CA) with diaminobenzidine used as the chromagen and hematoxylin as a counterstain. Only nuclear staining was accepted as evidence for tumor supressor protein expression as previously described (Geradts et al., 1995; Kratzke et al., 1995). Cell transduction Construction, propagation, and titering of the Adp16 and Adlacz viral vectors are described elsewhere (Grim et al., 1997). The Adp16 expresses a functional p16INK4a (derived from a plasmid that was a gift from Greg Otterson) that is translated from a start codon eight codons past the consensus start codon as ®rst reported (Serrano et al., 1993). The resulting protein is slightly smaller in size than the endogenous p16INK4a protein, yet remains fully functional in its capacity to inhibit cdk4-mediated phosphorylation and suppress cell growth (Kratzke et al., 1995; Parry and Peters, 1996; Serrano et al., 1993). In experiments involving gene transfer into mesothelioma cells, approximately 16105 cells were plated onto 100 mm culture dishes. The following day, Adp16 was added to cells in the presence of 2% FCS at a M.O.I. of 100 for 1 h. Media was then removed and fresh RPMI with 10% FCS was added to the cells. Cells for immunoblotting were harvested 48 h later as described. Growth suppression was evaluated at 72 h or at 7 days (depending on the experiment) by trypsinizing the culture plates and trypan blue staining the cells, followed by cell counting using a standard hematocytometer. Alternatively, in time course experiments cells were evaluated at 2, 4, and 6 days. All experiments were performed in triplicate except for cell counts of H630 and H2009 at 7 days. TUNEL assay for apoptosis The H2373 and H2461 mesothelioma cell lines were incubated with the Adp16 virus at an M.O.I. of 100 as previously described. After 96 h, the cells were rinsed twice in 16PBS, followed by a 30 min incubation with 16PBS

3093

p16INK4a expression in mesothelioma SP Frizelle et al

3094

with 1% BSA at room temperature. The cells were then incubated for 60 min at 378C in a humidi®ed chamber in the dark with either the TUNEL (In Situ Cell Death Detection Kit, Boehringer Mannheim) reaction mixture (TdT enzyme + FITC conjugated dUTP) or a negative control mixture (without TdT enzyme). Monolayers were washed three times with PBS and mounted using Fluoromont G. Fluoresecently labeled nuclei of apoptotic cells were visualized using a ¯uorescent microscope. Untreated cells were used for a negative control. As a positive control, cells were treated with the apoptosis inducing chemotherapeutic agent etoposide (Bristol-Meyers Oncology) at a concentration of 10 mg/ml (Chen et al., 1997). Flow cytometry 16105 cells were incubated with 161010 p.f.u./ml in RPMI 1640 media plus 2% calf serum for 60 min at 378C, 5% CO2 with intermittent shaking. After incubation, 10 ml of complete media was added to plates. Cells were harvested 72 h later by trypsinisation. The cells were washed twice in 16PBS. The cell pellet was resuspended in a 1 ml solution of sodium citrate (0.1%) and Triton-X (0.1%). Propidiun iodide (Sigma) (50 mg/ml) and DNAse free RNAse (Boehringer Mannheim) (1 mg/ml), were added to the suspension and incubated at 378C for 30 min. Cell suspension vials were wrapped in aluminum foil and kept at 48C for 1 ± 4 h. Flow cytometry and analysis were performed on a Becton Dickson FACscan. The results reported were repeated twice. Mouse xenografts 16107 mesothelioma cells (H2373) were used for all injections to generate both subcutaneous and intraperitoneal tumors. Experiments using mesothelioma cells that were transduced for 1 h prior to injection were performed by mixing the mesothelioma cell with either Adp16 or Adlacz at a M.O.I. of 100 for 1 h in RPMI with 2% fetal calf serum at room temperature, washing them once in RPMI with 10% calf serum, then twice in RPMI alone, and injecting them subcutaneously in 300 ml of RPMI into the rear ¯ank of athymic nude mice (Jackson Laboratory). An aliquot of the cells were assessed by trypan blue exclusion prior to injection to assure continued cell viability after Adp16 and Adlacz treatment. Adp16 treated cells were injected into the right ¯ank, while the left ¯ank received the Adlacz treated cells. A total of 10 mice were treated. When tumors developed to the size of approximately 1 cc3 (on the left ¯ank only as described in the

Results), the animals were sacri®ced and tumors were resected for both immunohistochemistry and immunoblotting. In experiments in which established tumors were treated, 10 mice were inoculated with 16107 mesothelioma cells 3 weeks prior to beginning treatment with the viral vectors. Adenoviral injections were carried out in a volume of 100 ml of sterile 16PBS directly injected into the tumor. All 10 mice received injections of Adp16 on the tumor on the right ¯ank, and Adlacz in the tumor on the right ¯ank. Injections consisted of 161010 p.f.u. per injection for both vectors. Injections occurred three times a week for two weeks. The mice were sacri®ced 1 week later. Tumor measurement was carried out in two dimensions using a micrometer after the ®rst and second week. Tumor size was evaluated against the size of the tumor at the beginning of treatment (i.e. standardized against 100% of the size of the tumor 4 weeks after implanting the mesothelioma cells). In a similar manner, intraperitoneal mesothelioma xenografts were produced by inoculation of 16107 mesothelioma cells. After 4 weeks, treatment of the established di€use intraperitoneal tumors was carried out by direct intraperitoneal injections of adenovirus. Prior to beginning the injections, however, two mice were sacri®ced to ensure the presence of di€use xenograft implantation (data not shown). Viral vector was diluted in 1 ml of sterile 16PBS and injected directly into the abdominal cavity of the mice. Injections consisted of 161011 PFU per injection for both vectors (Smythe et al., 1994). An identical schedule to the previously described experiment was used, consisting of six treatments over 2 weeks, with the mice being sacri®ced 2 weeks later. Intraperitoneal tumors were graded in accordance with the size and location as previously described by others (Smythe et al., 1994), with four regions (pancreas/stomach, retroperitoneo-diaphragm, mesenteric, and hepatic-portal) chosen to grade the presence and size of tumors. The presence of a tumor between 1 and 5 mm was given a score of 1. Tumors greater than 5 mm were graded as 2, while the presence of multiple or di€use tumors (thickening or matting, or 2 or more tumors greater than 5 mm in size in one location) were graded as 3.

Acknowledgements We would like to thank Gloria Niehans and Pat Rene for their assistance in the processing of pathology specimens, and to Claudine Fasching for her assistance with ¯ow cytometry. This work was supported in part by a Merit Grant from the Research Service of the Department of Veterans A€airs (RAK) and the National Institutes of Health -1 R01 CA 68245-01A1 (DTC).

References Antman KH, Pass HI and Recht A. (1989). In: Cancer: principle and practices of oncology, 3rd edition. DeVita VT, Hellman S and Rosenberg SA. (eds.) JB Lippincot: Philadelphia, pp. 1399 ± 1417. Bianchi AB, Mitsunaga SI, Cheng JQ, Klein WM, Jhanwar SC, Seizinger B, Kley N, Klein SA and Testa JR. (1995). Proc. Natl. Acad. Sci. USA, 92, 10854 ± 10858. Bodner SM, Minna JD, Jensen SM, D'Amico D, Carbone D, Mitsudomi T, Fedorko J, Buchhagen DL, Nau MM, Gazdar AF et al. (1992). Oncogene, 7, 743 ± 749. Carbone M, Rizzo P, Grimley PM, Procopio A, Mew DJY, Shridhar V, De Batolomeis A, Esposito V, Giuliano MT, Steinberg SM, Levine AS, Giordano A and Pass HI. (1997). Nature Med., 3, 908 ± 912. Carbone M, Pass HI, Rizzo P, Marinetti M, Di MM, Mew DJ, Levine AS and Procopio A. (1994). Oncogene, 9, 1781 ± 1790.

Chen WD, Otterson GA, Lipkowitz S, Khleif SN, Coxon AB and Kaye FJ. (1997). Oncogene, 14, 1243 ± 1248. De Luca A, Baldi A, Esposito V, Howard CM, Bagella L, Rizzo P, Caputi M, Pass HI, Giordano GG, Baldi F, Carbone M and Giordano A. (1997). Nature Med., 3, 913 ± 916. DeCaprio JA, Furukawa Y, Ajchenbaum F, Grin JD and Livingston DM. (1992). Proc. Natl. Acad. Sci. USA, 89, 1795 ± 1798. DeCaprio JA, Ludlow JW, Figge J, Shew JY, Huang CM, Lee WH, Marsilio E, Paucha E and Livingston DM. (1988). Cell, 54, 275 ± 283. Fujiwara T, Cai DW, Georges RN, Mukhopadhyay T, Grimm EA and Roth JA. (1994). J. Natl. Cancer. Inst., 86, 1458 ± 1462. Geradts J, Kratzke RA, Crush-Stanton S, Wen SF and Lincoln CE. (1996). Modern Pathol., 9, 339 ± 347.

p16INK4a expression in mesothelioma SP Frizelle et al

Geradts J, Kratzke RA, Niehans GA and Lincoln CE. (1995). Cancer Res., 55, 6006 ± 6011. Grim J, D'Amico A, Frizelle S, Zhou J, Kratzke RA and Curiel DT. (1997). Clin. Cancer Res., in press. Kratzke R, Otterson GA, Lincoln CE, Ewing S, Oie H, Geradts J and Kaye FJ. (1995). JNCI, 87, 1870 ± 1875. Kratzke RA, Greatens TM, Rubins JB, Maddaus MA, Niewoehner DE, Niehans GA and Geradts J. (1996). Cancer Res., 56, 3415 ± 3420. Metcalf RA, Welsh JA, Bennett WP, Seddon MB, Lehman TA, Pelin K, Linnainmaa K, Tammilehto L, Mattson K, Gerwin BI et al. (1992). Cancer Res., 52, 2610 ± 2615. Otterson GA, Kratzke RA, Coxon A, Kim YW and Kaye FJ. (1994). Oncogene, 9, 3375 ± 3378. Parry D and Peters G. (1996). Mol. Cell. Biol., 16, 3844-3852. Pass HI and Mew DJ. (1996). J. Cell Biochem. Suppl, 24, 142 ± 151. Pass HI, Stevens EJ, Oie H, Tsokos MG, Abati AD, Fetsch PA, Mew DJ, Pogrebniak HW and Matthews WJ. (1995). Ann. Thorac. Surg., 59, 835 ± 844. Sakaguchi M, Fujii Y, Hirabayashi H, Yoon HE, Komoto Y, Oue T, Kusafuka T, Okada A, Matsuda H. (1996). Int. J. Cancer, 65, 442 ± 445. Sandig V, Brand K, Herwig S, Lukas J, Bartek J and Strauss M. (1997). Nat. Med., 3, 313 ± 319.

Sekido Y, Pass HI, Bader S, Mew DJ, Christman MF, Gazdar AF and Minna JD. (1995). Cancer Res., 55, 1227 ± 1313. Serrano M, Hannon GJ and Beach D. (1993). Nature, 366, 704 ± 707. Shapiro GI, Edwards CD, Kobzik L, Godleski J, Richards W, Sugarbaker DJ and Rollins BJ. (1995). Cancer Res., 55, 505 ± 509. Sherr CJ. (1996). Science, 274, 1672 ± 1677. Smythe WR, Kaiser LR, Hwang HC, Amin KM, Pilewski JM, Eck SJ, Wilson JM and Albelda SM. (1994). Ann. Thorac. Surg., 57, 1395 ± 1401. Treat J, Kaiser LR, Sterman DH, Litzky L, Davis A, Wilson JM and Albelda SM. (1996). Hum. Gene Ther., 7, 2047 ± 2057. Ueki K, Ono Y, Henson JW, E®rd JT, von DA and Louis DN. (1996). Cancer Res., 56, 150 ± 153. Wolfel T, Hauer M, Schneider J, Serrano M, Wolfel C, Klehmann HE, De PE, Hankeln T, Meyer zum, Buschenfelde K and Beach D. (1995). Science, 269, 1281 ± 1284. Zuo L, Weger J, Yang Q, Goldstein AM, Tucker MA, Walker GJ, Hayward N and Dracopoli NC. (1996). Nat. Genet., 12, 97 ± 94.

3095