Adenovirus-Mediated P53 Gene Therapy Inhibits Human Sarcoma Tumorigenicity

Adenovirus-mediated p53 gene therapy inhibits human sarcoma tumorigenicity

Mira Milas,1 Dihua Yu,1 Aiqing Lang,1 Tong Ge,1 Barry Feig,1 Adel K. El-Naggar,2 and Raphael E. Pollock1 Departments of 1Surgical Oncology and 2Pathology, MD Anderson Cancer Center, Houston, Texas 77030.

Mutations of the p53 tumor-suppressor gene are the most frequent genetic abnormality in soft tissue sarcomas. Because these rare tumors also respond poorly to standard chemotherapy and bear a 50% 5-year mortality rate, we investigated the possible therapeutic benefits of p53 gene restoration in sarcomas. We constructed Ad5p53, which is an E1A-deleted, replication-deficient adenovirus expressing a cytomegalovirus promoter-driven wild-type p53 cDNA with a Flag sequence tag. SKLMS-1 human leiomyosarcoma cells containing a mis-sense p53 point mutation were effectively transduced with Ad5p53. Increasing levels of Flag-p53 protein, as well as dose-dependent p21Cip1 induction, were observed through a dose range of 10–500 plaque-forming units (PFU)/cell. In vitro administration of Ad5p53 as a single 100 PFU/cell dose caused 40–60% growth inhibition of SKLMS-1 cells at posttreatment days 4, 6, and 8 compared with untreated or viral control treated-cells (P Ͻ .05, Student’s t test). Relative to these same controls, in vivo treatment of SKLMS-1-bearing severe combined immunodeficient mice with 6 ϫ 109 PFU of Ad5p53 by intratumoral injection resulted in a 35-day tumor growth delay and complete tumor regression in 40% of mice (P Ͻ .05, Student’s t test). The expression of virally derived p53 mRNA in Ad5p53-treated tumor tissues was detected in treated tumor specimens by reverse transcriptase polymerase chain reaction. Reduced intratumoral cellularity and the presence of p53 staining in adjacent normal tissue, consistent with delivery of exogenous p53 to the tumor target, were evident only in Ad5p53-treated tumors after immunohistochemical staining for p53. These results indicate that wild-type p53 gene restoration in sarcomas retards tumor growth and may come to be usefully applied to the clinical treatment of this disease as a single regimen or in combination with conventional therapies. Cancer Gene Therapy (2000) 7, 422–429

Key words: p53; gene therapy; sarcomas.

ene-based treatment for human cancers is founded relevant to human soft tissue sarcomas, 30–65% of Gon the concept that functionally defective genes which harbor a p53 mutation.9,10 In contrast to other contribute to, if not directly cause, the malignant phe- tumors in which p53 biology has been studied, sarcomas notype.1 Over the preceding years, several strategies are an extremely rare and histologically diverse group of targeting cancer have evolved to administer therapeutic human cancers. Despite multidisciplinary treatment genes that inactivate oncogenes, restore tumor-suppres- combining surgery, radiation therapy, and chemother- 2,3 sor genes, and enhance the native immune response. apy, the overall prognosis for sarcoma patients remains A number of investigations have focused on p53 gene poor: only 30% of patients respond to the major, therapy, aiming to restore the important tumor-suppres- doxorubicin-based chemotherapy regimens, and the sor functions of this molecule in cancer cells with p53 5-year mortality rate is 50%.11 Perhaps because of the 3 gene abnormalities. The initial preclinical studies dem- rarity of this disease, advances in new, more effective onstrated suppressed tumor cell growth and reduced treatments have been slow, particularly in the area of tumorigenicity in nude mice in a number of different gene therapy. There are limited preclinical studies ad- tumor cell types stably transduced with wild-type (wt) 12–14 4–8 dressing p53 restorative therapy in sarcomas, and to p53. Some investigations have further advanced p53 our knowledge, no currently active clinical trials are gene therapy to an evaluation phase in clinical trials for speciﬁcally aimed at incorporating gene therapy for this several human cancers, including lung, head and neck, 2,3 2 malignancy. and colorectal cancer. In this study, the therapeutic efﬁciency of p53 gene The growth-regulatory functions of the p53 gene are restoration in sarcomas was investigated by treating mice bearing tumors induced by subcutaneous injection of the SKLMS-1 human leiomyosarcoma cells, which contain Received March 18, 1999; accepted June 27, 1999. Address correspondence and reprint requests to Dr. Raphael E. Pollock, mutant p53, with an adenoviral-mediated wt p53 gene Department of Surgical Oncology, Box 106, MD Anderson Cancer Center, delivery system such as could be applied in clinical 1515 Holcombe Boulevard, Houston, TX 77030. treatment. Previously, we have demonstrated that stable

422 Cancer Gene Therapy, Vol 7, No 3, 2000: pp 422–429 MILAS, YU, LANG, ET AL: p53 GENE THERAPY FOR SOFT TISSUE SARCOMAS 423

wt p53 transfectants derived from this leiomyosarcoma MgCl2, 10% glycerol). Aliquots of purified virus were stored as cell line exhibit reduced tumor growth and clonogenicity viral stock in phosphate-buffered saline (PBS) with 10% glycerol. Viral titer was calculated from an optical density 260 in vitro, decreased tumorigenicity in severe combined 10 15 reading of the viral stock solution and ranged from 1.45 ϫ 10 immunodeficient (SCID) mice, and enhanced chemo- ϫ 10 sensitivity and radiosensitivity.16,17 Here we report the to 4.5 10 plaque-forming units (PFU)/mL. effectiveness of using Ad5p53 in inhibiting sarcoma cell growth in vitro and in suppressing tumor growth in Identification of recombinant p53 adenovirus sarcoma-bearing mice in vivo. Viral stocks derived from single 293 cell plaques were analyzed for the presence or absence of appropriate p53-Flag insert in the adenoviral vectors. DNA was extracted from viral stock samples MATERIALS AND METHODS using the QIAamp Blood and Tissue Kit (Qiagen, Chatsworth, Calif). A panel of candidate clones were subjected to restriction Cell lines and cell culture enzyme mapping with HindIII to verify that a product of predicted The human leiomyosarcoma cell line SKLMS-1 and the human size was obtained, indicating successful DNA recombination. To embryonic kidney cell line 293 were obtained from the Amer- check for the presence of the p53-Flag insert, PCR amplification with Ј Ј ican Type Culture Collection (Manassas, Va). SKLMS-1 is the primers CMV-30-S-5 (5 -AATGTCGTAATAACCCCGC- Ј Ј Ј well-characterized and contains a known p53 mis-sense muta- CCCGTTGACGC-3 ) and p53-2–3 (5 -GGAACAAGAAGTG- Ј ϫ tion at codon 245 (glycine 3 serine).15,18 SKLMS-1 cells were GAGAATG-3 ) was performed for 32 cycles at 94°C 1 minute, ϫ ϫ maintained in Dulbecco’s modified Eagle’s medium/F12 me- 56°C 1.5 minutes, and 72°C 2 minutes. PCR products were dium (Life Technologies, Gaithersburg, Md) with 10% fetal resolved on a 1% agarose gel and photographed under ultraviolet bovine sera (Life Technologies) in a humidified 37°C incubator light. Finally, to check for the ability of the p53 adenovirus to induce expression of exogenous p53 protein, 293 cells were reinfected with with 5% CO2 in air. Cells were subcultured 1:3 by trypsinization upon reaching confluence. The 293 cells were also main- candidate vectors at doses of up to 100 PFU/cell, and the presence of tained in Dulbecco’s modified Eagle’s medium/F12 with 10% p53 in cellular protein extracts was assessed by Western blotting as fetal bovine serum and were used for experiments at passages described below. 35–40. In vitro growth assay Construction of Ad5p53 vector SKLMS-1 cells and 293 cells were infected in culture media A p53-Flag recombinant adenovirus was designed to contain a with Ad5p53 or control virus Ad5c4 at the following multiplic- cytomegalovirus (CMV) promoter, Flag-peptide tag, wt p53 ities of infection to determine transduction efficiency: 0, 10, 50, cDNA, and human growth hormone (hGH) polyadenylation 100, and 500 PFU/cell. The cells at 80% confluency in culture signal. It was generated by first isolating the 1.8-kb BamHI- were infected, and protein extracts were prepared 48 hours BamHI p53 cDNA fragment from the pCMV-Neo-Bam vector later for analysis of Flag protein, p53, p21Cip1, and actin. This (kindly provided by Dr. Bert Vogelstein, Johns Hopkins On- information allowed for the selection of an appropriate dose cology Center, Baltimore, Md) and removing the 5Ј untran- for the growth rate assays. Untreated SKLMS-1 cells and scribed region by NcoI partial digestion. This cDNA fragment SKLMS-1 cells treated with a one-time dose of 100 PFU/cell was inserted into the pFlag-CMV-2 expression vector (Kodak Ad5p53 or Ad5c4 in media were plated into 6-well plates at a Scientific Imaging Systems, New Haven, Conn) containing the density of 104 cells/well. Cells were harvested at posttreatment CMV promoter, Flag-peptide, and hGH poly(A). The p53- days 2, 4, 6, and 8 by trypsinization. Cell viability was assessed Flag expression cassette was generated after SpeI and PinAI by trypan blue dye exclusion, and cell count was determined double digestion and subsequently ligated to a shuttle vector, using a hemocytometer. Each timepoint for each treatment E1sp1A (Microbix Biosystems, Ontario, Canada). The above group had triplicate wells. The entire experiment was repeated DNA construct was cotransfected with the pJM17 plasmid, three times starting from viral infection of new SKLMS-1 which contains the entire E1-deleted Ad5 genome, into 293 cultures, and data were plotted from the averages of all cells for homologous recombination by N-(1-[2,3-dioleoy- experiments. loxy]propyl)-N,N,N-trimethylammonium methylsulfate liposomal-mediated transfection according to the manufacturer’s Detection of p53, p21Cip1, and Flag proteins protocol (Boehringer Mannheim, Indianapolis, Ind). The 293 cells were plated 24 hours before transfection at 60% conflu- Proteins were extracted from SKLMS-1 cells and 293 cells that ency in 60-mm dishes and transfected with 11 ␮g/60-mm dish were untreated or transduced with adenoviruses using ␮ of pJM17 and p53 shuttle vector (1:1 ratio) per 100 Lof PBSTDS lysis buffer (10 mM Na2HPO4, 155 mM NaCl, 1% N-(1-[2,3-dioleoyloxy]propyl)-N,N,N-trimethylammonium methyl- Triton X-100, 3 mM sodium dodecyl sulfate, 11 mM sodium sulfate liposome. After transfection, cells were fed culture deoxycholate, 30 mM NaN3, 0.9 mM NaF, pH 7.25), quantified media every 2–3 days until the onset of cytopathic effects. with the bicinchoninic acid protein assay (Pierce, Rockford, Aliquots of cell culture supernatants were collected from Ill), resolved on a 12% polyacrylamide gel (Bio-Rad, Hercules, plates with detectable cytopathic effects and stored as candi- Calif), and electroblotted onto a nitrocellulose membrane date positive p53 adenoviral stocks, which were further char- (Bio-Rad). The following primary antibodies (Abs) were used: acterized by restriction mapping, polymerase chain reaction monoclonal mouse anti-human p53 Ab-6 (Oncogene Science, (PCR), and Western blot analysis described below. A clone Uniondale, NY), monoclonal mouse anti-human WAF1 expressing wt p53 (Ad5p53) and a clone without the gene (Ab-1) (Oncogene Science), monoclonal mouse anti-human insert (Ad5c4) were then further amplified in 293 cells, col- actin AC-40 (Sigma, St. Louis, Mo), and monoclonal mouse lected by cell lysis after three freeze-thaw cycles at Ϫ80°C, anti-Flag M5 (Sigma). After incubation with corresponding purified by a CsCl gradient, and desalted in dialysis buffer (10 secondary Abs (Promega, Madison, Wis), the membranes were mM tris(hydroxymethyl)aminomethane-HCl (pH 7.5), 1 mM processed with enhanced chemiluminescence detection re-

Cancer Gene Therapy, Vol 7, No 3, 2000 424 MILAS, YU, LANG, ET AL: p53 GENE THERAPY FOR SOFT TISSUE SARCOMAS

agent (Amersham, Arlington Heights, Ill); signals were visual- RESULTS ized by autoradiography. Generation of Ad5p53 virus and transduction of human sarcoma cells Animals The Ad5p53 virus was constructed by first generating an Female SCID mice of 3–4 weeks of age (model ICRSC-M) were obtained from Taconic Farms (Germantown, NY). Ani- expression cassette that contained a CMV promoter, a mals were maintained on a standard laboratory diet ad libitum Flag tag, human wt p53 cDNA, and a hGH polyadenyl- in diurnal lighting conditions and received humane care in ation signal (Fig 1A). This expression cassette was then accordance with the Animal Welfare Act and the National ligated to the E1sp1A shuttle vector and cotransfected Institutes of Health “Guide for the Care and Use of Labora- with recombinant plasmid pJM17 into 293 cells for tory Animals.” The experiments were approved by the Institu- homologous recombination. Successful introduction of tional Animal Care and Use Committee at MD Anderson the Flag-tagged wt p53 into infected cells was confirmed Cancer Center. by restriction mapping and PCR detection of the exogenous p53. Restriction mapping using HindIII digestion In vivo tumorigenicity assay led to a panel of DNA products of expected size for a SKLMS-1 cells were grown to confluence in vitro, harvested by Flag-tagged p53 insert in the Ad5p53 adenovirus (data trypsinization, rinsed with PBS, and resuspended in Matrigel not shown). PCR amplification of a region extending 6 Ј Ј (Becton Dickinson, Bedford, Mass). An aliquot of 5 ϫ 10 cells from the 5 Flag end to the 3 wt p53 cDNA yielded the was injected subcutaneously into the flank of SCID mice, and expected 1.3-kb product from the Ad5p53 virus and the tumors allowed to develop to 0.5 mm diameter in size. Animals adenovirus clone 10 but not from the Ad5c4 virus (Fig then received intratumoral injections of saline (n ϭ 10), Ad5c4 1B). To further confirm that Ad5p53 infection led to the (n ϭ 10), or Ad5p53 (n ϭ 10) on a schedule of 109 PFU in 50 expression of Flag-tagged p53 protein, Western blot ␮ L injected every other day for a total of six injections (total analysis was performed on protein samples prepared ϫ 9 dose: 6 10 PFU). Daily weights and tumor size measure- from the 293 cell culture infected with this recombinant ments were obtained every 2 days from the first treatment virus clone, as well as on samples from 293 cells infected administration. Mice were sacrificed when the tumor diameter reached 2 cm. The experiment was performed in duplicate. with the adenoviruses clone 10 and Ad5c4. Figure 1C Tumor volume was calculated as ([a ϫ b2]/2), where a repre- shows the presence of Flag-tagged p53 proteins in 293 sents the largest tumor diameter and b represents the smallest cells infected with Ad5p53 using both anti-Flag and tumor diameter.19 Tumor growth delay was calculated as the anti-p53 Abs. Although the 1.3-kb p53 cDNA insert can difference between the time required for treated tumors to be detected in the adenovirus clone 10, expression of the reach 200 cm3 and the time for saline-treated tumors to reach Flag-tagged p53 protein was undetectable and clone 10 the same size. was not used for further investigation. Ad5c4 was an empty adenoviral vector construct with a deleted E1A Detection of Ad5p53 in tumor samples region that did not integrate the 1.3-kb p53 cDNA insert and did not express the Flag-tagged p53 protein. There- Tumors were harvested 72 hours after injection with saline, Ad5c4, or Ad5p53 (6 ϫ 109 PFU single treatment) and rinsed fore, this virus was used as a viral control in all subse- in PBS; portions were preserved in 3.8% formalin and embed- quent experiments. ded in paraffin. Immunostaining was performed on 4-␮m tissue To assess the ability of Ad5p53 virus to transduce sections using the avidin-biotin-peroxidase method of Hse et SKLMS-1 human leiomyosarcoma cells and to express al.20 Briefly, tissue sections were dewaxed and dehydrated, and the Flag-tagged p53 protein in these cells, Western blot endogenous peroxidase activity was blocked with 0.3% hydro- analysis was performed on protein extracts from gen peroxide. The sections were then treated with normal SKLMS-1 cells collected 48 hours after infection with horse sera followed by overnight incubation at 4°C with 0–500 PFU/cell Ad5p53. Figure 2 shows that Ad5p53 anti-p53 Ab (D01, 1/200 dilution; Novocastra Laboratories, virus effectively transduced the sarcoma cells, with exog- Newcastle on Tyne, UK). enous Flag-tagged p53 proteins present at viral doses of Fresh tumor tissue was used in the detection of exogenous p53 by reverse transcriptase (RT)-PCR. Tumors were treated 50 PFU/cell and increasing with higher doses. In addi- 9 tion, Ad5p53 caused a dose-dependent induction of every other day with saline, Ad5c4, or Ad5p53 at a dose of 10 Cip1 PFU per injection for six injections total. Tumors were then p21 in SKLMS-1 cells, indicating that the exogenous harvested 3 days after the third injection and 3 and 10 days p53 protein is functionally active in regulating its down- after the last injection. Total RNA was extracted from tumor stream target genes, such as p21Cip1. The expression of tissue using RNAzol B (Tel-Test, Friendswood, Tex), and Flag-tagged p53 proteins was lower in SKLMS-1 cells first-strain cDNA synthesis was performed with the Super- than in 293 cells, which are specifically engineered to Script Preamplification System (Life Technologies). The contain the adenoviral E1 region and thus allow efficient cDNA was amplified by PCR with p53 flanking primers; cycling viral replication; Ad5p53 infection in 293 cells led to conditions were as described above. The PCR products were peak p53 and Flag protein expression already at the 10 resolved on a 1.2% low-melting agarose gel. To intensify the Cip1 signal from the 1.3-kb p53 insert, the 1.3-kb p53 band was PFU/cell dose, although there was an absence of p21 isolated from the gel and melted at 45°C for 10 minutes; next, induction, possibly because other genetic alterations in 2 ␮L of the resulting product was used for a second PCR. The these cells may interfere with p53’s function as a tran- Cip1 PCR conditions and primers remained the same for the second scriptional activator for p21 . These results confirmed PCR, and products were resolved on a 1% agarose gel. that the Ad5p53 virus could be used to infect human

Cancer Gene Therapy, Vol 7, No 3, 2000 MILAS, YU, LANG, ET AL: p53 GENE THERAPY FOR SOFT TISSUE SARCOMAS 425

Figure 1. A: Schematic map of the p53 expression cassette. The Ad5p53 genome is ϳ35.4 kb in size, with the p53 expression cassette replacing the E1A region of the Ad5 genome and containing the DNA components shown. B: PCR ampliﬁcation of adenoviral p53 cDNA in 293 cells infected with different recombinant adenoviral constructs as indicated. The primers used in PCR analysis (shown in (A)) of the adenoviral vectors deﬁne a 1.3-kb product that includes the sequence encoding the Flag tag and the entire wt p53 cDNA. C: Western blot analysis for Flag-p53 protein on cell lysates of 293 cells infected with the indicated adenoviruses. Anti-Flag, anti-p53, and anti-␤-actin (control) Abs were used as primary Abs. Only the Ad5p53 clone demonstrated both exogenous p53 DNA integration and Flag-tagged p53 protein expression.

SKLMS-1 sarcoma cells and deliver multiple copies of growth rate assay was performed using untreated, the p53 gene, which leads to active translation of a Ad5p53-treated, or Ad5c4-treated SKLMS-1 cells. A functionally competent p53 protein. one-time dose of 100 PFU/cell Ad5p53 or Ad5c4 was used to maximize gene transfer efﬁciency and minimize Effect of exogenous p53 on sarcoma growth in vitro toxicity. Triplicate sets of untreated and treated cells To determine whether virally delivered p53 can exert were analyzed every other day for 8 days for cell viability growth-regulatory effects on SKLMS-1 cells, an in vitro and cell count, and the experiment was repeated three

Cancer Gene Therapy, Vol 7, No 3, 2000 426 MILAS, YU, LANG, ET AL: p53 GENE THERAPY FOR SOFT TISSUE SARCOMAS

Figure 2. Western blot analysis demonstrates the infection efﬁciency of Ad5p53 in SKLMS-1 human sarcoma cells. Exogenous p53 was detected by probing for the Flag protein tag using anti-Flag Ab. Western blot analyses using anti-p53 and anti-p21 Abs demonstrated a dose-dependent increase in p53 and p21Cip1 proteins at Ad5p53 treatment doses of Ն50 PFU/cell. The transduction of 293 cells engineered for viral replication is efﬁcient at doses as low as 10 PFU/cell. times. Ad5p53-treated SKLMS-1 cells showed 40–60% growth inhibition at posttreatment days 4, 6, and 8, relative to untreated and Ad5c4-infected cells (P Ͻ .05, Student’s t test; Fig 3). The Ad5c4-treated cells displayed a growth rate similar to that of untreated cells.

Figure 4. In vivo tumorigenicity assay showing the inhibition of leiomyosarcoma growth in SCID mice. A: SKLMS-1 cells subcutaneously injected into the flanks of SCID mice were injected intratu- morally with saline, Ad5p53, or Ad5c4 when tumor diameter reached 0.5 cm. The total viral dose administered was 6 ϫ 109 PFU within the time intervals shown. The tumor growth delay for Ad5p53- treated cells to reach 200 mm3 volume was 35 days (P Ͻ .05, Figure 3. In vitro growth curve of Ad5p53-treated human leiomyo- Student’s t test). B: One representative mouse from each group of sarcoma cells. SKLMS-1 cells were cultured in 100-mm dishes, mice of the tumorigenicity assay in (A); complete tumor regression untreated or infected with a single dose of 100 PFU/cell Ad5p53 or was demonstrated in 40% of mice treated with Ad5p53. vector control Ad5c4, and replated in triplicate at a density of 104 cells/well in 6-well plates. Cell viability ranged from 93% to 100% at all timepoints in each of the treatment groups. The curves represent the combined data of three experiments. The difference in growth Cell viability was assessed by trypan blue dye exclusion between Ad5p53-treated cells versus both untreated and viral and was found to be in the range of 93–100% for all control-treated cells was statistically significant (P Ͻ .05, Student’s treatment groups during the entire culture period. This t test) for days 4–8. suggested that the reduced growth rate might reflect

Cancer Gene Therapy, Vol 7, No 3, 2000 MILAS, YU, LANG, ET AL: p53 GENE THERAPY FOR SOFT TISSUE SARCOMAS 427

changes in cell proliferation rather than cell toxicity. In addition, no signiﬁcant induction of apoptosis was detected for all treatment groups under this culture con- dition. Overall, these results indicate that the virally delivered p53 gene signiﬁcantly inhibits the in vitro growth rate of SKLMS-1 sarcoma cells.

Inhibition of sarcoma tumorigenicity by Ad5p53 in vivo To examine whether the Ad5p53 virus may be used as a p53 gene therapy agent for p53 mutation-bearing human soft tissue sarcomas in a living host, we ﬁrst developed a tumor-bearing mouse model by injecting SCID mice subcutaneously with SKLMS-1 cells. We subsequently treated these mice with an intratumoral injection of saline, Ad5p53, or Ad5c4 when the SKLMS-1 tumor nodules reached a size of 0.5 cm in diameter. Each mouse received six injections either of 109 PFU virus in 50 ␮L saline or of saline alone, administered every 48 hours for 12 days. The total viral dose was 6 ϫ 109 PFU/mouse. The tumor growth rate for each of the treatment groups is depicted in Figure 4A. The tumors were measured at intervals evident on the graph begin- ning with the day of initial treatment and continuing until they reached maximum allowable size. Ad5p53 treatment of SKLMS-1 tumors signiﬁcantly reduced tumor growth, compared with both saline- and Ad5c4- treated tumors (P Ͻ .05, Student’s t test). At day 40, Ad5p53 tumors were 80% smaller in size than saline- treated tumors and 54% smaller than control-vector treated tumors. The tumor growth delay to reach 200 mm3 was 35 days; whereas the saline- and control vector-treated tumors enlarged to this volume within 5 days of initial injection, nearly 40% of Ad5p53 tumors completely regressed, and the remainder slowly grew to this size by day 40 postinjection (Fig 4B). Tumors harvested from SCID mice after treatment were analyzed for mRNA expression from the exogenous p53 gene and to determine the histological charac- teristics of the p53 treatment response. As illustrated in Figure 5A, Flag-p53 mRNA derived from Ad5p53 was demonstrable by RT-PCR in the tumors treated with Ad5p53, implying that the expression of exogenous wt p53 RNA occurred after gene delivery in vivo,asitdidin vitro. Figure 5B shows SKLMS-1 tumor sections immu- nohistochemically stained for p53 protein. Only tumor sections from the Ad5p53-treated mice displayed decreased cellularity within the nodule and p53 staining in the immediately adjacent normal tissues around the tumor and in the connective tissue elements within the

presence (ϩ) or absence (Ϫ) of reverse transcriptase. B: Immuno- histochemical analysis for p53 protein in leiomyosarcoma tumor sections from SCID mice treated differently. Only the Ad5p53- treated tumors displayed reduced intratumoral cellularity and the presence of p53 staining in adjacent normal tissue, consistent with delivery of exogenous p53 to the tumor target. All tumor cells Figure 5. A: Detection of exogenous p53 mRNA in Ad5p53-treated displayed p53 staining because of their endogenous mutant p53 tumor tissue by RT-PCR. RT denotes a reaction carried out in the status.

Cancer Gene Therapy, Vol 7, No 3, 2000 428 MILAS, YU, LANG, ET AL: p53 GENE THERAPY FOR SOFT TISSUE SARCOMAS tumor. The mutant p53 status of the endogenous p53 other sarcoma histological subtypes or delivery methods protein is reflected in the positive staining of tumor cells that have unique relevance to sarcoma treatment. For in samples from all the treatment groups. These results example, we have demonstrated previously in an animal indicate that Ad5p53 can retard the growth of SKLMS-1 model21 that it is feasible to deliver genetic constructs to sarcomas in the SCID mouse model in vivo. a tumor-bearing limb using isolated limb perfusion, a technique currently in clinical use for the treatment of advanced extremity sarcomas. The therapeutic effects of DISCUSSION p53 gene therapy, if confirmed in such clinical studies, would have tremendous practical ramifications, because In this study, we demonstrate the applicability of current there are significantly fewer choices for the effective gene therapy strategies to the investigation and treat- treatment of these rare tumors compared with other ment of human soft tissue sarcomas. The tumor-suppres- cancers. Only ϳ30% of patients respond to the most sor gene p53 is among those molecular therapies that frequently used chemotherapeutic agent, doxorubicin, may prove to be valuable in the clinical treatment of a and the prognosis for the majority of patients is poor, as number of human cancers, by restoring to tumors the reflected in overall 5-year mortality rates, which remain growth-regulatory capacity which had been lost during close to 50%.11 The current lack of new therapeutic the course of neoplastic transformation.2,5 This gene is choices for patients burdened by these tumors is a currently being delivered via recombinant adenovirus as significant reality. It is applications such as isolated limb the main gene delivery strategy in clinical trials for lung perfusion and the potential of combining p53 gene cancer and head and neck cancers.2 The results pre- therapy with current multimodality treatment options sented here show that adenoviral p53 therapy leads to that may improve the therapeutic response of sarcoma the successful delivery of wt p53 genes into human patients and their overall prognostic outlook. leiomyosarcoma cells harboring mutant p53, resulting in substantial growth inhibition of sarcoma cells both in ACKNOWLEDGMENTS vitro and in vivo, making p53 gene therapy a plausible new treatment strategy for this malignancy. This study was supported in part by Grants CA67802 (to The adenovirus used in our studies, Ad5p53, was R.E.P.) and CA60488 (to D.Y.) from the National Institutes of constructed to be replication-deficient and to contain Health (Bethesda, Md) and by MD Anderson Cancer National the human CMV promoter, Flag tag sequence, human Institutes of Health Core Grant CA16672. M.M. is supported wt p53 cDNA, and hGH polyadenylation signal. It is by National Institutes of Health T32 Training Grant CA09559 similar to previously reported p53 adenoviral gene ther- (to R.E.P.). apy vectors,2,6 except that the Flag tag is fused to the N terminus of p53 cDNA. Ad5p53 mediated a dose-depen- REFERENCES dent expression of p53 protein in human SKLMS-1 leiomyosarcoma cells and led to the induction of the cell Cip1 1. Knudson AG Jr. Genetics and etiology of human cancer. cycle regulator p21 in vitro, demonstrating that Advances in Human Genetics. 1997;8:1–66. Ad5p53 can deliver functionally potent wt p53. A single 2. Roth JA, Cristiano RJ. Gene therapy for cancer: what in vitro treatment of 100 PFU/cell achieved significant have we done and where are we going? J Natl Cancer Inst. sarcoma growth inhibition that persisted for at least 8 1997;89:21–39. days. A series of in vivo treatments by intratumoral 3. Nielsen LL, Maneval DC. p53 tumor suppressor gene injection of 6 ϫ 109 PFU likewise caused full regression therapy for cancer. Cancer Gene Ther. 1998;5:52–63. of some tumor nodules and substantially retarded 4. Chen PL, Chen Y, Bookstein R, et al. Genetic mechanisms growth in others, which lagged in size 35 days behind the of tumor suppression by the human p53 gene. Science. saline- and virus control-treated tumors. Thus, short- 1990;250:1576–1580. term, exogenous p53 administration appears to be effec- 5. Roemer K, Friedmann T. Mechanisms of action of the p53 tumor suppressor and prospects for cancer gene therapy by tive in diminishing the growth of sarcomas in a mono- reconstitution of p53 function. Ann N Y Acad Sci. 1994; layer environment and in reducing their tumorigenicity 716:265–282. in vivo. 6. Zhang WW, Fang X, Mazur W, et al. High-efficiency gene A major implication of these experiments is that transfer and high-level expression of wild-type p53 in possible therapeutic benefits may be derived when nor- human lung cancer cells mediated by recombinant adeno- mal, wt p53 is restored to human sarcomas. The preclin- virus. Cancer Gene Ther. 1994;1:5–13. ical studies reported here provide important information 7. Liu TJ, Zhang WW, Taylor DL, Roth JA, Goepfert H, about growth regulation by adenoviral p53 in sarcomas. Clayman GL. Growth suppression of human head and These findings, combined with earlier observations from neck cancer cells by the introduction of a wild-type p53 our laboratory that p53 enhances the chemosensitivity gene via a recombinant adenovirus. Cancer Res. 1994;54: 3662–3667. and radiosensitivity of sarcomas, justify the investigation 8. Fujiwara T, Grimm EA, Mukhopadhyay T, Zhang WW, of Ad5p53 as a single agent or in combination with Owen-Schaub LB, Roth JA. Induction of chemosensitivity chemotherapeutic drugs or radiation for the treatment in human lung cancer cells in vivo by adenovirus-mediated of sarcomas. Ad5p53 can further be evaluated for its transfer of the wild-type p53 gene. Cancer Res. 1994;54: effectiveness in sarcoma therapy by investigation using 2287–2291.

Cancer Gene Therapy, Vol 7, No 3, 2000 MILAS, YU, LANG, ET AL: p53 GENE THERAPY FOR SOFT TISSUE SARCOMAS 429

9. Andreassen A, Oyjord T, Hovig E, et al. p53 abnormalities suppress human soft tissue sarcoma by enhancing cell cycle in different subtypes of human sarcomas. Cancer Res. control. Clin Cancer Res. 1998;4:1985–1994. 1993;53:468–471. 16. Pollock R, Lang A, Ge T, Milas M, Yu D. Reintroduction 10. Latres E, Drobnjak M, Pollack D, et al. Chromosome 17 of wild type p53 enhances sarcoma chemosensitivity. Proc abnormalities and TP53 mutations in adult soft tissue Am Soc Clin Oncol. 1997:15. sarcomas. Am J Pathol. 1994;145:345–355. 17. Milas M, Yu D, Feig B, Lang A, Brock W, Pollock R. 11. Pisters PW, Fein DA, Somlo G. Soft-tissue sarcomas. In: Increased tumor radiosensitivity in human sarcoma cells Pazdur R, et al, eds. Cancer Management: A Multidisci- transduced with wild-type p53. Proc Assoc Acad Surg Annu plinary Approach. Huntington, NY: PRR; 1996:491–505. Meet. 1997:115. 12. Milas M, Yu D, Feig B, Ge T, Pollock R. p53 gene therapy 18. Stratton MR, Moss S, Warren W, et al. Mutation of the inhibits human sarcoma cell growth in vitro, and uses novel p53 gene in human soft tissue sarcomas: association with delivery methods in vivo. Keystone Symposium on Molecu- abnormalities of the RB1 gene Oncogene. 1990;5:1297– lar and Cellular Biology of Gene Therapy. 1998:80. 1301. 13. Gagnebin J, Kovar H, Kajava AV, et al. Use of transcrip- 19. Carlsson G, Gullberg B, Hafstrom L. Estimation of liver tion reporters with novel p53 binding sites to target tumour tumor volume using different formulas: an experimental cells expressing endogenous or virally transduced p53 study in rats. J Cancer Res Clin Oncol. 1983;105:20–23. mutants. Oncogene. 1998;6:685–690. 20. Hse SM, Raine I, Franger H. Use of avidin-biotin-perox- 14. Meye A, Lasch J, Wurl P, et al. Liposomal DNA transfec- idase complex (ABC) in immunoperoxidase techniques: a tion of sarcoma cells with alterations in p53: therapeutic comparison between ABC and unlabelled antibody (PAP) implications. Keystone Symposium Proceedings: Synthetic procedures. J Histochem Cytochem. 1981;29:577–580. Non-Viral Gene Delivery Systems. 1998:142. 21. Milas M, Feig B, Yu D, et al. Isolated limb perfusion in the 15. Pollock RE, Lang A, Ge T, Sun D, Tan M, Yu D. sarcoma-bearing rat: a novel preclinical gene delivery Wild-type p53 and a p53 temperature-sensitive mutant system. Clin Cancer Res. 1997;3:2197–2203.

Cancer Gene Therapy, Vol 7, No 3, 2000