Recombinant, refolded tetrameric p53 and gonadotropin-releasing hormone-p53 slow proliferation and induce apoptosis in p53-deficient cancer cells

The p53 tumor suppressor is a transcription factor that resides in the cytosol, and after activation by posttranslational modifications, it translocates to the nucleus, where it tetramerizes, binds DNA, and activates transcription of genes important in cell cycle regulation and DNA repair (1–3). Activation of p53 regulates over 100 cellular genes and results in cell cycle arrest, giving cells with damaged DNA an opportunity to make repairs or, in the case of irreparable DNA damage, to undergo apoptosis (4, 5). Somatic mutations in the p53 gene are found in many human tumors (6) and germ-line mutations in p53 are responsible for an inherited cancer predisposition, Li-Fraumeni syndrome (7). More recent publications have shown in mouse model that p53 restoration resulted in tumor regression and clearance (8–10), clearly validated the p53 replacement therapy concept. In addition, the antiangiogenesis effect of p53 has long been known, and one of the specific effects is the stimulation of extracellular release of collagen-derived peptides such as endostatin and tumstatin, strong inhibitors of angiogenesis (11).

Despite the fact that p53 is mutated in many human cancers and its role in tumor suppression is well characterized, there are no p53-based antineoplastic therapies available. Some of the current standard cancer therapies, including chemotherapy (e.g., cisplatin, carboplatin, and oxaliplatin) and radiation therapy, may partly depend on activation of p53-dependent pathways for induction of tumor cell death. However, in many tumors, p53 is inactivated by mutation, and these therapies are not as effective as cancer cells with wild-type p53. One approach used to treat p53 mutant-expressing tumors has been to use mutant p53 peptides as a vaccine to stimulate the immune system to destroy cells that express mutant variants of p53 (12, 13). To date, only a modest sensitization of the target tissue to second-line chemotherapy has been elicited from this treatment.

Two strategies have been used to reactivate mutant, inactive p53. The use of small molecules (e.g., nutlin, chalcones, and WMC-79) that stabilize or activate p53 function in p53-deficient cancer cells is one promising strategy, but it poses several challenges (14, 15). Few molecules have been identified that function to restore the activity of p53, and those that do may be effective for only a few p53 mutants. In addition, these molecules may stabilize dominant-active mutants of p53, an effect that could be detrimental to patients. COOH-terminal–derived p53 peptides have also been used to restore wild-type activity to mutant p53 (16), but this reactivation strategy may also be effective for only a few mutants.

For tumors in which p53 cannot be reactivated, or in tumors where p53 expression levels are low, alternative strategies that involve p53 replacement may be developed. Studies in cell culture and animal models have shown a selective sensitivity of p53-deficient tumor cells to the effects of exogenous p53 compared with normal cells (17). This suggests that p53 replacement therapy would have fewer side effects than conventional therapies that rely on exploiting active p53-dependent pathways in cells. Gene therapy is one replacement approach that has been used to restore p53 function to cancer cells in cell culture and animal models. In general, however, p53 gene therapy has met with limited success due to the production of adenovirus-neutralizing antibodies and random integration of the p53 gene. Therefore, many of the trials have been discontinued (18). Protein therapy using p53 may overcome several of the limitations of gene therapy encountered thus far, specifically the safety, toxicity, immune response, and random integration of the transgene.

A successful p53-based protein therapeutic candidate should be wild-type to reduce the immunogenicity, tetrameric to be efficient, and able to enter the targeted cancer cells. To date, the attempts to produce wild-type, full-length tetrameric p53 that is stable in the absence of chemical cross-linking have been unsuccessful. There are several reasons why isolation of p53 as a stable tetramer is important for a protein therapeutic application. First, the active form of p53 is a tetramer. p53 monomers bind DNA in a cooperative manner and the affinity for DNA is increased up to 100-fold by tetramerization (19). Second, the tetramerization domain is important for protein-protein interactions, and tetramerization may regulate binding of some proteins to p53. The tetrameric structure may also allow for binding of multiple protein partners at the same time (20–22). Third, some posttranslational modifications (e.g., phosphorylation and ubiquitination) require p53 to be oligomerized (23–26). Fourth, the nuclear export signal is located in the tetramerization domain (27). It is exposed when p53 is monomeric but buried in the tetramer, allowing p53 to remain in the nucleus. Finally, and perhaps most importantly, tetrameric p53 may not heterotetramerize with endogenous mutant p53, an interaction that may decrease the effectiveness of a p53 therapeutic. In this report, we describe the production of a wild-type tetrameric p53 that binds DNA and induces apoptosis when introduced into p53-deficient cancer cell lines. In addition, we have also produced a stable tetrameric gonadotropin-releasing hormone-p53 fusion protein (GnRH-p53) and characterized the protein as a potential novel protein therapeutic for the treatment of cancer.

Wild-type p53 (accession number NM_000546) was used as a PCR template for cloning of both wild-type and the fusion protein, GnRH-p53. Oligonucleotides were designed to introduce convenient restriction endonuclease sites and, for the fusion protein, the GnRH sequence at the 5′-end. Wild-type p53 was cloned using a Gateway cloning system (GeneCopoeia, Inc.) with a short peptide NH₂-terminal fusion to facilitate high-level expression and inclusion body formation. For GnRH-p53, a polyglycine linker was installed between the coding region of GnRH and p53 to minimize possible structural interference between these two domains. The designed expression genes were cloned into a modified pET11a vector, and the resulting expression plasmids are called pET11a-p53 and pET11a-GnRH-p53, respectively. The following are the NH₂-terminal fusion sequences of the wild-type p53 and GnRH-p53: wild-type p53, MASMTGGQQMGRGSSTSLYKKAGS-p53; GnRH-p53, MHWSYGLRPGGGGS-p53. The GnRH sequence is underlined. GnRH-p53 PCR primers are described in Supplementary Fig. S1.³

Escherichia coli BL21(DE3) cells were transformed with pET11a-p53 and pET11a-GnRH-p53, and the expression and inclusion body purification were done essentially as described (28–30). Briefly, inclusion bodies were solubilized in buffer containing 8 mol/L urea, 0.1 mol/L Tris, 1 mmol/L glycine, 1 mmol/L EDTA, 10 mmol/L β-mercaptoethanol, 10 mmol/L DTT, and 1 mmol/L reduced glutathione (pH 10.5) at a final concentration of 2 mg/mL. Inclusion bodies were diluted 20-fold into refolding buffer containing 20 mmol/L Tris, 0.2 mol/L arginine, and 0.1 mmol/L ZnCl₂ followed by slowly adjusting pH to 8.0. Refolded proteins were concentrated using tangential flow filtration and purified using size exclusion chromatography with a Sephacryl S-300 column (XK50 × 850-mm; Amersham) in refolding buffer with 0.1 mol/L NaCl to separate soluble aggregates from refolded protein. SDS-PAGE was done with a NuPAGE 4% to 12% Bis-Tris gel (Invitrogen) in MES buffer. Gels were stained with Coomassie blue and destained in water. Western blot analysis was done using BP53-12 mouse monoclonal p53 antibody (Genetex) as the first antibody (1:1,000 dilution) and stained using One-Step Western Blot kit from Columbia-Bio.

The physical properties of the refolded p53 and GnRH-p53 were analyzed using SDS-PAGE and dynamic light scattering (DLS), and the resulting data were used as initial criteria for proper refolding of the proteins into stable tetramers. DLS was done on purified samples (1–2 mg/mL) using Brookhaven Instruments Corporation 90 Plus particle size analyzer and data were collected using the Zeta PALS Particle Sizing Software.

The TransAm p53 DNA-binding kit (Active Motif) was used according to the manufacturer's instructions to assess p53 DNA-binding activity. A dose-response curve of each purified p53 and GnRH-p53 was established, and the specific activity of each protein preparation was determined. Specificity of DNA binding was determined using competitive wild-type and mutant oligos in the DNA-binding assay. Activity was compared with commercially available p53 (Active Motif). Experiments were carried out with samples in triplicate and the SE was determined. Each experiment was done at least thrice, and Student's t test was used to evaluate the statistical significance of the data.

DU145, MDA-MB-231, and OVCAR3 cells from the American Type Culture Collection were maintained in RPMI 1640 supplemented with 10% fetal bovine serum and penicillin/streptomycin. For experimental treatments, cells were plated at a density of 60,000/mL in 96-well plates (100 μL), eight-chamber slides (200 μL), or 6-cm dishes (5 mL) 24 h before treatments. Cells were treated with 100 μmol/L nutlin from Calbiochem overnight and then treated with 50 μg/mL methyl methanesulfonate (Sigma) for 2 to 3 h before terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) staining. For GnRH-treated cells, cells were treated with indicated concentration of recombinant human GnRH from Axxora for 72 h.

The ability of the fusion proteins to induce growth arrest was measured by 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS) proliferation assay using CellTiter 96 Aqueous One (Promega) according to the manufacturer's instructions. Plates were read using a SpectraMax plate reader from Molecular Devices. Experiments were carried out with samples in triplicate and the SE was determined. Each experiment was done thrice, and the statistical significance of the data was evaluated using Student's t test.

Nucleosome ELISA was done using a Nucleosome ELISA kit from Calbiochem according to the manufacturer's protocol. Cells were harvested from 5-cm plates and nuclear lysates were prepared. A 1:5 dilution of lysate was used and data were normalized for sample protein concentration.

The ability to induce apoptosis was analyzed using TUNEL staining with the FragEL kit from EMD Biosciences according to the manufacturer's instructions. Experiments were carried out with samples in triplicate and the SE was determined. Each experiment was done thrice, and Student's t test was used to evaluate the statistical significance of our data.

After plating for 24 h, cells were treated with GnRH-p53 or control samples for 10, 60, and 120 min at 37°C. Treatment was terminated by removal of medium and washing thrice with ice-cold PBS. The cells were fixed in 4% paraformaldehyde for 20 min at 4°C, washed once in cold PBS, and stored in cold PBS until immunostaining procedure was started (3–5 h). Next, the cells were permeabilized for 1 h in 5% human serum, 5% donkey serum, 0.1% Triton X-100, 0.01% saponin, and 1% milk in PBS. Antibodies were diluted in PBS/0.5% milk/0.01% saponin/2% serum and stained with BP53-12 mouse monoclonal p53 antibody (1:200) overnight and rinsed thrice for 15 min with PBS and stained with secondary antibody (Alexa Fluor 555 donkey anti-mouse IgG, 1:500; Molecular Probes) for 1 h. The stained cells were then washed (thrice for 20 min) in PBS followed by TOTO3 incubation for nuclear staining (Molecular Probes). The cells were visualized on a Nikon Eclipse E800 fluorescent microscope with 40× objective and images were captured with ACT-1 software. Confocal Z-stack images (100×) were acquired on a Zeiss LSM510 equipped with argon, 543 HeNe and 633 HeNe lasers, and LSM510 (3.5 SP1.1) software. Stacks were analyzed and colocalization statistics were generated using Bitplane Imaris Suite (4.5.2).

To test the possibility of developing p53 as a cancer therapeutic, we initially expressed wild-type, full-length p53 protein in E. coli. We designed the expression construct with additional NH₂-terminal peptide to facilitate the high-level expression and inclusion body formation (see Supplementary Fig. S1A).³ The expressed p53 formed insoluble inclusion bodies, which were purified and tested for different refolding conditions. Size exclusion chromatography in native refolding buffer from one of the refolding conditions showed a peak in the tetrameric position (Fig. 1A). Surprisingly, the tetramer is still stable when dissolved in a 0.1% SDS sample buffer and run on a SDS-PAGE (Fig. 1B). The identity of the p53 bands on the SDS-PAGE was confirmed by Western blot analysis (results not shown). A DNA-binding assay was used to access the in vitro activity of our recombinant p53 (PTI-p53). Native p53 oligomerizes in a concentration-dependent manner, and DNA binding involves cooperativity between the subunits of the active tetramer. The specific DNA-binding affinity of the recombinant p53 compared with commercial p53 is shown in Fig. 1C and D, which results in several interesting observations. First, the degree to which binding is dependent on concentration of p53 is lower for PTI-p53 compared with commercially available p53. This is presumably due to the fact that tetramerization is concentration dependent, and PTI-p53 is already tetrameric at these concentrations. This characteristic may be important for a p53 therapeutic because tetrameric p53 may be more effective at a lower dose than monomeric p53. Second, at a relatively high concentration of p53 (7.5 nmol/L, 390 ng/mL), PTI-p53 is 0.85 times as active as commercially available recombinant p53. However, at low concentrations (0.1 nmol/L, 5 ng/mL), PTI-p53 is 12.5 times as active as the commercially available p53 (Fig. 1D). Again, this is presumably due to the preoligomerized state of PTI-p53.

Figure 2A (lane 1) shows that, at extended incubation at 37°C, a small amount of the tetramer dissociates into monomer, dimer, and trimer and some polymerized into multimers. High concentrations of urea disrupted the tetramer (lane 2). Further, EDTA enhanced the disruption, favoring formation of monomer (lane 3). On the other hand, both a negatively charged detergent, 0.25% N-lauroylsarcosine, and a neutral detergent, 2.5% Tween 20, showed a protective or stabilizing effect for the p53 tetramer (lanes 4 and 5). This “protection” effect is not as apparent when using three other detergents: trimethylamine N-oxide, Zwittergent, and NP40 (lanes 6–8). DLS was done to confirm the presence of tetramer in solution at a relatively low concentration of p53 (0.25 mg/mL, 4.7 μmol/L) and to determine stability of the tetramer under stressed conditions. The effective diameter of monomeric p53 is ∼9 nm (31). The DLS experiments showed that the effective diameter of our p53 is ∼45 nm (Fig. 2B). This roughly corresponds to the diameter of a tetrameric particle. The p53 tetramer has an increased Stokes' radius when compared with globular proteins of known molecular weight (31). On heating to 55°C in a 10°C increments, the protein did not undergo any increase in particle size that would correspond to unfolding or aggregation of unfolded protein. Therefore, tetrameric p53 is stable up to 55°C under the experimental conditions. A different batch of p53 (Fig. 2C) prepared under refolding conditions that favor formation of soluble, misfolded p53 (judged by SDS-PAGE and DNA-binding activity) has a higher polydispersity value (data not shown) and underwent heat-induced aggregate formation at 40°C. The sample in Fig. 2C shows a slightly larger effective diameter than the sample in Fig. 2B, with an average diameter of 53 nm compared with 45 nm. The difference in particle size observed may be explained by the fact that oxidized p53 is 18% larger than reduced p53, a value that corresponds with our experimental DLS data for Fig. 2C. Size exclusion chromatography experiments show that mild oxidation of p53 results in formation of monomers and high molecular weight species (32), consistent with our results in Fig. 2C.

Preliminary experiments were carried out to determine if PTI-p53 causes growth arrest or apoptosis when delivered into p53-deficient cells. We transfected PTI-p53 into p53-deficient PC3 prostate cancer cells or SaOS2 osteosarcoma cells using the protein transfection reagent Chariot. This reagent relies on noncovalent binding of the Chariot reagent amphipathic peptide carrier to the protein of interest (33). The efficiency of transfection was ∼30% as judged by a parallel transfection of β-galactosidase followed by 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside staining (see Supplementary Fig. S2F).³ Cells were treated with nutlin to stabilize p53 and methyl methanesulfonate to induce a p53-dependent genotoxic response (34, 35). Several methods were used to assay cellular p53 activity, including a MTS proliferation assay, nucleosome formation assay, and TUNEL staining. Results showed that PTI-p53 induced apoptosis when transfected into p53-deficient cell lines (Supplementary Fig. S2D).³

To develop p53 protein as a cancer therapeutic, the protein needs to be delivered intracellularly. One of our delivery strategies is targeting specific cancer cells though receptor-mediated endocytosis. Accordingly, we designed a GnRH-p53 fusion protein to target GnRH-receptor (GnRH-R)-positive cancer cells (protein sequence shown in Supplementary Fig. S1B).³ The expression and refolding of the GnRH-p53 fusion protein, as well as the tetramer formation and stability, were comparable with that of wild-type p53 (Fig. 3A and B). In addition, GnRH-p53 shows a similar DNA-binding profile with commercial p53 (Fig. 3C and D).

We chose three p53-defficient, GnRH-R–expressing cancer cell lines to test the function of our GnRH-p53. These cell lines are DU145 (prostate cancer), OVCAR3 (ovarian cancer), and MDA-MB-231 (breast cancer). We first tested the effect of GnRH-p53 treatment on cell proliferation, which is shown in Fig. 4A to C. The figure shows that, after 72 h of treatment, increased concentration of GnRH-p53 results in increased reduction in cellular proliferation. No effect of the same concentration of wild-type p53 treatment on proliferation was observed, clearly indicating that the effect on proliferation of these cell lines is due to the GnRH-p53 fusion protein. In a separate experiment, no effect of commercial GnRH treatment on proliferation was observed even at >40 times molar concentrations (25 μg/mL or 22 μmol/L; see Supplementary Fig. S3),³ clearly indicating that the effect on proliferation of these cell lines is not due to GnRH but to the GnRH-p53 fusion protein.

To determine if the effect on proliferation was due to apoptosis induced by GnRH-p53, two apoptosis assays were done: a nucleosome formation assay (early apoptotic event) and TUNEL staining (late apoptotic event). Results of the nucleosome formation ELISA assay are shown in Fig. 4D. In this figure, when cells were treated for 48 h with GnRH-p53 (20 μg/mL), nucleosome formation was stimulated 3- and 10-fold in OVCAR3 and DU145 cells, respectively. The MDA-MB-231 cells did not show apparent apoptosis in the concentrations of GnRH-p53 tested, probably reflecting the different sensitivity of various cell lines under condition of the assay. DNA fragmentation, a late apoptotic event, was detected most robustly in the OVCAR3 cells by TUNEL staining (see Supplementary Fig. S4)³ but not in the DU145 or MDA-MB-231 cells with the concentration of GnRH-p53 tested. OVCAR3 cells were treated for 48 h with GnRH-p53 (25 μg/mL), and a fluorescent TUNEL assay was done. Treatment with GnRH-p53 induced positive TUNEL staining in ∼30% of the cells, but wild-type p53 had no effect (Supplementary Fig. S4).³

Internalization and translocation of GnRH-p53 were detected by immunocytochemistry. The p53 antibody used in this study detects both mutant and wild-type p53; therefore, it will recognize both endogenous mutant p53 and the GnRH-p53 fusion protein (Fig. 5A and B). Cells were treated with GnRH-p53 (20 μg/mL) for 10, 60, or 120 min to capture receptor binding (10 min) and internalization and nuclear translocation (60 and 120 min) of the GnRH-p53 fusion protein. Endogenous p53 is localized exclusively in the nucleus before treatment. After 10 min of GnRH-p53 treatment, the fusion protein is localized to the plasma membrane and shows punctuate staining in the cytoplasm (Fig. 5A, arrows). By 60 min, the fusion protein is localized to the nuclear membrane and nucleus and cytoplasm as seen by punctuate staining in both (Fig. 5A, arrows). At 120 min, all p53 is localized to the nucleus exclusively. Results from confocal microscopy (Fig. 5B) are consistent with Fig. 5A. Figure 5C shows graphically the colocalization statistics from two individual fields (normalized for background staining and analyzed by volume) of GnRH-p53 with TOTO3 nuclear stain generated using Bitplane Imaris Suite (4.5.2). At 0 min, ∼90% of p53 colocalized to the nucleus, indicating a near-exclusive nuclear localization of the endogenous mutant p53. At 10 min, the colocalization decreased to ∼60% in the GnRH-p53 samples, indicating the cell membrane localization of the exogenous GnRH-p53. The internalization of the GnRH-p53 is indicated at the 60-min point, which increased to ∼70% colocalization. At 120 min, nearly all the GnRH-p53 was transported to the nucleus, evidenced by ∼94% colocalization.

The overall aim of this study is to show the feasibility of using tetrameric p53 as a cancer therapeutic. To achieve this goal, we first expressed and refolded full-length, wild-type p53 into a stable tetramer. There are 10 cysteines in the p53 protein, all of which are located in DNA-binding core domain and reduced (36). It is well known that the p53 protein tetramerizes through a noncovalent dimer of dimer model (37); therefore, it is surprising that the refolded tetramer can be stable in high concentrations of strong detergents, such as SDS and N-lauroylsarcosine, even in the presence of 8 mol/L urea (Fig. 2A, lane 4), conditions that would disrupt the quaternary, tertiary, and even secondary structures of most proteins. The structural basis for the stability of the tetramer is interesting but beyond the scope of this report.

To develop an effective p53 protein therapeutic that works intracellularly, efficient strategies for delivering the protein to the cytoplasm and nucleus need to be devised. Intracellular delivery strategies include conjugation of the protein to a chemical or peptide entity or liposome encapsulation of the protein, both strategies allowing for passage of the protein across the plasma membrane and into the cytoplasm. However, for pharmaceutical application, a fusion protein approach may be more practical.

Although there are other fusion approaches such as using HIV-1 TAT protein transduction domain for general intracellular delivery (38), we chose a targeting fusion peptide as our initial model system for proof of the concept. In general, the targeted approach should be less toxic than nonselective delivery of protein drugs. The peptide we chose is GnRH, which has been shown to efficiently deliver small molecules, peptides, and proteins into cells. GnRH is a secreted decapeptide that binds GnRH-R (39), which is expressed in the pituitary gland, ovary, placenta, breast, and prostate tissues (40). GnRH-R is also highly expressed in solid tumors and hormone-responsive cancer cell lines (41). Therefore, this strategy would allow targeting of a p53 protein therapeutic to tumors in these hormone-responsive tissue types. A GnRH fusion with the 66-kDa Pseudomonas exotoxin protein resulted in growth inhibition and killing in a wide variety of hormone-responsive cancer cell lines, including ovarian and breast cancer lines (42, 43). The fused GnRH peptide also efficiently delivered Pokeweed antiviral protein (44) and proapoptotic proteins Bik, Bax, Bak (45), and DFF40 (46) to several adenocarcinoma cell lines where they were able to significantly inhibit growth. Therefore, the GnRH fusion protein strategy has been proven to deliver functional proapoptotic proteins that can inhibit cell proliferation of the hormone-responsive cells.

In this study, we observed that GnRH-p53 has different effects on proliferation of each of the cell lines tested. GnRH-p53 slows proliferation of OVCAR3 cells and DU145 cells; on the other hand, less effect is observed on proliferation in MDA-MB-231 breast cancer cells (Fig. 4A–C; results not shown). This is consistent with nucleosome formation and TUNEL staining experiment (Fig. 4D), in which the most sensitive cell line is OVCAR3, the second is DU145, and the least sensitive is MDA-MB-231. This observation indicates that the efficiency of wild-type p53 fusion protein to rescue p53 mutants may depend on the level of activity of the mutant and the regulatory mechanisms of p53 that the mutation affects and the expression level of the GnRH-R in target cells. In addition, the efficiency of the p53 fusion protein to rescue the p53 mutant phenotype will depend on whether the wild-type fusion protein interacts with the inactive mutant p53 inside the cells. As indicated above, because our p53 is already a stable tetramer, it is unlikely that it will heterotetramerize with endogenous mutant p53. However, more detailed studies need to be done to validate this hypothesis.

Although GnRH has been shown to mediate intracellular delivery of several molecules, in this study, we showed directly the binding and delivery of tetrameric p53 into GnRH-R–positive cells. Our immunocytochemistry results indicate that GnRH-p53 is rapidly bound to the GnRH-R on the cell surface (10 min) and then internalized and translocated to the nucleus (60–120 min). After 120 min of treatment, we observe an elongation of the nuclei, which may indicate an immediate cellular effect of p53 on transcriptional regulation. The delayed effect on proliferation and cell death corresponds to the time required for initiation of p53-specific cell death and cell cycle gene expression. In addition, p53 requires extensive posttranslational modifications in order for its activity to be regulated. Although tetramerization is required for some posttranslational modifications, it is not known if all of the necessary modifications can be made to p53 in its tetrameric form. Previous work has shown that transduction of a p53-scFv antibody fusion protein resulted in cell death (47), suggesting that exogenous p53 delivered in this way undergoes appropriate posttranslational modifications inside the cell. The results shown here indicate that GnRH-p53 is functional to slow proliferation and induce apoptosis in p53-deficient cancer cell lines. In addition, wild-type p53 delivered to PC3 cells by a commercially available protein transfection reagent induces apoptosis, suggesting that the necessary posttranslational modifications are present after intracellular delivery.

The results shown in this study support the conclusion that a p53 fusion protein designed to cross the plasma membrane and be delivered into cells is a valid approach for developing a p53-based protein therapeutic. This therapeutic could be used to treat individuals with p53-deficient cancers in GnRH-responsive tissue types (ovarian, prostate, and breast) as a primary or secondary therapy. In addition, the antiangiogenesis activity of p53 (11) may extend the treatment option to more general solid tumor types.

In this report, we show that GnRH-p53 can enter GnRH-R–positive cancer cells, inhibit proliferation, and induced apoptosis of the target cells. The efficiency of the functional intracellular delivery shown in this study may not be enough for therapeutic development. On the other hand, the experiments in this study measure only the growth-suppressing effect of p53 but not the antiangiogenesis effect, which requires separate set of experiments beyond the scope of this report. Nevertheless, it is well known that efficient intracellular delivery of large proteins requires more than internalization. Although experimental evidence shows cytosolic delivery of functional fusion proteins and delivery of fusion proteins to endosomes, it is possible that some proteins that are internalized can remain trapped in the endosomes because escape from these vesicles often is inefficient (48). To overcome this potential limitation, strategies need to be devised for efficient “escape” of internalized proteins from endosomes. To do this, one of the strategies is to incorporate a peptide containing the 20 NH₂-terminal amino acids from the influenza virus hemagglutinin protein (HA2) into the fusion protein. Influenza virus uses a mechanism of escape from endosomes that involves a pH-sensitive conformational change in the HA2 protein that destabilizes lipid membranes (49, 50). Fusion of a TAT domain to the 20 NH₂-terminal amino acids of HA2 enhances the release of the fusion protein into the cytosol (48). Fusion of GnRH-p53 to this peptide may allow efficient endosomal escape and nuclear localization of the fusion protein. This strategy will be a natural extension of the present studies reported here.

The authors own stocks of ProteomTech, Inc., as part of the employment options. The authors may therefore derive financial benefits if the drug could eventually be commercialized.

We thank Constant Marks, Wayne Liu, Michael Tuan, and Edward Wong for their technical assistance; Dr. Shahrooz Rabizadeh for his helpful discussions; and Danielle Crippen for help in the immunostaining and confocal microscope images and analysis.