Cellular responses to γ-irradiation exposure are controlled by phosphatidylinositol 3-kinase–related kinases (PIKK) in the nucleus, and in addition, cytosolic PIKKs may have a role in such responses. Here, we show that the expression of tripeptidyl-peptidase II (TPPII), a high molecular weight cytosolic peptidase, required PIKK signaling and that TPPII was rapidly translocated into the nucleus of γ-irradiated cells. These events were dependent on mammalian target of rapamycin, a cytosolic/mitochondrial PIKK that is activated by γ-irradiation. Lymphoma cells with inhibited expression of TPPII failed to efficiently stabilize p53 and had reduced ability to arrest proliferation in response to γ-irradiation. We observed that TPPII contains a BRCA COOH-terminal–like motif, contained within sequences of several proteins involved in DNA damage signaling pathways, and this motif was important for nuclear translocation of TPPII and stabilization of p53. Novel tripeptide-based inhibitors of TPPII caused complete in vivo tumor regression in mice in response to relatively low doses of γ-irradiation (3–4 Gy/wk). This was observed with established mouse and human tumors of diverse tissue backgrounds, with no tumor regrowth after cancellation of treatment. These TPPII inhibitors had minor effects on tumor growth as single agent and had low cellular toxicity. Our data indicated that TPPII connects signaling by cytosolic/mitochondrial and nuclear PIKK-dependent pathways and that TPPII can be targeted for inhibition of tumor therapy resistance. [Cancer Res 2007;67(15):7165–74]
An understanding of cellular stress sensing and signal transduction at the molecular level is crucial for development of therapies against many pathogenic conditions, including cancer, ischemia, and viral infections (1–3). Several types of stress activate phosphatidylinositol 3-kinase–related kinases (PIKK), a family that includes several kinases, DNA-dependent protein kinase catalytic subunit (DNA-PKcs), ataxia-telangiectasia mutated (ATM), and ATM and Rad3-related (ATR) in the nucleus, as well as mammalian target of rapamycin (mTOR) that resides in the cytosol and mitochondria (4, 5). PIKKs play a role as signal transducers from sensor molecules in response to stress and phosphorylate a network of regulatory factors to initiate DNA repair and cell cycle arrest, pathways often constitutively activated in transformed cells (6, 7). Cells and tissues failing to express PIKKs have deficient p53 stabilization, cell cycle checkpoints, and an increased susceptibility to apoptosis induced by γ-irradiation, as present in genetic diseases with failure to express ATM and related diseases (5, 8).
Cancer therapy frequently depends on the induction of DNA damage (e.g., treatment with γ-irradiation or DNA topoisomerase inhibitors). Nuclear PIKKs (i.e., ATM, ATR, and DNA-PKcs) are therefore also possible targets to increase the efficiency of such therapy. Incubation of tumor cells with inhibitors of these PIKKs blocks DNA repair responses, which increases susceptibility to γ-irradiation–induced apoptosis in vitro (9). This occurs despite that PIKK inhibitors prevent stabilization of p53, suggesting that apoptosis of such γ-irradiated cells is p53 independent. However, also normal tissues require PIKKs for protection against DNA damage (4, 5, 8). Experimental in vivo studies of mice treated with inhibitors of nuclear PIKKs have shown significant effects on tumor growth, but tumor regression responses are usually not observed using such protocols (10, 11). mTOR, a cytosolic PIKK family member, has a crucial role in the integration of signals from nutritional sensing, regulation of protein translation, and control of Akt kinase activation (12). Inhibitors of mTOR sensitize tumors to γ-irradiation in mice, with the occasional observation of tumor regression, and such inhibitors show promising results in trials against some forms of cancer (12–15). In addition, inhibition of Chk1 and Chk2 kinases, downstream targets of PIKK signaling pathways, increases the susceptibility to tumor treatment, an effect that may include targeting of cancer stem cells (16, 17). It is yet unclear which PIKKs, or PIKK-dependent pathways, represent targets for efficient cancer therapy.
Several stimuli that cause cellular stress, proteotoxic, oncogenic, and nutrient stress, increase expression of tripeptidyl-peptidase II (TPPII; refs. 18–21). In this article, we have explored the potential role of TPPII in γ-irradiation–induced DNA damage responses in vitro and in resistance to γ-irradiation–based cancer therapy in vivo. TPPII is built from a unique 138-kDa subunit expressed in multicellular organisms from Drosophila to Homo sapiens (22–24). Data from Drosophila suggest that the TPPII complex consists of repeated subunits forming two twisted strands with a native structure of ∼6 MDa (24). TPPII degrades cytosolic polypeptides (18–20), generates certain MHC class I ligands (25, 26), and complements the proteasome in protein turnover. However, other roles of TPPII may also exist, which may be unrelated to protein turnover. TPPII regulates transduction of apoptotic signals as well as centrosome homeostasis by unclear mechanisms (21, 27–29). We here found that the expression of TPPII is controlled by mTOR and that TPPII is rapidly translocated into the nucleus in response to γ-irradiation. TPPII was required for efficient stabilization of p53 and for efficient cell cycle arrest at low γ-irradiation doses (2–4 Gy). Akt activation and resistance to low serum concentrations was also impaired in cells with inhibited TPPII expression. We observed a dramatic in vivo radiosensitization of mouse and human tumors by the tripeptide-based TPPII inhibitor [Z-Gly-Leu-Ala-OH (Z-GLA-OH)], with complete regression in response to repeated 3 to 4 Gy doses of γ-irradiation. We found that TPPII has an essential role in responses to γ-irradiation and that this peptidase can be targeted for treatment of tumors.
Materials and Methods
Cells and culture conditions. EL-4 is a benzpyrene-induced lymphoma cell line derived from the C57BL/6 mouse strain. EL-4.wt and EL-4.TPPIIi are EL-4 cells transfected with the pSUPER vector (30), empty versus containing the small interfering RNA (siRNA) directed against TPPII. EL-4.TPPIIwt and EL-4.TPPIIwt/G725E are EL-4 cells transfected with pSUPER-TPPIIi and pcDNA3-TPPIIwt versus pcDNA3-TPPIIwt/G725E (see plasmids). HeLa cells are human cervical carcinoma cells. YAC-1 is a Moloney leukemia virus–induced lymphoma cell line derived from the A/Sn mouse strain. ALC is a T-cell lymphoma induced by radiation leukemia virus D-RadLV, derived from the C57BL/6 mouse strain. For induction of stress, cells were γ-irradiated 5 to 10 Gy and incubated at 37°C and 5.3% CO2.
Enzyme inhibitors. NLVS is an inhibitor of the proteasome that preferentially targets the chymotryptic peptidase activity and efficiently inhibits proteasomal degradation in live cells. Ala-Ala-Phe-chloromethylketone (AAF-CMK) is a serine peptidase inhibitor that efficiently inhibits TPPII (19), and butabindide is a specifically designed inhibitor of TPPII (30). Z-GLA-OH is an inhibitor of subtilisin (Bachem), a bacterial enzyme with an active site that is homologous to that of TPPII (31). Wortmannin is an inhibitor of PIKK family kinases (Sigma), and rapamycin is a specific inhibitor of mTOR (Cell Signaling Technology). All inhibitors were dissolved in DMSO and stored at −20°C until use.
Plasmids and gene transfection. TPPII siRNA-expressing pSUPER (30) plasmids were constructed as follows. Nonphosphorylated DNA oligomers (Thermo Hybaid) were resuspended to a concentration of 3 μg/mL. Each oligo pair (1 μL) was mixed with 48 μL of annealing buffer [100 mmol/L potassium acetate, 30 mmol/L HEPES-KOH (pH 7.4), 2 mmol/L magnesium acetate] and heated to 95°C for 4 min and 70°C for 10 min and then slowly cooled to room temperature. Annealed oligomers (2 μL) were mixed with 100 ng of pSUPER plasmid (digested with BglII and HindIII), ligated, transformed, and plated on Amp/LP plates as described previously (30). Colonies were screened for the presence of inserts by EcoRI-HindIII digestion and DNA sequencing. Annealed oligomer pairs were as follows: pSUPER-TPPIIi, 5′-GATCCCCGATGTATGGGAGAGGCCTTTCAAGAGAAGGCCTCTCCCATACATCTTTTTGGAAA-3′ (forward) and 5′-AGCTTTTCCAAAAAGATGTATGGGAGAGGCCTTCTCTTGAAAGGCCTCTCCCATACATCGGG-3′ (reverse).
For generation of stable transfectants, 5 × 106 cells were washed in PBS and then resuspended into 500 μL PBS in a Bio-Rad gene pulser and pulsed with 10 μg DNA and 250 V at 960 μF and selected by resistance to G418. EL-4.TPPIIwt and EL-4.TPPIIwt/G725E are EL-4 cells transfected with pSUPER-TPPIIi (10 μg) and pcDNA3-TPPIIwt or pcDNA3-TPPIIwt/G725E (1 μg) and selected for resistance to G418 (resistance gene in pcDNA3). To allow expression of the pcDNA3-TPPIIwt plasmids in EL-4.TPPIIi cells, we inserted three silent mutations in the 3′ region of TPPII among the nucleotides that interact with the pSUPER-TPPIIi–encoded siRNA (this plasmid was denoted TPPIIwt), in addition to the mutation in position 725 (denoted TPPIIwt/G725E).
Antibodies and antisera. The following molecules were detected by the antibodies specified: Akt by rabbit anti-Akt serum (Cell Signaling Technology); ATM by the mouse monoclonal 2C1 (Genetex, Inc.); phosphorylated Akt (Ser473) by 193H2 rabbit anti–phosphorylated Akt serum (Cell Signaling Technology); γ-H2AX by rabbit anti-γ-H2AX (Cell Signaling Technology); p21 by SX118 (R&D Systems); p53 (R&D Systems); Rae-1 by monoclonal Rat anti-mouse Rae-1, 199215 (R&D Systems); and X-linked inhibitor of apoptosis protein (XIAP) by monoclonal mouse anti-human XIAP, 117320 (R&D Systems). For detection of TPPII, we used chicken anti-TPPII serum (Immunsystem). Western blotting was done by standard techniques. Protein concentration was measured by bicinchoninic acid protein assay reagent (Pierce Chemical Co.). Protein (5 μg) was loaded per lane for separation by SDS-PAGE unless stated otherwise.
Immunocytochemistry. Cells were attached to glass coverslips through cytospin and fixed in acetone/methanol (1:1) for 1 h; then, the slides were rehydrated in balanced salt solution (BSS) buffer for 1 h. The first antibody was added and remained for 1 h until a brief wash in BSS, after which a secondary conjugate (anti-rabbit FITC) was added and incubated for 1 h. Then, the slides were washed and stained with Hoechst 333258 for 30 min. Finally, the slides were mounted with DABCO mounting buffer and kept at 4°C until analysis.
Flow cytometry. For staining of cell surface Rae-1 antigens, we incubated 0.5 × 106 to 1.0 × 106 cells with 50 μL of Rae-1 monoclonal antibody 199215 at 20 μg/mL and incubated on ice for 30 min. After washing in PBS, we sequentially incubated with biotinylated polyclonal rabbit anti-rat Ig (DakoCytomation) and streptavidin-FITC (PharMingen), with washing in PBS after each step. Fluorescence was quantified by a FACSCalibur. Dead cells were excluded by propidium iodide staining.
Enzyme purification and peptidase assay. Cells (100 × 106) were sedimented and lysed by vortexing in glass beads and 500 to 1,000 μL of homogenization buffer [50 mmol/L Tris base (pH 7.5), 250 mmol/L sucrose, 5 mmol/L MgCl2, 1 mmol/L DTT]. Cellular fractions generated by differential centrifugation were as follows: nuclear fraction (pellet of 14,000 × rpm, 15 min), microsomal fraction (pellet of 100,000 × g centrifugation of supernatant from nuclear fraction, 1 h), and high molecular weight cytosolic fraction (pellet of 100,000 × g centrifugation of supernatant of microsomal fraction, 3 to 5 h). The resulting high molecular weight pellet was dissolved in 50 mmol/L Tris base (pH 7.5), 30% glycerol, 5 mmol/L MgCl2, and 1 mmol/L DTT, and 1 μg of high molecular weight protein was used as enzyme in peptidase assays or in Western blotting for TPPII expression. To test the activity of TPPII, we used the substrate Ala-Ala-Phe-aminomethylcoumarin (AAF-AMC; Sigma) at 100 μmol/L concentration in 100 μL of test buffer composed of 50 mmol/L Tris base (pH 7.5), 5 mmol/L MgCl2, and 1 mmol/L DTT. Inhibitors were preincubated with the enzyme for 1 h at 37°C before adding the substrate. Cleavage activity was measured by emission at 460 nm in a LS50B luminescence spectrometer (Perkin-Elmer).
Tumor growth experiments. Tumor cells were washed in PBS and resuspended in a volume of 200 μL/inoculate. The cells were then inoculated into the right flank at 106 per syngeneic C57BL/6 mouse, and growth of the tumor was monitored by measurement twice weekly. The initiation of antitumor treatment of the mice was to some extent individualized according to when tumor growth started in each mouse. The mice were irradiated with 4 Gy before tumor inoculation to inhibit antitumor immune responses. The tumor volume was calculated as the mean volume in mice with tumor growth according to (a1 × a2 × a3) / 2 (the numbers a1, a2, and a3 denote tumor diameter, width, and depth, respectively). The time of first palpation varied between different mice, although the general pattern of growth was similar in virtually all of the mice. In most diagrams, a log scale is used to better visualize the therapeutic effects against small tumors (i.e., the presence of complete rejections). For inhibition of TPPII in vivo, we made i.p. injections with 13.8 mg/kg body weight (14 μL of a 50 mmol/L solution/mouse) of the subtilisin inhibitor Z-GLA-OH twice weekly, diluted into 200 μL PBS. All γ-irradiations were full body exposures.
Cell cycle arrest at low doses of γ-irradiation depends on TPPII. TPPII expression is increased by several stimuli that result in cellular stress, but which signals that cause this increase is not clear (18–21). We therefore tested whether this increase was controlled by PIKKs and whether mTOR was involved. To trigger a cellular stress response where members of the PIKK family members control signal transduction, we used γ-irradiation. By Western blotting analysis of cellular lysates from the T-cell lymphoma line EL-4 with chicken anti-TPPII serum, we found that TPPII expression was increased by γ-irradiation, as found with other types of stress (18–21). Further, this increase was not present in γ-irradiated EL-4 cells treated with 1 μmol/L wortmannin (PIKK inhibitor) or 100 nmol/L rapamycin (mTOR inhibitor) that instead reduced TPPII expression (Fig. 1A). Treatment with NLVS, a proteasomal inhibitor, reduced down-regulation of TPPII in wortmannin-treated γ-irradiated EL-4 cells, suggesting that TPPII is degraded at least in part by the proteasome in the absence of PIKK signaling (Fig. 1A). Increased levels of TPPII were previously found in EL-4 cells growing in the presence of inhibited proteasomal activity (18–20).
To further study whether TPPII had any role in cellular responses mediated by PIKKs, we generated stable EL-4 transfectants expressing siRNA directed against TPPII, encoded by the pSUPER vector (denoted EL-4.TPPIII; ref. 30). EL-4.TPPIIi cells had both inhibited expression and activity of TPPII in comparison with EL-4.wt cells (transfected with empty pSUPER vector; Supplementary Data I). We first analyzed the subcellular distribution of TPPII in untreated versus γ-irradiated EL-4 cells using EL-4.TPPIIi cells as specificity control of the immunocytochemical staining with chicken anti-TPPII serum. TPPII was reported as a soluble cytosolic peptidase (22, 25), but we here found rapid translocation of TPPII into the nucleus of most γ-irradiated EL-4 cells (Fig. 1B). This was evident already 1 h following γ-irradiation 5 Gy exposure of EL-4 cells as detected by immunocytochemical analysis of TPPII, and this translocation was inhibited by treatment with 100 nmol/L rapamycin (Fig. 1B). A similar translocation was observed in ALC and YAC-1 lymphoma as well as Lewis lung carcinoma (LCC) cells.3
Activation of PIKKs is required to halt cellular proliferation in response to DNA damage (4, 5). We observed that proliferation of EL-4.wt cells was halted after exposure to 2 to 4 Gy of γ-irradiation, but this response was much weaker in EL-4.TPPIIi cells that continued proliferation (Fig. 1C). Proliferation of both EL-4.TPPIIi and EL-4.wt cells was blocked in response to higher doses of γ-irradiation (10 Gy; Fig. 1C). Further, EL-4.wt control cells showed both G1 and G2-M arrest, whereas EL-4.TPPIIi cells failed to arrest in G1 but accumulated in G2-M after exposure to 10 Gy of γ-irradiation. It is presently not clear if this reflects the presence of an active G2-M checkpoint mechanism in EL-4.TPPIIi or mitotic failure (Fig. 1D). In addition, EL-4.TPPIIi cells displayed a reduced sub-G1 peak compared with EL-4.wt cells, suggesting a reduced level of acute EL-4.TPPIIi cell death in response to γ-irradiation (Fig. 1D). Initial detection of DNA damage was still present in γ-irradiated EL-4.TPPIIi cells as measured by immunocytochemical detection of γ-H2AX (Ser139-phosphorylated H2AX; Supplementary Data II). H2AX is phosphorylated in response to ATM activation, which triggers the formation of DNA repair foci (4). Thus, TPPII expression depends on mTOR and was required to efficiently halt proliferation of EL-4 cells in response to γ-irradiation but not for detection of DNA damage.
Failure to stabilize p53 in cells with inhibited TPPII expression. The transcription factor p53 initiates cell cycle arrest in response to many types of stress, and its expression is controlled through direct phosphorylation by PIKKs; also, mTOR may have a role in this process (12, 15). By Western blotting analysis in cellular lysates of γ-irradiated EL-4.wt cells, we found increased levels of p53, whereas those of EL-4.TPPIIi cells showed low levels (Fig. 2A). However, treatment with NLVS increased p53 expression of γ-irradiated EL-4.TPPIIi cells, suggesting that p53 was still synthesized but degraded by the proteasome in EL-4.TPPIIi cells (Fig. 2A). We designed three additional pSUPER-expressed siRNAs, of which two inhibited p53 stabilization, following γ-irradiation of transiently transfected EL-4 cells (Supplementary Data III). p21, a transcriptional target of p53, was very weakly expressed in EL-4.TPPIIi cells following exposure to γ-irradiation compared with EL-4.wt control cells (Fig. 2B). Further, EL-4.pcDNA3-TPPII cells that stably overexpress TPPII showed increased levels of p53 following exposure to γ-irradiation in comparison with EL-4.pcDNA3 cells (Fig. 2C; ref. 32). In addition, treatment of EL-4.pcDNA3-TPPII cells with the specific TPPII inhibitor butabindide caused substantially reduced p53 expression (Fig. 2C; ref. 33). We found that p53 expression was also TPPII dependent in γ-irradiated YAC-1 and ALC lymphoma cells, where virtually no p53 was detectable following stable expression of pSUPER-TPPIIi (Supplementary Data IV). We failed to find expression of p53 in LLC cells, also supporting that nuclear translocation of TPPII did not require p53 (Supplementary Data IV).3 We noted substantial levels of p53 in some of our control tumor cell lines also before exposure to γ-irradiation (e.g., YAC-1 and ALC), a phenomenon in line with the frequent presence of a constitutive DNA damage response in transformed cells (6, 7). To test if p53 and TPPII were physically linked, we next did coimmunoprecipitation experiments using an anti-serum directed against the NH2 terminus of p53 followed by Western blot analysis for TPPII. In p53 immunoprecipitates from lysates of EL-4.wt cells, we detected TPPII, levels that were much increased by γ-irradiation (Fig. 2D, at 8 h). We also detected ATM in p53 immunoprecipitates from EL-4.wt, but not from EL-4.TPPIIi cells, as measured by Western blotting (Fig. 2D). Further, NLVS-treated EL-4.TPPIIi cells, which accumulated p53, also failed to show ATM in p53 immunoprecipitates (Fig. 2D). These data indicate that TPPII is required for a physical link between p53 and ATM following γ-irradiation.
A BRCA COOH-terminal–like motif of TPPII required for efficient p53 stabilization in response to γ-irradiation. BRCA COOH-terminal (BRCT) repeat domains are often contained within proteins controlling DNA damage signaling pathways, where they control interactions with ATM substrates (34–36). We found one region of TPPII centered around the GG doublet at position 725, which matched most, but not all, requirements of a BRCT motif (Fig. 3A; ref. 34). We did site-directed mutagenesis of the characteristic Gly-Gly doublet present in many BRCT sequences (Fig. 3A,, asterisk), mutating it into Gly-Glu in our pcDNA3-TPPII vector. To allow expression of this plasmid in EL-4.TPPIIi cells, we inserted three silent mutations in the 3′ region of TPPII among the nucleotides that interact with the pSUPER-TPPIII–encoded siRNA (this plasmid was denoted TPPIIwt), in addition to the mutation in position 725 (denoted TPPIIwt/G725E). We found that both TPPIIwt as well as TPPIIwt/G725E mutant molecules were stably expressed in EL-4 cells cotransfected with pSUPER-TPPIIi, with TPPII expression levels similar to those observed in EL-4.wt control cells (Fig. 3B). The expression of p53 was analyzed in EL-4.wt, EL-4.TPPIIwt, and EL-4.TPPIIwt/G725E transfectant cells exposed to γ-irradiation. EL-4.TPPIIwt/G725E cells showed minor expression of p53 compared with EL-4.TPPIIwt and EL-4.wt control cells, further supporting the role for TPPII in γ-irradiation–induced p53 stabilization (Fig. 3C). The G725E mutation had no significant effects on the serine peptidase activity of TPPII as measured by fluorimetric substrate cleavage by partially purified TPPII enzyme (Supplementary Data V). The subcellular localization of TPPII was examined in EL-4.TPPIIwt and EL-4.TPPIIwt/G725E cells by immunocytochemical analysis before and after γ-irradiation. EL-4.TPPIIwt cells showed a predominantly cytosolic staining, with a subsequent relocalization into the nucleus following γ-irradiation (Fig. 3D). However, this failed to occur in EL-4.TPPIIwt/G725E cells, where TPPII instead remained mainly cytosolic following exposure to 5 Gy of γ-irradiation (Fig. 3D). These data indicated that a nuclear localization of TPPII was important for γ-irradiation–induced p53 stabilization and that this depended on a BRCT-like domain of TPPII.
TPPII controls cellular requirements for growth factors during proliferation in vitro. Because the expression of TPPII was controlled by mTOR, we examined whether EL-4.wt versus EL-4.TPPIIi cells responded differently to changes in nutrition. mTOR-Rictor complexes control Ser473 phosphorylation of Akt kinase, an important step for Akt activation and for transduction of certain growth factor signals (although other PIKKs are also claimed for this role; ref. 37–39). We detected substantial levels of phosphorylated Ser473 Akt in lysates of EL-4.wt cells. However, EL-4.TPPIIi cells displayed very low levels of phosphorylated Ser473 Akt, whereas the total expression of Akt protein was similar to EL-4.wt cells (Fig. 4A). In addition, we find increased Ser473 phosphorylation of Akt in EL-4.pcDNA3-TPPII, in comparison with EL-4.pcDNA3 control cells, further supporting that TPPII expression correlates with Akt Ser473 phosphorylation (Fig. 4A). These data were in line with the serum requirements during in vitro cell growth of EL-4.wt versus EL-4.TPPIIi cells. In normal medium (5% serum), EL-4.TPPIIi cells showed an increased rate of proliferation, compared with EL-4.wt, but also an increased accumulation of dead cells (Fig. 4B). By lowering serum concentrations to 1%, this accumulation was accelerated compared with EL-4.wt cells (Fig. 4B). Further, TPPII was strongly induced in EL-4.wt on days 5 to 7 (following seeding of cells at 100,000/mL), and replenishment of medium down-regulated TPPII expression (Fig. 4C,, arrow). In addition, EL-4.pcDNA3-TPPII cells were able to do limited growth in 0.5% serum, which EL-4.pcDNA3 cells did not (Fig. 4D). These results indicated that TPPII expression was important for Akt Ser473 phosphorylation and cell survival during in vitro culture. We previously reported that up-regulation of TPPII causes increased expression of c-IAP-1 and XIAP molecules (endogenous caspase inhibitors) in EL-4.pcDNA3-TPPII cells (21), and XIAP stability is controlled by Akt (40). By treatment with etoposide, we found that expression of XIAP was substantially higher in EL-4.wt cells compared with EL-4.TPPIIi cells, with a slower rate of degradation, in line with previous studies (Supplementary Data VI; ref. 21). Activation of PIKKs also mediates expression of cell surface NKG2D ligands (41). By flow cytometric measurements, we detected expression of Rae-1 (belonging to a mouse family of NKG2D ligands) on EL-4.wt and YAC-1 cells, whereas ALC failed to express Rae-1 ligands. In contrast, EL-4.TPPIIi and YAC-1.TPPIIi cells expressed minor amounts of Rae-1 (Supplementary Data VII). Overall, these data suggest the requirement for TPPII in several stress-induced pathways that respond to γ-irradiation or changes in nutrition.
TPPII expression controls γ-irradiation resistance of EL-4 tumors in vivo. PIKKs, including ATM, DNA-PKcs, and mTOR, as well as their targets Chk1 and Chk2, are possible target molecules for novel cancer therapies (16, 17, 42). To address whether TPPII-mediated growth regulation was important for in vivo tumor growth, we inoculated 106 EL-4.wt control or EL-4.TPPIIi cells into syngeneic C57BL/6 mice. We found that EL-4.wt and EL-4.TPPIIi cells grew at an approximately equal rate, suggesting that TPPII was not important for growth of EL-4 tumors in vivo (Fig. 5A, and B, Control). However, we also treated mice carrying either tumors of EL-4.wt or EL-4.TPPIIi cells with 2 to 4 doses of 4 Gy (400 Rads) γ-irradiation. We found that this had minor effects on the size of tumors of EL-4.wt cells that continued to grow despite γ-irradiation (Fig. 5A, γ-irradiation indicated with arrow). In contrast, mice carrying tumors of EL-4.TPPIIi cells responded to γ-irradiation treatment with complete regression of established tumors (Fig. 5B). These data resembled those obtained with tumors of EL-4.ATMi or EL-4.TPPIIwt/G725E cells because these also failed to resist γ-irradiation in vivo, whereas control tumors resisted γ-irradiation doses of 3 to 4 Gy (Fig. 5C). EL-4.ATMi are EL-4 cells with inhibited ATM expression due to a stable expression of an ATM-directed pSUPER plasmid (Supplementary Data VIII). These data may suggest TPPII as a possible target to increase in vivo γ-irradiation susceptibility of tumor cells in vivo.
Tripeptide-based TPPII inhibitors radiosensitize tumors in vivo. To invent an experimental therapy from these findings, a low molecular weight compound may be required, but it is unclear whether a catalytic TPPII inhibitor causes effects resembling those of siRNA. Further, several of the available catalytic inhibitors of TPPII, such as butabindide and AAF-CMK (a serine peptidase inhibitor that efficiently inhibits TPPII), are unsuitable for use in vivo due to insufficient stability and specificity of these compounds. TPPII is a subtilisin-type serine peptidase, with a catalytic domain that is homologous to bacterial subtilisins (31). We found that the tripeptide subtilisin inhibitor Z-GLA-OH efficiently inhibited TPPII, slightly less efficient than observed for butabindide (which has a Ki50 of 7 nmol/L; ref. 33), as observed by inhibited TPPII cleavage of the substrate AAF-AMC (Fig. 6A). Moreover, Z-GLA-OH was relatively stable in serum compared with both AAF-CMK and butabindide.3
We next examined the toxicity of Z-GLA-OH to cell lines in vitro and compared with AAF-CMK. TPPII inhibitors down-regulate the activity of the ubiquitin-proteasome pathway and inhibit growth of certain tumor cells in vitro (18, 20, 43). We therefore compared inhibitors of TPPII for their ability to accumulate an unstable green fluorescent protein (GFP)-based proteasomal reporter substrate by treatment of EL-4.Ub-R-GFP cells (44). We found that AAF-CMK inhibited the ubiquitin-proteasome pathway as indicated by R-GFP accumulation in EL-4.Ub-R-GFP cells measured by flow cytometry (44). However, neither Z-GLA-OH nor butabindide did so (even at 100 μmol/L to 1 mmol/L; Fig. 6A). Further, Z-GLA-OH failed to inhibit growth of two different Burkitt lymphoma lines, in contrast to AAF-CMK that showed substantial toxic effects (Supplementary Data IX). This discrepancy may in part be due to the particular choice of cell line for these in vitro assays, but the acute toxicity and ubiquitin-proteasome pathway inhibition caused by 100 μmol/L AAF-CMK are unlikely due to TPPII inhibition because other TPPII inhibitors fail to mediate these effects. We concluded that Z-GLA-OH inhibits TPPII and has low toxicity to several tumor cell lines in vitro.
To test the effects of catalytic TPPII inhibition during tumor γ-irradiation in vivo, we exposed C57BL/6 mice with established EL-4 tumors to γ-irradiation doses of 4 Gy (one dose per week) and injections with Z-GLA-OH twice weekly (13.8 mg/kg body weight). As observed previously, weekly γ-irradiation doses of 4 Gy had minor effects on growth of established EL-4 tumors in C57BL/6 control mice. In contrast, following injection with Z-GLA-OH, we observed complete tumor regression after three to four doses of 4 Gy γ-irradiation in all mice tested (in total >50 mice, one representative experiment 7–8 mice/group in Fig. 6B). When these tumors were no longer palpable, the treatment was canceled, and no regrowth of tumors was observed for the entire period of observation (over 3 months). These therapeutic effects showed a dose-dependent response both for γ-irradiation and amounts of Z-GLA-OH, with complete tumor rejections observed also in some mice treated with reduced doses (Supplementary Data X).
One common reason behind tumor therapy resistance, including in vivo resistance to γ-irradiation, is p53 mutations (45). To test whether also p53-mutated tumors responded to γ-irradiation in the presence of TPPII inhibitors, we similarly inoculated 106 LLC cells in syngeneic C57BL/6 mice (see p53 expression in Supplementary Data IV). We found that LLC tumors were virtually insensitive to repeated γ-irradiation doses of 4 Gy, and Z-GLA-OH only (in the absence of γ-irradiation) gave no effect (Fig. 6C). In contrast, we observed complete regression of established LLC tumors to γ-irradiation in mice injected with Z-GLA-OH (Fig. 6C). We found that the protected dipeptide analogue Z-GL-OH was ineffective both in terms of TPPII inhibition and radiosensitization of LLC tumors (Fig. 6A and C). Further, butabindide was also an ineffective in vivo as a radiosensitizer of LLC tumors, in line with its poor stability in serum (25). TPPII is an evolutionary conserved enzyme with an identity of 96% at the amino acid level between human and mouse, and we observed strong tumor regression also of human HeLa cervical carcinoma cells in Z-GLA-OH–treated severe combined immunodeficient (SCID) mice in response to γ-irradiation (Fig. 6D). A reduced dose of γ-irradiation (1.5 Gy/dose) was used because SCID mice have substantially reduced radioresistance. From these results, we concluded that a low molecular weight compound designed to inhibit a bacterial enzyme has antitumor effects in vivo in mice, most likely through an inhibitory effect against its mammalian homologue, TPPII.
Toxicity studies show that Z-GLA-OH had minor effects in vivo as single agent in doses up to 100 mg/kg in a preliminary study.4
U. Höglund, unpublished data.
This article describes an essential role for TPPII in cellular responses to γ-irradiation. The observations of complete in vivo tumor regression in mice injected with peptide-based TPPII inhibitors, in response to relatively low doses of γ-irradiation, suggest that our findings have therapeutic potential. Our data indicated that mTOR activity was required for translocation of a large proportion of cytosolic TPPII into the nucleus of γ-irradiated cells. TPPII controlled a γ-irradiation–induced physical link between ATM and p53 and may thereby have a role, direct or indirect, in the recruitment of regulatory factors to PIKKs, a role that is unclear at this moment. ATM, ATR, and DNA-PKcs have a certain degree of redundancy in stabilization of p53, with multiple NH2-terminal sites for p53 phosphorylation and with more than one PIKK targeting the same site (46). A failure of regulatory factor recruitment (e.g., to DNA repair foci) may deny p53 activation by several PIKKs, which would explain the strong reduction of p53 stabilization that we observe in pSUPER-TPPIII–expressing cells. Our in vivo data may support a model were γ-irradiation exposure in the absence of TPPII expression in vivo leads to a failure to respond to stress, resulting in an irreversible state of tumor damage. Radiosensitization by in vivo targeting of the DNA damage response in γ-irradiated tumor cells makes sense from a cell biological perspective, but the high efficiency of TPPII targeting in this experimental tumor therapy was unexpected. A large number of established and experimental cancer therapies inhibit the growth of tumors in mice (e.g., inhibitors of PIKKs and Chk proteins; refs. 10–17). However, reported data fail to support that these cause consistent complete tumor regressions as observed here. Inhibitors of PIKKs and Chk proteins cause a substantially increased acute toxicity in response to γ-irradiation in vitro, which was not observed here with EL-4.TPPIIi cells (see Fig. 1D). In addition to direct damage on the tumor γ-irradiation also has an antiangiogenic effect, mTOR signaling is required for protection of tumor vasculature exposed to γ-irradiation (47). However, tumor regression following treatment with angiogenesis inhibitors regularly results in tumor relapses after termination of treatment (48). Still it will be important to determine if TPPII has a role in mTOR-Raptor or mTOR-Rictor signaling. The exact mechanism for in vivo tumor protection of this TPPII-dependent stress response pathway remains to be explained.
Previous observations have shown that TPPII activity is up-regulated in several lymphoma cell lines with low proteasomal activity during cellular adaption to growth in the presence of proteasomal inhibitors and overexpression of c-myc. This was associated with a TPPII-mediated compensation for insufficient proteasomal substrate degradation (18–20). In view of the present data, increased TPPII activity may have been a general response to stress, including activation of PIKKs. TPPII was also studied in relation to centrosomal duplication errors during mitosis, where it was reported that butabindide inhibited such events in vitro (43). TPPII inhibitors were proposed as agents to suppress mitotic infidelity in tumors (28, 29, 43). Lymphoma cells with constitutively overexpressed TPPII have overexpression of IAP family proteins, whereby errors during mitosis may have resulted from inhibition of apoptosis at the mitotic checkpoint (21, 29). These deregulations may also be caused by altered signal transduction in pathways controlled by PIKKs because several PIKKs as well as their substrates localize to centrosomes during mitosis (49).
The degree of homology between the NH2-terminal catalytic domain of TPPII and subtilisin of Bacillus subtilis (381 amino acids) is 25% to 30%, including a conserved pentapeptide encompassing the catalytically active Ser449 of TPPII (Asn-Gly-Thr-Ser-Met; ref. 31). Bacterial subtilisins are thoroughly studied enzymes, with numerous reports on crystal structure and enzymatic function, which may further assist development of TPPII inhibitors (50). TPPII may potentially be a useful molecular target during the acute phases of ischemia, or HIV infections, because pathogenesis of these diseases depends on signaling by PIKKs (2, 3). Future clinical testing of these tripeptide-based TPPII inhibitors may give further information about whether they are useful for therapy in patients.
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).
Grant support: Swedish Cancer Society, Swedish Research Council, Swedish Foundation for Strategic Research, and af Jochnick stiftelsen.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
We thank H-G. Ljunggren and C. Söderberg-Nauclér for support.