The Kinase Inhibitor PP1 Blocks Tumorigenesis Induced by RET Oncogenes¹ Free

Rearrangements of the RET tyrosine-kinase receptor (1) with different genes are found frequently in PTCs,³ in particular, in radiation-induced childhood PTC. RET/PTC1 and -3, generated by the fusion of RET to the H4 or RFG genes, respectively, are the most prevalent rearrangements (2, 3, 4, 5). RET/PTC oncogenes cause thyroid PC Cl 3 cells to proliferate in the absence of TSH (6) and induce PTC in transgenic mice (7). Germ-line mutations of RET cause MEN2 (8, 9). The MEN2A subtype is characterized by MTC, pheochromocytoma, and parathyroid hyperplasia; MEN2B by MTC, pheochromocytoma, and ganglioneuromas of the intestinal tract. MTC is the only feature of familial MTC . In virtually all MEN2A and in several familial MTC cases, there are substitutions of cysteines of the extracellular RET domain, whereas most MEN2B cases are caused by the M918T mutation in the tyrosine kinase RET domain. M918T is also found in sporadic MTC (10), with M918T mutation-positive tumors often displaying a more aggressive phenotype (11).

RET/PTC and RET/MEN2A oncoproteins have constitutive kinase activity consequent to ligand-independent dimerization (12, 13). The M918T mutation modifies the structure of the kinase, thereby switching on the enzymatic function and altering substrate specificity of RET/MEN2B (12, 14).

Protein kinases can be inhibited by ATP or substrate mimics (15, 16). Their low molecular weight, selectivity, bioavailability, and favorable pharmacokinetics properties make these signal transduction inhibitors successful in the clinic (17, 18, 19, 20). Here we demonstrate that the pyrazolo-pyrimidine PP1 inhibits the enzymatic activity and the transforming effects of RET oncoproteins in NIH3T3 fibroblasts and thyroid carcinoma cell lines.

The 12 different compounds used in this study included cinnamomalonitriles (AG82, AG213, AG490, AG555, and AG556), quinoxalines (AG1295), quinazolines (AG1478), and pyrazolo-pyrimidines (PP1, AGL2137, AGL2164, AGL2174, and AGL2189), and were synthesized in one of our laboratories (A. L.; Refs. 15, 16). Stock solutions (50 mm) were made in 100% DMSO. Equivalent DMSO concentrations served as vehicle controls.

NIH3T3 and NIH-EGFR, a gift of P. P. Di Fiore (21), NIH-RET/PTC3 (3), NIH-MEN2A, NIH-MEN2B (12), and NIH-RAF (22) were cultured in DMEM supplemented with 10% calf serum (Life Technologies, Inc., Paisley, PA). After overnight starvation, NIH-EGFR was stimulated or not with 100 ng/ml EGF (Upstate Biotechnology Inc., Lake Placid, NY) for 15 min. PC Cl 3, PC-RET/PTC3 (6), and PC-MOS (22) cell lines were cultured in Coon’s modified Ham F12 medium (Sigma Chemical Co., St. Louis, MO) supplemented with 10% calf serum (Life Technologies, Inc.) and a mixture of 6H (TSH, insulin, transferrin, somatostatin, hydrocortisone, and glycil-histidyl-lysine; Sigma Chemical Co.). Human thyroid-carcinoma-derived cell lines TPC1 (23), FB2⁴, derived from papillary carcinomas harboring the RET/PTC1 rearrangement, and ARO (Ref. 24; a gift of J. Fagin), derived from an anaplastic carcinoma negative for RET/PTC rearrangements, were cultured in RPMI 1640 supplemented with 10% FCS (Life Technologies, Inc.). The 293 cells were from American Type Culture Collection and were grown in DMEM supplemented with 10% FCS. Transient transfections were carried out with 5 μg of DNA by using the LipofectAMINE reagent according to the manufacturer’s instructions (Life Technologies, Inc.).

Cell lysates containing comparable amounts of proteins, estimated by a modified Bradford assay (Bio-Rad, Munchen, Germany), were immunoprecipitated with the required antibody or subjected to direct Western blot. Immune complexes were detected with the enhanced chemiluminescence kit (Amersham Pharmacia Biotech, Little Chalfort, United Kingdom). Antiphosphotyrosine (4G10) and anti-EGFR were from Upstate Biotechnology Inc. Anti-MAPK and anti-phospho-MAPK were from New England Biolabs (Beverly, MA). Anti-RET is a polyclonal antibody raised against the tyrosine kinase protein fragment of human RET (25). Secondary antibodies coupled to horseradish peroxidase were from Santa Cruz Biotechnology (Santa Cruz, CA).

For the autokinase assay, subconfluent cells were solubilized in lysis buffer [50 mm HEPES (pH 7.5), 150 mm NaCl, 1% glycerol, 1% Triton X-100, 1.5 mm MgCl₂, and 5 mm EGTA] without phosphatase inhibitors. Proteins (200 μg) were immunoprecipitated with anti-RET; immunocomplexes were recovered with protein A-Sepharose beads, washed 5 times with kinase buffer [20 mm HEPES (pH 7.5), 150 mm NaCl, 10% glycerol, 0.1% Triton X-100, 15 mm MgCl₂, and 15 mm MnCl₂] and incubated (20 min at room temperature) in kinase buffer containing 2.5 μCi [γ-³²P]ATP and unlabeled ATP (20 μm). Samples were separated by SDS-PAGE. Gels were dried and exposed to autoradiography or to phosphorimager (GS525; Bio-Rad, Hercules, CA). For the phosphorylation of the synthetic substrate, immunocomplexes prepared in the presence of phosphatase inhibitors were incubated (20 min at room temperature) in kinase buffer containing 200 μm poly-GT (Sigma Chemical Co.), 2.5 μCi [γ-³²P]ATP, and unlabeled ATP (20 μm). Samples were spotted on Whatman 3 MM paper (Springfield Mill, United Kingdom), and ³²P incorporation was measured with a β-counter scintillator (Beckman). The GST-RET/TK plasmid was generated by PCR amplification of the intracellular RET domain (residues 718-1072) and fusion to the GST coding sequence into the pEBG vector, a kind gift of S. Meakin (26). GST-RET/TK was purified from 293 cell lysates using glutathione Sepharose according to standard protocols.

For cell proliferation assays, NIH3T3 (10,000/plate) and human thyroid carcinoma cells (50,000/plate) were seeded on 60-mm dishes in the appropriate medium. One day after (day 0), 5 μm of PP1 or vehicle alone were added, medium was changed every 2 days, and cells were counted every 1 (fibroblasts) or 2 days (thyroid cells). For cytofluorometric analysis, cells were grown to subconfluence, serum starved for 24 h, and then subjected or not to 5 μm of PP1 treatment for an additional 24 h. After harvesting, cells were fixed in cold 70% ethanol in PBS. Propidium iodide (25 μg/ml) was added, and samples were analyzed with a FACScan flow cytometer (Becton Dickinson, San Jose, CA) interfaced with a Hewlett Packard computer (Palo Alto, CA). For the soft-agar colony assay, cells were seeded on 60-mm plates (10,000 cells/plate) in 0.3% agar in complete medium on a base layer of 0.5% agar with or without different concentrations of the inhibitor; 500 μl of the compound solution were added every 3 days to the top layer. Colonies were counted 15 days later.

NIH-RET/PTC3 cells (50,000/mouse) and NIH-RAF (250,000/mouse) were inoculated s.c. into the right dorsal portion of 6-week-old male BALB/c-nu/nu mice (Jackson Laboratories). PP1 (200 μg/mouse/day) or vehicle alone was injected in the tumor area starting 1 day after cell injection. Tumor diameters were measured with calipers every 2–3 days. Tumor volumes (V) were calculated by the rotational ellipsoid formula: V = A × B²/2 (A = axial diameter; B = rotational diameter). Animal studies were conducted in accordance with the Italian regulation for experimentations on animals. No mice showed signs of wasting or other signs of toxicity.

RET-derived oncoproteins (Fig. 1,A) autophosphorylate in vitro in the absence of ligand. We evaluated the tyrosine kinase inhibiting activity of several classes of inhibitors, i.e., cinnamomalonitriles (AG82, AG213, AG490, AG555, and AG556), quinoxalines (AG1295), quinazolines (AG1478), and pyrazolo-pyrimidines (PP1; Refs. 15, 16) using an in vitro autophosphorylation assay on NIH3T3 cells expressing the RET/PTC3 oncogene. PP1, reported previously as an efficient c-Src-related kinase and PDGFR inhibitor (27), inhibited RET/PTC3 autophosphorylation in a dose-dependent manner. No other compound significantly inhibited RET/PTC3 kinase activity (Fig. 1,B, top panel; data not shown). Three other pyrazolo-pyrimidines, AGL2137, AGL2164, and AGL2174, inhibited RET/PTC3 autophosphorylation but to a lesser extent compared with PP1, whereas AGL2189 was a poor inhibitor of the kinase (Fig. 1 B, bottom panel).

Then, we quantitated the potency of the RET/PTC3 protein to phosphorylate the synthetic peptide poly-GT in the presence of the different inhibitors. Again, PP1 was found to be the most effective compound (Fig. 1,C). We also evaluated the inhibitory effect of PP1, AGL2137, AGL2164, and AGL2174 on poly-GT phosphorylation mediated by RET/MEN2A and RET/MEN2B. At a concentration of 0.5 μm, PP1 inhibited RET/MEN2A and RET/MENB more than the other pyrazolo-pyrimidine derivatives (Fig. 1,D). Finally, we generated the isolated RET kinase domain fused to a GST tag and measured the IC₅₀ of PP1 for RET/PTC3 and GST-RET/TK using the poly-GT in vitro phosphorylation assay. PP1 IC₅₀ was found to be ∼80 nm for both kinases (Fig. 1 E).

We tested the effects exerted by PP1 on RET/PTC3 autophosphorylation and signal transduction in living cells. Parental and EGFR-expressing NIH3T3 cells served as controls. After 24 h of serum deprivation, NIH-RET/PTC3 and NIH-EGFR were treated with 5 μm of PP1 for 0, 2, or 6 h. In EGFR-transduced cells, EGFR activation was induced with 100 ng/ml EGF for 10 min before harvesting. RET/PTC3 and EGFR were immunoprecipitated, and phosphotyrosine content was analyzed by immunoblot. RET/PTC3 showed constitutive phosphorylation, whereas phosphotyrosine was detected in EGFR only after EGF stimulation. PP1 completely abolished RET/PTC3 autophosphorylation as early as 2 h after treatment, whereas it had no effect on EGFR phosphorylation (Fig. 2 A).

Oncogenic versions of RET constitutively activate the Ras/MAPK pathway via recruitment of Shc-Grb2-Sos and Frs2-Grb2-Sos complexes (28, 29). Thus, after serum starvation, NIH-RET/PTC3 but not parental cells maintained high levels of phosphorylated MAPK, as revealed by immunoblotting with an anti-phospho-MAPK-specific antibody (Fig. 2,B). We determined whether PP1 affected RET/PTC3- and EGFR-induced signaling. As early as 2 h after exposure to 5 μm of PP1, there was a dramatic reduction of RET/PTC3- but not EGFR-dependent MAPK phosphorylation (Fig. 2 B).

RET/PTC3 induces morphological transformation, serum- and anchorage-independent proliferation, and tumorigenicity in nude mice of NIH3T3 cells (3). PC Cl 3, a continuous line of Fischer rat thyroid cells, requires a 6H mixture, which includes TSH and insulin, for proliferation (22). PC Cl 3 cells transduced with RET/PTC3 become independent from 6H for proliferation and lose their differentiated morphology (6). We treated RET/PTC3-expressing NIH3T3 and PC Cl 3 cells with 5 μm of PP1 for 24 h and analyzed the morphological changes induced by the drug. As controls, we used parental or either RAF-(NIH-RAF) or MOS-(PC-MOS) transformed counterparts. As shown in the top panels of Fig. 3, PP1 caused a complete morphological reversion of NIH-RET/PTC3 cells, whereas neither parental nor NIH-RAF cells were affected by PP1. Similarly, PC-RET/PTC3 cells reverted to a flat and polygonal morphology and started to grow in clusters on PP1 treatment. PP1 had no effect on parental PC Cl 3 or PC-MOS cells (Fig. 3, bottom panels).

We then studied the effects exerted by PP1 on growth rate of RET/PTC3-expressing NIH3T3. A remarkable inhibitory effect on cell growth was seen in NIH-RET/PTC3 cells treated with 5 μm of PP1. As expected, PP1-induced proliferative inhibition corresponded to an increase of the G₀/G₁ cell fraction as shown by flow cytometry analysis. In contrast, PP1 had no effect on NIH-RAF cells (Fig. 4,A). The TPC1 and FB2 human thyroid carcinoma cell lines bear the RET/PTC1 rearrangement. Treatment of TPC1 and FB2 with 5 μm of PP1 arrested cell growth. Accordingly, flow cytometry scan showed a dramatic increase of the G₀/G₁ fraction and a marked reduction of the S fraction (Fig. 4,B). PP1 also induced morphological reversion of TPC1 and FB2 cells; however, they are rather flat and, for this, such effects were less sharp than those observed in in vitro transfected cells. In contrast, PP1 had a very modest effect on the proliferation and the morphology of another human thyroid carcinoma cell line, ARO, which does not contain a RET/PTC rearrangement (Fig. 4 B). Such a modest effect is likely explained by the inhibition of other cellular tyrosine kinases by PP1.

Finally, we tested PP1-dependent inhibition of the capacity of NIH-RET/PTC3 to grow in a semisolid medium. At a concentration of 5 μm, PP1 inhibited NIH-RET/PTC3 colony formation by >10-fold. A representative microscopic field is shown in Fig. 5,A, and the average results of three independent experiments are shown in the bar graphs of Fig. 5,B. NIH-RAF-anchorage-independent growth was unaffected by PP1. We used the soft agar colony formation assay to compare the RET/PTC3 inhibitory effect of PP1 with that of the pyrazolo-pyrimidines AGL2137, AGL2164, and AGL2174. Again, PP1 was the most active compound (Fig. 5 B).

NIH-RET/PTC3 and NIH-RAF cells form tumors within a few days when injected s.c. in nude mice. To investigate the potential of PP1 as an anti-RET/PTC3 cancer drug, we injected two groups of 10 mice each with 50,000 NIH-RET/PTC3 or 250,000 NIH-RAF cells. One day after cell injection, 5 mice from each group were treated with PP1 (200 μg/mouse/day) and the other 5 with DMSO. Both treatments were administered under skin in the cell injection area. Among NIH-RET/PTC3 injected mice, all of the control animals developed a tumor (of ∼200 mm³) within 10 days after cell injection. Only 1 PP1-treated animal developed a tumor until day 10 after injection. Thereafter, tumors appeared also in PP1-treated mice but their growth rate was much lower than in DMSO-treated animals (Fig. 6, top panel), remaining <100 mm³ even at day 20 after injection (data not shown). Importantly, NIH-RAF induced tumors were not affected at all by PP1 treatment (Fig. 6, bottom panel); in two weeks, both DMSO and PP1 treated animals developed tumors of sizes similar to those of NIH-RET/PTC3 untreated mice.

The first demonstration of the validity of “signal transduction therapy” was obtained by the systematic analysis of a series of low molecular weight protein tyrosine kinase inhibitors directed toward the substrate site of EGFR kinase domain. (30, 31, 32). Since then, several protein tyrosine kinase inhibitors, natural and synthetic, have been tested for their anticancer properties, and some are being tested in clinical trials or have already been approved for human cancer therapy (19).

RET-derived oncoprotein kinase inhibitors might be beneficial for the treatment of RET oncogenes bearing human tumors and especially of MTCs that respond very poorly to chemotherapeutic agents. In addition, anti-RET agents could be used in MEN2 carriers to delay parafollicular cell hyperplasia formation and, consequently, preventive thyroidectomy. We tested the RET-blocking capacity of several classes of tyrosine kinase inhibitors. We found that some compounds with a pyrazolo-pyrimidine moiety had the greatest inhibitory effect. Among the pyrazolo-pyrimidines studied, PP1 was the most effective. Notably, PP1 exerted powerful growth inhibitory effects on human thyroid carcinoma cell lines harboring RET/PTC rearrangements. PP1 is not selective for RET, being a potent inhibitor also of Hck, lck, and fynT kinases (IC₅₀ of 5 nm), and a good inhibitor of c-Src (IC₅₀ of 200 nm) and PDGFR (IC₅₀ of 100 nm; Ref. 27). Therefore, in addition to the direct effect on the RET kinase in vitro, we cannot exclude indirect effects mediated in vivo by inhibition of other kinases and mainly of c-Src, a pivotal downstream RET effector (30). Should this be the case, a single molecule can be used for “multiple-signal transduction therapy” of RET-dependent tumor formation. Similarly, PP1 has successfully been proposed as an inhibitor of both PDGFR and c-Src to prevent vascular remodeling and restenosis (33).

The crystal structure of PP1-bound Hck kinase has been elucidated (34). PP1 binds to the kinase by inserting the methylphenyl group into a hydrophobic pocket adjacent to the ATP binding site. In most kinases, a conserved bulky amino acid (methionine) lies at the bottom of such a pocket, and mutation of this residue to a small amino acid like glycine greatly enhances sensitivity to PP1 (35). RET possesses a valine residue at that position. It would be interesting to mutagenize this residue in RET to determine whether it is possible to modify RET kinase sensitivity to PP1.

2-Indolidone derivatives have been described as RET/PTC1 inhibitors (36). However, they seem weaker than PP1 in inhibiting RET, the most powerful (cpd 1) of them having an IC₅₀ for RET/PTC1 of ∼30 μm. We believe that that PP1 efficacy on RET/PTC3-induced transformation is probably attributable not only to its efficient activity on the RET kinase but also to the simultaneous inhibition of c-Src. Whatever the case, the potent inhibitory effect of PP1 is encouraging for the treatment of human tumors in which RET is mutated.

We thank James A. Gilder for editing the text, James A. Fagin for ARO cells, Fulvio Basolo for FB2 cells, Susan O. Meakin for the pEBG-GST vector, and Pier Paolo Di Fiore for the NIH-EGFR cells. We also thank Marc Billaud, Sergie Manie, Rosa Marina Melillo, Fortunato Ciardiello, and G. Tortora for useful suggestions. Additionally, we thank Giovanni Sequino and Salvatore Sequino for animal care.