NSC114792, a novel small molecule identified through structure-based computational database screening, selectively inhibits JAK3

Abstract

Background: Human or animals lacking either JAK3 or the common gamma chain (gammac) expression display severe combined immunodeficiency disease, indicating the crucial role of JAK3 in T-cell development and the homeostasis of the immune system. JAK3 has also been suggested to contribute to the pathogenesis of tumorigenesis. Recent studies identified activating JAK3 mutations in patients with various hematopoietic malignancies, including acute megakaryoblastic leukemia. Importantly, functional analyses of some of those JAK3 mutations have been shown to cause lethal hematopoietic malignancies in animal models. These observations make JAK3 an ideal therapeutic target for the treatment of various human diseases. To identify novel small molecule inhibitors of JAK3, we performed structure-based virtual screen using the 3D structure of JAK3 kinase domain and the NCI diversity set of compounds.

Results: We identified NSC114792 as a lead compound. This compound directly blocked the catalytic activity of JAK3 but not that of other JAK family members in vitro. In addition, treatment of 32D/IL-2Rbeta cells with the compound led to a block in IL-2-dependent activation of JAK3/STAT5 but not IL-3-dependent activation of JAK2/STAT5. Consistent with the specificity of NSC114792 for JAK3, it selectively inhibited persistently-activated JAK3, but failed to affect the activity of other JAK family members and other oncogenic kinases in various cancer cell lines. Finally, we showed that NSC114792 decreases cell viability by inducing apoptosis through down-regulating anti-apoptotic gene expression only in cancer cells harboring persistently-active JAK3.

Conclusions: NSC114792 is a lead compound that selectively inhibits JAK3 activity. Therefore, our study suggests that this small molecule inhibitor of JAK3 can be used as a starting point to develop a new class of drugs targeting JAK3 activity, and may have therapeutic potential in various diseases that are caused by aberrant JAK3 activity.

Figure 1

Schematic diagrams of modeling of structure-based computational database screening. (A) The chemical structure of NSC114792 (C₂₆H₃₂N₄O₂S; M.W., 464.6). Figures were generated by Pymol http://pymol.sourceforge.net. (B) Stereo structures of overlaid staurosporine analog AFN941 (4ST) and NSC114792 with JAK3 kinase domain (PDB ID: 1YVJ). 4ST is drawn in yellow and NSC114792 in cyan. (C) Predicted binding model of NSC114792 to the JAK3 kinase domain. Only the residues that contact NSC114792 are labeled. These contacts were selected only when two atoms in protein and ligand are located within 3.5Å.

Figure 2

NSC114792 selectively inhibits JAK3 kinase activity in vitro. JAK1 (A), JAK2 (B), JAK3 (C), or TYK2 (D) immunoprecipitates were used for in vitro kinase assay. Each immunoprecipitate was pre-incubated with either vehicle (DMSO) alone, NSC114792 at different concentrations or the pan-JAK kinase inhibitor AG490 (100 μmol/L) for 1 hour. Kinase reactions were subsequently performed by the addition of recombinant STAT3α protein as a substrate and 2 μmol/L ATP for 30 minutes at 30°C. The reaction products were processed for Western blotting and probed with antibodies specific for phospho-STAT3, STAT3, JAK1, JAK2, JAK3, and TYK2. Note that NSC114792 blocks JAK3 kinase activity in a dose-dependent manner but failed to inhibit the kinase activity of other JAK family members. STAT3, JAK1, JAK2, JAK3, and TYK2 serve as a loading control.

Figure 3

NSC114792 affects the viability of BaF3 cells expressing an activating allele of JAK3. BaF3-JAK3^V674A cells were transformed by introducing an activating allele of JAK3 (V674A) [23]. (A) Treatment of BaF3-JAK3^V674A cells with NSC114792 results in a decrease in phosphorylated JAK3 and STAT5 levels in a dose-dependent manner. (B) NSC114792 decreases the viability of BaF3-JAK3^V674A cells in a time- and dose-dependent manner. BaF3-JAK3^WT cells were incubated for the indicated time periods in the absence or presence of IL-3, and with or without 20 μmol/L NSC114792. BaF3-JAK3^V674A cells were also cultured with NSC114792 at various concentrations for the indicated time periods. BaF3-JAK3^WT cells showed IL-3-dependent survival, and the viability of these cells was barely affected by 20 μmol/L NSC114792. By contrast, this dose of compound significantly decreased the viability of BaF3-JAK3^V674A cells. These data suggest that the effect of NSC114792 on BaF3-JAK3^V674A cell viability is not caused by the non-specific cytotoxicity of the compound. (C and D) NSC114792 blocks both JAK3 and STAT5 phosphorylation, and causes a significant decrease in cell viability in BLNK^-/- mice-derived BKO84 cells, which express constitutively-active JAK3/STAT5 due to the lack of BLNK inhibition of JAK3 [18]. (B) *, p < 0.001, indicates statistical significance compared to DMSO-treated BaF3-JAK3^V674 cells. #, p < 0.01, indicates statistical significance compared to IL-3-treated BaF3-JAK^WT cells. (D) *, p < 0.001, indicates statistical significance compared to DMSO-treated BKO84 cells.

Figure 4

NSC114792 inhibits IL-2-induced JAK3/STAT5 but not PRL- or IL-3-induced JAK2/STAT5 signaling. Effects of NSC114792 on JAK2/STAT5 and JAK3/STAT5 signaling induced by PRL (prolactin) (A) and IL-2 (B), respectively, in the rat pre-T lymphoma cell line Nb2. Effects of NSC114792 on JAK2/STAT5 and JAK3/STAT5 signaling induced by IL-3 (C) and IL-2 (D), respectively, in the murine myeloid progenitor cell line 32D stably expressing IL-2Rβ (32D/IL-2Rβ). These cells were starved for 16 hours in the presence of either vehicle (DMSO) alone, NSC114792 at different concentrations, or the pan-JAK inhibitor AG490 (100 μmol/L). Nb2 cells and 32D/IL-2Rβ cells were subsequently stimulated by PRL (100 ng/mL) or IL-2 (100 ng/mL) for 10 minutes, and IL-3 (5 ng/mL) or IL-2 (100 ng/mL) for 30 minutes, respectively. In Nb2 cells, NSC114792 decreased IL-2-induced phospho-JAK3 and -STAT5 levels in a dose-dependent manner, but failed to decrease PRL-induced phospho-JAK2 and -STAT5 levels. Similarly, in 32D/IL-2Rβ cells, this compound blocked IL-2-induced phospho-JAK3 and -STAT5, but was defective in inhibiting IL-3 induced phospho-JAK2 and -STAT5. AG490, a pan-JAK inhibitor, non-selectively blocked both phospho-JAK2 and -JAK3 in all cells.

Figure 5

NSC114792 selectively inhibits persistently-active JAK3/STAT signaling in cancer cells. The Hodgkin's lymphoma cell lines L540 (A) and HDLM-2 (B), the breast cancer cell line MDA-MB-468 (C), and the prostate cancer cell line DU145 (D) were cultured for 24 hours in the presence of either vehicle (DMSO) alone, NSC114792 at various concentrations up to 20 μmol/L, or the pan-JAK inhibitor AG490 (150 μmol/L). Whole-cell extracts were processed for Western blot analysis using antibodies specific for the molecules indicated. Treatment with NSC114792 did not alter the levels of phospho-JAK1 and -JAK2 (B-D, lanes 1-4), but it did potently inhibit phospho-JAK3 in a dose-dependent manner (A, lanes 5 and 6). Consistently, the compound decreased phospho-STAT3 only in L540 cells, which exhibit persistent JAK3 activation (lanes 7 and 8). In contrast, AG490 non-selectively decreased the phosphorylation of JAK kinases and thus led to a block in phospho-STAT3 in all cell lines tested. NSC114792 did not alter the levels of other activated oncogenic kinases, including phospho-Src (lanes 9 and 10), phospho-Lyn (lanes 11 and 12), phospho-Akt (lanes 13 and14), and phospho-ERK1/2 (lanes 15 and 16). None of these kinases were affected by treatment with NSC114792 at concentrations up to 20 μmol/L.

Figure 6

NSC114792 decreases cell viability and induces apoptosis in cancer cells with persistent JAK3 activation. (A) NSC114792 decreases cell viability only in L540 cells that express persistently-active JAK3. L540, HLDM-2, MDA-MB-468 and DU145 cells were treated with either vehicle (DMSO) alone, NSC114792 at different concentrations or the pan-JAK inhibitor AG490, and incubated for the indicated time periods. The trypan blue exclusion assay was performed to count viable cells. Results are shown as the mean of three independent experiments (± SD indicated by error bar). *, p < 0.001, indicates statistical significance compared to DMSO-treated cells. (B) TUNEL assay shows that NSC114792 induces apoptosis in L540 cells. Cells were treated with vehicle (DMSO) alone or NSC114792 at various concentrations for 72 hours. Cells were then stained with FITC-conjugated BrdU antibody and propidium iodide (PI), and subsequently subjected to flow cytometry. The percentages of TUNEL-positive cells are indicated. (C and D) Treatment with NSC114792 increases cleaved fragments of Poly (ADP-ribose) polymerase (PARP) and caspase-3, and decreases the expression of anti-apoptotic genes in a dose-dependent manner in L540 cells. Cells were treated with vehicle (DMSO) alone or NSC114792 at various concentrations for 48 hours. Whole-cell extracts were processed for Western blot analysis using antibodies specific for the molecules indicated. GAPDH serves as a loading control.