Targeting TUBG1 in RB1-negative tumors

Abstract

The disruption of microtubule dynamics serves as a pivotal strategy for eliminating tumor cells, despite its accompanying toxicities affecting non-tumor cells. This study investigates the potential of selectively targeting γ-tubulin1 (TUBG1) as a therapeutic strategy in cancer treatment. By elucidating the TUBG1-E2F1-retinoblastoma protein (RB1) network, we introduce a novel compound, 4-(6-((3-Methoxyphenyl)amino)pyrimidin-4-yl)-N,N-dimethylbenzenamine, (L12). L12 treatment enhanced RB1 expression and selectively targeted cells with impaired RB1 signaling, while reduced E2F1 expression attenuated its cytotoxicity. Furthermore, L12-mediated cytotoxicity depends on an E2F1-mediated upregulation of procaspase 3 expression, highlighting the role of E2F1 in the apoptotic response. Unlike traditional tubulin-targeting agents, L12's specificity for tumor cells lies in its inhibitory effects on TUBG1, without affecting the second human isoform of TUBGs, TUBG2. Despite its interaction with specific kinases, the concentrations required for antitumor effects are 100-fold lower than those influencing kinase activities. Subsequent investigations underscore L12's reduced neuronal axonal toxicity compared to vincristine. Lastly, L12 demonstrates promising results in inhibiting tumor growth in xenografted small cell lung cancer models, demonstrating potential specificity toward tumor cells while minimizing adverse effects on healthy tissues. This research emphasizes the potential of TUBG1 inhibitors as a promising advancement in personalized chemotherapy approaches and their potential as a groundbreaking treatment for various cancers.

Keywords: RB1; TUBG1; cancer; chemotherapy.

FIGURE 1

Cytotoxic effect of L12 treatment and its dependency on the RB1 pathway functionality. (A) Depiction of the structure of L12. (B) Western blotting (WB) analysis of RB1 and phosphorylated Ser⁷⁸⁰ RB1 proteins using an anti‐RB1 (RB1) and anti‐phospho‐Ser⁷⁸⁰ RB1 (pRB1) antibodies. Total lysates from various cell lines were probed, including the RB1‐deficient human retinoblastoma Y79, the osteosarcoma U2OS with a constitutively phosphorylated RB1 at Ser⁷⁸⁰ (pRB1), the mammary gland epithelial cell line MCF10A, and the adenocarcinoma alveolar basal epithelial A549, both with functional RB1 pathways, as well as the RB1‐deficient small cell lung carcinoma U1690. α‐tubulin (αTubulin) served as a loading control. Cells treated with different L12 concentrations for 24 h were assessed for cell accumulation in the sub‐G1 fraction (indicating dead cells). The histograms depict mean ± SD of the percentages of cells in the sub‐G1 fractions (N = 3; *p < .05). (C) Examination of L12's impact on the downstream gene target of TUBG, RB1, was conducted by WB analysis using an anti‐RB1 antibody on MCF10A and A549 cell lysates. α‐tubulin and actin were utilized as loading controls. Numbers on the western blot indicate variations in RB1 expression compared to the vehicle control. To account for protein loading discrepancies, RB1's protein concentration was normalized to its ratio with the respective loading control for each treatment. The graph displays relative RB1 protein expression across MCF10A (N = 4) and A549 (N = 3) cell lines when treated with varying L12 concentrations, with data presented as mean ± SD. Please note that the intensity of the RB1 bands may vary between figures as a result of adjusted exposure times for western blot analysis. These adjustments were made to prevent overexposure and to accurately capture differences in RB1 expression levels across experimental conditions.

FIGURE 2

Influence of TUBG1, E2F1, and RB1 protein levels on the cytotoxic effect of L12 treatment. (A and B) The MCF10A cell lines used include: Control (MCF10A, non‐modified parental cells), MCF10A sh TUBG (stably expressing TUBG shRNA) and MCF10A shTUBG TUBG1 (stably co‐expressing TUBG shRNA and a sh‐resistant TUBG1 gene). (A) MCF10A cells were treated with DMSO (vehicle) or the indicated concentrations of L12 for 24 h. DNA content was measured by nuclear counter to determine the cell cycle profile and the percentage of cells in the sub‐G1 fraction (indicative of dead cells). Histograms display the results, and graphs summarize the mean ± SD percentages of sub‐G1 cells (N = 3; two‐way ANOVA, ****p < .0001). (B) Western blotting (WB) was performed to analyze TUBG and RB1 protein levels in total lysates using anti‐TUBG and anti‐RB1 antibodies. Actin served as the loading control. Graphs illustrate relative protein expression (Student's t test, N = 3, *p < .05, **p < .01). (C and D) The U2OS cell lines used include: Control (U2OS, non‐modified parental cells), U2OS shE2F1 (transiently expressing E2F1 shRNA) and U2OS E2F1 sh E2F1 (transiently co‐expressing E2F1 sgRNA and a E2F1 gene). U2OS cells were treated with DMSO (vehicle) or 50 nM L12 for 24 h. (C) DNA content was measured to determine the cell cycle profile and the percentage of cells in the sub‐G1 fraction. Histograms show representative data, and graphs summarize the mean ± SD percentages of sub‐G1 cells (N = 4; Student's t test, *p < .01, **p < .01). (D) WB was used to analyze E2F1 and procaspase 3 protein levels in total lysates using anti‐E2F1 and anti‐procaspase 3 antibodies. GAPDH served as the loading control. Graphs display relative protein expression (Student's t test, N = 4, *p < .05, **p < .01). The numbers above the blots (WB) represent the normalized intensity of the protein bands. (E and F) The A549 cell lines used include: Control (A549, non‐modified parental cells), A549 sgRB1 (stably expressing RB1 sgRNA) and A549 RB1 sgRNA RB1 (stably co‐expressing RB1 sgRNA and a sg‐resistant RB1 gene). (E) A549 cells were treated with DMSO (vehicle) or the indicated concentrations of L12 for 24 h. The cell cycle profile and percentage of sub‐G1 cells were determined. Histograms represent the results, and graphs show mean ± SD percentages of sub‐G1 cells (N = 3; two‐way ANOVA, ****p < .0001). (F) WB analyzed TUBG and RB1 protein levels in total lysates using anti‐TUBG and anti‐RB1 antibodies. Actin served as the loading control. Graphs depict relative protein expression (Student's t test, N = 3, *p < .05, ****p < .0001). To ensure accurate comparisons of RB1 protein levels under different conditions, Western blot exposure times were optimized for each experiment to balance signal detection and prevent overexposure, enabling the detection of subtle differences in RB1 expression.

FIGURE 3

Variable cytotoxic effects of L12 in cells expressing TUBG1 vs. TUBG2. (A) U2OS cells, U2OS cells stably expressing Flag‐tagged TUBG2 (TUBG2‐Flag), or TUBG1 single guide (sg) RNA (sgTUBG1, resulting in TUBG1 knockout) co‐expressing either sg‐resistant TUBG1 or a sg‐resistant TUBG2 were treated with DMSO (vehicle) or varying concentrations of L12 for 24 h. Total lysates were analyzed by western blotting (WB; N = 3). Antibodies targeting the C‐terminus (T3320, rabbit) or N‐terminus (T6557, mouse) of TUBG were used to detect endogenous and recombinant TUBG proteins, with actin serving as a loading control. DNA content was measured using a nuclear counter, and histograms display cell cycle profile changes, specifically in the sub‐G1 fraction. Graphs present normalized TUBG1 and TUBG2 levels (from WB) relative to actin and mean ± SD percentages of cells in the sub‐G1 fraction (N = 3). A schematic highlights the GTPase domain (residues 5–255) and DNA‐binding domain (DBD; residues 334–451) of the human TUBG1 (h‐TUBG1) gene. Amino acid differences between TUBG1 and TUBG2 are shown, with gray and blue indicating TUBG1‐specific residues and magenta denoting TUBG2‐specific residues. (B) WB analysis of cytosolic (Cytosol) and chromatin fractions from the indicated U2OS cells. Antibodies targeting TUBG's C‐terminus (T3320, rabbit) or N‐terminus (T6557, mouse) were used to detect endogenous and recombinant TUBG proteins. T3320 preferentially detects TUBG1, while T6557 preferentially detects TUBG2, especially in fractionated samples, where T6557's specificity for TUBG1 improves. Densitometric analysis of TUBG1 and TUBG2 levels are shown, normalized to α‐tubulin (cytosolic marker) or histone (chromatin marker). An anti‐flag antibody was used to detect TUBG2–Flag. (C) WB analysis of total lysates shows RB1, TUBG1, and TUBG2 protein levels using anti‐RB, T3320, and T6557 antibodies. An α‐tubulin (αTubulin) served as a loading control. The graphs depict relative RB1, TUBG1, and TUBG2 expression across the indicated cell lines (RB1: TUBG1‐sgRNA‐U2OS‐TUBG1 and TUBG1‐sgRNA‐U2OS‐TUBG2, N = 8; U2OS‐TUBG2 Flag, N = 3; TUBG1: U2OS and TUBG1‐sgRNA‐U2OS‐TUBG1, N = 3; TUBG2: U2OS and TUBG1‐sgRNA‐U2OS‐TUBG2, N = 3). (D) Confocal microscopy images of U2OS‐TUBG2‐Flag cells stained with an anti‐TUBG and anti‐Flag antibodies. Images highlight the location of TUBG2‐Flag at γ‐tubules and centrosomes. Colocalization pixel maps (CPM) of magenta/red and green channels are included. White regions signify colocalization, with arrowheads and arrows indicating γ‐tubules and centrosomes, respectively. Scale bars: 10 μm. Graphs show the fluorescence intensity of T3320 (TUBG1) or anti‐Flag (TUBG2) found at γ‐tubules and centrosomes (Student's t test, ****p < .0001; γ‐tubules: N = 127; centrosomes: N = 108). Note that the issue of antibody specificity encountered in WB analysis does not affect immunofluorescence assays, as the proteins maintain a different conformation that is not influenced by SDS treatment. (E) Confocal microscopy images of TUBG1‐sgRNA‐U2OS‐TUBG1 and TUBG1‐sgRNA‐U2OS‐TUBG2 cells stained with anti‐TUBG antibodies. Graphs display the mean percentages of cells with γ‐tubules (N = 5; Student's t test, **p < .01) Hoechst was used for nuclear staining. Scale bars: 10 μm.

FIGURE 4

L12 stabilizes γ‐tubule without affecting axonal integrity. (A) Confocal microscope images display U2OS cells exposed to cold treatment, immediately fixed and stained (0 min), or treated with warm medium containing either DMSO (vehicle) or 50 nM L12. The images demonstrate that γ‐tubules disassemble during the cold treatment and are reassembled after a 20‐min incubation at 37°C. Cells were immunofluorescently labeled using anti‐TUBG and α‐tubulin antibodies. The key finding is the mean percentage of cells with γ‐tubules post 30‐min cold treatment in the presence of 50 nM L12, followed by a 20‐min incubation at 37°C. This is illustrated in the accompanying graph (N = 5; Student's t test, **p < .01). For each sample, a minimum of 100 cells was assessed. Scale bars: 10 μm. Arrowheads and arrows serve to denote γ‐tubules and centrosomes, respectively. (B) Z‐stacks images present an average intensity projection of sequential 1‐μm interval images from human H9 neural stem cells (NSC) differentiated into neurons over 14 days. These cells were treated with 50 nM L12 or 50 nM vincristine for 24 h. After fixation, they were immunofluorescently labeled using an anti‐β3‐tubulin antibody, which acts as a neural axon and neural differentiation marker. The graph illustrates the mean value of the β3‐tubulin signal in axons (N = 5; Student's t test,****p < .0001).

FIGURE 5

L12 stabilizes TUBG1 against thermal denaturation. (A) Cellular thermodynamic stabilization of TUBG, α‐tubulin, and actin were assessed in live Y79 cells. These cells underwent a temperature gradient, and the levels of soluble TUBG, α‐tubulin, and actin were subsequently examined using western blot analysis with the specific antibodies. (B) After a 30 min pre‐treatment of Y79 cells with DMSO or varying L12 concentrations, ligand‐binding alterations in heat‐induced TUBG precipitation were observed via western blotting. The specificity of L12 binding to TUBG1 was confirmed by using anti‐α‐tubulin and ‐actin antibodies. Note that the protein levels of α‐tubulin and actin were not significantly affected by L12. Graphs show densitometric analysis of the changes in the TUBG and α‐tubulin levels post heat‐induced precipitation, normalized to the 25°C control set as 1 (N = 4). (C) Live U2OS cells stably expressing TUBG1 sgRNA (sgTUBG1) and co‐expressing either a sg‐resistant TUBG1 (TUBG1‐sgRNA‐U2OS‐TUBG1) or TUBG2/TUBG1‐sgRNA‐U2OS‐TUBG2), were subjected to a temperature gradient. Soluble levels of TUBG1, TUBG2, and actin post‐gradient were evaluated using anti‐TUBG (T3320, recognized TUBG1; T6557, recognized TUBG2) and anti‐actin antibodies in western blot analysis. (D) Following a 30 min pre‐treatment of TUBG1‐sgRNA‐U2OS‐TUBG1 and TUBG1‐sgRNA‐U2OS‐TUBG2 cells with DMSO or various concentrations of L12, ligand‐induced changes in heat‐induced TUBG precipitation were monitored. Graph illustrates variations in soluble TUBG1 and TUBG2 levels post heat‐induced precipitation as measured by densitometric analysis of the TUBG protein content in the western blot membranes. The data were normalized to untreated cells (set as 1; N = 4; Student's t test, *p < .05, **p < .01).

FIGURE 6

Influence of leucine 321, glutamine 167, and asparagine 251 on L12 binding to TUBG. (A and B) Cartoon representation of the following TUBG1 structure: Simulated (mustard‐yellow), X‐ray (dark blue) and Alpha‐Fold (AF; light blue). The structures depict the predicted binding pocket of GTP and colchicine with a view down the L12‐binding pocket, highlighting the residues predicted to interact with L12. (B) L12 (blue) is shown in a stick representation at the colchicine‐binding site. A close view of the binding pocket, with approximately the same orientation, shows the side chains of residues predicted to interact with L12 (best energy pose). The C‐alpha atoms, when visible, are colored orange, while the remaining side chain atoms are colored cyan. The mutated residues are shown in red. (C) The illustration provided delineates the GTPase domain (residues 5–255) and the DNA‐binding domain (DBD; residues 334–451) of the human TUBG1 (h‐ TUBG1) gene. The magenta letters in the depiction denote residues comprised in the colchicine binding pocket interacting with L12 (Q167, N251 and L321). TUBG1‐sgRNA‐U2OS‐TUBG1, TUBG1‐sgRNA‐U2OS‐TUBG1‐Ala³²¹ (LA), TUBG1‐sgRNA‐U2OS‐TUBG1‐Ala¹⁶⁷ (QA), and TUBG1‐sgRNA‐U2OS‐TUBG1‐Ala²⁵¹ (NA) cells were treated with 50 nM L12 for 24 h. Changes in the cell cycle profiles and the accumulation of cells in the sub‐G1 phase were analyzed by determining the DNA content of cells using a nuclear counter (histograms). The graph presents the mean ± SD of the percentage of cells in the sub‐G1 fraction (N = 4; Student's t test, ****p < .0001). (D and E) The thermodynamic stabilization of TUBG and actin was examined in live TUBG1‐sgRNA‐U2OS‐TUBG1‐Ala³²¹ (LA), TUBG1‐sgRNA‐U2OS‐TUBG1‐Ala¹⁶⁷ (QA), and TUBG1‐sgRNA‐U2OS‐TUBG1‐Ala²⁵¹ (NA) cells. Cell populations underwent a temperature gradient (°C) or were exposed to specific temperatures post 30‐min L12 pre‐treatment (D). Subsequently, soluble levels of TUBG and actin were determined using western blot analysis with respective antibodies. Graphs in section D depict changes in the soluble TUBG levels post heat‐induced precipitation, with data normalized to a 25°C control set as 1 (N = 4).

FIGURE 7

Impact of L12 on microtubule outgrowth. (A–C) U2OS cells (A) or those stably co‐expressing TUBG‐sgRNA and either TUBG1 (B) or a TUBG2 (C) sg‐resistant were subjected to cold treatment. Subsequently, these cells were either immediately stained (0 min) or stained 2 min after the addition of warm medium supplemented with: 2 μL/mL DMSO (vehicle), 10 μM colcemid (Colc.; known to inhibit microtubule regrowth), or in the presence of various concentrations of L12, as specified. After treatment, cells were fixed and subjected to immunofluorescence labeling with an anti‐α‐tubulin antibody. Scale bars are indicated as 10 μm. Arrows indicate astral microtubule regrowth from centrosomes. The graphs display the average aster length measured with ImageJ (Fiji, RRID:SCR_003070) software in U2OS cells (A; N = 100) and in those stably co‐expressing TUBG‐sgRNA with either TUBG1 (B; N = 100) or TUBG2 (C; N = 100). Statistical analysis was conducted using Student's t test, with significance denoted as ****p < .0001.

FIGURE 8

L12 effect on tumor growth of xenografted U1690 mice. Left represents the changes in body weight of mice xenografted with U1690 cells during treatment. The middle shows tumor growth in xenografted U1690 cells treated three days per week with a dosage of 20 mg/kg L12. L12 treatment started on day 12 when the tumors reached a mean tumor volume of 100 mm³. (two‐way ANOVA, *p < .05, **p < .01, ***p < .001, N = 8 mice per group). As the xenografted tumor surpassed a tumor volume of 1 cm³, three mice from the control group and one mouse in the L12‐treated group were sacrificed, as indicated.