HKDC1 functions as a glucose sensor and promotes metabolic adaptation and cancer growth via interaction with PHB2

Abstract

Glucose sensing and metabolic adaptation to glucose availability in the tumor microenvironment are critical for cancer development. Here we show that HKDC1, a hexokinase highly expressed in cancer associated with poor prognosis, functions as a glucose sensor that alters its stability in response to environmental glucose. The glucose-sensing domain is located between amino acids 751-917, with Ser896 as a key residue that regulates HKDC1 stability by affecting Lys620 ubiquitination. This sensing mechanism enables cellular adaptation to glucose starvation by promoting mitochondrial fatty acid utilization. Furthermore, HKDC1 promotes tumor growth by sequestering prohibitin 2 (PHB2) to disable its suppressive effect on SP1, thus promoting the expression of pro-oncogenic molecules. Abrogation of HKDC1 by genetic knockout or by glucose depletion releases PHB2, leading to suppression of cancer cell proliferation and inhibition of tumor growth. Our study reveals a previously unrecognized role of HKDC1 in glucose sensing and metabolic adaptation, and identifies HKDC1 as a potential therapeutic target.

Fig. 1. HKDC1 protein level changes in response to glucose via a regulatory domain within amino acids 751-917.

A Western blot analysis of HKDC1 protein in the indicated lung cancer cell lines cultured with or without glucose for 24 h. B Western blot analysis (left panel) of hexokinases and qRT-PCR analysis (right panel) of HKDC1 mRNA expression in A549 cells cultured in glucose-free medium for the indicated time points. Data are shown as mean ± SD of 3 independent experiments, Student’s t test: *P, < 0.05, ***, P < 0.001. C Change of HKDC1 protein level in response to glucose availability. A549 cells in normal culture were first switched to glucose-free RPMI 1640 medium for 24 h, and then switched back to standard RPMI 1640 medium with glucose for the indicated time. The levels of HKDC1, HK1, HK2, HK4 and β-actin were analyzed by Western blotting. D Western blot analysis of HKDC1, HK1, HK2, and HK4 in A549 cells cultured in standard or glucose-free RPMI 1640 medium for the indicated time with or without 10 µM MG132. E Western blot analysis of HKDC1, HK1, HK2, and HK4 in A549 cells cultured in standard or glucose-free RPMI 1640 medium with or without 5 mM 2-DG for the indicated time. F Study design to analyze the key domain in HKDC1 that mediates glucose response. HKDC1 structure was predicted using I-TASSER web-tool, and its potential glucose binding was modeled by molecular docking using the PatchDock web tool and further refined by FireDock. Two possible glucose-binding domains were identified (shown in red and green color). Based on this structural information, six series truncation mutants (Δ1 to Δ6) of HKDC1 were constructed, and transfected into A549 cells with pre-knockout of HKDC1 or into HEK293T without endogenous HKDC1 expression. The cells were then cultured in medium with or without glucose for 24 h, and cell lysates were analyzed by Western blotting for HKDC1 protein levels. G HKDC1 in A549 cells was first knocked out by CRISPR/Cas9 (sgRNA #1), and the cells were then transfected with full-length HKDC1 (wild-type, WT) plasmid or the indicated truncated HKDC1 plasmids for 24 h. The cells were then cultured with or without glucose for 24 h. Levels of HKDC1 protein were detected by western blotting using HKDC1 polyclonal antibodies.

Fig. 2. Ser896 is critical for regulation of HKDC1 protein stability by glucose.

A Molecular docking of glucose binding to HKDC1 using the AutoDock Vina program (https://vina.scripps.edu/). The amino acid residues with potential interaction with glucose are labeled in red color. B Study design to test the effect of different HKDC1 mutations on response to glucose depletion. Wild-type and mutant HKDC1 expression vectors were constructed, and transfected into A549 cells with pre-knockout of HKDC1 or into HEK293T cells (without endogenous HKDC1 expression). The cells were cultured with or without glucose, and HKDC1 protein levels were detected by western blot. C DNA sequences confirming the site-specific mutations of HKDC1 (D532A, R539A, T679A, S896A). D A549 cells with or without pre-knockout of HKDC1 were transfected with wild-type (W.T.) or mutant HKDC1 expression vectors or the control vector (NC) for 24 h. The cells were then cultured with or without glucose for 24 h, and cell lysates were analyzed by Western blotting for HKDC1. E HEK293 T cells (without endogenous HKDC1 expression) were transfected with wild-type or mutant HKDC1 vectors or the control plasmid as indicated. The cells were cultured with or without glucose for 24 h and HKDC1 expression was analyzed by Western blotting.

Fig. 3. Glucose modulates HKDC1 protein stability by affecting its ubiquitination at K620.

A Proteosome inhibitor MG132 enhanced the overall levels of protein ubiquitination and partially restored glucose depletion-induced degradation of HKDC1 protein. A549 cells were cultured in standard or glucose-free RPMI 1640 medium in the presence or absence of MG132 (10 μM) for 16 h, and cell lysates were analyzed by immunoblotting using antibodies against HKDC1, ubiquitin, or β-actin as indicated. B Effect of glucose and MG132 on ubiquitination of the endogenous HKDC1. A549 cells were cultured in standard or glucose-free RPMI 1640 medium in the presence or absence of MG132 (10 μM) for 16 h, and the cell lysates were immunoprecipitated with HKDC1-specific antibody or with control IgG, followed by immunoblotting using of antibodies against ubiquitin, HKDC1, or β-actin as indicated. C Effect of glucose on the ubiquitination of exogenously expressed HKDC1. HKDC1-Flag plasmid was co-transfected with either HA-ubiquitin plasmid or vector into HEK293T cells for 24 h. The cells were then cultured in standard or glucose-free RPMI 1640 medium for 24 h, and the cell lysates were immunoprecipitated by anti-flag magnetic beads. The immune complexes were immunoblotted with anti-HA and anti-Flag antibodies as indicated. D Effect of glucose depletion on the protein stability of different HKDC1 mutants. A549 cells with pre-deletion of HKDC1 were transfected with expression vectors containing either wild-type (wt) or mutant HKDC1 as indicated, and the cells were cultured with or without glucose for 24 h. HKDC1 protein levels were analyzed by Western blotting and quantified using their band density. E K620 mutation abolished HKDC1 response to glucose deprivation. A549 cells with pre-knockout of HKDC1 were transfected with either wild-type (WT) or K620R mutant HKDC1 for 24 h, and then cultured in medium with or without glucose for 24 h. Cell lysates were analyzed by Western blotting. F Effect of different lysine mutations on glucose depletion-induced ubiquitination of HKDC1. Flag-HKDC1 expression vector containing either a wild-type or a mutant HKDC1 (K620R, K784R, or K865R) was co-expressed with a HA-Ubiquitin or a control vector into HEK293T cells for 24 h, and the cells were then cultured in medium with or without glucose (6 h) as indicated. The cell lysates were immunoprecipitated by anti-flag antibody, and the pulled-down immune complexes were subjected to western blot analysis using anti-HA and anti-flag as indicated. G S896A mutation in HKDC1 reduced its ubiquitination and increased HKDC1 protein level regardless of glucose depletion. HKDC1-WT, or HKDC1-S896A mutant were transfected into the HKDC1-deleted A549 cells for 24 h, and the cells were cultured with or without glucose for 6 h in the presence or absence of MG132 as indicated. Cell lysates were immunoblotted using anti-ubiquitin, anti-β-actin, and anti-HKDC1 antibodies. H Molecular modeling showing the narrowing of the HKDC1 glucose binding pocket (green circle) induced by S896A mutation (red color). This protein conformation change altered the location of lys620 (blue color) and hindered its ubiquitination.

Fig. 4. HKDC1 promotes fatty acid utilization and cell survival under glucose starvation.

A, B Effect of HKDC1 knockout on glucose uptake and lactate production. HKDC1 in A549 cells was knocked out using sgRNA-mediated gene deletion. Western blot analysis was performed to detect HKDC1 protein expression (A); glucose uptake and lactate production (B) were measured in A549 WT and HKDC1-knockout cells using a YSI 2950D Biochemistry Analyzer (YSI Inc., Yellow Springs, Ohio 45387 USA). C Real-time measurement of oxygen consumption rate (OCR) in A549 cells with or without HKDC1 knockout, using the Seahorse extracellular flux (XF24e) analyzer. D Effect of HKDC1 knockout on cell proliferation. HKDC1 in A549 cells was knocked out using two sgRNAs (sg-HKDC1-1 or sg-HKDC1-2), and cell proliferation was measured by MTS assay. E The HKDC1-knockout cells were transfected with a HKDC1-overexpression (OE) vector or the control vector, and cell proliferation was measured by MTS assay. F Effect of HKDC1 knockout on OCR of A549 cells under the conditions of low glucose (1.8 mM) with a supplement of palmitate (167 μM) and in the presence or absence of the CPT1 inhibitor etomoxir (40 μM); G Quantification of basal OCR, maximum OCR and ATP-linked OCR from experiment in (F). Fatty acid-related OCR is defined as [OCR_palm+BSA – OCR_BSA+ETO]; H Scatter plot showing relative gene expression in tumor tissues derived from xenografts of A549 cells with wt-HKDC1 versus HKDC1 knockout cells. Genes significantly upregulated are shown in red color, while the downregulated genes are shown in green color. Genes that remain unchanged are shown in black. The pathways with most upregulated genes by HKDC1 knockout are listed on the right panel. I Heatmap of differentially expressed mitochondria-related genes in tumor tissues with wt-HKDC1 vs HKDC1 knockout (cut-off: fold change >2 or <0.5 with a P value of <0.05). J Western blot analysis of mitochondria electron chain complex components in A549 with wt-HKDC1 or with HKDC1 knockout by sgRNA. K Western blot analysis of CPT1A, CPT1B and CPT1C protein levels in A549 cells with or without HKDC1 knockout by sg RNA, or in HCC827 lung cancer cells with or without HKDC1 knockdown by shRNA. L A549 cells with wt-HKDC1 or HKDC1-knockout were treated with the indicated concentrations of etomoxir for 72 h, and apoptosis was measured using annexin-V/7AAD double staining followed by flow cytometry analysis. M A549 cells with or without HKDC1 knockout were incubated with the indicated concentrations of etomoxir, and cell viability was analyzed using MTS assay. For panels (D, E, G, M) data are mean ± SD; *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Fig. 5. HKDC1 enhances cancer cell stemness and tumor growth in vivo.

A Schematic illustration of study design to evaluate the impact of HKDC1 on cancer cell stemness in vitro and in vivo. B A549 cells with wt-HKDC1 or HKDC1-knockout were further transfected with HKDC1 expression vector or the control plasmid. Cells were culture for 10 days to form tumor-spheres/cell clusters. Representative images are shown. C Effect of sgRNA-mediated HKDC1 knockout and shRNA-mediated HKDC1 knockdown on the expression of stem cell markers in A549 cells and HCC827 cells, respectively. Expression of ABCG2, OCT4, SOX2, CD13 and CD133 were analyzed by Western blotting. D A549 and HCC827 cells with sgRNA- or shRNA-mediated abrogation of HKDC1 were transfected with HKDC1 over-expression (OE) vector or the control vector. The expression of stem cell markers was measured by Western blotting. E, H Effect of HKDC1 knockout on tumor formation and growth in vivo. Nude mice were inoculated with the same numbers of A549 cells with or without HKDC1 knockout, tumor incidence (E), gross morphology (F), and growth (G, H) are shown. I Western blot analysis of HKDC1 protein expression in tumor tissues from mice bearing xenografts from the control or HKDC1-knockout A549 cells. The statistical significance between tumor incidences in (E) was evaluated using Chi-square test; The statistical significance in tumor growth between two groups was tested using Two-way ANOVA. Results are presented as mean ± SEM, *, P < 0.05; **, P < 0.01, ***, P < 0.001.

Fig. 6. HKDC1 physically interacts with PHB2 to regulate gene expression via Sp1.

A Schematic illustration of the strategy to identify potential molecule that mediates the effect of HKDC1 on cancer stemness using a tandem immunoprecipitation and LC-MS/MS analysis, leading to identification of PHB2 as the top candidate. B A549 cell lysates were immunoprecipitated (IP) using rabbit anti-PHB2 (left panel) or rabbit anti-HKDC1(right panel), and rabbit IgG as a control, and IP products were then immunoblotted with anti-HKDC1 or anti-PHB2 antibody. C A549 cells with or without sgRNA-mediated HKDC1 knockout (by sgHKDC1#1) were cultured in the presence or absence of glucose as indicated. Cells were then fixed and stained with MitoTracker Red CMXRos to show mitochondria (red color), FITC-conjugated PHB2 antibody to reveal PHB2 in green, and DAPI to show the nuclei in blue. Representative images taken by a confocal microscope (NIKON CSU-W1 Spinning Disc Confocal Microscopy) are shown. Scale bar, 5 μm. D Effect of glucose on cytosolic and nuclear HKDC1 and PHB2 revealed by Western blotting. GAPDH and PARP were also blotted as the indicators for the purity of cytosolic and nuclear fractions, respectively. E HKDC1 knockout promoted nuclear localization of PHB2. The cytosolic and nuclear fractions were prepared from A549 cells with or without HKDC1 knockout, and PHB2, GAPDH, and PARP were detected by Western blot analysis. F Molecular interaction between HKDC1 (cyan) and PHB2 (orange) revealed by space structure docking. The structural information of PHB2 and HKDC1 was obtained from AlphaFold Protein Structure Database (https://alphafold.ebi.ac.uk/), and the docking between the two molecules was analyzed using the HDOCK web-tool (http://hdock.phys.hust.edu.cn/). The key amino acids involved in molecular interactions are shown as sticks, whereas the rest of the molecules shown as ribbons. H-bones are displayed in red dash lines with the bond distance labeled in Å. The binding confidence score (BCS) was calculated as 0.9014. G Effect of PHB2 knockdown by siRNA on the expression of HKDC1, CD13, ABCG2, CD133, OCT4 in A549 cells. H Physical binding of PHB2 with Sp1 but not with TFAP2C in cells cultured in the presence or absence of glucose as indicated. Protein interaction was assayed by co-immunoprecipitation.

Fig. 7. Working model: HKDC1 functions as a glucose sensor and alters its protein stability in response to glucose availability.

The presence of glucose induces HKDC1 conformational change and covers the key ubiquitination residue Lys620 and thus stabilize HKDC1, which sequesters PHB2 and abolishes its ability to suppress SP1, leading to elevated expression of pro-oncogenic molecules. Glucose starvation causes the exposure K620 for ubiquitination, leading to HKDC1 degradation and an increase in mitochondrial fatty acid utilization due in part to an increase in expression of fatty acid transporter CPT1. This enables metabolic adaptation to low glucose microenvironment. Abrogation of HKDC1 by genetic knockout or glucose starvation releases PHB2, which then suppresses cancer stemness and inhibits tumor growth in vivo.