GTP binding protein 2 maintains the quiescence, self-renewal, and chemoresistance of mouse colorectal cancer stem cells via promoting Wnt signaling activation

Colorectal cancer (CRC) is one of the most common cancers and the second most deadly cancer across the globe. Colorectal cancer stem cells (CCSCs) fuel CRC growth, metastasis, relapse, and chemoresistance. A complete understanding of the modulatory mechanisms of CCSC biology is essential for developing efficacious CRC treatment. In the current study, we characterized the expression and function of GTP binding protein 2 (GTPBP2) in a chemical-induced mouse CRC model. We found that GTPBP2 was expressed at a higher level in CD133+CD44+ CCSCs compared with other CRC cells. Using a lentivirus-based Cas9/sgRNA system, GTPBP2 expression was ablated in CRC cells in vitro. GTPBP2 deficiency caused the following effects on CCSCs: 1) Significantly accelerating proliferation and increasing the proportions of cells at G1, S, and G2/M phase; 2) Impairing resistance to 5-Fluorouracil; 3) Weakening self-renewal but not impacting cell migration. In addition, GTPBP2 deficiency remarkably decreased β-catenin expression while increasing β-catenin phosphorylation in CCSCs. These effects of GTPBP2 were present in CCSCs but not in other CRC cell populations. The Wnt agonist SKL2001 completely abolished these changes in GTPBP2-deficient CCSCs. When GTPBP2-deficient CCSCs were implanted in nude mice, they exhibited consistent changes compared with GTPBP2-expressing CCSCs. Collectively, this study indicates that GTPBP2 positively modulates Wnt signaling to reinforce the quiescence, self-renewal, and chemoresistance of mouse CCSCs. Therefore, we disclose a novel mechanism underlying CCSC biology and GTPBP2 could be a therapeutic target in future CRC treatment.


Introduction
Colorectal cancer stem cells (CCSC) are crucial for the initiation, progression, recurrence, and metastasis of colorectal cancer (CRC).Their self-renewal maintains the CCSC pool and they can generate all relevant cancer cells with various differentiation statuses [1].CCSCs are likely in a relatively quiescent state, making them resistant to radiochemotherapy [1].Until recently, several cell surface proteins, such as CD133, CD44, epithelial cell adhesion molecule (EpCAM), and CD24, have been identified as CCSC-related markers [2].Particularly, CD133 + CD44 + CRC cells were considered as CCSCs owing to their cancer-initiating ability [3][4][5].However, the distribution of CCSC markers differs between patients and cell lines.Some studies argue that CD44 + CD133 − cells have CCSC features while CD133 high CD24 low tumors correlate with the worst prognosis in patients [6][7][8].Therefore, the genuine CCSC markers remain elusive.
The canonical Wnt signaling is important for CRC development.Without Wnt ligands, intracellular β-catenin is phosphorylated by a destruction complex consisting of Casein kinase 1 (CK1), glycogen synthase 3β (GSK-3β), Axin, and adenomatous polyposis coli (APC).Phosphorylated β-catenin is easily degraded and thus kept at a low level.When Wnt ligands bind to the Frizzled receptor, β-catenin is released from the destruction complex to avoid degradation and subsequently enters the nucleus to trigger the transcription of target genes such as c-myc, c-jun, survivin, and OCT4, etc [9].Cumulative evidence has demonstrated aberrant activation of Wnt signaling as an indispensable factor in CRC carcinogenesis [10].Particularly, Wnt signaling is critical for maintaining stemness, chemoresistance, and metastasis of CCSCs [11][12][13][14].Therefore, understanding the regulatory mechanisms of Wnt signaling is critical for studying CCSC biology.
GTP-binding proteins are proteins with GTP hydrolase activity and participate in cell signal transduction, cytoskeletal rearrangement, gene expression, and other biochemical processes.GTP binding protein 2 (GTPBP2) is a member of the G protein superfamily and is expressed in most organs and tissues, especially in the cerebral cortex, cerebellum, thymus, smooth muscle, liver, pancreas, and digestive tract [15,16].Recent research reports that GTPBP2 is a positive regulator of the invasion, migration, and proliferation of non-small cell lung cancer [17].Notably, GTPBP2 is necessary for canonical Wnt signaling in Xenopus embryos, via suppressing the accumulation of Axin, which is a component of the β-catenin destruction complex [18].Since Wnt signaling is crucial for CCSC development, GTPBP2 likely regulates CCSC biology by modulating Wnt signaling.However, the significance of GTPBP2 to the development of CCSCs has not been disclosed.In this research, we characterized the role of GTPBP2 in regulating CCSC proliferation, self-renewal, and chemoresistance in a chemical-induced primary CRC model.

Primary CRC model
This animal research was approved by the Wuhan Third Hospital Animal Care and Use Committee (Approval ID: WTH20210146) and conducted under the Animal Research: Reporting of In Vivo Experiments (ARRIVE) guidelines.Eight-week-old male C57BL/6J mice were purchased from Wanqian Animal Technology Inc. Azoxymethane (AOM, Cat# A5486) and dextran sulfate sodium (DSS, Cat# D8906-50G, molecular weight = 36-50 kDa) were purchased from Sigma-Aldrich.AOM was diluted in sterile saline at a concentration of 2.5 mg/ml.The model was established according to previous reports with minor modifications [19,20].Briefly, mice were injected intraperitoneally with 10 mg AOM per kg mouse weight.After 6 days, mice were treated with three cycles of 2% DSS for 7 days in drinking water, followed by 2 weeks of regular drinking water (Supplementary Fig. 1).Three months after the AOM injection, mice were euthanized with CO 2 and tumor formation was assessed by gross examination.Mice with visible tumor(s) in the colon and rectum were used in the study (Supplementary Fig. 1).The control mice did not receive AOM injection and were fed with regular drinking water.

Isolation of tumor cells from primary CRC tissues
According to previous reports with modifications [21,22], the tumors were taken from mouse colons and rectums, minced into 1-mm 3 pieces and incubated in EDTA buffer (2 mM EDTA, 43.4 mM sucrose, 54.9 mM D-sorbitol, 0.5 mM DL-dithiothreitol in PBS) for 60 min on ice.After two washes with PBS, the tissues were dissociated in digestion buffer (DMEM supplemented with 5% fetal calf serum, 2.5 ng/ml of amphotericin B, 200 U/ml collagenase IV, and 0.1 mg/ml type II dispase) for 45 min at 37 • C with agitation every 15 min.The supernatants and tissues were filtered through a 70-μm cell strainer to prepare single cells, followed by centrifugation at 200g for 3 min and washing once with 5 ml of PBS before further experiments.The reagents were purchased from Sigma-Aldrich.

Lentivirus preparation and transduction
The Lenti-Cas9-gRNA-GFP vector (Cat# 124770) encoding Cas9, sgRNA, and green fluorescence protein (GFP) was purchased from Addgene.The Gtpbp2 sgRNA (5′-ATTATGACAGTGACGTGCCC-3′) and scrambled sgRNA (5′-GCACTACCAGAGCTAACTCA-3′) were designed by Integrated DNA Technologies.Predicted off-target genes of the Gtpbp2 sgRNA are Thrb, Cndp1, Plaa, and Fer.However, we did not observe alterations of mRNAs of these genes in our pilot experiments (Data not shown), suggesting that the Gtpbp2 gene is the only target.Vector cloning and lentivirus packaging/titration were achieved by AtaGenix Biotech Inc.To infect primary CRC cells, CRC cells were suspended in DMEM supplemented with 10% fetal calf serum (FCS).Cells were seeded in 6-well culture plates at a density of 5 × 10 4 cells/well.After that, polybrene (Maokang Biotech) was added at a final concentration of 6 μg/ml.Lentiviral particles were added at a multiplicity of transduction (MOI) of 50, followed by centrifugation at 700g for 1 h at 32 • C and then incubated overnight.The supernatants were replaced with fresh 10% FCS-DMEM to incubate cells for 2 days.The infection efficiency was determined by assessing the frequency of GFP + cells (i.e.successfully infected cells).GFP + cells were sorted by flow cytometry, followed by detecting their phenotype and function.
To determine the effect of Wnt signaling, sorted GFP + CCSCs were cultured at a density of 5 × 10 5 /ml in DMEM supplemented with 10% FCS.SKL2001 (also known as Wnt agonist II, Cat# 681667, Sigma-Aldrich) was added into the culture at a final concentration of 20 μM to incubate cells for 24 h.Detection of phosphorylated β-catenin, total β-catenin, and Ki67 was then performed by flow cytometry.
To overexpress GTPBP2, GTPBP2 lentiviral vector (22819065) and the control pLenti-GIII-CMV vector (LV587) were purchased from Applied Biological Materials.Lentivirus preparation and infection procedures were the same as above.Two days after infection, the cells were treated with 2 μg/ml puromycin (Sigma-Aldrich) for 24 h to select infected cells.Viable cells were subjected to further analysis.

Cell cycle
Cells were fixed in ice-cold 70% ethanol for 2.5 h, followed by incubation in Hank's balanced salt solution (HBSS) containing 2 mg Hoechst 33342 (94403, Sigma-Aldrich) and 4 mg Pyronin Y (P9172, Sigma-Aldrich) for 20 min.Afterward, cells were washed with PBS once and the intensities of Hoechst 33342 and Pyronin Y were analyzed by flow cytometry.
Sphere formation.CRC cells were suspended at a density of 2.5 × 10 3 /ml in DMEM/F12 supplemented with 1 × N2 supplement, 1 × B27 supplement, 1 mM N-acetylcysteine, 10 ng/ml EGF, 20 ng/ml basic FGF, and 1 mg/ml heparin (All from Stem Cell Technologies). 1 ml of cell suspension was placed in each well of a 6-well ultra-low attachment culture plate (Corning) and cultured for 7 days to form spheres (Passage 1).The spheres were digested with 0.5% Trypsin-EDTA (Sigma-Aldrich) for 5 min at 37 • C, followed by vigorous pipetting to prepare single cells.The single cells were pelleted by centrifugation and cultured in supplemented DMEM/F12 in the same manner as above to form passage-2 spheres.The above procedures were repeated 7 days later to generate passage-3 spheres.To count the spheres, the culture was centrifuged at 150 g for 5 min, followed by aspirating the supernatant and gently re-suspending the sphere pellet in 500 μl of DMEM/F12.After that, 100 μl of the sphere suspension was added into a well (marked with four quadrants on its underside) of a 96-well plate.The spheres (>50 μm in diameter) in each quadrant were counted and summed on a Zeiss AXIO Imager.M1 phase contrast motorized microscope (Zeiss).To determine the effect of Wnt signaling, 20 μM SKL2001 was added at the start of passage-2 sphere culture and was present for 7 days.

Cell migration
5 × 10 4 CRC cells were suspended in serum-free DMEM and seeded on the upper compartment of Matrigel-coated invasion chambers (Corning) with an 8-μm pore polycarbonate membrane.The lower compartment was filled with 10% FCS-DMEM.After 24-h incubation, cells on the upper side of the membrane were removed using a cell scraper.Migrated cells on the lower side of the membrane were fixed with 95% ethanol, stained with 0.1% crystal violet solution, and counted with a light microscope.
C. Ke et al.
Absorbance at 490 nm was measured on a SpectraMax® ABS Plus microplate reader (Molecular Devices).In chemoresistance analysis, CRC cells were seeded at a density of 5 × 10 3 cells per well and treated with 5-Fluorouracil (5-FU) at the indicated concentrations for 48 h, followed by the MTT assay as described above.The percentage of cell viability was computed using the equation: % Viability = Mean OD sample /Mean OD blank × 100.
Sorted GFP + CRC cells were suspended in 10% FCS-DMEM and seeded in triplicate in 96-well plates at 5 × 10 3 cells/well, flowed by exposure to 10 μM 5-FU for 48 h.Apoptosis and necrosis were analyzed by flow cytometry.To evaluate the relationship between Wnt signaling and GTPBP2, SKL2001 was added into the culture at 20 μM.

Tumor implantation
Eight-to-ten-week-old nude mice were purchased from Wanqian Animal Technology Inc. 1 × 10 6 CSCs were mixed with 50 μl of Matrigel matrix (Corning) and injected into the right flank.Six weeks later, the mice were euthanized by CO 2 and the volumes of subcutaneous tumors were measured as (length × width 2 ) × 0.5.Following that, tumor cells were isolated as previously described.

Hematoxylin and eosin staining
Normal mouse colon, CRC tumors, and tumor implants were fixed in formalin, embedded in paraffin, and cut into 10-μmm sections following the routine procedures.Hematoxylin and eosin staining was carried out following the standard protocol.

Immunohistochemistry
Tissue sections were prepared as described above.The sections were blocked with 10% normal goat serum (ThermoFisher), followed by incubation with the anti-GTPBP2 polyclonal antibody (Cat# PA5-75936, 1:100) overnight at 4 • C before incubation with the horseradish peroxidase-labeled goat anti-rabbit IgG (31460, Invitrogen) for 1 h at room temperature.After PBS washes, the DAB substrate kit (ab64238, Abcam) was used for detection.The samples were observed on an AXIO Imager Microscope (Zeiss).

RNA purification and real-time PCR
Cellular RNAs were purified using the PureLink RNA Mini Kit (Invitrogen).Tissue RNAs were extracted using TRIzol (Thermo-Fisher) following the manufacturer's instructions.cDNAs were synthesized using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems).The PowerUp™ SYBR™ Green Master Mix (Applied Biosystems) was used to quantify gene expression on a 7300 real-time PCR platform (Applied Biosystem).The expression of target genes was normalized to β-actin and calibrated using the 2 − ΔΔCt method.Primers are shown in Supplementary Table 1.

Statistics
All data are presented as mean ± SD.Student's t-test or One-way ANOVA was applied to compare mean values among different groups.The GraphPad Prism 9 was used for statistical analysis.All experiments were repeated two or three times.A P value < 0.05 is considered statistically significant.

GTPBP2 is highly expressed in CSCs of primary CRC
To determine GTPBP2 expression in CRC, we established a chemical-induced CRC model.The colon tissue in the AOM/DSS group exhibited loss of goblet cells and epithelial cell destruction, indicating CRC formation (Supplementary Fig. 1).GTPBP2 was moderately expressed in granular cells of normal colons but strongly expressed in CRC cells after AOM/DSS treatment (Supplementary Fig. 2A).The relative mRNA level of GTPBP2 in the CRC tissue negatively correlated with mouse body weight and positively correlated with mean tumor volume but did not correlate with either tumor number or colon length (Supplementary Figs.2B-2E), suggesting that GTPBP2 might enhance CRC growth.CRC tumors were harvested from the colon and rectum of each mouse, followed by enzyme digestion and mechanical dissociation.The resultant single cells were subjected to flow cytometry analysis.CD45 + leukocytes, CD31 + vascular endothelial cells, MCT4 + stromal cells, and propidium iodide-positive dead cells were excluded (Supplementary Fig. 3).The rest cells, i.e.CRC cells, were divided into subsets based on the expression of CD133 and CD44: CD133 − CD44 − , CD133 − CD44 + , CD133 + CD44 + , and minor CD133 + CD44 − cells (Fig. 1A).CD133 − CD44 − , CD133 − CD44 + , and CD133 + CD44 + cells were sorted by flow cytometry (Supplementary Fig. 4).Only CD133 + CD44 + cells produced abundant tumor spheres in vitro, suggesting that they were CCSCs (Supplementary Fig. 5A).Next, GTPBP2 expression was determined in each subset.CCSCs expressed significantly higher GTPBP2 than CD133 − CD44 − and CD133 − CD44 + cells (Fig. 1B and C and Supplementary Fig. 5B).

Discussion
CCSCs fuel CRC growth, metastasis, relapse, and chemoresistance, rendering them a promising therapeutic target [25].GTPBP2 belongs to the G protein superfamily.G proteins, capable of binding guanosine triphosphate (GTP) and guanosine diphosphate (GDP), act as molecular switches and signal transducers inside cells [26].The role of GTPBP2 in either normal cellular processes or malignant transformation is largely unknown.GTPBP2-deficient mice do not exhibit remarkable phenotypic anomalies [15].In Xenopus embryos, GTPBP2 was reported to interact with Axin to positively regulate Wnt signaling and participate in BMP/SMAD1 signaling [18,27].Considering the essential role of the conserved Wnt signaling in CCSC biology, it is plausible to deduce that GTPBP2 might regulate Wnt signaling in CCSCs.
We enriched primary CCSC according to the expression of CD133 and CD44, which are considered valid CCSC markers in human patients and mice [1,22,28].However, the distribution of CCSC markers differs between patients and tumor cell lines.For example, CD133 − CD44 + cells correlate with CCSC features in the SW620 cell line and CRC patients [6,8].CD133 high CD24 low cells are probably CCSC in patients with CRC since they are associated with the worst prognosis [7].Therefore, the universal set of markers to identify CCSCs in CRC remains ambiguous.Our data indicated that CD133 − CD44 + cells did not express abundant stemness markers and failed to generate spheres, implying that these cells were unlikely CCSCs in our experimental setting.In our study, CD133 + CD44 + CCSCs expressed higher GTPBP2, signifying the potential importance of this molecule in CCSC activity.However, the cause of GTPBP2 up-regulation is not clear.To our knowledge, no previous study has delineated the transcriptional or translational mechanisms of GTPBP2 expression.Therefore, future investigations are necessary to reveal whether stemness-associated signaling pathways, such as the Hedgehog pathway and Notch pathway, are responsible for GTPBP2 up-regulation.Moreover, it would be interesting to check whether normal colonic stem cells, i.e. stem cells in the bottom of the colonic crypt [29], have high GTPBP2 expression.
Intriguingly, GTPBP2 ablation significantly altered the proliferation, chemoresistance, and self-renewal of CCSCs rather than CD133 − CD44 − (Subset I) and CD133 − CD44 + (Subset II) CRC cells, implying that GTPBP2 is essential for CCSC activity but not differentiated CRC cells.GTPBP2 is probably not important for differentiated CRC cell growth due to its low expression in Subset I and Subset II.Another possibility is that GTPBP2 deficiency might be compensated by other G proteins or pathways in differentiated CRC cells.Therefore, the remarkable role of GTPBP2 seems to be CCSC-specific.
Our data indicate that GTPBP2 ablation accelerates CCSC division but impairs CCSC self-renewal.There are two types of stem cell division: symmetric division and asymmetric division.The former means that a stem cell produces two identical stem cells, while the latter indicates that a stem cell produces one differentiated cell and one stem cell [30].Therefore, symmetric division is more efficient in maintaining the self-renewal capacity, whereas asymmetric division incurs a gradual decrease in the self-renewal capacity of a group of stem cells.Although we did not assess the types of Gtpbp2 − /− and Gtpbp2 +/+ CCSCs division, the weaker sphere formation of Gtpbp2 − /− CCSCs suggests that GTPBP2 ablation promotes asymmetric CCSC division and subsequently diminishes CCSCs' self-renewal capacity.Moreover, the slower division of Gtpbp2 +/+ CCSCs indicates that GTPBP2 is essential for CCSC quiescence that  reinforces chemoresistance.Therefore, the higher resistance of Gtpbp2 +/+ CCSCs to 5-FU could be attributed to the quiescent status.In the future, bulk or single-cell transcriptome sequencing should be performed to answer whether Gtpbp2 − /− and Gtpbp2 +/+ CCSCs differentially express self-renewal-or quiescence-associated genes.A recent study applied single-cell whole-exome sequencing and bulk-cell targeted exome sequencing to demonstrate that CD133 + CD44 − cells are likely the original CCSC population [8].However, in another recent research, transcriptome profiling of human CRC samples by SMART-seq2 indicates that CSCs are enriched from both CD44 + CD133 + and CD44 − CD133 + CRC cells [31].These studies exhibit the complexity and heterogeneity of CCSCs and provide valuable data to analyze the stemness and quiescence of putative CCSCs.
Another important finding is the down-regulation of Wnt signaling in Gtpbp2 − /− CCSCs, which is consistent with a previous study reporting that GTPBP2 positively regulates Wnt signaling [18].Wnt signaling activation enhances stemness, self-renewal, and chemoresistance [11,12,32,33].Therefore, GTPBP2 ablation attenuates Wnt signaling activation and consequently harms the self-renewal and chemoresistance of CCSCs.Regarding cell proliferation, although Wnt signaling appears to promote the proliferation of CRC cells [10], it enforces the quiescence of hematopoietic stem cells and neural stem cells [34,35].Furthermore, SKL2001 facilitates colon cancer spheroid formation but suppresses spheroid growth by inducing G0/G1 phase arrest, implying that the Wnt signaling facilitates CCSC self-renewal but inhibits CCSC proliferation [36].In our study, SKL2001 also inhibited CCSC proliferation, strongly indicating that Wnt signaling maintains CCSC quiescence.SKL2001 disrupts the interaction between β-catenin and Axin to prevent β-catenin phosphorylation and subsequent degradation.Indeed, we observed lower β-catenin phosphorylation and elevated β-catenin levels in SKL2001-treated Gtpbp2 − /− CCSCs.Although we did not perform co-immunoprecipitation to determine the potential interaction between GTPBP2 and the destruction complex owing to the rarity of primary CCSCs, this assay could be done on human or mouse CRC cell lines in the future.
Metastatic CRC cells release proteolytic enzymes like matrix metalloproteinase 2 (MMP-2) and matrix metalloproteinase 9 (MMP-9) to digest the extracellular matrix (ECM), facilitating irregular cancer cords to penetrate adjacent tissues [37].Other factors, such as E-cadherin down-regulation, N-cadherin up-regulation, and enhanced motility of cancer cells also contribute to CRC invasion [38].Previous research has suggested that the Wnt signaling might promote the expression of MMP-2 and MMP-9 in liver cancer [39].Besides, the Wnt signaling increases the invasion of CRC [40].Therefore, GTPBP2-induced β-catenin stabilization would reinforce CRC invasion.However, our in vitro and in vivo data imply that GTPBP2 does not affect CRC cell migration or invasiveness.Perhaps GTPBP2 activates other signal pathways to suppress β-catenin-mediated invasiveness.This puzzle needs to be unraveled in future investigations.
Apart from the above findings, we found that stemness markers such as OCT4, SOX2, and NANOG were highly expressed in CCSCs relative to other subpopulations.GTPBP2 ablation only down-regulated these markers in CCSCs, strongly implying that GTPBP2 upregulates the expression of stemness markers in CCSCs.Among these markers, NANOG is a transcription factor key to the stemness of cancer stem cells because it induces self-renewal, invasiveness, and chemoresistance of cancer cells through multiple pathways [41].Importantly, NANOG induces the dormancy, migration, and chemoresistance of colorectal cancer cells [42,43].Therefore, GTPBP2 likely activates the Wnt signaling to induce high NANOG expression, and NANOG promotes CCSC stemness and maintains the malignant nature of CCSCs.However, GTPBP2 ablation or overexpression could not affect the expression of NANOG and other stemness markers in CD133 − CD44 − and CD133 − CD44 + cells, suggesting that GTPBP2 no longer regulates stemness in non-CCSC cancer cells probably due to altered intracellular microenvironment.
In summary, our study demonstrates that GTPBP2 positively modulates Wnt signaling to reinforce the quiescence, self-renewal, and chemoresistance of CCSCs.We disclose a novel mechanism underlying CCSC biology and GTPBP2 could be a therapeutic target in future CRC treatment.

Fig. 1 .
Fig. 1.GTPBP2 expression in primary CRC cells.(A) Representative dot plots showing CRC cell subsets.CCSC: colorectal cancer stem cells.(B and C) GTPBP2 expression in sorted CRC cell subsets.Representative Immunoblotting images are shown in (B) and statistics are shown in (C).N = 5 independent samples per group.Each sample contains cells pooled from five mice.***: P < 0.001.One-way ANOVA.