Cystathionine β-synthase affects organization of cytoskeleton and modulates carcinogenesis in colorectal carcinoma cells

Background Cystathionine β-synthase (CBS), one of three enzymes that endogenously produce hydrogen sulfide, is extensively studied for its relevance in the cells of various tumors. In our previous work, we observed that the immunofluorescence pattern of CBS is very similar to that of tubulin and actin. Therefore, we focused on the potential interaction of CBS with cytoskeletal proteins β-actin and β-tubulin and the functional relevance of the potential interaction of these proteins in colorectal carcinoma cell lines. Methods To study the potential interaction of CBS with cytoskeletal proteins and its functional consequences, a CBS-knockout DLD1 (DLDx) cell line was established by using the CRISPR/Cas9 gene editing method. The interaction of the selected cytoskeletal protein with CBS was studied by immunoprecipitation, Western blot analysis, immunofluorescence, and proximity ligation assay. The functional consequences were studied by proliferation and migration assays and by generation of xenografts in SCID/bg mice. Results We have found that CBS, an enzyme that endogenously produces H2S, binds to cytoskeletal β-tubulin and, to a lesser extent, also to β-actin in colorectal carcinoma-derived cells. When CBS was knocked out by the CRISPR/Cas9 technique (DLDx), we observed a de-arranged cytoskeleton compared to the unmodified DLD1 cell line. Treatment of these cells with a slow sulfide donor GYY4137 resulted in normal organization of the cytoskeleton, thus pointing to the role of CBS in microtubule dynamics. To evaluate the physiological importance of this observation, both DLD1 and DLDx cells were injected into SCID/bg mice, and the size and mass of the developed xenografts were evaluated. Significantly larger tumors developed from DLDx compared to the DLD1 cells, which correlated with the increased proliferation of these cells. Conclusions Taken together, in colorectal cancer DLD1 cells, CBS binds to the cytoskeleton, modulates microtubule dynamics, and thus affects the proliferation and migration in the colorectal carcinoma stable cell line.


Introduction
Cystathionine b-synthase (CBS) is one of three enzymes that endogenously produce hydrogen sulfide (H 2 S). H 2 S is a relatively novel gaseous transmitter, to which several regulatory roles have already been attributed. CBS is localized both in the cytoplasm and mitochondria (1) and is regulated by a variety of physiological factors, e.g., 1,25-dihydroxyvitamin D3, hypoxia-inducible factors, and estradiol-17b (for a review, see 2), depending on the cell type. Using cryo-electron microscopy, McCorvie and co-workers (3) have shown that human CBS reveals a filamentous architecture. This full-length enzyme polymerizes as an active filament that changes conformation due to S-adenosyl-L-methionine. Interactions due to filamentation result in a shift of the residues 417-423 on the C-terminal regulatory domain. The physiological role of CBS is still not completely clear, although its detrimental effect on the cardiovascular system is evident (4). Permanent CBS gene knockdown in immortalized human adipose-derived mesenchymal stem cells promoted a cellular senescence phenotype with increased adipogenic potential and excessive lipid storage (5). The regulation of immune and inflammatory responses by CBS has also been described (6).
The role of CBS in carcinogenesis has been studied on various types of tumors by several laboratories using different approaches. Increased levels of CBS were detected in colon/colorectal (7,8), breast, and prostate cancer (9). This enzyme was also found to be increased in thyroid malignancies (10). In ovarian cancer cells, CBS regulates bioenergetics by regulating mitochondrial ROS production, oxygen consumption, and ATP generation (11). In breast cancer, CBS-derived H 2 S might play a role in the protection of breast cancer cells against activated macrophages (12). Inhibition of CBS can improve ovarian cancer treatment as well (13). In colon cancer-derived epithelial cell lines, a role for endogenous H 2 S in tumor angiogenesis has been shown (7). However, decreased CBS expression was determined in clear cell renal cell carcinoma that was spontaneously hypoxic compared to the matched controls, and this decrease was dependent on the grade of tumor (14). A question remains as to whether higher CBS levels in tumors together with overexpression of other H 2 S-synthesizing enzymes could be due to a lower amount of H 2 S in tumors, as suggested by Dongsoo et al. (15). A bell-shaped model has been proposed to explain the role of H 2 S in cancer development. Specifically, endogenous H 2 S or a relatively low level of exogenous H 2 S may exhibit a pro-cancer effect, whereas exposure to H 2 S at a higher amount or for a long period may lead to cancer cell death (16). In favor of this hypothesis might be the fact that the slow sulfide donor GYY4137 can induce apoptosis in stable cancer cell lines in vitro and also in xenografts induced in immunodeficient mice treated with GYY4137 (17,18).
CBS has been shown to be localized mainly in cytosol and to a lesser extend in mitochondria (19,20). However, we have recently shown a specific CBS pattern by immunofluorescent staining (14), similar to staining of the cytoskeleton. Thus, we produced a hypothesis that CBS might bind to cytoskeletal proteins.
Microtubules are major cytoskeletal components in the eukaryotic cells. Microtubules and microtubule dynamics are involved in several important activities, such as maintenance of cell shape and cell motility, cell division and accurate chromosome segregation during mitosis, general intracellular communication, and intracellular tracking of macromolecules and organelles in the interphase. In cancer cells, a high expression of several b-tubulin isotypes correlates with aggressive clinical behavior, chemotherapy drug resistance, and poor prognosis (21). Cytotoxic action of microtubule-targeting drugs (such as taxanes or vinca alkaloids) is based on the variations of microtubule dynamics. Modulation of microtubules can significantly affect the fate of cells. Chaudhuri and co-workers (22) have shown that tubulin disulfides may play a role in tubulin folding and that thiol-disulfide exchange in tubulin could be a key regulator in microtubule assembly and dynamics of tubulin in vivo. Under physiological conditions, tubulins are sulfhydrated, which affects microtubule assembly (23). Hosono et al. (24) showed the effect of diallyl trisulfide (DATS) on the modification of cysteines Cys-12b and Cys-354b in b-tubulin, which is assumed to be a cause for the disruption of microtubule network formation. Nevertheless, microtubule assembly is dependent on a variety of other factors (e.g., acetylation, GTP, and phosphorylation), and a mutual interplay of all these factors should be also considered.
Based on the current knowledge on CBS and also on our previous experiments, we aimed to study the potential interaction of CBS and some cytoskeletal proteins, i.e., b-actin and b-tubulin (this type of tubulin has binding sites for taxanes and vinca alkaloids, etc. (25)). Since the cytoskeleton is crucially involved in carcinogenesis and is a target of some groups of chemotherapeutics, cancer cell growth and proliferation due to the potential interaction with CBS was studied as well.

Generation of CBS-knockout DLD1 cell line
The CBS-knockout DLD1 cell line, hereafter referred to as DLDx, was established by using the CRISPR/Cas9 (CRISPR (clustered, regularly interspaced, short palindromic repeats)/Cas9 (CRISPR-associated protein 9)) gene editing method. The CBS CRISPR guide RNA sequences (GATTTCGTTCTTCAGCCGCC and TGTGCCCTCAGGGATCGGGC) were designed by the laboratory of Feng Zhang at the Broad Institute in order to efficiently target the CBS gene with minimal risk of off-target Cas9 binding elsewhere in the genome (28,29). The lentiviral t r a n s f e r p l a s m i d s l e n t i C R I S P R v 2 _ C B S -1 a n d lentiCRISPRv2_CBS-3 (GenScript, Leiden, Netherlands) contained a lentiCRISPRv2 backbone and single above-mentioned oligo cloned into the single-guide RNA (sgRNA) scaffold. To produce the lentiviral particles, the transfer plasmids lentiCRISPRv2_CBS-1 or lentiCRISPRv2_CBS-3 were cotransfected into HEK293T cells with the packaging plasmids pMD2.G (Addgene, Watertown, MA, USA) and psPAX2 (Addgene, Watertown, MA, USA). The virus-containing medium was collected after 48, 60, and 72 h and passed through a 0.45 mm low protein-binding filter. Lentiviruses were concentrated using PEG-it (System Biosciences, Palo Alto, CA, USA) and sedimented by centrifugation (1500 × g, 4°C for 30 min). As a positive control to monitor transduction efficiency, CRISPR-lenti human EMX1 positive control transduction particles (CRISPR11V-1EA, Sigma-Aldrich, St. Louis, MO, USA) were used. Similarly, as a negative control, CRISPR-lenti non-targeting control transduction particles (CRISPR12V-1EA, Sigma-Aldrich, St. Louis, MO, USA) were used. This control includes a guide RNA sequence that does not target known human, mouse, and rat genes. DLD1 cells, plated the day before at a density of 0.25×10 5 cells per 6 cm plate, were infected with each lentivirus or their combination. DLD1 cells transduced by control lentivirus particles were called DLD-PC (positive control) and DLD-NC (negative control). Twenty-four hours after transduction, the cells were selected by puromycin (Puromycin, InvivoGen, USA), and then, the CBS protein knockout was confirmed by immunofluorescence (IF) and Western blot analysis (WB).

Proximity ligation assay
The proximity ligation assay (PLA) was used for in situ detection of the co-localization between CBS and b-tubulin or bactin. The assay was performed in a humid chamber at 37°C according to the manufacturer's instructions (Olink Bioscience, Uppsala, Sweden). The cells were seeded on glass coverslips and further cultured for 24 h. Afterwards, the cells were fixed with methanol, blocked with 3% PBS-BSA for 30 min, incubated with a mixture of antibodies against CBS and b-tubulin or b-actin for 1 h, washed three times with 1 x PBS, and incubated with plus and minus PLA probes for 1 h. Then, the cells were washed (3 x 5 min), incubated for 40 min with ligation mixture containing connector oligonucleotides, washed again, and incubated with an amplification mixture containing a fluorescently labeled DNA probe for 100 min. After a final wash, the samples were mounted, and the signal was analyzed using a Zeiss LSM 510 Meta confocal microscope with a Plan Neofluar 40_/1.3 oil objective. The following antibodies were used: human CBS (1:100 dilution, AP6959c, Abgent, San Diego, CA, USA), b-tubulin (1:1000 dilution, ab231082, Abcam, Cambridge, UK), and b-actin (1:250 dilution, ab6276, Abcam, Cambridge, UK).

Immunoprecipitation
The appropriate antibody was incubated with 60 µL washed magnetic beads, as described previously (31) (Dynabeads M-280), and coated with M-280 sheep anti-rabbit IgG (Invitrogen Dynal AS, Oslo, Norway) overnight at 4°C on a rotator (VWR International, Radnor, PA, USA). The beads with an attached antibody were washed (twice, 200 µL) with phosphate-buffered saline (PBS) with BSA. The proteins were immunoprecipitated from 0.5 mg of detergent-extracted total protein by incubation for 4 h at 4°C with antibody-bound beads. The bead complexes were washed four times with PTA solution (145 mmol/L NaCl, 10 mmol/L NaH2PO4, 10 mmol/L sodium azide, and 0.5% Tween020, pH 7.0). Immunoprecipitated proteins were then extracted with 60 µL of 2x Laemmli sample buffer (Bio-Rad, Hercules, CA, USA) and boiled for 5 min.

Detection of H 2 S by H 2 S fluorescent probe, P3
The cells were plated on black 24-well plates or on glass coverslips. Following treatment with the 10 µmol/L GYY4137, the cells were incubated with 10 µmol/L of H 2 S Fluorescent Probe, P3 (Calbiochem, San Diego, CA, USA) for 30-60 min at 37°C in a watersaturated atmosphere with 5% CO 2 . Afterwards, the cells were twice washed with 500 µL of PBS. Fluorescence was measured on a Synergy H1 Reader (BioTek, Bad Friedrichshall, GE) at excitation 375 nm and emission 505 nm. The results were expressed as the arbitrary units of fluorescence. The fluorescent signal in the cells on the coverslips was visualized by a Zeiss Axiolab 5 FL microscope (Zeiss, Jena, Germany). The ZEN2 software was used to process the images and to evaluate the resultant pictures (Zeiss, Jena, Germany).

Proliferation assay
The relative viability of the cells was determined by the CellTiter-Glo ™ Luminescent Cell Viability Assay (Promega Corporation, Madison, WI, USA) on a 96-well plate using 3000 cells per well and evaluated by the LumiStar GALAXY reader (BMG Labtechnologies, Germany) 4 days after plating and treatment with GYY4137.

Cell migration assay
Fifty thousand DLD1, DLDx, DLD-NC, and DLD-PC cells per well were plated on ImageLock 96-well plates (Essen BioScience, Ann Arbor, MI, USA) and left to adhere for 24 h. Confluent monolayers were then wounded with a wound needle (IncuCyteWoundMaker, Essen BioScience, Ann Arbor, MI, USA), washed twice, and supplemented with a fresh culture medium or fresh culture medium with GYY4137. Images were taken every 2 h for the next 24 h via the IncuCyte ZOOM ™ kinetic imaging system (Essen BioScience, Ann Arbor, MI, USA). Cell migration was evaluated by the IncuCyte ZOOM ™ 2016A software (Essen BioScience, Ann Arbor, MI, USA) based on the relative wound density measurements.
In vivo experiments SCID/bg mice (male, 17 weeks old at the beginning of experiment) were bilaterally subcutaneously injected by 1x10 6 DLD1, DLDx, DLD-PC, or DLD-NC cells resuspended in 100 µL of serum-free cultivation medium. Visible xenografts developed after 4 days. The mice were then randomly divided into either a treatment group (intraperitoneal injection of 20 mg/kg GYY4137 diluted in saline, given daily) or a control group (saline). Each group contained 13-22 tumors. Xenografts were measured every 3 days with a caliper, and the tumor volume (V) was calculated according to the formula: V=(length×width × 2)/2, the width being the greatest transverse diameter and length the greatest longitudinal diameter. The mice were kept on standard pelleted food and water ad libitum, monitored daily for weight loss or other signs of possible toxic effects of the drug. At the experimental endpoints (after 14 days of therapy), the mice were sacrificed. The xenografts were resected, weighed, and cryopreserved at -80°C for further experiments or stored in buffered formalin.

Ethics approval
The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Institutional Ethic Committee and the national competence authority-State Veterinary and Food Administration of the Slovak Republic (project registration No. Ro-2032-3/2020-220)-in compliance with Directive 2010/63/EU and Regulation 377/2012 on the protection of animals used for scientific purposes. The project was conducted in the approved animal facility (license No. SK UCH 02017).

In silico analysis
The Tumor online Prognostic analysis Platform (ToPP) (32), which collects multiomics and clinical data from 55 types of tumor datasets from The Cancer Genome Atlas (TCGA), International Cancer Genome Consortium (ICGC), and the Clinical Proteomic Tumor Analysis Consortium (CPTAC), was used for CBS gene expression analysis. The tumor versus normal expression and overall survival were analyzed using the TCGA-READ and TCGA-COAD datasets.

Statistical analysis
The results are presented as mean ± S.E.M. Each value represents an average of at least three wells from at least three independent cultivations of each type of cell. Statistical differences among the groups were determined by one-way ANOVA. Statistical significance was considered to be significant at * or + -p < 0.05, ** or ++p < 0.01, and *** or +++p < 0.001.

CBS has a similar immunofluorescence pattern to b-tubulin and b-actin in variety of cells
Immunofluorescence with the anti-CBS antibody revealed similar staining to b-tubulin and b-actin in variety of cells ( Figure 1A). This staining pattern was visible in colorectal carcinoma cell lines DLD1 and HCT116, in the cell line derived from the clear cell renal cell carcinoma (ccRCC), and in the noncancerous cell line derived from the EA.hy926 epithelial cells.

Co-localization of CBS b-tubulin and/or b-actin
We built a hypothesis that CBS might co-localize with btubulin and/or b-actin. To verify this hypothesis, we performed immunoprecipitation with the anti-CBS antibody and subsequent immunodetection with b-tubulin and b-actin antibodies in the DLD1 and HCT116 cells ( Figure 1B). We observed a clear signal with both cytoskeletal antibodies and also in the presence of slow sulfide donor GYY4137 (GYY). Furthermore, we decided to show the co-localization of these proteins by a proximity ligation assay. In Figure 1C, the red dot signal shows the co-localization of CBS with both b-actin and b-tubulin in DLD1. To demonstrate the CBS i n t e r a c t i o n w i t h b -t u b u l i n , w e p e r f o r m e d d o u b l e immunofluorescence staining with CBS (Figure 2A, red signal) and b-tubulin (green signal) in the control DLD1 group in a group of DLD1 cells treated with vincristine (which prevents polymerization of tubulin) or paclitaxel (which prevents depolymerization of tubulin). In the vincristine-treated DLD1 group, the b-tubulin signal diminished, similarly to the CBS signal. In the paclitaxel-treated DLD1 group, both the b-tubulin signal and CBS signal were much stronger compared to the untreated group (Figure 2A), thus proving the interaction of these two proteins. The specificity of the CBS and b-tubulin signals was verified by the negative control (Figure 2A, inset), where the primary antibodies were omitted. However, the b-tubulin net was detectable also in the DLDx cells (Figure 2A), although the CBS signal was not present. This would suggest that CBS is not crucial for the cytoskeleton, and it might just have a modulatory function.

Functional role of CBS in colorectal carcinoma cells
To study the functional role of CBS in colorectal carcinoma cells, we prepared a DLD1/CBS_del (DLDx) cell line with an inactive CBS gene through the CRISPR/Cas9 gene editing method. These clones were selected by puromycin selection, and CBS protein knockout was confirmed by WB analysis ( Figure 2B). Absence of CBS in DLDx was demonstrated also by a proximity ligation assay with b-tubulin and CBS antibodies (not shown) and immunofluorescence with the CBS antibody ( Figure 2C). We also prepared positive (DLD-PC) and negative (DLD-NC) CRISPR controls to eliminate possible false results due to CRISPR manipulations ( Figure 2B). When comparing b-tubulin net in DLD1 and DLDx cells, we observed marked differences between these two groups of cells ( Figure 3). The control group of DLD1 cells contained very well-defined microtubules with regular distribution in the cells (without accumulation in some cell parts). Individual microtubules showed homogenous structure. Compared to the control, the signal of b-tubulin in the DLDx cells was significantly weaker. The structure of microtubules was non-homogenous, with a presence of much brighter "spots" in their structure (Figure 3, arrows). In these cells, the microtubules showed accumulation mainly around the nuclei, and on the contrary, they were almost missing at the periphery of the cells, where their spot-like structure was very well evident. Application of GYY4137 led to the restoration of a homogeneous distribution of microtubules and to the homogeneity of individual microtubules (without spot-like structure, Figure 3).

Determination of intracellular H 2 S
The intracellular H 2 S level was determined in both cell lines by detecting fluorescence on a fluorescent reader ( Figure 4A), namely, H 2 S Fluorescent Probe, P3. In the DLDx cells, the level of intracellular H 2 S was significantly higher compared to the DLD1 cells ( Figure 4A). When both these cell lines were treated with a slow sulfide donor GYY4137 (GYY), an increase in the intracellular level of H 2 S compared to DLD1 was shown ( Figure 4A). However, no changes in the intracellular H 2 S levels were observed when comparing the DLDx and DLDx/GYY groups ( Figure 4A). Images from the fluorescence microscope revealed cytoplasmic localization of the P3 fluorescence signal (not shown).

Detection of protein levels of other enzymes that endogenously produce H 2 S
Discrepancy between the increased level of intracellular H 2 S and the missing CBS protein in the DLDx versus DLD1 cells might be due to changes of other enzymes endogenously producing H 2 Scystathionine-g-lyase (CSE) and 3-mercaptopyruvate sulfurtransferase (MPST). The level of CSE protein was significantly increased in the DLDx cells compared to the DLD1 cells ( Figures 4B, C), while the MPST protein level was not changed (Figures 4B, C).

In vivo experiments on immunodeficient SCID/bg mice
Both the DLD1 and DLDx cells were injected subcutaneously into the immunodeficient SCID/bg mice. When small tumors started to be Effect of H 2 S on microtubule structure. We compared b-tubulin morphology (green) in the unmodified DLD1 cells, DLD1/CBS_del (DLDx) cell, and DLDx cells treated with slow sulfide donor GYY4137 (GYY). In the DLDx cells, we observed partial re-modulation of the b-tubulin skeleton (showed by arrows). After GYY treatment, this skeleton became the same as in the untreated cells. The nuclei were stained with DAPI. Scale bar represents 7.5 µm in both magnifications. detectable, GYY4137 treatment was applied to two groups of mice (n = 13-22 tumors; Figures 4D, E). After 14 days, the mice were sacrificed, and the volume and the tumor's weight were determined. Representative tumors of each group are shown in Figure 4E. The volume ( Figure 4D upper graph, E) and weight ( Figure 4D lower graph) of tumors were approximately twice as big in the mice inoculated with DLDx than in those with DLD1 cells. Interestingly, when the mice with DLDx tumors were simultaneously treated with GYY4137, the volume and weight of the tumors were significantly smaller and remained at values from the DLD1-inoculated mice ( Figure 4D). Injection of positive and/or negative CRISPR cell lines (DLD-PC and DLD-NC) in the SCID/bg mice resulted in the same size of tumors as inoculation with DLD1 (not shown).

Impact of CBS knockout on cell proliferation and migration
Proliferation and migration were also significantly higher in DLDx compared to DLD1 ( Figure 5). Nevertheless, while proliferation was not affected by GYY4137 treatment in the DLD1 cells, a significant decrease was visible in the DLDx cells ( Figure 5A). GYY4137 decreased migration in both the DLD1 and DLDx cells ( Figure 5B). Control positive and negative CRISPR did not affect these processes (Figures 5C-E).

Discussion
Using the cystathionine b-synthase antibody, we observed "fibred"-like immunofluorescent staining in the DLD1, HCT116, ccRCC and EA.hy926 cells, which suggests the binding of CBS to cytoskeletal proteins. Immunoprecipitation with the CBS antibody and a subsequent Western blot analysis and proximity ligation assay proved that CBS binds to b-tubulin and b-actin. Moreover, we compared the immunofluorescence of CBS and b-tubulin by using two microtubule targeting agents, both affecting b-tubulinpaclitaxel and vincristine [for a review, see (25,33)]. Paclitaxel was shown to inhibit human cervical cancer cell division at low Determination of levels of intracellular H 2 S, other enzymes endogenously producing H 2 S, and xenograft formation in SCID/bg mice inoculated by DLD1 and DLDx cells. The intracellular level of H 2 S was significantly increased in the DLDx cells compared to DLD1, as determined by a fluorescent reader (A), using H 2 S Fluorescent Probe, P3. Treatment of both cell lines with a slow sulfide donor, GYY4137 (GYY), also resulted in increased intracellular H 2 S levels compared to the untreated controls (A). In the DLDx cells, the level of cystathionine-g-lyase (CSE) was significantly higher compared to the DLD1 cells, while the level of 3-mercaptopyruvate sulfurtransferase (MPST) was not changed (B, C). Immunodeficient SCID/bg mice were inoculated by either DLD1 or DLDx (with knocked-out CBS). On the fourth day, when small tumors started to become visible, the mice were divided into four groups. One of two DLD1 groups and one of two DLDx groups were simultaneously treated daily with GYY4137 (GYY). On termination day 14, the animals were sacrificed, and the xenograft's volume (D upper graph, E) and weight of tumors (D, lower graph) were estimated. Xenografts from the DLDx cells were approximately twice as big in volume and weight compared to those from the DLD1 cells. When the mice were simultaneously treated with GYY, a significant decrease in volume and mass occurred in tumors from these cells. The volume and mass of tumors in the GYY-treated DLDx mice was significantly smaller (D). Typical tumors from each group are shown in part (E). Positive and negative CRISPR controls did not affect the size and/or volume of the tumors compared to the DLD1 group (not shown). The results in graphs (A-C) are displayed as mean ± S.E.M. and represent an average of 6-11 parallels from at least three independent cultivations. The results from the nude mice are displayed as mean ± S.E.M., and the number of tumors evaluated was n=13-22. Statistical significance * -p< 0.05, ** -p< 0.01, and *** represent p< 0.001 compared to the control DLD1 group. Statistical significance +++ represents p< 0.001 compared to the DLD1 group treated with GYY4137 (A) or to the DLDx group (D).
concentrations of the drug (0.25 mmol) and to block human cervical cancer cells in the G2/M phase (34). Blocking cells in the G2 phase was determined also in three colorectal carcinoma cell lines (35). Immunofluorescence of the DLD1 cells showed significantly increased staining of both b-tubulin and CBS in paclitaxel-treated cells compared to the control untreated ones. Additionally, when we used vincristine, a chemotherapeutic agent that prevents polymerization of b-tubulin, a very small signal of CBS and tubulin occurred compared to the controls. These results together with those previously discussed strongly suggest that CBS binds to cytoskeletal proteins in DLD1 cells.
It is obvious that the binding of CBS to cytoskeletal proteins might result in functional consequences. Changes in CBS binding might participate in re-organization of the cytoskeleton and potentially to altered tumorigenicity. There are several possibilities to explain the function of CBS on the cytoskeletoneither directly by stabilizing the cytoskeleton due to protein-protein interaction or indirectly through H 2 S. One possibility is that CBS stabilizes the cytoskeleton through binding to b-tubulin or b-actin. Another possible explanation lies in the local production of H 2 S. Although H 2 S can freely diffuse in the cell, it is probable that H 2 S first reacts with the targets in the close vicinity of its production that can ensure certain specificity of the H 2 S (as is the case of some other signaling molecules, e.g., calcium). Exogenous H 2 S is applied to the cells in much higher concentrations than is an endogenous production; therefore, it can diffuse to longer distances. This proposal is based on previous observations that cysteine residues in tubulin are actively involved in regulating ligand interactions and microtubule formation both in vivo and in vitro. These cysteine residues are sensitive markers in determining the conformation of tubulin. Tubulin dimer possesses 20 cysteine residues, from which twelve are localized in a-tubulin, and eight cysteines are in btubulin. Nevertheless, while the function of the b-tubulins in microtubule assembly is known, the function of b-tubulin isotypes in microtubule dynamics remains unknown (33). Cysteine oxidation is usually accompanied by loss of polymerization competence. Thus, it is not clear whether these modifications are only the result of oxidative stress or whether they also function as regulators under normal conditions. Certain cysteine residues of tubulin might regulate the dimer/microtubule equilibrium, and the thioredoxin system might, in turn, regulate this equilibrium (36).
Production of endogenous H 2 S is not performed solely by CBS but also by other enzymes, mainly CSE and MPST. CBS, CSE, and MPST are differently expressed in various tissues and organs, but their levels can also be up or downregulated under diverse conditions and in different tumor types. Different enzymes (or combinations of these enzymes) may play a role in H 2 S overproduction (for a review, see 37). We have found that CBS knocked out the DLDx cell line overexpressed CSE, which might be a compensatory mechanism for H 2 S production. Indeed, the level of cytosolic H 2 S was significantly higher in DLDx compared to the DLD1 cell line. Expression of CBS differs in individual types of cancer and is definitely dependent on the cancer grade. The negative correlation of CBS expression with the pathologic parameters in hepatocellular carcinoma (HCC) indicates its potential as a prognostic marker in HCC (38). Moreover, levels of CBS were decreased by the higher grade of clear cell renal cell carcinoma (14). On the other hand, CBS, CSE, and MPST increased with the rise of malignant degrees in human bladder tissues and human UCB cell lines (39). Therefore, further studies on the role of CBS in tumorigenesis must consider tumor type and also grade.
In order to explore the expression and prognostic value of CBS in colon and rectal adenocarcinomas, we performed a comprehensive in silico analysis using the Tumor online Prognostic analysis Platform (ToPP) (32). Based on the expression status of CBS (CBS high/low), the overall survival presented as a Kaplan-Meier (KM) survival plot is depicted in Figure 6B. Kaplan-Meier curves show that although generally low CBS expression is associated with colon adenocarcinoma, reduced overall survival is significantly (p=0.017) associated with high CBS expression. In contrast, rectal adenocarcinomas show higher CBS expression compared to nonneoplastic tissue. However, low expression of CBS is associated with shorter overall survival, although it is statistically insignificant due to the lower number of evaluated patients (n=94). From the longterm survival, it is apparent that low CBS expression affects rectal carcinoma rather than colon adenocarcinoma.
Epigenetic factors (such as hypermethylation) should also be part of these studies. Downregulation of CBS through promoter methylation has been observed in multiple gastric cancer cell lines and four colon cancer cell lines (including HCT116) (40). More information is available about the relation of CBS expression in gastric cancer. The CpG island methylator phenotype (CIMP) is an epigenetic molecular subtype, which is observed in multiple malignancies and associated with the epigenetic silencing of tumor suppressors. A novel association between CBS epimutations and the CIMP subtype in gastric cancer was discovered, with in vitro models of CBS deficiency resulting in abnormal DNA methylation and inflammatory response (41).
To evaluate the physiological relevance of CBS in colorectal carcinoma cells, we used the CRISPR/Cas9 gene editing method (28,29) to prepare the DLD1 cell line with knocked-out CBS (DLDx). A B FIGURE 6 Analysis of CBS gene expression in the TCGA-COAD and TCGA-READ datasets. (A) Differential expression between tumor and normal tissue is presented as a boxplot for the two groups. The Wilcox test was performed to test whether there was a significant difference between the two groups. *** p < 0.001. The univariate analysis presented as a Kaplan-Meier (KM) survival plot (B) includes the p-value and hazard ratio with 95% confidence interval information.
Efficiency of the CRISPR was determined by Western blot and hybridization with the CBS antibody and by immunofluorescence and proximity ligation assay with the CBS and b-tubulin antibodies. To eliminate false results due to possible interference of the CRISPR technique, we also prepared a positive (DLDPC) and negative (DLDNC) CRISPR control. These controls normally expressed the CBS.
The b-tubulin immunostaining of DLDx revealed significant changes in cytoskeletal morphology compared to the unmodified DLD1 cells. In the DLDx cells, we observed partial destruction of the tubulin microfilaments, thus suggesting an important role of CBS and sulfide signaling, possibly through cysteine residues. It appears that CBS has a protective effect on microtubules. It was already shown that substitution of cysteine residues yields tubulin that is not capable of assembling into microtubules (42). However, reduced tubulin is polymerization competent, forming normal microtubules (36). Thus, a role for the cysteines in tubulin remains unresolved. Although there are several papers discussing the different roles of individual isotypes of b-tubulin in carcinogenesis (for a review, see 33), these isotypes have 85-95% similarity, and therefore, from the point of CBS binding, it is diffi cult to distinguish between them. Nevertheless, polymerization of b-tubulin depends on a wide variety of factors (e.g., phosphorylation and acetylation), and all these processes might be targeted by H 2 S produced by CBS. Further research is required to clarify this mechanism (which might be distinct from the production of H 2 S).
Moreover, we determined the levels of intracellular H 2 S in the DLD1 and DLDx cells. We found a significant increase of the intracellular H 2 S level in the DLDx cells compared to DLD1. This is in contrast to the eliminated CBS; therefore, we determined the protein levels of other endogenously producing enzymes, CSE and MPST. We observed a significant increase in the levels of CSE (and not of MPST), which might point to the compensatory effect of CSE in producing H 2 S. In various types of cancer, CBS and CSE expressed differently (16), so their mutual interaction and/or modulation might differ. Inhibition of CBS has shown anti-tumor activity, particularly in colon cancer, ovarian cancer, and breast cancer, whereas the consequence of CSE or 3MST inhibition remains largely unexplored in cancer cells (16). This observation uncovers a field for further studies.
The physiological relevance of the CBS/tubulin complex was studied by determining xenografts on immunodeficient SCID/bg mice and also proliferation and migration on DLD1 and DLDx cell lines. The xenografts obtained from the DLDx cells were significantly larger than those from the DLD1 cells. Parallel treatment of mice with GYY4137 resulted in the prevention of the DLDx tumor's rapid growth, suggesting the involvement of H 2 S in this process. GYY4137 is a slow sulfide donor that releases H 2 S slowly and steadily, either in aqueous solution or administered to the animals, because of low toxicity (43). The effect of GYY4137 on tumor suppression has already been demonstrated on DLD1 and HepG2 cells (17,18).
In cell cultures, DLDx also exhibited increased proliferation compared to DLD1, and GYY4137 significantly decreased this process in the DLDx cells but not in the DLD1 cells. This is in line with the observation of Zhang and co-workers (44), who demonstrated that endogenous overexpression of CBS and exogenous H 2 S could inhibit the proliferation and migration of colorectal carcinoma cells both in vivo and in vitro. Thus, it is apparent that expression of the CBS might be an important regulator of cell proliferation and migration. The mechanism behind how CBS, through cytoskeletal proteins, can affect tumor growth, proliferation, and migration needs to be studied in detail. As stated by Zuhra et al. (2), "even when the involvement of CBS in a given biological process is undisputable, it is often difficult to determine if the observed biological effects related to CBS are, in fact, due to upstream alterations (e.g., homocysteine accumulation due to CBS inhibition), downstream alterations (e.g., lack of production of cytoprotective cystathione or H 2 S after CBS inhibition) or global cellular changes (e.g., alterations in cellular glutathione levels and compensatory changes in redox balance)". Likewise, protein-protein interaction should be taken into consideration regarding the function. It has already been demonstrated that silencing of the CBS in HCT116 cells decreases cell proliferation and attenuates HCT116 xenograft growth and vascularization in female Balb/c nude mice. These observations were performed either by silencing techniques or by CBS blockers, such as S-adenosyl-L-methionine (45) or aminooxyacetic acid (46), which did not allow the complete blockade of CBS. Using the CRISPR/Cas9 approach, one can ensure that CBS is permanently deleted. Furthermore, we used positive and negative CRISPR controls to eliminate possible false results caused by CRISPR technology.

Conclusions
In summary, we have shown that in DLD1 colorectal carcinoma cells, CBS binds to the cytoskeletal proteins b-actin and b-tubulin, which has an impact on tumor growth, proliferation, and migration. When DLD1 cells with knocked-out CBS were inoculated into immunodeficient mice, larger tumors were obtained compared to the mice inoculated by the normal DLD1 cell line. Exogenous supplementation of H 2 S to the animals decreased the size of the tumors. When a slow sulfide donor, GYY4137, was added to the CBS knocked-out cell line, a decrease in proliferation and migration was shown.

Data availability statement
The original contributions presented in the study are included in the article/supplementary material. Further inquiries can be directed to the corresponding author.

Ethics statement
The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Ethic Committee and by the national competence authority -State Veterinary and Food Administration of the Slovak Republic (project registration No. Ro-2032-3/2020-220) in compliance with the Directive 2010/63/EU and the Regulation 377/2012 on the protection of animals used for scientific purposes. Project was conducted in the approved animal facility (license No. SK UCH 02017). Written informed consent was obtained from the owners for the participation of their animals in this study.