Skip to main content
NCBI home page
BMC Cancer logoLink to BMC Cancer
. 2025 Aug 22;25:1356. doi: 10.1186/s12885-025-14717-5

Regulation of EMT-MET and chemoresistance by the Lc3Cer-synthase B3GNT5

Laura E Clark 1,#, Katherine Hylton Rorie 1,#, Amanda J G Dickinson 1, Santiago Lima 1,2,✉
PMCID: PMC12372232  PMID: 40847295

Abstract

Background

Glycosphingolipids (GSL) are essential components of the plasma membrane where they are known to play key structural and functional roles and are known to influence molecular processes involved in cancer malignancy, including multi-drug chemoresistance, the epithelial to mesenchymal transition (EMT), and the activation or receptor tyrosine kinases (RTK). Thus, investigating and understanding how GSLs are regulated in cancer and the impact they have on malignancy have important therapeutic potential. In the GSL biosynthetic pathway, one critical regulator of two of the four major branches of GSLs is the gene product of B3GNT5, which produces the precursor for all GSLs in the lactoside and neolactoside series.

Methods

Publicly available data was mined to determine the types and prevalence of genetic lesions at the B3GNT5 locus in various cancers, and to assess the impact of increased expression on patient outcomes. HeLa cells in which B3GNT5 was partially depleted using CRISPR-Cas9 approaches were used to determine how its expression levels impacted several phenotypic properties associated with cancer malignancy. Mass spectrometry was used to assess the effect of B3GNT5 on the levels of the Lc3Cer precursors glucosylceramide (GlcCer) and lactosylceramide (LacCer).

Results

B3GNT5 copy number gain and overexpression are widespread across human cancers and are significantly associated with poor prognosis. Partial depletion of B3GNT5 in HeLa cells led to accumulation of GlcCer and LacCer, increased chemoresistance, altered EMT marker expression, and decreased activation of multiple RTKs following stimulation with serum—suggesting broad signaling and phenotypic shifts.

Conclusions

B3GNT5 is frequently altered in human cancers and correlates with adverse clinical outcomes. Functional depletion reveals its key role in regulating glycosphingolipid metabolism, signaling, and malignant phenotypes. These findings support B3GNT5 as an important target for therapeutic intervention in cancer.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12885-025-14717-5.

Keywords: B3GNT5, Glycosphingolipid, Cancer, Chemoresistance, EMT, Lactosylceramide, Glucosylceramide

Background

GSLs are components of the outer layer of the eukaryotic plasma membrane where they serve as mediators of membrane-associated protein functions, cell-cell interactions, and other vital cellular dynamics and physiological processes [1–3]. Alterations in lipid metabolic pathways are an emerging hallmark of cancer [4], and in many cancers the metabolism of sphingolipids has been shown to be significantly altered [5–7]. The basic building blocks of all GSLs, GlcCer and LacCer, are significantly elevated in the tumors of patients and are known impact disease malignancy and correlate with poor outcomes [5, 7–9]. Additionally, GlcCer has also been well-established as a key player in acquired and intrinsic chemoresistance in cancer [10]. Therefore, because of their involvement in various molecular processes that contribute to cancer malignancy, understanding the mechanisms that regulate the levels of cellular GSLs, and their building blocks, in transformed cells is crucial.

Cellular GSL levels are regulated by a host of enzymes that sequentially glycosylate and chemically modify GlcCer and LacCer to form diverse glycolipid structures [11]. Early steps in GSL biosynthesis involve the galactosylation of GlcCer by LacCer-synthase to produce LacCer, which is the precursor of four major classes of cellular GSLs including gangliosides, globosides, lactosides, and neolactosides. To synthesize GSLs in the lactoside and neolactoside series, LacCer is modified in the trans-Golgi by β-1,3-N-acetylglucosaminyltransferase 5 (B3GNT5) to produce Lc3Cer. Therefore, B3GNT5 is a key enzyme in the regulation of GSL biosynthesis as it controls flux into two of the four major classes of mammalian cellular GSLs.

There is growing evidence that suggests aberrant levels of B3GNT5 play a role in human cancers. For example, increased expression of B3GNT5 was observed in uterine endometrial cancer-derived cells [12], in glioma stem cells and glioma cell lines [13], in basal-like breast cancer tumors [14], and in tissues of patients with non-small cell lung cancer [15]. Moreover, high tumor B3GNT5 expression was also associated with poor prognosis in breast cancer [14] and glioblastoma multiforme [13]. B3GNT5 has also been shown to promote malignancy; in non-small cell lung cancer tumor derived cell lines increased expression of B3GNT5 increased their invasiveness in vitro and tumor growth in vivo [15]. Conversely, in glioma stem cells, decreased expression of B3GNT5 resulted in reduced growth of neurospheres suggesting this enzyme could be required for tumor growth [13]. The enzymatic product of B3GNT5, Lc3Cer, is also associated with cancer and its levels were elevated in the bone marrow of patients with acute myeloid leukemia [16]. Furthermore, various species of lactosides and neolactosides, which are built upon Lc3Cer, were elevated in the tumors and serum of patients with non-small cell lung cancer [15]. Serum levels of Lc3Cer have also been associated with disease progression and poor outcomes [15]. Finally, LacCer, the substrate for B3GNT5, was also significantly elevated in the tumors of patients with lung adenocarcinoma, lung squamous cell carcinoma, endometroid endometrial carcinoma, head & neck squamous cell carcinoma, and colorectal adenocarcinoma [5–7].

Thus, it is becoming clear that alterations in the expression levels of B3GNT5, and sphingolipid metabolites used or synthesized by this enzyme, are associated with cancer. However, it is unclear how this important mediator of GSL biosynthesis could contribute to cancer phenotypes. Based on previous work and known roles of GSLs it was hypothesized that elevated B3GNT5 in cancer cells may alter membrane structure by indirectly regulating the levels of GlcCer and LacCer, which in turn could lead to changes in cell-cell interactions and the function membrane protein receptors and transporters. To test this hypothesis, B3GNT5 was genetically depleted in the human tumor derived squamous cell carcinoma cell line HeLa. It was predicted that this approach may reveal the influence of B3GNT5 on malignancy associated phenotypes. Results show that HeLa cells with decreased expression of B3GNT5 accumulated GlcCer and LacCer, which would alter membrane structure. Moreover, B3GNT5 partial depletion altered the expression various markers of the EMT and decreased their invasive capacity suggesting changes in cell-cell and cell-matrix interaction changes. Reduction in B3GNT5 expression also increased the chemoresistance of HeLa cells to various cytotoxic chemotherapeutic agents and altered their RTK activation profile. The work presented here is novel in that it is the first to dissect the mechanisms by which B3GNT5 could participate in cancer progression and treatment resistance, which is of important clinical relevance.

Methods

Cell culture

HeLa (CCL-2) cells were from the American Tissue Type Collection and were cultured in Eagle Modified Essential Medium (112-018-101, Quality Biological) supplemented with 10% fetal bovine serum (FBS, 35-011-CV, Corning, Corning, NY). All cell culture was performed by growing cells at 37° C and 5% CO2 in a humidity saturated incubator.

Cell viability and proliferation assays

Cell viability assays were performed as described with WST-8 formazan [17]. Briefly, for viability assays, cells were seeded at a density of 10,000 per well in 96-well plates (standard tissue culture-treated plates) using complete growth medium supplemented with 10% FBS. Following a 24-hour attachment period, cells were exposed to either chemotherapeutic compounds or vehicle control and incubated for an additional 72 h. Drug solutions were prepared as follows: paclitaxel, vinorelbine, and doxorubicin were dissolved in DMSO; cisplatin was diluted in water and used immediately, or discarded if stored at 4 °C beyond one week. Cell viability was assessed using a colorimetric assay based on WST-8 conversion to a water-soluble formazan product, with absorbance measured at 450 nm using a microplate reader (Biotek Instruments, Winooski, VT, USA). The WST-8 reagent was obtained from Domino Molecular Technologies (CCK-8 kit, CK04).

To assess proliferation, 2,500 cells were plated per well in 96-well plates and allowed to adhere for 18 h in complete medium. Cell growth was monitored at 24-hour intervals using the WST-8 assay as described above.

Immunoblotting

Cells extracts were prepared by sonication in buffer containing 20 mM HEPES, pH 7.4, 250 mM NaCl, 1% Triton X-100, 20% glycerol, and containing Halt protease plus phosphatase inhibitors (78440, ThermoFisher Scientific, Waltham, MA). For western blot analysis, SDS-PAGE gels loaded with equal amounts of extracts were resolved, and then proteins transferred to 0.2 μm polyvinylidene fluoride membranes (1620177, Bio-Rad, Hercules, CA). After transferring, to maximize the number of proteins that could be analyzed per gel, membranes were cut in strips as appropriate. Briefly, to facilitate cutting, the first and last lane of gels were loaded with molecular weight marker ladder (BioRad #1610374). Strips were cut using a small paper trimmer (Tonic Studios #4496EUS) by cutting parallel to molecular weight markers on either side of the membrane. Cuts were determined based on the molecular weight of the target protein. Strips probing B3GNT5 were cut either through or above 75 kDa marker and below the 37 kDa marker; GAPDH blots were cut just below the 50 kDa marker and below the 25 kDa marker; SNAIL and SLUG blots were only cut through the 37 kDa marker; ZEB-1 or ZO-1 blots were cut only through the 100 kDa marker; blots probing claudin-1 were cut above 37 Kda and below 20 kDa markers; blots probing N-cadherin or E-cadherin were cut just below the 100 kDa marker; vimentin blots were cut through the 75 kDa marker and below the 50 kDa marker; cytokeratin blots were cut through the 75 kDa marker and below 37 kDa; β-catenin blots cut only through the 75 kDa marker. Antibodies used for immunoblotting were diluted in a 2% solution of bovine serum albumin (MilliporeSigma, 126609) dissolved in 50 mM Tris-Cl, pH 7.5, 150 mM NaCl, 0.1% Tween-20 and were: B3GNT5 (1:1000, 20422-1-AP, Proteintech), MRP1/ABCC1 (1:2000, Cell Signaling Technology, 72202), MDR1/ABCB1 (1:2000, Cell Signaling Technology 13342) GAPDH (1:3000, 2118, Cell Signaling Technology, Danvers, MA), vimentin (1:1000, Cell Signaling Technology, 5741), N-cadherin (1:1000, Cell Signaling Technology, 13116), β-Catenin (1:1000, Cell Signaling Technology, 8480), SNAIL (1:1000, Cell Signaling Technology, 3879), SLUG (1:1000, Cell Signaling Technology, 9585), ZEB1 (1:1000, Cell Signaling Technology, 3396), E-Cadherin (1:1000, Cell Signaling Technology, 3195), pan-keratin (1:1000, Cell Signaling Technology, 4545). HRP-conjugated secondary antibodies were used to visualize blots by chemiluminescence (1:5000, anti-rabbit 111035045, anti-mouse 115035166; Jackson Immuno Research Labs, West Grove, PA).

Mass spectrometry

Mass spectrometry was performed as described [17]. Briefly, cells were seeded at 350,000 per well in six-well tissue culture-treated plates and cultured for 48 h in medium supplemented with 10% fetal bovine serum. After incubation, media were removed and cells were washed twice with ice-cold phosphate-buffered saline (PBS; ThermoFisher Scientific). Cells were then scraped in 200 µL of PBS containing protease and phosphatase inhibitors (Halt, ThermoFisher Scientific) and collected on ice. A portion of the suspension (100–150 µL) was added to 1 mL of cold LC–MS-grade methanol, followed by the addition of chloroform to achieve a final solvent ratio of 2:1:0.1 methanol: chloroform: water. Samples were stored at − 80 °C until further processing.

Prior to lipid extraction, internal standards (prepared in ethanol: methanol: water, 7:2:1) were added as previously described [17]. The remaining cell lysate was sonicated, clarified by centrifugation, and used for protein quantification by Bradford assay (Bio-Rad). Sphingolipids were quantified by LC–ESI-MS/MS (5500 QTRAP, AB Sciex) using retention time standards and methods described in prior studies [5, 7, 17]. All reagents and solvents used were of LC–MS grade.

B3GNT5 knockout cells

To establish a B3GNT5 knockout cell line, cells were infected with a predesigned lentivirus (Sigma-Millipore) delivering a vector (LV01: U6-gRNA: EF1α-puro-2 A-Cas9-2 A-tGFP; HSPD0000119326) encoding Cas9, a guide RNA (5′–3′: TAATCAAGTATTGGTAGCG) targeting the B3GNT5 locus (NM_032047.2), and a puromycin resistance cassette and GFP that were used for selection. The parental wild-type cell line was used as the comparator in functional assays. We acknowledge that this approach does not control for the effects of lentiviral infection, Cas9 expression, or antibiotic selection; which is a limitation of the current study. The day before infection, 3 × 106 cells were seeded in a 10 cm dish. The following day, media was aspirated, and new media was added containing 8 µg/mL polybrene and lentivirus at multiplicity of infection of 1. After 24 h the media was replaced, the cells allowed to expand for 3 days, and then single cells were isolated into 96-well plates by sorting green fluorescent protein positive cells using fluorescence-activated cell sorting at the VCU Flow Cytometry Core. Outgrowing clones were examined for expression of B3GNT5 using western blotting with two different antibodies.

Statistical analyses

Statistical analyses were performed using Prism 8 (GraphPad, Boston, MA). Experiments were repeated three times and representative data shown in figures. A P-value ≤ 0.05 was considered statistically significant. For western-blotting and RTK proteome profiler arrays, immunopositive bands were analyzed with Fiji/ImageJ 1.53 [18].

Phospho-RTK proteome profiler analysis

Proteome Profiler Human Phospho-RTK Array Kit (ARY001B; R&D Systems, Minneapolis, MN) experiments were performed as described [17]. Briefly, Receptor tyrosine kinase (RTK) phosphorylation was assessed using the Proteome Profiler Human Phospho-RTK Array Kit (ARY001B; R&D Systems, Minneapolis, MN, USA) according to the manufacturer’s instructions with minor modifications. For each condition, 3 × 10⁶ cells were seeded onto 10 cm tissue culture dishes and cultured for 48 h. GBA knockout and control cells were processed in parallel. To evaluate RTK activation following serum stimulation, cells were serum-starved in FBS-free medium for 3 h at 37 °C and 5% CO₂. Serum was then added to a final concentration of 10%, and cells were incubated for an additional 30 min. Cells were placed on ice, washed twice with cold PBS, and lysed in 1.5 mL of lysis buffer 17 supplemented with protease and phosphatase inhibitors (Halt, ThermoFisher Scientific, 78440). Lysates were rocked at 4 °C for 1 h, and total protein concentration was determined. Equal amounts of protein (300 µg in 1 mL of lysis buffer mixed with 500 µL of array buffer) were incubated with RTK array membranes overnight at 4 °C with gentle agitation.

Membranes from each condition were developed simultaneously using chemiluminescent detection and imaged in parallel on a ChemiDoc MP system (Bio-Rad). Signal intensity for duplicate antibody spots was quantified using densitometry in Fiji/ImageJ 1.53 1.53 [19]. Statistical analysis was performed in GraphPad Prism 8. All experiments were independently repeated three times.

Invasion assays

Assays were performed as described [17]. Briefly, all components used to handle Matrigel (Corning 356234), including pipette tips, inserts, and tubes, were prechilled at − 20 °C. Matrigel was thawed overnight at 4 °C on ice prior to use. On the day of the assay, it was diluted to 0.5 mg/mL in serum-free medium, kept cold, and 100 µL of the diluted matrix was applied to each 8.0 μm pore, 12 mm diameter transwell insert (MilliporeSigma PI8P01250). Inserts were incubated for 3 h at 37 °C and 5% CO₂ to allow gel solidification. After incubation, inserts were transferred into 24-well plates (Grenier Bio-One) containing 750 µL of pre-warmed complete medium (with 10% FBS) in the lower chamber.

Cells were serum-starved for 2 h in serum-free medium, trypsinized, and resuspended at 5 × 10⁵ cells/mL. Then, 100,000 cells in 200 µL of serum-free medium were added to the upper chamber of each Matrigel-coated insert. Cells were incubated for 48 h at 37 °C and 5% CO₂ to allow invasion. Following incubation, inserts were washed in PBS and transferred to a fresh 24-well plate. The inner and outer surfaces of each insert were rinsed, and invading cells on the outer surface were fixed with 4% paraformaldehyde (5 min), permeabilized in 100% methanol (20 min), and stained using 0.5% crystal violet (filtered, in water). Excess stain was removed with PBS washes. Cells and Matrigel remaining on the inner surface were gently removed with a PBS-moistened cotton swab. Stained cells on the outer membrane were imaged using a stereoscope against a white background and manually counted. Statistical comparisons were performed using unpaired t-tests in GraphPad Prism 8.

Results

B3GNT5 genomic alterations in human cancers

Given that increased levels of B3GNT5 have been observed in various tumor derived cells and is associated with malignancy in several cancers, an unbiased examination of B3GNT5 expression across many cancers utilizing the TNMPlot transcriptomic cancer database and analysis tool was conducted [20]. In comparison to normal tissues, B3GNT5 expression levels were significantly increased in 21 out of 22 cancer types evaluated (Fig. 1a), suggesting that elevated transcript levels of this enzyme is a common feature in cancer. Next, to better understand the source of the increased levels of B3GNT5 expression in cancer, genomic alterations at the B3GNT5 loci in the tumors of patients with cancer using genomic data available in The Cancer Genome Atlas (TCGA) were examined. Results indicated that a substantial proportion of patients with various cancer histologies had copy number variant (CNV) gain mutations at the B3GNT5 loci (Fig. 1b). Specifically, B3GNT5 CNV gain mutations were observed in 87.2% of patients with lung squamous cell carcinoma, 72.9% of patients with cancers of the cervix, 71.7% with cancers of the ovaries, 69.6% of patients with head & neck cancer, 62.3% of patients with cancers of the esophagus, 47.0% of patients with bladder cancer, and others (Fig. 1b). On the other hand, simple somatic mutations and CNV loss mutations in B3GNT5 were uncommon in these cancers (Fig. 1b). As it has been previously shown that there is a correlation between CNV gain mutations and elevated gene expression in cancer [21], these results suggest that the elevated transcript levels in Fig. 1a may be primarily due to CNV gain genetic lesions at the B3GNT5 loci.

Fig. 1.

Fig. 1

Open in a new tab

Increased expression and copy number amplifications of B3GNT5 are highly prevalent in human cancers. a The TNMplot online database was queried for the expression of B3GNT5 in the indicated cancers. Cancers where tumor levels of B3GNT5 were significantly higher than in normal tissues are marked with an asterisk (*). b For the indicated cancers, TCGA data was analyzed for B3GNT5 (ENSG00000176597) genomic alterations. Plots b-g show the percentage of cases altered and colored by CNV gain alterations (red segments), CNV loss alterations (green segments), and simple somatic mutations (SSM, blue segments). c–g TCGA data for the indicated cancers was analyzed for CNV gain alterations in genes of sphingolipid, lactoside, neolactoside, ganglioside, isogloboside, and globosides metabolic pathways

In cancer, CNV alterations in sphingolipid metabolism genes are common and whether B3GNT5 CNV gain mutations are more or less common that other genes in sphingolipid metabolism was queried [8, 17]. For this analysis, TCGA data was queried for the number of patient cases with alterations in genes listed in the Kegg database [22] for the: core sphingolipid pathway (Kegg hsa00600); glycosphingolipid lacto and neolacto series (Kegg hsa00601); ganglio (Kegg hsa00604), isoglobo and globo series (Kegg hsa00603). Results of this analysis showed that in lung squamous cell carcinoma (Fig. 1c), cervical squamous cell carcinoma (Fig. 1d), and head & neck squamous cell carcinoma (Fig. 1f), B3GNT5 was the gene within sphingolipid metabolic pathways with the highest percentage of patients with CNV gain mutations. In ovarian serous cystadenocarcinoma (Fig. 1e) and esophageal carcinoma (Fig. 1g) B3GNT5 alterations were ranked the second-most altered relative to other sphingolipid metabolism genes.

While it was observed that B3GNT5 locus CNV-gain mutations and elevated transcript levels were commonplace in many cancers, this analysis did not address whether these alterations could impact patient survival. To address this, associations between outcomes and high and low tumor transcript levels of B3GNT5 were evaluated in specific cancer types using the Human Protein Atlas survival analysis tool [23], and with data obtained through the CBioPortal cancer genomics portal [24]. Results of this analysis indicated that high transcript levels of B3GNT5 were associated with significantly decreased overall survival in cancers of the thyroid (Fig. 2a), liver (Fig. 2b), head & neck (Fig. 2c), endometrium (Fig. 2d), lung (Fig. 2e), pancreas (Fig. 2f), kidney (Fig. 2g), stomach (Fig. 2h), and urothelium (Fig. 2i). To determine if B3GNT5 CNV alterations are associated with patient outcomes in all cancers, a pan-cancer analysis that included all TCGA projects was performed. Results showed that patients with any B3GNT5 mutation were at significantly higher risk for decreased survival (Fig. 2j) and disease-free survival (Fig. 2k) when compared to patients with no genomic B3GNT5 alterations. These results suggest that CNV gain alterations of B3GNT5 indeed impact patient outcomes across many cancer histologies and in pan-cancer.

Fig. 2.

Fig. 2

Open in a new tab

Increased expression of B3GNT5 is significantly associated with poor outcomes in human cancers. a–i The Human Protein Atlas online tool was used to generate Kaplan-Meier plots and analyze the outcomes of patients with high and low tumor expression of B3GNT5. j, k Kaplan-Meier plots generated using pan-cancer TCGA genomic data for cases in which the B3GNT5 (ENSG00000176597) loci was altered (any type of mutation) or unaltered (no mutation in B3GNT5) and the indicated outcome measures for overall survival (OS; panel j) or disease-free survival (DFS; panel k) determined (Prism 8)

Collectively, these findings suggest that increased transcript levels of B3GNT5 in human patients could be explained by genomic alterations at the B3GNT5 locus, and specifically CNV-gain genetic lesions. Also, that such alterations are widespread features of tumors from various types of cancer histologies, and that these genetic changes in B3GNT5 significantly impact survival. Because of these findings suggesting a high prevalence of increased expression of B3GNT5 and impact on survival, further analysis of the role of B3GNT5 in cancer cells was conducted.

Reduction of B3GNT5 expression in HeLa cells led to the accumulation of LacCer and GlcCer

To assess the role that B3GNT5 expression levels have on cancer phenotypes, CRISPR-Cas9 was used to genetically deplete its levels in the human cervical squamous cell carcinoma cell line HeLa. Following selection of clones expressing the transfection associated cassette marker, B3GNT5 expression in putative knockout cells was assessed using western blotting. For further experiments, a clone was selected that had significant but partial reduction (approximately 50%) of B3GNT5 protein levels when compared to control cells (Fig. 3a and b; B3GNT5 partial knockout HeLa cells henceforth referred as B3GNT5PKO). Full B3GNT5 knockout cells were not observed in several rounds of experiments performed to delete it in HeLa cells using CRISPR-Cas9.

Fig. 3.

Fig. 3

Open in a new tab

Partial depletion of B3GNT5 resulted in the accumulation of GlcCer and LacCer. a, b B3GNT5 was targeted in HeLa cells using CRISPR-Cas9. An isolated clone was analyzed using immunoblotting with the indicated antibodies. Blots were quantified by densitometry using ImageJ 1.53. The y-axis shows band intensity normalized to GAPDH and expressed as fold change relative to wild-type. n = 5. c, d Mass spectrometry analysis of the indicated lipids per mg of input protein in HeLa/HeLa B3GNT5PKO cells. n = 3. Bar graphs with mean ± SEM shown; T-test analysis (Prism 8): ns, not significant; *, P ≤ 0.05; **, P ≤ 0.01; ****, P ≤ 0.0001

In the biosynthesis of GSLs, B3GNT5 uses LacCer as a substrate to generate Lc3Cer, which is the precursor to lactosides and neolactosides. LacCer is synthesized from GlcCer. Therefore, it was of interest to examine if depletion of B3GNT5 led to changes in the levels of LacCer and GlcCer as they are precursors to Lc3Cer. Mass spectrometry was used to assess their levels, and results showed that B3GNT5PKO cells had a significant accumulation of LacCer with 24:0 and 24:1 acyl chain-lengths (Fig. 3c). Mass spectrometry analysis also showed that HeLa B3GNT5PKO cells had a significant accumulation of monohexosylceramides of 16:0, 24:1, and 24:0 acyl chain lengths (Fig. 3d). GlcCer is a monohexosylceramide, and it is the precursor of LacCer. Together, these results confirm that HeLa B3GNT5PKO cells have decreased B3GNT5 protein levels and accumulate GlcCer and LacCer, which are the precursors of Lc3Cer that is made by B3GNT5.

Partial depletion of B3GNT5 altered the HeLa chemosensitivity profile

Alterations of plasma membrane lipids have been shown to impact the responses of cancer cells to chemotherapy, and there is a strong correlation between GSL accumulation and chemoresistance in cancer [10]. It was also shown that partial depletion of the lysosomal glucosylceramidase β1 (GBA), which is the enzyme that catalyzes the last step in GSL degradation, resulted in the accumulation of various GSL species and changed the chemoresistance profile of cells in which it was depleted. Therefore, because increased expression of B3GNT5 is a common feature of human cancers, it was predicted that knockdown of B3GNT5, and the resulting changes to the cellular GSL profile, may alter the sensitivity of cells to chemotherapy agents used to treat patients. To test this prediction, assays examined viability after control HeLa cells and HeLa B3GNT5PKO cells were treated with cytotoxic agents. Results indicated that partial depletion of B3GNT5 significantly increased the resistance (compared to wild-type control HeLa cells) of cells when they were treated with vinorelbine (Figs. 4a and 10-fold increased IC50), and paclitaxel (Figs. 4b and 5-fold increased IC50). However, there were no significant changes in viability in response to doxorubicin (Fig. 4c; 3-fold increased IC50) or cisplatin (Fig. 4d.56-fold decreased IC50). To verify that HeLa B3GNT5PKO cells were resistant to these agents via a cell reproductive death dependent mechanism, clonogenic assays were also performed. For these assays the concentrations of each drug were selected near the IC50 as determined in viability assays. Consistent with the viability assays, clonogenic experiments showed that HeLa B3GNT5PKO had significantly more surviving colonies than wild-type HeLa cells when treated with 20 nM vinorelbine (Fig. 4e), and 7.5 nM paclitaxel (Fig. 4f). Finally, because mass spectrometry analysis revealed significant accumulation of GlcCer in B3GNT5PKO cells, and given the established association between elevated GlcCer and increased expression of multidrug resistance transporters such as MDR1 [25], we examined MDR1 and MRP1 expression by immunoblotting. As shown in Fig. 4g, MDR1 levels were increased more than 20-fold in B3GNT5PKO cells compared to wild-type, whereas MRP1 expression remained unchanged. Together, these two experiments support the concept that the accumulation of GSLs like GlcCer in transformed cells are strongly correlated with chemoresistance phenotypes.

Fig. 4.

Fig. 4

Open in a new tab

Reduction of B3GNT5 levels significantly increased the chemoresistance of HeLa cells to vinorelbine and paclitaxel. a–d HeLa and HeLa B3GNT5PKO cells were treated with the indicated drugs and concentrations for 72 h, and viability assayed with WST-8. n = 3. e, f HeLa and HeLa B3GNT5PKO cells were seeded at low density in 10 cm tissue culture plates, allowed to attach for 18 h, and then treated with the indicated chemotherapeutic agents for 72 h. The media was then exchanged, and cells allowed to grow until colonies contained > 200 cells. Cells were then fixed, the colonies stained, and colonies with > 200 cells counted. n = 3. g) Wild-type and B3GNT5PKO cells were analyzed by immunoblotting for the indicated antibodies, and quantified by densitometry using ImageJ 1.53. The y-axis shows band intensity normalized to GAPDH and expressed as fold change relative to wild type. n = 5. All cells were grown at 37 °C and 5% CO2. X-axis in panels a-d is in log2 scale. T-test analysis (panels e, f), IC50 fitting, and ANOVA (panels a-d) were performed with Prism 8; ns, not significant; *, P ≤ 0.05; **, ≤ 0.01, P ***, P ≤ 0.005; ****, P ≤ 0.0001

Fig. 5.

Fig. 5

Open in a new tab

HeLa B3GNT5PKO cells had significantly reduced invasive capacity. a Representative images of the invasive capacity of HeLa and HeLa B3GNT5PKO cells as examined by quantifying the number of cells that migrated through a Matrigel matrix and were stained with crystal violet. Before assays, cells were serum-starved for 2 h prior to seeding in the upper chamber of a transwell insert containing Matrigel. Cells were allowed to grow for 48 h at 37 °C and 5% CO2. The lower chamber contained media containing 10% FBS. Plot is bar graph with mean and ± SEM. T-test analysis (Prism 8): ***, P ≤ 0.005; n = 3

Reduction of B3GNT5 expression impacted the invasive capacity of HeLa cells

One of the hallmarks of cancer is metastatic potential by detachment from neighboring cells and invasion of adjoining and distant tissues by passing through extracellular matrices. Cell membrane physiology and cell-cell interactions play an important role in these processes. As various GSLs have been well correlated with the invasiveness properties of cancer cells [26–28], and as reduction of B3GNT5 expression significantly altered the levels of GSLs in HeLa cells, it was predicted that these cells would have altered invasive properties. To test this prediction, standard invasion assays with an artificial extracellular matrix (Matrigel) were conducted. Results indicated that B3GNT5PKO cells had a significantly reduced (~ 40% less) invasive capacity than their otherwise isogenic parental cells (Fig. 5a).

Effects of B3GNT5 partial depletion on EMT-MET

Cancer cells detach from their neighbors and gain motility and invasive phenotypes in part by a process called EMT. Since HeLa B3GNT5PKO cells were significantly less invasive than their parental control cells, it was predicted that B3GNT5 could have a role in modulating cellular regulators of EMT. In support of this idea is that a reduction of GBA levels, which altered the cellular GSL landscape in cancers cells, significantly reduced the expression levels of EMT markers, and increased the levels of markers indicative of a mesenchymal to epithelial transition (MET) phenotype [17]. Therefore, the expression levels of EMT and MET markers were examined in HeLa B3GNT5PKO cells with the prediction that reducing B3GNT5 would result in a more epithelial-like phenotype. Results demonstrated that HeLa cells with reduced B3GNT5 had a significant increase in E-cadherin protein levels (Fig. 6a and b) indicative of a more epithelial phenotype. Consistent with these results, in the B3GNT5PKO cells there was also a significant increase in the tight junction proteins ZO-1 and claudin-1, indicative of a more epithelial character (Fig. 6c and d). However, it was also determined that there was also a significant increase in N-cadherin in these B3GNT5PKO HeLa cells when compared to control cells (Fig. 6a and b), which suggest that the effects of B3GNT5 partial depletion may be more complex, beyond inducing a simple cadherin switching. Other proteins that are upregulated in motile mesenchymal-like cells and are markers of EMT [29] are the intermediate filament proteins, vimentin and cytokeratin, but these were not altered in B3GNT5PKO HeLa cells (Fig. 6a and b).

Fig. 6.

Fig. 6

Open in a new tab

Decreased expression of B3GNT5 altered the expression of EMT-MET markers. The indicated proteins in HeLa and HeLa B3GNT5PKO cells were analyzed by western blotting, and representative immunoblots (a, c) and corresponding densitometry (b, d) analysis shown. n = 3. Plots in panels b and d are bar graphs with mean and ± SEM, with y-axis showing the relative band intensity normalized to the wild-type. Blots were quantified with ImageJ 1.53. T-test (Prism 8): ns, not significant; *, P ≤ 0.05; ****, P ≤ 0.0001

Various transcription factors can modulate EMT by regulating the expression of junctional and cytoskeletal proteins and include SNAIL, SLUG, and ZEB1 [30]. Given that B3GNT5 partial depletion resulted in a mixed or hybrid effect on the junctional proteins, the levels of regulatory transcription factors to determine if this could provide an explanation for these observations were examined. Consistently, HeLa cells that had depleted levels of B3GNT5 had mixed changes in the expression of these transcription factors (Fig. 6c, d). Compared to control cells, HeLa B3GNT5PKO cells had significantly increased protein levels of SNAIL and ZEB-1, whereas SLUG levels were significantly decreased (Fig. 6c and d). Beta-catenin is an integral adherens junction protein as well as a core member of canonical Wnt signaling, and it has a demonstrated role in cancer progression [31]. B3GNT5 depleted HeLa cells had a significant decrease in the levels of β-catenin (Fig, 6c, 6d) consistent with a less mesenchymal-like phenotype. Together these results reveal that B3GNT5 has a role in modulating cellular proteins that are important players in cell-cell contacts and metastasis.

Effect of partial B3GNT5 depletion on RTK activation

The regulation of EMT is complex, involving multiple cellular signaling pathways, many of which are regulated by receptor tyrosine kinases (RTK) [32]. Additionally, as a major part of the cell membrane, GSLs are known modulators of membrane receptors including RTKs [32]. Moreover, depletion of GBA, which regulates the global balance of GSLs significantly reduced activation of various RTKs in cancer cells when they were stimulated with serum [17]. Therefore, it was of interest to examine the serum-stimulated RTK profile of HeLa cells that had decreased levels of B3GNT5 expression. For these experiments, human RTK proteome profiler arrays were utilized as in previous work [17]. These arrays assess phosphorylation and thus are surrogate indicators of activity for a wide range of RTKs. Results indicated that when cells that had reduction in B3GNT5 expression were starved of serum and then returned to media containing serum for 30 min, this resulted in decreased phosphorylation of PDGFR-β, c-RET, the insulin receptor, the ephrin A5 receptor, HGFR, the axl-DTK and axl receptors (Fig. 7a-c). Importantly, not all receptors in the profiler panel showed significant changes, as EGFR, the insulin growth factor receptor, and the RYK receptor were not significantly different between HeLa and HeLa B3GNT5PKO cells (Fig. 7a-c). Together these results reveal that B3GNT5 has a role in modulating some RTKs which could explain the changes in some EMT markers and decreased invasive capability.

Fig. 7.

Fig. 7

Open in a new tab

HeLa with decreased B3GNT5 expression had significantly reduced activation of RTKs upon stimulation with FBS. a HeLa and HeLa B3GNT5PKO cells were cultured in media lacking fetal bovine serum (FBS) for 3 h, and then incubated with media containing 10% FBS for 30 min. Extracts were run in triplicate. Shown are representative image of Proteome Profiler Human RTK arrays. Spots outlined are shown in panel b. Shown are also immunoblots and quantification with the indicated antibody of input extracts used to incubate RTK arrays. n = 3. b Spots from panel a that were quantification using densitometry (ImageJ 1.53), with data shown as relative normalized to the wild-type and plotted in panel c. T-tests were performed with Prism 8. ns, not significant; *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.0005

Discussion

B3GNT5 is essential in the production of a vast repertoire of cellular GSLs since it generates Lc3Cer, the precursor for all sphingolipids in the neolactoside and lactoside series. In addition, it was established in this report that increased levels of B3GNT5 in cancer are commonplace. Therefore, it is unsurprising that alterations in this gene can have profound effects on cell physiology. In this report, data show that B3GNT5 may modulate chemoresistance, invasive qualities, RTK signaling, and junctional remodeling. Analysis of human tumor data also show that in patients B3GNT5 CNV gain genetic lesions at the B3GNT5 loci are highly common features of many types of human cancers, and that patients with tumors that have high levels of B3GNT5 are at significantly increased risk of poor outcomes. This data is consistent with previous reports demonstrating that alterations in B3GNT5 expression contribute to malignancy in cell lines and cancers of the breast [14], endometrium [12], colon [33], glioma [13], and lung [15]. Similarly, increased levels of Lc3Cer has also been observed in leukemia [16] and endometrial adenocarcinoma [34].

B3GNT5 partial depletion and spatiotemporal regulation of chemoresistance

Alterations in sphingolipid metabolism that lead to increased levels of the GSL GlcCer have long been associated with intrinsic and acquired multi drug chemoresistance (MDR) in patients with cancer and in tumor derived cell lines [10, 35]. It has also been shown that the MDR phenotype is driven in part by the increased expression of Golgi-associated transporters, which are required to translocate GlcCer and other GSLs to the plasma membrane [36–41]. Therefore, this suggests that the MDR phenotype is strongly associated with alterations in GSL biogenesis defects within the Golgi that lead to the accumulation of GlcCer and other metabolites. Recently, we demonstrated that genetic depletion of the GlcCer-glycosidase GBA in two tumor derived cell lines resulted in the accumulation of GlcCer but decreased chemoresistance [17]. We proposed that in these cells the loss of GBA activity resulted in the accumulation of GlcCer by a degradation defect in the endo-lysosomal system, rather than a biosynthesis defect in the Golgi where GlcCer is synthesized. This led us to speculate that GSL-associated MDR may be spatiotemporally dependent, and specifically, that Golgi-localized disruptions in GSL metabolism promote MDR, while lysosomal disruptions may not.

Importantly, in this report we observed that B3GNT5PKO cells had significantly higher levels of GlcCer and markedly increased expression of MDR1, a key efflux transporter associated with GlcCer-mediated drug resistance. This provides further mechanistic support for the link between Golgi-associated GSL alterations and MDR. Together, these data strengthen the model that the subcellular origin of GlcCer accumulation—Golgi versus endo-lysosomal—plays a critical role in shaping the drug resistance phenotype in transformed cells.

Partial B3GNT5 depletion and hybrid EMT-MET phenotype

Alterations in GSL metabolism are associated with a shift in the balance of proteins associated with EMT-MET [17, 42, 43]. GSLs also have a modulatory effect on RTKs that regulate the transcription of genes involved in these processes [2, 44–47]. Some of the key players in EMT include the proteins E- and N-cadherin, and the transcription factors SNAIL, SLUG, and ZEB-1, which control what is considered to be the core EMT regulatory network. SNAIL and SLUG are repressors of E-cadherin expression, which is typically associated with epithelial phenotypes. As cells undergo full EMT, the balance of E-cadherin and N-cadherin shifts, and cells expressing low E-cadherin and high N-cadherin are considered mesenchymal-like. In cancer, mesenchymal phenotypes are linked with poor prognosis and aggressive disease [30].

In this study, it was observed that partial deletion of B3GNT5 in HeLa cells resulted in significant changes in the levels of E-cadherin, N-cadherin, SNAIL, SLUG, claudin-1, and ZO-1, along with decreased activation of various RTK receptors when cells were stimulated with FBS. However, although B3GNT5PKO cells had significantly increased levels of E-cadherin there was also a significant increase in N-cadherin, along with increased expression of SNAIL. Typically, as SNAIL levels decrease, E-cadherin levels concomitantly increase, because SNAIL acts as a repressor of E-cadherin expression [48]. Therefore, it was unexpected to find increased levels of both E-cadherin and SNAIL in the HeLa B3GNT5PKO cells. Another important regulator of E-cadherin expression is the SLUG transcription factor. In B3GNT5PKO cells, SLUG levels were significantly lower than in parental cells. Therefore, we speculate that the loss of B3GNT5 results in the accumulation of GlcCer and LacCer that modulate the activity of RTKs and lead to a “hybrid” EMT-MET state [49]. This state is the most consistent with our observations as it is characterized by changes induced by medium expression of ZEB and high expression of SNAIL [49]. It has been proposed that hybrid EMT-MET phenotypes, because of their association with stemness, are strongly correlated with chemoresistance [49]. Indeed, B3GNT5PKO cells were significantly more chemoresistant than the parental cells to vinorelbine and paclitaxel.

Study strengths and limitations

A central strength of this study is the demonstration that partial disruption of B3GNT5 expression and function is sufficient to produce robust, cancer-relevant phenotypes. These include increased epithelial marker expression, decreased invasiveness, altered chemoresistance, and altered receptor tyrosine kinase (RTK) signaling, all of which support a role for B3GNT5 in modulating tumor cell plasticity. Notably, the phenotypes observed arose in the context of a partial knockout cell line, rather than a full gene ablation, reinforcing the concept that even partial inhibition of B3GNT5 may be therapeutically impactful—a scenario that more realistically reflects what is achievable with targeted inhibitors.

From a genomic standpoint, we provide strong evidence that B3GNT5 is recurrently amplified across a wide range of human cancers. Copy number gain at the B3GNT5 locus is among the most frequent alterations observed in TCGA tumor datasets, and high B3GNT5 expression correlates with significantly poorer patient survival in multiple cancer types. These clinical associations not only highlight the oncogenic potential of B3GNT5 but also support its candidacy as a therapeutic target. Functionally, partial depletion of B3GNT5 in cancer cells led to reduced invasiveness, increased E-cadherin expression, and altered activation of RTKs involved in growth factor signaling. These phenotypic changes consistent with a less aggressive, more epithelial-like state. Therefore, we speculate that these effects on pathways associated with invasion and metastasis likely contribute to the poor outcomes observed in patients with high B3GNT5 expression.

Together, the convergence of clinical and experimental data positions B3GNT5 as a critical regulator of tumor progression and a potential point of therapeutic vulnerability. However, our findings also reveal a potential challenge: cells with partial B3GNT5 depletion accumulate GlcCer and exhibit increased chemoresistance, accompanied by elevated expression of MDR1. This suggests that therapies targeting B3GNT5 may inadvertently promote MDR unless GlcCer accumulation is simultaneously addressed. To mitigate this risk, co-treatment with GlcCer synthase inhibitors—several of which are FDA-approved—may be necessary to prevent or overcome drug resistance in tumors with reduced B3GNT5 activity. This combinatorial strategy may enhance the therapeutic efficacy of B3GNT5-targeted interventions while reducing the likelihood of resistance.

We acknowledge that our study used a CRISPR-edited PKO HeLa cell line, in which the gRNA targets a region near the N-terminus of B3GNT5, while the antibodies used recognize epitopes downstream (~ aa118–378). Thus, while protein depletion was confirmed by western blotting, we cannot fully exclude the presence of residual non-functional protein fragments. To address this, we employed targeted mass spectrometry, which demonstrated accumulation of upstream substrates—GlcCer and LacCer—providing orthogonal evidence of impaired enzymatic activity. We also note that extended glycosphingolipid species in the lactoside and neolactoside series were not quantified in this study, representing an area for future investigation.

Another limitation of this study is that RTK activation profiling was performed using a phospho-RTK array, which, while useful for detecting broad shifts in signaling dynamics, may not reliably quantify specific receptor activation. As such, these results are interpreted descriptively, and further validation using targeted assays would be required to confirm the involvement of individual RTKs.

Conclusion

GSLs play essential roles in many important cellular processes. However, in human cancers, there is ample evidence that these molecules are involved in the etiology of EMT-MET and MDR, which are significant barriers to overcome to achieve good outcomes in patients with cancer. In this report, it was established that the Lc3Cer-synthase B3GNT5 is highly amplified in the tumors of patients, associated with poor survival, and a modulator of EMT-MET markers, chemoresistance, MDR1 levels, and RTK activation. Data presented in this work also significantly contributed to our understanding of GSL-mediated mechanisms of chemoresistance. As MDR in cancer therapy continues to be an unresolved therapeutic target and no approved therapies have been developed, these studies, as they shed light on the role Golgi-specific sphingolipid metabolism defects on MDR, have important clinical relevance as they reveal that it is important to target specific cellular physiology pathways when developing pharmacological therapies designed to target chemoresistance.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Material 1 (1.3MB, pdf)

Acknowledgements

None.

Abbreviations

GSL

Glycosphingolipids

GlcCer

Glucosylceramide

LacCer

Lactosylceramide

B3GNT5

β-1,3-N-acetylglucosaminyltransferase 5

EMT

Epithelial-to-mesenchymal transition

MET

Mesenchymal to epithelial transition

RTK

Receptor tyrosine kinases

FBS

Fetal bovine serum

CNV

Copy number variant

B3GNT5PKO

B3GNT5 partial knockout HeLa cells

GBA

Lysosomal glucosylceramidase β1

MDR

Multi drug chemoresistance

TCGA

The Cancer Genome Atlas

Author contributions

Conceptualization, S.L; methodology, S.L., A.D.; formal analysis, S.L., L.C., K.H.R., A.D.; investigation, S.L., L.C., A.D. K.H.R.; resources, S.L.; data curation,, S.L., L.C., A.D., K.H.R.; writing—original draft preparation,, S.L., L.C., A.D.; writing—review and editing,, S.L., L.C., A.D., K.H.R.; visualization, S.L., A.D.; supervision, S.L.; project administration, S.L.; funding acquisition, S.L. All authors have read and agreed to the published version of the manuscript.

Funding

This work was supported by National Institutes of Health Grant R21CA232234 (S.L.), and L.C. was supported in part by a Cayman Biomedical Research Institute (CABRI) award. Services in support of the research project were provided by the VCU Massey Cancer Center Lipidomics and Metabolomics Shared Resource, and the Virginia Commonwealth University Flow Cytometry Shared Resource, and supported, in part, with funding from NIH-NCI Cancer Center Support Grant P30 CA016059.

Data availability

All data generated or analyzed during this study are included in this published article [and its supplementary information files].

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Laura E. Clark and Katherine Hylton Rorie have contributed equally to this work.

References

  • 1.Varki A, Cummings RD, Esko JD, Stanley P, Hart GW, Aebi M, et al. editors. Essentials of glycobiology. cold spring harbor (NY): cold spring harbor laboratory press copyright 2015–2017 by the consortium of glycobiology.Editors, La jolla, California. All rights reserved.; 2015.
  • 2.Julien S, Bobowski M, Steenackers A, Le Bourhis X, Delannoy P. How Do Gangliosides Regulate RTKs Signaling? Cells. 2013;2(4):751–67. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Kaucic K, Liu Y, Ladisch S. Modulation of growth factor signaling by gangliosides: positive or negative? Methods Enzymol. 2006;417:168–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Koundouros N, Poulogiannis G. Reprogramming of fatty acid metabolism in cancer. Br J Cancer. 2020;122(1):4–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Boyd AE, Grizzard PJ, Hylton Rorie K, Lima S. Lipidomic profiling reveals biological differences between tumors of Self-Identified African Americans and Non-Hispanic Whites with Cancer. Cancers [Internet]. 2023; 15(8). [DOI] [PMC free article] [PubMed]
  • 6.Rohrbach TD, Boyd AE, Grizzard PJ, Spiegel S, Allegood J, Lima S. A simple method for sphingolipid analysis of tissues embedded in optimal cutting temperature compound. J Lipid Res. 2020;61(6):953–67. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Boyd AE, Allegood J, Lima S. Preparation of human tissues embedded in optimal cutting temperature compound for mass spectrometry analysis. J Vis Exp. 2021;(170). [DOI] [PubMed]
  • 8.Ogretmen B. Sphingolipid metabolism in cancer signalling and therapy. Nat Rev Cancer. 2018;18(1):33–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Zhuo D, Li X, Guan F. Biological roles of aberrantly expressed glycosphingolipids and related enzymes in human Cancer development and progression. Front Physiol. 2018;9:466. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Morad SAF, Cabot MC. Chapter Nine - The onus of sphingolipid enzymes in cancer drug resistance. In: Chalfant CE, Fisher PB, editors. Advances in Cancer research. Volume 140. Academic; 2018. pp. 235–63. [DOI] [PubMed]
  • 11.Hakomori SI. Structure and function of glycosphingolipids and sphingolipids: recollections and future trends. Biochim Biophys Acta. 2008;1780(3):325–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Yoshiki J, Kubushiro K, Tsukazaki K, Udagawa Y, Nozawa S, Iwamori M. High expression of uridine diphosphate-galactose: Lc3Cer beta 1–3 galactosyltransferase in human uterine endometrial cancer-derived cells as measured by enzyme-linked immunosorbent assay and thin-layer chromatography-immunostaining. Japanese J cancer Research: Gann. 1997;88(7):669–77. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Jeong HY, Park SY, Kim HJ, Moon S, Lee S, Lee SH, et al. B3GNT5 is a novel marker correlated with stem-like phenotype and poor clinical outcome in human gliomas. CNS Neurosci Ther. 2020;26(11):1147–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Miao Z, Cao Q, Liao R, Chen X, Li X, Bai L, et al. Elevated transcription and glycosylation of B3GNT5 promotes breast cancer aggressiveness. J Experimental Clin Cancer Res. 2022;41(1):169. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Meng Q, Hu X, Zhao X, Kong X, Meng Y-M, Chen Y, et al. A circular network of coregulated sphingolipids dictates lung cancer growth and progression. EBioMedicine. 2021;66:103301. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Wang Z, Wen L, Ma X, Chen Z, Yu Y, Zhu J, et al. High expression of lactotriaosylceramide, a differentiation-associated glycosphingolipid, in the bone marrow of acute myeloid leukemia patients. Glycobiology. 2012;22(7):930–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Clark LE, Dickinson AJG, Lima S. GBA regulates EMT/MET and chemoresistance in squamous cell carcinoma cells by modulating the cellular glycosphingolipid profile. Cells. 2023;12(14). [DOI] [PMC free article] [PubMed]
  • 18.Schindelin J, Arganda-Carreras I, Frise E, Kaynig V, Longair M, Pietzsch T, et al. Fiji: an open-source platform for biological-image analysis. Nat Methods. 2012;9(7):676–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Schindelin J, Rueden CT, Hiner MC, Eliceiri KW. The ImageJ 1.53 ecosystem: an open platform for biomedical image analysis. Mol Reprod Dev. 2015;82(7–8):518–29. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Bartha Á, Győrffy B. TNMplot.com: A web tool for the comparison of gene expression in normal, tumor and metastatic tissues. Int J Mol Sci [Internet]. 2021; 22(5). [DOI] [PMC free article] [PubMed]
  • 21.Shao X, Lv N, Liao J, Long J, Xue R, Ai N, et al. Copy number variation is highly correlated with differential gene expression: a pan-cancer study. BMC Med Genet. 2019;20(1):175. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Kanehisa M, Furumichi M, Sato Y, Kawashima M, Ishiguro-Watanabe M. KEGG for taxonomy-based analysis of pathways and genomes. Nucleic Acids Res. 2023;51(D1):D587–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Uhlen M, Zhang C, Lee S, Sjöstedt E, Fagerberg L, Bidkhori G, et al. A pathology atlas of the human cancer transcriptome. Science. 2017;357:6352. [DOI] [PubMed] [Google Scholar]
  • 24.Cerami E, Gao J, Dogrusoz U, Gross BE, Sumer SO, Aksoy BA, et al. The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data. Cancer Discov. 2012;2(5):401–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Gouazé-Andersson V, Yu JY, Kreitenberg AJ, Bielawska A, Giuliano AE, Cabot MC. Ceramide and glucosylceramide upregulate expression of the multidrug resistance gene MDR1 in cancer cells. Biochimica et biophysica acta (BBA) - Molecular and cell biology. Lipids. 2007;1771(12):1407–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Todeschini AR, Dos Santos JN, Handa K, Hakomori SI. Ganglioside GM2-tetraspanin CD82 complex inhibits Met and its cross-talk with integrins, providing a basis for control of cell motility through glycosynapse. J Biol Chem. 2007;282(11):8123–33. [DOI] [PubMed] [Google Scholar]
  • 27.Kundu M, Mahata B, Banerjee A, Chakraborty S, Debnath S, Ray SS, et al. Ganglioside GM2 mediates migration of tumor cells by interacting with integrin and modulating the downstream signaling pathway. Biochim Biophys Acta. 2016;1863(7 Pt A):1472–89. [DOI] [PubMed] [Google Scholar]
  • 28.Ishihara S, Aoki K, Mizutani T, Amano M, Nishimura SI, Haga H. Glycosphingolipid GM2 induces invasiveness in Irradiation-tolerant lung Cancer cells. Cell Struct Funct. 2018;43(2):177–85. [DOI] [PubMed] [Google Scholar]
  • 29.Kuburich NA, den Hollander P, Pietz JT, Mani SA. Vimentin and cytokeratin: good alone, bad together. Sem Cancer Biol. 2022;86:816–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Yang J, Antin P, Berx G, Blanpain C, Brabletz T, Bronner M, et al. Guidelines and definitions for research on epithelial–mesenchymal transition. Nat Rev Mol Cell Biol. 2020;21(6):341–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Yu F, Yu C, Li F, Zuo Y, Wang Y, Yao L, et al. Wnt/β-catenin signaling in cancers and targeted therapies. Signal Transduct Target Therapy. 2021;6(1):307. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Loh CY, Chai JY, Tang TF, Wong WF, Sethi G, Shanmugam MK et al. The E-Cadherin and N-Cadherin switch in Epithelial-to-Mesenchymal transition: signaling, therapeutic implications, and challenges. Cells. 2019;8(10). [DOI] [PMC free article] [PubMed]
  • 33.Holmes EH, Hakomori S, Ostrander GK. Synthesis of type 1 and 2 lacto series glycolipid antigens in human colonic adenocarcinoma and derived cell lines is due to activation of a normally unexpressed beta 1----3 N-acetylglucosaminyltransferase. J Biol Chem. 1987;262(32):15649–58. [PubMed] [Google Scholar]
  • 34.Takehara K, Kubushiro K, Kiguchi K, Ishiwata I, Tsukazaki K, Nozawa S, et al. Expression of glycolipids bearing Lewis phenotypes in tissues and cultured cells of human gynecological cancers. Japanese J cancer Research: Gann. 2002;93(10):1129–37. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Gouaze-Andersson V, Cabot MC. Glycosphingolipids and drug resistance. Biochimica et biophysica acta (BBA). - Biomembr. 2006;1758(12):2096–103. [DOI] [PubMed] [Google Scholar]
  • 36.Liu Y-Y, Gupta V, Patwardhan GA, Bhinge K, Zhao Y, Bao J, et al. Glucosylceramide synthase upregulates MDR1 expression in the regulation of cancer drug resistance through cSrc and β-catenin signaling. Mol Cancer. 2010;9(1):145. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Eckford PD, Sharom FJ. The reconstituted P-glycoprotein multidrug transporter is a flippase for glucosylceramide and other simple glycosphingolipids. Biochem J. 2005;389(Pt 2):517–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Shabbits JA, Mayer LD. P-Glycoprotein modulates Ceramide-mediated sensitivity of human breast Cancer cells to Tubulin-binding anticancer Drugs1. Mol Cancer Ther. 2002;1(3):205–13. [PubMed] [Google Scholar]
  • 39.Lala P, Ito S, Lingwood CA. Retroviral transfection of Madin-Darby canine kidney cells with human MDR1 results in a major increase in globotriaosylceramide and 105- to 106-Fold increased cell sensitivity to verocytotoxin: ROLE OF P-GLYCOPROTEIN IN GLYCOLIPID SYNTHESIS*. J Biol Chem. 2000;275(9):6246–51. [DOI] [PubMed] [Google Scholar]
  • 40.De Rosa MF, Sillence D, Ackerley C, Lingwood C. Role of multiple drug resistance protein 1 in neutral but not acidic glycosphingolipid biosynthesis**. J Biol Chem. 2004;279(9):7867–76. [DOI] [PubMed] [Google Scholar]
  • 41.Morjani H, Aouali N, Belhoussine R, Veldman RJ, Levade T, Manfait M. Elevation of glucosylceramide in multidrug-resistant cancer cells and accumulation in cytoplasmic droplets. Int J Cancer. 2001;94(2):157–65. [DOI] [PubMed] [Google Scholar]
  • 42.Guan F, Handa K, Hakomori S-i. Specific glycosphingolipids mediate epithelial-to-mesenchymal transition of human and mouse epithelial cell lines. Proc Natl Acad Sci. 2009;106(18):7461–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Zhang J, van der Zon G, Ma J, Mei H, Cabukusta B, Agaser CC, et al. ST3GAL5-catalyzed gangliosides inhibit TGF-β-induced epithelial-mesenchymal transition via TβRI degradation. Embo J. 2023;42(2):e110553. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Zhuo D, Guan F. Ganglioside GM1 promotes contact Inhibition of growth by regulating the localization of epidermal growth factor receptor from glycosphingolipid-enriched microdomain to caveolae. Cell Prolif. 2019;52(4):e12639. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Miljan EA, Bremer EG. Regulation of growth factor receptors by gangliosides. Sci STKE. 2002;2002(160):re15. [DOI] [PubMed] [Google Scholar]
  • 46.Coskun Ü, Grzybek M, Drechsel D, Simons K. Regulation of human EGF receptor by lipids. Proc Natl Acad Sci. 2011;108(22):9044–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Thiery JP, Sleeman JP. Complex networks orchestrate epithelial-mesenchymal transitions. Nat Rev Mol Cell Biol. 2006;7(2):131–42. [DOI] [PubMed] [Google Scholar]
  • 48.Cano A, Pérez-Moreno MA, Rodrigo I, Locascio A, Blanco MJ, del Barrio MG, et al. The transcription factor snail controls epithelial–mesenchymal transitions by repressing E-cadherin expression. Nat Cell Biol. 2000;2(2):76–83. [DOI] [PubMed] [Google Scholar]
  • 49.Jolly MK, Boareto M, Huang B, Jia D, Lu M, Ben-Jacob E, et al. Implications of the hybrid epithelial/mesenchymal phenotype in metastasis. Front Oncol. 2015;5:155. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Material 1 (1.3MB, pdf)

Data Availability Statement

All data generated or analyzed during this study are included in this published article [and its supplementary information files].

Articles from BMC Cancer are provided here courtesy of BMC

RESOURCES