This article is mentioned in:

Introduction

Breast cancer is the most common malignancy in women and remains a major global challenge. Approximately 1.4 million cases occur per year, worldwide (1,2). Surgery, chemotherapy and targeted therapy are the standard treatments for breast cancer. Numerous studies have reported that changes in cell metabolism were associated with transformation and tumor progression, which are associated with electron transport (3,4). In cancer metabolism, electron acceptor deficiency affects NAD+ regeneration from NADH, which indicates that the chemical disposal of excess electrons to synthesize nucleotides could limit proliferation (5). In addition, electron donor and electron acceptor deficiencies lead to an increase in reactive oxygen species (ROS) by affecting ROS detoxification and mitochondrial electron transport chain function, which is consistent with tumor growth (6).

Glyoxal (GO) is the smallest dialdehyde formed in the oxidation-reduction reaction and is associated with electron transfer and metabolism (7,8). A previous study has indicated that GO levels increase in patients with diabetes and reacts with several proteins to form advanced glycation end products via a Maillard-like reaction (9); however, few studies have focused on cancer. In addition, GO has been shown to reduce keratinocyte migration, downregulation of SNAI2 and inhibition of EGF-dependent proliferation (10). Recently, bis(thiosemicarbazones), derived from 1,2-diones, such as GO, have been recognized as potential therapeutic agents against cancer, by suppressing the cell cycle in the human neuroblastoma cell line BE(2)-M17(11). Meanwhile, in a separate study, the Cu(GTSC) complex, derived from GO, GO-bis(4-methyl-4-phenyl-3-thiosemicarbazonato) copper(II), inhibited tumor growth by 95.0±3.9% in the HCT116 xenograft mouse model (12). Thus, GO could be an important agent for disrupting tumor cell metabolism and a potential anti-cancer target. However, similar to arsenic trioxide, GO has been associated with strong intracellular, digestive and respiratory toxicity, that lead to usage limitations (13,14). Once the balance between the safety and the effects of GO has been achieved, GO might become an effective agent against breast cancer by disrupting cancer metabolism.

The present study aimed to investigate the biofunctions of GO and investigate its molecular mechanisms in breast cancer progression.

Materials and methods

Cell lines and culture

The MDA-MB-231, SUM149 and SUM159 cell lines were used as representative triple-negative breast cancer cell models, while the EMT6 and MCF-7 cell lines were used as estrogen receptor-positive breast cancer cell models, and the MCF-10A cell line was used to represent a normal breast cell line. All the cells were purchased from the Shanghai Institutes for Biological Sciences and cultured at 37˚C in a humidified incubator with 5% CO2 in DMEM (HyClone; Cytiva) supplemented with 10% FBS (Natocor) and antibiotics (Sigma-Aldrich; Merck KGaA) (15).

GO preparations

GO (Sinopharm Chemical Reagent Co., Ltd.) is an organic compound with the chemical formula OCHCHO, and the Chemical Abstract Service number, 107-22-2. GO was filtered using a 0.22 µm filter (Millipore Sigma) and stored at 26˚C in the dark.

MTT assay

Cell viability was performed using the MTT assay (Sigma-Aldrich; Merck KGaA). The MDA-MB-231, SUM149, SUM159, EMT6, MCF-7 and MCF-10A cell lines (3x103 cells/well) were seeded in 96-well plates and treated with different concentrations of GO (0.129, 0.258, 0.516, 1.032, 2.065, 4.13 and 8.25 mmol/l for MDA-MB-231, SUM149 and SUM159; and 1.032, 2.065 and 4.13 mmol/l for EMT6, MCF-7 and MCF-10A). After 24 h, 20 µl 5 mg/ml MTT solution was added to each well and the plates were further incubated at 37˚C for 4 h. Following which, the medium was aspirated, and 200 µl dimethyl sulfoxide was added to each well. After the purple formazan crystals had dissolved, the absorbance was determined at 492 nm using an INFINITE F50 microplate reader (Tecan Group, Ltd.). According to the MTT data, IC50 values were computed by GraphPad Prism (v8.0; GraphPad Prism Software, Inc.). Meanwhile, the IC50 for GO was used for 24 h, and the morphology was captured by Nikon TS100 microscope (Nikon Corporation). The results were obtained from three independent experiments.

Crystal violet assay

The MDA-MB-231, SUM149, SUM159, EMT6, MCF-7 and MCF-10A cell lines were seeded in 24-well plates, at a density of 1x103 cells/well and incubated for 24 h. The cells were then treated with different concentrations of GO (0.516, 1.032, 2.065, 4.13 and 8.25 mmol/l) continuously for 5 days. After fixation with 4% paraformaldehyde for 30 min at 26˚C, the cells were stained with crystal violet solution for 2 h at 26˚C. The results were obtained from three independent experiments.

Cell migration assay

Cell migration ability was performed using a wound-healing assay. Approximately 2x104 cells were seeded into each well of a 6-well plate without serum and a light microscope was used the following day to confirm that each well was coated with cells at ~90% confluence. A 1.0 ml pipette tip was used to remove the cells in the wound-healing region and the plates were washed with PBS three times to remove the displaced cells. The cells were treated with different concentrations of GO (3.550/4.130 mmol/l for MDA-MB-231 and SUM149; 1.437/2.065 mmol/l for SUM159; 1.85/4.13 mmol/l for EMT6; and 1.032/4.13 mmol/l for MCF-7 and MCF-10A). The control cells were incubated with serum-free medium at 37˚C with 5% CO2. At 0, 24 and 48 h, images were captured using a Nikon TS100 microscope (Nikon Corporation), at x40 magnification and the wound was measured using ImageJ software (v1.52a; National Institutes of Health). The experiment was repeated three times.

3D laser scanning confocal microscope and transmission electron microscope (TEM)

The MDA-MB-231, SUM149, SUM159 and EMT6 cell lines were treated with different concentrations of GO (3.550, 3.550, 1.437 and 2.22 mmol/l, respectively) for 24 h. For TEM, a total of 1x107 cells were pelleted by centrifugation at 2,683 x g for 5 min at 26˚C, then washed three times with PBS. The cells were then fixed in 2.5% glutaraldehyde at 4˚C for 24 h. Next, the cells were washed with PBS three times and post-fixed in 1% osmium tetroxide for 60 min at 4˚C, encapsulated in 1% agar and stained with uranyl acetate and phosphotungstic acid for 60 min at 4˚C. The cells were then dehydrated in a graded ethanol series and subsequently incubated in propylene oxide for 35 min at 26˚C. The TEM images were captured using a Hitachi TEM system (Hitachi High-Technologies Corporation).

For 3D micro-morphology, the volume and height of the SUM149, MDA-MB-231, SUM159 and EMT6 cell lines were measured using a VK-V150 laser microscopy system (Keyence Corporation). Phase-contrast observations of the cells were performed using an Olympus IX71 microscope (Olympus Corporation). The results were obtained from three independent experiments.

Western blot analysis

The protein expression level of ERK, phosphorylated (p)-ERK, MEK, p-MEK, AKT, p-AKT-Ser473, p-AKT-Thr308, mTOR and p70-S6k was measured using western blot analysis in the MDA-MB-231, SUM149 and SUM159 cell lines, which were each treated with the IC50 of GO. All cells were lysed in RIPA lysis buffer (cat. no. P0013B; Beyotime Institute of Biotechnology) and then centrifuged at 15,702 x g for 15 min at 4˚C. Protein concentrations were determined using a BCA kit (Beyotime Institute of Biotechnology). A total of 20 µg protein was separated on 6-10% gels using SDS-PAGE and transferred to PVDF membranes (MilliporeSigma). The membranes were blocked for 1 h at 26˚C with 5% bovine serum albumin containing 0.1% Tween-20. Immunoblotting was performed using the following primary antibodies: ERK (cat. no. 13-6200, 1:1,000), p-ERK (cat. no. 44-680G; 1:500), MEK (cat. no. PA5-116802; 1:500), p-MEK (cat. no. 44-452, 1:1,000), AKT (cat. no. MA191204; 1:1,000), mTOR (cat. no. A301-144A-T; 1:1,000), p70-S6k (cat. no. MA5-36267; 1:1,000) (all Invitrogen; Thermo Fisher Scientific, Inc.) and Tubulin (cat. no. AF1216; 1:1,000; Beyotime Institute of Biotechnology) overnight at 4˚C. The membranes were then washed with 1% TBS-Tween-20 three times and incubated with the corresponding secondary antibodies (cat. no. A0208; goat anti-rabbit; 1:5,000; Beyotime Institute of Biotechnology) at 37˚C for 2 h. The membranes were washed again with TBS, and the proteins were visualized using an enhanced chemiluminescence assay kit (Beyotime Institute of Biotechnology). Images were captured using a Bio-Rad Chemodoc XRS+ system and the Image-lab software (Version 6.0; Bio-Rad Laboratories, Inc.). The test was repeated three times.

Cell cycle analysis using flow cytometry

The MDA-MB-231, SUM149 and SUM159 cell lines (5x104 cells/well) were seeded in 6-well plates and treated with GO (IC50: 3.78, 1.85 and 1.60 mmol/l, respectively) for 24 h. Next, the cells were collected and stored in pre-cooled alcohol overnight at 4˚C, then stained with PI (Shanghai Yeasen Biotechnology, Co., Ltd.) for 15 min in the dark at 4˚C. The samples were tested using a Guava EasyCyte Plus flow cytometer (Merck KGaA) and FlowJo VX (Becton-Dickinson and Company) was used to analyze the results. The test was repeated three times.

Cell apoptosis analysis using flow cytometry

The MDA-MB-231, SUM149 and SUM159 cell lines (5x104 cells/well) were seeded in 6-well plates and treated with GO (IC50: 3.78, 1.85 and 1.60 mmol/l, respectively) for 24 h. The cells were then collected and stained using the Annexin V/PI kit (Shanghai Yeasen Biotechnology, Co., Ltd.) for 15 min in the dark at 4˚C. The samples were tested using a Guava EasyCyte Plus flow cytometer (Merck KGaA) and FlowJo VX (Becton, Dickinson and Company) was used to analyze the results. The test was repeated three times.

Statistical analysis

The data were expressed as the mean ± SD, and unpaired t-tests or repeated measures one-way ANOVA followed by Dunnett's multiple comparisons test were performed for statistical analyses using GraphPad Prism (v8.0; GraphPad Prism Software, Inc.). P<0.05 was considered to indicate a statistically significant difference.

Results

GO inhibits breast cancer cell proliferation

MTT and crystal violet assays were used to investigate the biofunctions of GO on cancer cell proliferation in different breast cancer cell lines. The results demonstrated that cell proliferation was inhibited in a concentration-dependent manner. As the GO concentration increased, a greater inhibitory effect was exerted in the breast cancer cell lines. Inhibition rates were up to 78.95±0.05, 88.83±1.35, 87.49±1.11, 88.98±9.90 and 71.77±1.29% for the MDA-MB-231, SUM149, SUM159, EMT6, and MCF-7, respectively (P<0.05 compared with cells without GO). However, GO only slightly reduced the proliferation rate in the MCF-10A normal breast cell lines, as the inhibition rate was <38.26%. The IC50 values of the MDA-MB-231, SUM149, SUM159, EMT6, MCF-7 and MCF-10A cell lines were 3.78, 1.85, 1.60, 1.29, 2.22, and 4.39 mmol/l, respectively (Fig. 1A). In addition, cellular morphology indicated cell death after treatment with GO for 24 h (Fig. 1B). Furthermore, under the above general tendency, the crystal violet assay showed that cell proliferation was notably suppressed (Fig. 2A).

Figure 1 Inhibition ability at different doses of GO was analyzed using the MTT and crystal violet assays in the MDA-MB-231, SUM149, SUM159, EMT6, MCF-7 and MCF-10A cell lines. (A) Under the above general tendency, the results from the MTT assay showed that GO inhibited MDA-MB-231, SUM149, SUM159, EMT6 and MCF-7 cell proliferation in a concentration-dependent manner. GO at 4.13 mmol only notably inhibited the MCF-10A normal breast cell lines. (B) The MDA-MB-231, SUM149, SUM159, EMT6 and MCF-7 cell lines were treated with GO for 24 h and changes in cell morphology and cell death were observed. However, the cell morphology of the MCF-10A cell line was almost normal. GO, glyoxal.

Figure 2 (A) GO notably reduced the colony numbers in the MDA-MB-231, SUM159, MCF-7, MCF-10A and EMT6 cell lines following treatment with different concentrations. (B) The result of Western blot assay. a, TNBC cell lines were treated with GO. The protein expression level of AKT1 decreased in the SUM149 and SUM159 groups, and the expression level of mTOR and P70-S6K proteins decreased in all three cell lines. Meanwhile, p-ERK protein expression increased in all three cell lines, and p-MEK protein expression increased in MDA-MB-231 and SUM149 cell lines. b, Analysis of Western blot bands in each cell line by Image J software (National Institutes of Health, USA). GO, glyoxal; CTR, control; p, phosphorylated.

To further elucidate the mechanisms underlying the action of GO, the protein expression level of the downstream kinases of the MAPK and AKT/mTOR pathways were investigated using western blot analysis. Consistent with the aforementioned results, GO was found to be involved in the regulation of the MEK-ERK and AKT/mTOR pathways. The results indicated that the protein expression level of AKT1 was suppressed in SUM149 and SUM159 group, and the expression level of mTOR and P70-S6K proteins was suppressed in MDA-MB-231, SUM149, and SUM159 cells (Fig. 2B). By contrast, GO also increased p-ERK protein expression in the same cell lines, and p-MEK protein expression increased in MDA-MB-231 and SUM149 cell lines (Fig. 2B). In summary, the results indicated that GO suppressed breast cancer cell proliferation by acting on the MAPK and AKT/mTOR pathways.

GO inhibits breast cancer cell migration

To investigate the effect of GO on breast cancer and normal breast cell migration, differences in the wound healing rate in the MDA-MB-231, SUM149, SUM159, EMT6, MCF-7 and MCF-10A cell lines 48 h following GO treatment were observed. The results showed that the relative scratch width of the GO group was significantly wider compared with that in the control group at the 24 and 48 h time points. In the MDA-MB-231 group, the migration inhibition rates were 84.18 and 86.32% at 3.550 and 4.130 mmol/l, respectively. In the SUM149 group, the migration inhibition rates were 81.19 and 82.94% at 3.550 and 4.130 mmol/l, respectively. In the SUM159 group, the migration inhibition rates were 67.65 and 80.00% at 1.437 and 2.065 mmol/l, respectively. In the EMT6 group, the migration inhibition rate was 52.86 and 81.88% at 1.85 and 4.13 mmol/l, respectively. In the MCF-7 cells, the migration inhibition rate was 77.33 and 80.42% at 1.032 and 4.13 mmol/l, respectively. In the MCF-10A normal breast cells, the migration inhibition rate was 44.20 and 44.92% at 1.032 and 4.13 mmol/l, respectively. Under the above general tendency, the data indicated that GO suppressed cell migration in a concentration-dependent manner at both 24 and 48 h (P<0.05) (Fig. 3).

Figure 3 Wound healing assay using the MDA-MB-231, SUM149, SUM159, EMT6, MCF-7 and normal MCF-10A breast cell lines. The migration distance in the glyoxal groups was significantly shorter compared with that in the control group. Images were captured at 0, 24 and 48 h after scratching. *P<0.05, **P<0.01, ***P<0.001. Magnification, x40.

GO induces cellular ultrastructure and changes morphology

The cellular ultrastructure was observed using TEM. A previous study has indicated that typical morphological features of apoptosis include chromatin condensation, nuclear fragmentation and the disappearance of surface microvilli (16). As shown in Fig. 4, GO treatment for 24 h notably altered the ultrastructure of the mitochondria, nucleus and microvilli. Mitochondria appeared as enlarged organelles. The cellular morphology became irregular and cytoplasmic vacuolization was observed. In addition, rich microvilli, which were around the cells, almost disappeared in the GO group, which is consistent with cell migration inhibition. Chromatin also dissociated and appeared around the edge of the nucleus after treatment with GO. However, the cell membranes remained intact, and chromatin condensation and nuclear fragmentation were not observed in either group.

Figure 4 TEM analysis in the SUM149, SUM159 and MDA-MB-231 cell lines. Compared with that in the control group, glyoxal induced cellular surface microvillus disappearance (blue arrow), but the cell membrane remained intact (yellow arrow). Magnification, x5,000 and x20,000.

To improve the understanding into the changes in cellular morphology following GO treatment, a 3D laser scanning confocal microscope was used, which is a valuable tool for obtaining high-resolution images and 3D reconstructions. Treatment with GO for 24 h notably altered cellular morphology, height and volume (Fig. 5A). Compared with that in the control group, the cellular height of the SUM149, MDA-MB-231, SUM159 and EMT6 cell lines decreased 67.22 (4.69±0.23 vs. 1.54±0.29 µm; P<0.001), 74.22 (4.73±0.86 vs. 1.22±0.15 µm; P<0.001), 21.78 (2.96±0.41 vs. 2.32±0.49 µm; P <0.05) and 33.43% (1.98±0.48 vs. 1.32±0.12 µm; P<0.05) (Fig. 5B). In addition, the cellular volume of the SUM149, MDA-MB-231, SUM159 and EMT6 cell lines decreased by 40.29 (797.28±66.43 vs. 476.06±65.02 µm3; P<0.05), 69.90 (991.09±305.26 vs. 298.30±13.42 µm3; P<0.001), 72.83 (982.57±218.70 vs. 266.92±54.35 µm3; P<0.001) and 63.29% (1,028.85±203.27 vs. 377.96±15.55 µm3; P<0.001) (Fig 5C), respectively. As a result, the cellular morphology was altered, becoming more circular and flatter, and the microvilli on the cell surface also disappeared, which suggested that the cell skeleton was affected and cell apoptosis occurred.

Figure 5 (A) The SUM149, SUM159, MDA-MB-231 and EMT6 cell lines were analyzed using 3D laser scanning confocal microscope. The cell morphology was changed to circular and flat structure following glyoxal treatment. Magnification, x400. Cellular (B) height and (C) volume decreased. *P<0.05, ***P<0.001.

GO induces cell apoptosis and arrests the cell cycle

Next, the effects of GO on cell apoptosis and the cell cycle were investigated. As shown in Fig. 6, the proportion of Annexin V (+)/PI (+) apoptotic cells increased significantly following treatment with GO. The percentage of AnnexinV (+)/PI (+) cells in the GO group was 86.97±0.89, 31.8±0.28 and 31.15±3.61% compared with that in the control 5.43±1.57, 15.6±7.21 and 14.65±2.76% for the MDA-MB-231 (P<0.001), SUM149 (P<0.05) and SUM159 (P<0.05) cell lines, respectively (Fig. 6A). In addition, GO arrested the cell cycle. A higher percentage of cells in the G1 phase was accompanied by a decrease in the proportion of cells in the G2/M phase following treatment with GO for 24 h (7.01±1.73 vs. 23.82±2.24% for MDA-MB-231; P=0.014; 10.75±2.33 vs. 83.50±4.10% for SUM149; P=0.002; 42.60±1.41 vs. 67.40±5.94% for SUM159; P=0.029). Thus, these results demonstrated that GO inhibited cell proliferation by arresting the cell cycle in the G2 phase (Fig. 6B).

Figure 6 Cell apoptosis and cell cycle was analyzed using flow cytometry. (Aa) Glyoxal increased the positive rate of Annexin V+/PI+ cells in the MDA-MB-231, SUM149 and SUM159 cell lines and the results were (Ab) statistically analyzed. (Ba) G2/M phase was arrested in the glyoxal group compared with that in the control group and the results were (Bb) statistically analyzed. *P<0.05, **P<0.01, ***P<0.001.

Discussion

The morbidity rate of breast cancer has surpassed that of lung cancer so that breast cancer is now the most malignant tumor in the world. Thus, besides chemotherapy, targeted treatment and endocrine treatment, more treatments are required for breast cancer. Cytotoxic agents are no longer the only potential novel anticancer drugs. The ‘Warburg effect’ suggests that even under oxygen-sufficient conditions, tumor cells still take advantage of glycolysis metabolism, using oxidative phosphorylation, which is associated with the respiratory chain rather than producing ATP (17). Therefore, changes in cancer tissue metabolite levels can have important implications for cell physiology (18), and the majority of previous studies have focused on the role of mitochondria in the regulation of cell proliferation and apoptosis (19-21). During tumor metabolism processes, cytochrome c oxidase and succinate dehydrogenase are increased, and the activity of glucose transporters is enhanced. Furthermore, mitochondrial aerobic oxidation is inhibited, disrupting the tricarboxylic acid cycle (22). Zhou et al (23) indicated that mitochondrial function plays an important role in the ‘bystander effect’ mediated by radiation therapy, which relies on the NF-κB/iNOS/COX-2/prostaglandin E2 pathways of mitochondria. In addition, the cell energy metabolism level can be reduced by avicins, which are triterpene compounds that act on the outer mitochondrial membrane by closing voltage-dependent negative ion channels (24).

In addition, tumor cell proliferation is significantly inhibited by dichloroacetic acid, which inhibits mitochondrial pyruvate dehydrogenase kinase and activates potassium channels in all cancer cells, thereby inducing apoptosis (25). In addition, drugs that suppress microvillus development and arrest the cell cycle at the G0/G1, S or G2 phases, prevent cancer proliferation (19,26-30). In addition, King et al (31) summarized and discussed the fact that the activities of succinate dehydrogenase (SDH) and fumaric acid hydratase were associated with mitochondrial dysfunction in malignant tumor cells. Notably, SDH is an important component of the tricarboxylic acid cycle, that plays a key role in the transition of ubiquinone to ubiquinol in the mitochondrial electron transport chain (32). Thus, changes in the electron transfer chain can alter the biological behavior of tumors and interrupting electron transportation in cell metabolism contributes to tumor progression, which is a potential therapeutic target.

Furthermore, metabolic reprogramming is an important hallmark of cancer cell proliferation (33). The preferential use of glycolysis unavoidably generates methylglyoxal (MGO), which is associated with the glycolysis reaction via the spontaneous dephosphorylation of GAP and dihydroxyacetone phosphate (34). A dual role has been previously demonstrated for MGO, which is favorable to neuron viability and excitability at low levels, while high levels are cytotoxic (33). At high concentrations, MGO suppressed breast cancer cell proliferation via glycolysis, but low doses of MGO promoted tumorigenesis even in the same tissue (33,35,36). Accordingly, the expression balance of MGO is more critical than its expression level. By contrast, GO was associated with glycolysis reactions; however, it is the smallest dialdehyde and contains two adjacent reactive carbonyl groups, which form during the oxidation-reduction reaction. These are referred to as reactive electrophilic species and they are more effective than MGO (13,37). Furthermore, GO-induced cytotoxicity and protein carbonylation were more severe than for MGO (37); therefore, GO concentration must be finely adjusted and maintained within a safe range in glycolytic cancer cells. Therefore, GO was selected as the focus of the present study, even though GO can cause mitochondrial toxicity (14,38). In addition, GO plays important roles in cellular responses to energy metabolism and T-cell activation by regulating cell signaling pathways (39,40). In addition, GO impairs the electron transport chain, mitochondrial function and energy metabolism. For example, GO leads to advanced lipoxidation and glycation end products, which are associated with aging and age-related chronic diseases via mitochondrial dysfunction (41).

Therefore, the present study aimed to further investigate the effects and mechanisms of GO stress on breast cancer cells. The results of functional analyses showed that GO treatment notably induced a decrease in cell proliferation, increased cell apoptosis, arrested the cell cycle in G2 phase and altered cellular ultrastructure. The western blot results indicated that GO treatment was involved in the MAPK and mTOR signaling pathways in breast cancer. Taken together, these findings suggest that GO is a potential anti-cancer therapeutic agent that inhibits breast cancer progression by regulating the MAPK and AKT/mTOR signaling pathways. Notably, measuring the increase in cellular levels of p-ERK has been shown to be an indirect indicator of the increase in bioavailable copper that causes cell apoptosis (42,43).

In conclusion, several compounds affect breast cancer cell lines via glycolysis, such as MGO and GO. As GO is the smallest dialdehyde and contains two adjacent reactive carbonyl groups, GO can cause increased cytotoxicity, and protein carbonylation than MGO. However, GO causes severe side-effects and toxicity than MGO; therefore, the concentration of GO must be strictly adjusted to be kept in a safe range. The results from the present study showed that GO could be a potential therapeutic agent for breast cancer; however, additional research is required to gain a more in-depth understanding of its mechanisms.

Acknowledgements

Not applicable.

Funding

Funding: No funding received.

Availability of data and materials

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Authors' contributions

PR and LY designed the study. GN, LY and LQ contributed to the cell culture and experiments. PR and LX analyzed the data. LX and LY wrote the original paper. PR, LY, GN, LQ and LX confirm the authenticity of all the raw data. All authors have read and approved the manuscript.

Ethics approval and consent to participate

Not applicable.

Patient consent for participation

Not applicable.

Competing interests

The authors declare that they have no competing interests.

References