Screening and validation of nickel ion cytotoxicity biomarkers based on transcriptomic and proteomic technology

Abstract The aim of this study was to screen cytotoxicity biomarkers of nickel ions (Ni2+) using transcriptomic and proteomic approaches combined with molecular biology validation. First, the MTT method was used to evaluate cytotoxicity in L929 cells treated with Ni2+ at different concentrations. Ni2+ at both 100 μM and 200 μM affected cell proliferation. Then, transcriptomic and proteomic technology was used to study the effects of Ni2+ on the expression of genes/proteins in cells. It was found that 1490, 789, 652 and 729 genes (12, 24, 48 and 72 h, respectively) and 177, 2191 and 2095 proteins (12, 24 and 48 h, respectively) were differentially expressed after treatment with 100 μM Ni2+. In total, 1403, 963, 916 and 1230 genes (12, 24, 48 and 72 h, respectively) and 83, 1681 and 2398 proteins (12, 24 and 48 h, respectively) were differentially expressed after treatment with 200 μM Ni2+. Then, four target gene/protein biomarkers were filtered by combined screening using gene/proteomic experimental data and biological pathway analyses. Further expression level validation of all these target biomarkers and functional validation of selected gene/protein biomarkers were carried out, and a final gene/protein biomarker (UQCRB) was identified.


Introduction
Biomarkers are indicators that can be used to objectively monitor and evaluate normal biological processes, pathological processes and drug intervention reactions [1]. Biomarkers are also important early indicators that warn of damage to an organism, including changes in the cell molecular structure and function; changes in biochemical metabolic processes; the abnormal performance of physiological activities; and abnormal changes in individuals, groups or the entire ecosystem [2]. With the advantages of fast, simple and rapid detection, biomarkers have been applied in medical diagnosis and clinical research, drug screening and evaluation, environmental assessment (ecological risk assessment) and toxicity assessment of biomaterials.
Because of the shape memory effect, nickel and titanium alloys are often used in artificial bones, orthodontic wires, etc. However, nickel-titanium alloys have a high nickel content; thus, after implantation into the human body, nickel ions (Ni 2þ ) are released, which leads to certain toxicity. Therefore, it is necessary to study biomarkers that can be used to assess the cytotoxicity of Ni 2þ early. Clemens et al. studied the expression of the Ect2 protooncogene and protein in mouse fibroblasts after treatment with Ni 2þ for 48 h by RT-PCR and northern blot analysis. The results showed that Ect2 can be used as an effective biomarker to monitor changes in cell morphology or tumorigenesis induced by Ni 2þ or nickel compounds [3]. Kwon et al. identified 20 genes (ACTB, AFAP1, etc.) and 16 proteins (UTS2, ACTB, etc.) as potential biomarkers that can be used to test nickel toxicity by a gene expression microarray, functional proteomics and pathway analysis [4].
The above-mentioned studies investigating Ni 2þ toxicity biomarkers were carried out by either traditional molecular biology methods or a single biomics method, and joint analyses based on multiomics methods are lacking; this has led to the problem that biomarkers have been inconsistently screened at different levels. Furthermore, the above studies only describe bioinformatics analyses of the experimental results of biomics ('dry method') and reveal the theoretical functions of biomarkers and do not report a series of verification experiments ('wet method') to validate the screened biomarkers. Therefore, biomarker investigation is still at the 'on-paper' stage.
Regarding the issue above, the aim of this study was to use transcriptomic and proteomic technologies combined with bioinformatics analyses to screen cytotoxicity biomarkers of Ni 2þ to avoid the inconsistency resulting from the screening of biomarkers using different biomics platforms. Then, a series of molecular biology experiments were carried out to verify the expression levels and function levels of the screened biomarkers to ensure the reliability of the biomics experiments and bioinformatics analysis. First, transcriptomic and proteomic methods were used to study the effects of Ni 2þ on the expression of genes/ proteins in L929 cells. Then, the transcriptomic/proteomic data were jointly screened, and genes/proteins that were differentially expressed in multiple experimental groups with expression patterns that were basically consistent at the mRNA and protein levels were selected as candidate gene/protein biomarkers ('dry method'). Furthermore, a biological pathway analysis of these candidate biomarkers was performed to identify cytotoxicityrelated pathways enriched in the target gene/protein biomarkers. Finally, qRT-PCR and western blot analysis were used to verify the expression levels of the target gene/protein biomarkers. The functions of the target gene/protein biomarkers were verified by assays to determine the cell proliferation rate, oxidative stress and the mitochondrial membrane potential (MMP) ('wet method'), and the final gene/protein biomarkers of Ni 2þ cytotoxicity were determined.
Cell culture L929 mouse fibroblast cells were purchased from the Shanghai Cell Bank of the Chinese Academy of Sciences and cultured in complete RPMI-1640 medium at 37 C under 5% CO 2 and saturated humidity. All experiments were performed with cells in the logarithmic growth phase.
MTT assay to determine the cell cytotoxicity of Ni 21 Nickel chloride hexahydrate powder (2.376 g) was weighed and dissolved in 100 ml of complete RPMI-1640 medium to obtain 0.1 M Ni 2þ . Then, 0.1 ml of the stock solution was diluted 1000-fold and 500-fold with complete RPMI-1640 medium to obtain 100 lM and 200 lM Ni 2þ , respectively.
The in vitro cytotoxicity of 100 lM and 200 lM Ni 2þ in L929 cells after 12, 24, 48 and 72 h of treatment was examined using a MTT assay as described in a previous study [5]. Cells that were not treated with Ni 2þ served as a negative control, and cells that were treated with 0.7% acrylamide in complete RPMI-1640 medium served as a positive control. The sample size was six.

Biomics experiments
Transcriptomics experiments L929 cells were cultured in a cell culture flask with a bottom surface area of 75 cm 2 . The number of seeded cells was 4 Â 10 6 , 3 Â 10 6 , 1.5 Â 10 6 or 10 6 in the experimental groups (cells treated with 100 lM or 200 lM Ni 2þ for 12, 24, 48 or 72 h) and 10 6 in the control group (cells not treated with Ni 2þ and cultured for 72 h). After 24 h, the culture medium in the experimental groups was replaced with medium containing 100 lM or 200 lM Ni 2þ , and the cells were cultured for another 12, 24, 48 or 72 h.
After culture for a given duration, the cells were lysed, total RNA was extracted and transcriptome sequencing experiments were performed by Shanghai Omicsspace Biotech Co., Ltd. (Shanghai, China). The experimental process is briefly described as follows: mRNA purification and reverse transcription were performed, and then, the constructed library was used for quality inspection. Then, sequencing experiments were carried out, and the gene expression in each group was analyzed. Genes with a log 2 fold change (log 2 FC) > 1 were defined as upregulated, and those with a log 2 FC < -1 were defined as downregulated. The sample size was three.

Proteomics experiments
iTRAQ-based proteomics technology was used to study the protein expression profile in L929 cells after treatment with 100 lM Ni 2þ for 12, 24 or 48 h in our previous research, and the differentially expressed proteins in each Ni 2þ -treated group were identified [6]. In this study, the effect of 200 lM Ni 2þ on protein expression profiles was further investigated. The sample size was three.

Screening of candidate gene/protein biomarkers
The differentially expressed genes and proteins whose screening is described in the 'Biomics experiments' Section were analyzed, and the genes/proteins that were coexpressed in five or more groups of Ni 2þ -treated L929 cells were selected. Further joint analysis of the differentially expressed genes and proteins was performed, and those genes/proteins that were differentially expressed at both the mRNA and protein levels with basically consistent expression patterns were screened as candidate biomarkers.
Determination of target gene/protein biomarkers DAVID (http://david.abcc.ncifcrf.gov/) was used to perform a KEGG biological pathway analysis of the candidate gene/protein biomarkers, and those enriched in Ni 2þ cytotoxicity-related pathways were determined to be target gene/protein biomarkers.
Validation of the expression level of target gene/ protein biomarkers qRT-PCR After L929 cells were treated with 100 or 200 lM Ni 2þ for 12, 24, 48 or 72 h, total RNA was extracted, and the expression levels of four genes (Uqcrb, Csf1, Cth and Acot2) were detected with qRT-PCR. Cells that were not treated with Ni 2þ and cultured for 72 h served as the control. The sample size was three.

Western blot analysis
After L929 cells were treated with 100 or 200 lM Ni 2þ for 12, 24, 48 or 72 h, the total protein was extracted, and the expression levels of four proteins (ubiquinol-cytochrome c reductase-binding protein (UQCRB), colony stimulating factor 1 (CSF1), cystathionase (CTH) and ACOT2) were detected by western blot analysis. GAPDH was used as an internal control. Cells that were not treated with Ni 2þ and cultured for 72 h served as the control. The sample size was three.

Construction of stable gene interference/overexpression cell lines
Primers were designed based on the sequences of Uqcrb and Csf1 and synthesized, and then, target gene lentivirus interference/ overexpression vectors were constructed. The lentiviruses were then subjected to packaging and viral titer determination, followed by lentivirus infection and drug screening. qRT-PCR and western blot analysis were used to detect the expression of the target genes and proteins to verify the effectiveness of target gene interference or overexpression. The sample size was three. Cell lines with stable gene interference were constructed by Shanghai Omicsspace Biotech Co., Ltd., and cell lines with stable gene overexpression were constructed by Zhenjiang Aibimeng Biotechnology Co., Ltd.

Validation experiments
Cell proliferation assay.
The MTT method was used to evaluate the effects of 100 and 200 lM Ni 2þ on cell proliferation after different durations (12, 24, 48 and 72 h) in different cells (normal L929 cells, Csf1-silenced L929 cells and Csf1-overexpressing L929 cells). The experimental process was the same as that described in the Section 'MTT assay to determine the cell cytotoxicity of Ni 2þ '. Normal L929 cells that were not treated with Ni 2þ served as the control. The sample size was six.

Oxidative stress analysis.
Normal L929 cells, Uqcrb-silenced L929 cells and Uqcrb-overexpressing L929 cells were seeded in 96-well plates and then treated with 100 or 200 lM Ni 2þ for 12, 24, 48 or 72 h. Normal L929 cells that were not treated with Ni 2þ served as the control. After the cells were cultured for a given time, Dihydroethidium (DHE) and Hoechst staining were performed in sequence, the fluorescence intensity of ethidium bromide was measured by a high-content cell analysis system (Cellomics ArrayScan VTI HCS Reader, Thermo Fisher, USA) and the cells were counted. The average fluorescence intensity of ethidium bromide was used to evaluate the level of superoxide anion in the cells [7]. The sample size was six.

MMP detection.
Ni 2þ treatment was applied as described for the oxidative stress analysis experiment [8]. After the cells had been cultured for a given time (12,24,48 or 72 h), JC-1 and Hoechst staining was performed in turn, and a high-content cell analysis system was used to determine the fluorescence intensities of the JC-1 monomer (green fluorescence) and polymer (red fluorescence). The ratio of green to red fluorescence was used to evaluate the MMP. Normal L929 cells that were not treated with Ni 2þ served as the control. The sample size was six.

Statistical analysis
The data are expressed as the mean 6 SD of at least three independent experiments. Unless stated otherwise, statistical analysis was performed using Student's t test. The results were considered statistically significant when P < 0.05 and highly significant when P < 0.01.

Results and discussion
Results of the cell cytotoxicity assay Figure 1 shows the results of the cell proliferation assay of L929 cells treated with 100 or 200 lM Ni 2þ for different durations (12,24,48 or 72 h). The cell proliferation rate of the Ni 2þ -treated groups decreased as the Ni 2þ concentration and action time increased. Grade 1 cytotoxicity occurred after 72 h of Ni 2þ treatment, indicating that Ni 2þ impacted cell proliferation. This result is consistent with our previous studies [5,9].

Results of the transcriptomics experiments
The differentially expressed genes in L929 cells after 12, 24, 48 or 72 h of treatment with 100 or 200 lM Ni 2þ compared to control L929 cells are shown in Table 1, and detailed information is shown in Supplementary Tables S1 and S2.

Results of the proteomics experiments
Using proteomics technology, it was found that 177, 2191 and 2095 proteins were differentially expressed after 12, 24 and 48 h of exposure to 100 lM Ni 2þ , respectively, in our previous study [6]. Here, 83, 1681 and 2398 proteins were differentially expressed after treatment with 200 lM Ni 2þ for 12, 24 and 48 h, respectively ( Table 2). Detailed information is shown in Supplementary  Table S3.

Screening of candidate gene/protein biomarkers
The differentially expressed genes and proteins identified in the 'Biomics experiments' Section were screened. A total of 127 genes were coexpressed in any five of the 8 Ni 2þ -treated L929 cell groups, 106 genes were coexpressed in any six of the 8 Ni 2þtreated L929 cell groups, 100 genes were coexpressed in any seven of the 8 Ni 2þ -treated L929 cell groups and 147 genes were coexpressed in all eight Ni 2þ -treated L929 cell groups (as listed in Table 1), yielding a total of 480 genes. Detailed information is shown in Supplementary Table S4. A total of 119 proteins were coexpressed in any five of the 6 Ni 2þ -treated L929 cell groups and 31 proteins were coexpressed in all six Ni 2þ -treated L929 cell groups (as listed in Table 2), for a total of 150. Detailed information is shown in Supplementary Table S5.
The above-mentioned 480 differentially expressed genes and 150 differentially expressed proteins were further compared, and in total, 12 gene/protein pairs that were differentially expressed at both the mRNA and protein levels were screened. Their relative expression values are shown in Table 3. The expression patterns of CSF1, Acyl-CoA thioesterase 2 (ACOT2), FAH, LBH and HYOU1 were the same at the mRNA and protein levels and were   Table 3, Nos. [8][9][10][11][12]. This might be because mRNA produced after gene transcription is finely regulated in terms of its processing, maturation, degradation and translation initiation. Among them, a series of small RNA molecules, such as microRNAs and small interfering RNAs, have been found to have a regulatory role in the process by which mRNA is translated into protein, which results in the absence of a quantitative relationship or a weak or even an opposite correlation in expression [10]. Therefore, CSF1, ACOT2, FAH, LBH, HYOU1, UQCRB and CTH were selected as candidate gene/protein biomarkers.

Determination of target gene/protein biomarkers
To further screen the target gene/protein biomarkers of Ni 2þ cytotoxicity, a KEGG biological pathway analysis of the seven obtained candidate gene/protein biomarkers was performed ( Table 3, Nos. 1-7). Detailed information is shown in Supplementary Table S6. According to previous research related to the cytotoxicity mechanism of Ni 2þ by our group, Ni 2þ mainly exerts its toxic effects by affecting the following factors: (i) the cytoskeleton and cell adhesion, (ii) cell growth and proliferation, (iii) the expression of related genes and proteins in cells, (iv) intracellular energy metabolism and (v) oxidative stress and inflammation [5,9]. Therefore, the genes/proteins involved in the pathways related to Ni 2þ cytotoxicity were screened according to the above factors. In total, four of the seven candidate gene/protein biomarkers in Table 3 were found to participate in pathways related to Ni 2þ cytotoxicity (Table 4), while the other three did not participate in such pathways; thus, these four candidate gene/protein biomarkers (CSF1, CTH, ACOT2 and UQCRB) were determined to be target biomarkers. Table 4 shows that CSF1 participates in the following six pathways related to Ni 2þ cytotoxicity: (i) the cytokine-cytokine receptor interaction pathway (cytokines are important cell regulators and drivers of cell biological activities that induce responses by binding specific receptors on the surface of target cells and have functions such as regulating cell growth, differentiation, immune activity and inflammatory responses) [11], (ii) Rap1 signaling pathway (Fig. 2), (iii) Ras signaling pathway, (iv) PI3K-Akt signaling pathway, (v) Mitogen-activated protein kinase (MAPK)  Rap1 is a small GTPase that controls various cellular processes, such as cell adhesion and cell polarity. Rap1 controls cell-cell and cell-cell-matrix interactions by regulating the functions of integrins and other adhesion molecules in different cell types. Furthermore, Rap1 can regulate MAP kinase activity [12]. Ras and PI(3)K are protein kinases that are closely related to cell proliferation. Ras and PI(3)K signaling plays a key role in regulating cell growth by regulating the downstream molecule mTOR [13]. The basic MAPK pathway is structured as a three-level kinase model consisting of MAP kinase, MAPK kinase and Mitogen-activated protein kinse-ERK kinase (MEK). The activation of these three kinases together regulates several important cellular physiological/ pathological processes, including cell growth, differentiation, adaptation to environmental stress and the inflammatory response.
As an important cytokine, TNF can induce the activation of various intracellular signaling pathways, including apoptosis, cell survival, inflammation and immunity. Therefore, the above indicates that CSF1 met the screening conditions for Ni 2þ cytotoxicity markers. CTH participated in three cytotoxicity-related pathways (biosynthesis of amino acids pathway (Fig. 3); cysteine and methionine metabolism pathway; and glycine, serine and threonine metabolism pathway), all of which are amino acid metabolism pathways. Amino acid metabolism is very important for biological activities. In this process, digested and decomposed amino acids from ingested proteins are used to synthesize tissue proteins, and amino acids can be used to synthesize other nitrogenous substances useful for the human body or decomposed to provide energy.
ACOT2 was found to be involved in two cytotoxicity-related pathways (biosynthesis of unsaturated fatty acids pathway and fatty acid elongation pathway). Fatty acids are the main components of neutral fats, phospholipids and glycolipids. With the help of various related enzymes, fat is digested, absorbed, synthesized, decomposed and processed into the substances needed by the body to ensure normal physiological function. Lipids are important substances for energy storage and energy supply in the body and important structural components of biological membranes. Therefore, fat metabolism is of great importance to biological activities.
UQCRB was found to participate in one cytotoxicity-related pathway (the oxidative phosphorylation pathway), which belongs to the energy metabolism pathway. Oxidative phosphorylation is the coupling of energy released when substances are oxidized in Table 4. Target gene/protein biomarkers involved in pathways related to Ni 2þ cytotoxicity

No.
Target gene/ protein biomarker  Figure 2. The Rap1 signaling pathway in which CSF1 is involved [14].
the body to supply ADP and inorganic phosphorus to synthesize ATP through the respiratory chain.
In conclusion, CSF1, CTH, ACOT2 and UQCRB were all shown to be involved in Ni 2þ cytotoxicity-related pathways; thus, they were identified as target gene/protein biomarkers of Ni 2þ cytotoxicity. The other three candidate gene/protein biomarkers (FAH, LBH and HYOU1) did not participate in a Ni 2þ cytotoxicity-related pathway, so they were not used as target biomarkers and are not discussed further.

Validation of the expression of target gene/ protein biomarkers
qRT-PCR Figure 4 shows the results of qRT-PCR analysis of the expression levels of the four target gene markers (Csf1, Cth, Acot2 and Uqcrb) in L929 cells after 12, 24, 48 or 72 h of treatment with 100 or 200 lM Ni 2þ , and the results were compared with the transcriptome sequencing results. The qRT-PCR results showed that the four genes were downregulated in six or more Ni 2þ -treated groups, and the Csf1 gene was downregulated in all groups. Therefore, the expression patterns of these four target biomarkers at the mRNA level were consistent, so further experiments to verify their expression patterns at the protein level were carried out.  Supplementary Fig. S1 and Table S7-S10. Figure 3. The biosynthesis of amino acids pathway in which CTH is involved [14].

Western blot analysis
The expression patterns of three proteins (CSF1, CTH and UQCRB) determined by western blot analysis were basically consistent with the results of the proteomics experiment, which showed downregulation. However, the expression patterns of ACOT2 determined by the two experiments were inconsistent. Therefore, the three target protein biomarkers (CSF1, CTH and UQCRB) whose expression patterns were consistent are discussed next.
CSF1 is a cytokine that is expressed in many cell types and plays a crucial role in cell survival, proliferation and differentiation [15]. CSF1 usually binds CSF1R, and the CSF1/CSF1R signaling pathway can activate the MAPK signaling pathway to promote cell proliferation or differentiation [16]. Therefore, we further verified the function of CSF1 by a cell proliferation assay.
CTH is mainly expressed in cells as cystathionine c-lyase (CSE). In the liver, CSE is the source of H 2 S, which can regulate the production of glucose, in the blood vessels [17]. Since the function of CTH is not directly related to Ni 2þ cytotoxicity, the function of CTH was not verified.
UQCRB is a part of the mitochondrial respiratory chain and plays an important role in electron transport and redox-related proton transport. Jung et al. studied the effect of UQCRB on the oxygen-sensing mechanism under conditions of overexpression/ interference and found that a change in the UQCRB expression level could regulate the ROS content in cells and affect electron transfer [18]. Therefore, the function of UQCRB was further verified by detection of the ROS level and MMP.

Functional validation of the gene/protein biomarkers
The above experiments to verify expression patterns ensured that the results of experiments used to determine expression of the selected target gene/protein biomarkers were correct and reliable. However, whether the changes in their expression levels and cell function are related could not be determined by purely theoretical analysis. Therefore, it was necessary to combine the 'dry method' (theoretical analysis) with the 'wet method' (experimental verification). The possible molecular functions of the biomarkers determined by bioinformatics analysis were verified by cellular/molecular biology experiments to determine whether they play a specific role in Ni 2þ -induced cytotoxicity and their mechanisms. Since RNA interference and overexpression can significantly downregulate and upregulate the expression levels of specific genes in cells, cell biology experiments can be used to study whether such downregulation and upregulation lead to corresponding changes in cell function to determine whether the gene can regulate cell function; thus, this technology is often used to study and verify the functions of specific genes [19][20][21].
In this study, RNA interference/overexpression technology was first used to construct stable cell lines with Csf1 and Uqcrb interference/overexpression. Then, through detecting cell proliferation, oxidative stress and the MMP, we investigated whether the downregulation and upregulation of these two genes were related with the changes in cell function. The functions of the two target biomarkers were finally verified.

Construction of stable gene interference/overexpression cell lines
The expression of Csf1 and Uqcrb in normal L929 cells, L929 cells with stable gene interference and L929 cells with stable overexpression was verified by qRT-PCR, and the results are shown in Fig. 6.
It is generally believed that when target gene expression levels in cells after RNA interference and overexpression decrease by 50% or increase by more than 2-fold, respectively, interference and overexpression have been successful. Figure 6 shows that the Csf1 expression value in the interference group was 60% lower than that in the normal group and that the Csf1 expression value in the overexpression group was 450% higher than that in the normal group, implying that the Csf1-silenced and Csf1-overexpressing cell lines had been successfully constructed. The expression level of Uqcrb in the interference group was 60% lower than that in the  normal group, but the expression level of Uqcrb in the overexpression group was only 30% higher than that in the normal group, suggesting that the efficiency of gene overexpression was not satisfactory. Therefore, the protein expression level of UQCRB in normal L929 cells and L929 cells with stable Uqcrb overexpression was further determined by western blot analysis (Fig. 7). The UQCRB protein content in the overexpression group was more than two times higher than that in the normal cells. Combined with the results of the qRT-PCR and western blot experiments, these results suggested that the Uqcrb-silenced and Uqcrb-overexpression cell lines had been successfully constructed.

Validation experiments
Cell proliferation assay. Figure 8 shows the cell proliferation rates of three types of L929 cells (normal L929 cells, Csf1-silenced L929 cells and Csf1-overexpressing L929 cells) after treatment with 100 or 200 lM Ni 2þ for different durations (12,24,48 or 72 h). As the Ni 2þ concentration increased, the cell proliferation rate of each Ni 2þ -treated group gradually decreased in the same type of cells, and all experimental groups showed evidence of cytotoxicity after 72 h of Ni 2þtreatment (P < 80%). These results showed that Ni 2þ could affect cell proliferation and was cytotoxic and were consistent with our previous conclusions [5,9]. Further analysis showed that the cell proliferation rates of the Csf1-silenced/Csf1-overexpressing L929 cells were lower than those of the normal cells. Thus, both the downregulation and upregulation of the Csf1 gene led to a decrease in the proliferation rate of the L929 cells, indicating that the biological function of the Csf1 gene is not necessarily related to cell proliferation. Shen et al. indicated that Ni 2þ toxicity and processes such as cell adhesion and migration may be connected [22]. Therefore, the CSF1 gene/protein was not selected as a biomarker for assessing Ni 2þ cytotoxicity.
Oxidative stress analysis. Figure 9 shows the mean fluorescence intensity of ethidium bromide in three types of L929 cells (normal L929 cells, Uqcrb-silenced L929 cells and Uqcrb-overexpressing L929 cells) after treatment with 100 or 200 lM Ni 2þ for different durations (12,24,48 or 72 h). The higher the fluorescence intensity was, the stronger the oxidative stress was. As the Ni 2þ concentration and exposure time increased, the degree of oxidative stress in each Ni 2þ -treated group increased in the same type of L929 cells. These results showed that Ni 2þ could induce oxidative stress, which is also consistent with our previous conclusions [9]. Further analysis revealed that after treatment with Ni 2þ at two concentrations for different durations, the oxidative stress level in the Uqcrb-silenced L929 cells was significantly higher than that in normal cells (P < 0.05), while the oxidative stress level in the Uqcrb-overexpressing L929 cells was significantly lower than that in the normal cells (P < 0.05). These results indicated that the downregulation and upregulation of the Uqcrb gene led to an increase or decrease, respectively, in the level of intracellular ROS, suggesting that the Uqcrb gene plays a role in regulating intracellular ROS.
MMP detection. Figure 10 shows the ratio of the green/red fluorescence intensity of JC-1 in three types of L929 cells (normal L929 cells, Uqcrb-silenced L929 cells and Uqcrb-overexpressing L929 cells) after treatment with 100 or 200 lM Ni 2þ for different durations (12, 24, 48 or 72 h); the higher the ratio was, the lower the MMP was. As the Ni 2þ concentration increased, the MMP in each Ni 2þ -treated group decreased (the ratio of green/red fluorescence intensity increased) in the same type of cells. These results showed that Ni 2þ could lead to a decrease in the MPP.
Further analysis revealed that after treatment with Ni 2þ at two concentrations for different durations, the MMP in the Uqcrbsilenced L929 cells was significantly (P < 0.05) or highly significantly (P < 0.01) lower than that in the normal cells, while the MMP in the Uqcrb-overexpressing L929 cells was significantly (P < 0.05) or highly significantly (P < 0.01) higher than that in the normal cells. These results indicated that the downregulation/ upregulation of the Uqcrb gene could lead to a decrease/increase in the MMP, proving that the Uqcrb gene plays a role in regulating the MMP.
The above indicated that the Uqcrb gene regulates the ROS content and MMP in cells, and these biological functions are related to the biological process involved in Ni 2þ cytotoxicity.
Hong et al. and Kim et al. demonstrated that UQCRB expression in colorectal cancer cells and tissues was correlated with colorectal cancer progression; thus, UQCRB is a biomarker for diagnosing colorectal cancer [23,24]. Based on the above results of expression to verify the expression levels and function of this biomarker, the UQCRB gene/protein could serve as a final biomarker for evaluating Ni 2þ cytotoxicity.

Conclusion
This study establishes an innovative technical method for the screening of biomarkers of biomaterial cytotoxicity by combining the 'dry method' and 'wet method' (Fig. 11). This technique was used to conduct a joint multiomics screening of Ni 2þ cytotoxicity  biomarkers for the first time through combined analysis of transcriptomics and proteomics data ('dry method'). Four target gene/protein biomarkers were first identified, and one final gene/ protein biomarker (UQCRB) was identified through a series of cytomolecular experiments ('wet method'). This study provides new insight into investigating cytotoxicity biomarkers of other biomaterials.

Supplementary data
Supplementary data are available at REGBIO online.

Funding
This paper was supported by the National Natural Science Foundation of China (31971254).
Conflicts of interest statement. None declared.