RETRACTED ARTICLE: Ras-ERK1/2 signalling promotes the development of osteosarcoma through regulation of H4K12ac through HAT1

Abstract We, the Editors and Publisher of the journal Artificial Cells, Nanomedicine, and Biotechnology, have retracted the following article: Jingzhe Zhang, Meihan Liu, Wanguo Liu & Wenjun Wang (2019) Ras-ERK1/2 signalling promotes the development of osteosarcoma through regulation of H4K12ac through HAT1, Artificial Cells, Nanomedicine, and Biotechnology, 47:1, 1207–1215, DOI: 10.1080/21691401.2019.1593857 Since publication, concerns have been raised about the integrity of the data in the article. When approached for an explanation, the authors checked their data and confirmed there are fundamental errors present. Therefore, they have agreed to the retraction of this article. The authors apologise for this oversight. We have been informed in our decision-making by our policy on publishing ethics and integrity and the COPE guidelines on retractions. The retracted article will remain online to maintain the scholarly record, but it will be digitally watermarked on each page as ‘Retracted’.


Introduction
Osteosarcoma (OS) is a common primary malignant tumor, usually occurs in adolescents and children, and has the characteristics of high malignancy and early metastasis [1,2]. Previously, amputation is the major treatment for OS, but, owing to occurance of lung metastasis, the 5-year survival rate of these patients is <20% [3]. With the advancement of surgery and the development of adjuvant chemotherapy, the current managements for OS have been changed [4]. However, poor prognosis and drug resistance remain unavoidable. It is urgently to seek a novel and effective therapeutic strategy for OS.
Ras (rat sarcoma viral oncogene homolog) is known to be a kind of conserved oncogene, which can activate the downstream signalling pathway by increasing expression, gene point mutation, gene insertion and translocation [5,6]. Activation of Ras has been found in various cancers, which is implicated in mediating different biological processes, including cell proliferation, apoptosis and metastasis [7,8]. With regard to OS, Ras activation has been reported to regulate WISP-1-enhanced cell migration and matrix metalloproteinases (MMPs) expression in OS [9]. More interestingly, an orthotopic mouse model experiment demonstrated that the Ras-ERK signalling pathway was closely associated with lung metastasis of OS [10]. It is known to all that ERK1/2 is an important downstream signalling of Ras, which plays an important regulatory role in tumorigenesis [11]. However, it is still lack enough studies especially focused on whether Ras-ERK1/2 is involved in the pathogenesis of OS.
Recent studies have demonstrated that abnormal epigenetic information is closely related to the occurrence and development of human malignant tumors [12,13]. Histone acetylation is an important part of epigenetics, the dynamic equilibrium between histone acetyltransferase (HAT) and histone deacetylase (HDAC) is associated with the chromatin structure and gene expression [14]. As an important histone acetylase, histone H4 acetylation at lysine 12 (H4K12ac) has been reported to be mediated by estrogen receptor-a, which is linked to BRD4 function and inducible transcription [15]. Additionally, H4K12ac has been confirmed to be associated with the poor prognosis in pancreatic cancer [16]. However, little is known about the involvements of H4K12ac in the pathogenesis of OS.
HDAC6 is the largest protein ever found in humans, which mainly locates in the cytoplasm [17]. Increasing pieces of evidence demonstrated that HDAC6 was highly expressed in the heart, liver, kidney, brain and other organs, which was involved in regulating diverse important biological processes, including cell migration, microtubule stability, intracellular transport, immune synapse formation, and cell autophagy [18,19]. In this study, we aimed to corroborate the involvements of H4K12ac in the Ras-ERK1/2 affecting cell proliferation and migration in OS cells. Simultaneously, the regulatory effects of HDAC6, HAT1 and murine double minute-2 (MDM2) were also explored in the processes to uncover the underlying mechanism. This study discussed the importance of H4K12ac in the tumorigenic signal transmission of Ras-ERK1/2 in OS cells, which might be a new target for the diagnosis and treatment of OS.

Cell transfection
After incubation overnight, cells (5 Â 10 5 cell/well) were transfected with the plasmids constructed above or siRNA against HDAC6 (si-HDAC6-1 and si-HDAC6-2) and MDM2 (si-MDM2) by using Lipofectamine 2000 (Invitrogen, Carlsbad, CA) based on the manufacturer's protocol. After transfection for 48 h, cells were harvested and utilized in the following experiments.

Cell viability
After transfection for 48 h, MG-63 cells were collected and were incubated in 96-well plates with a density of 5 Â 10 3 cells/well. For detection of cell viability, 20 mL 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2Htetrazolium bromide solution (MTT; Sigma-Aldrich, St. Louis, MO) was added to each well and incubated for another 4 h at 37 C. After incubation, 100 lL of dimethyl sulfoxide (DMSO, Sigma-Aldrich) was used to lyse the formazan crystals. The absorbance of each well was measured by using a Microplate Reader (Bio-Rad, Hercules, CA) at 570 nm.

Cell migration assay
Cell migration of MG-63 cells was detected by transwell assay with 8-lM pore size filters (Costar, Boston, MA). After filling the lower chamber with complete culture medium, the upper chamber was supplemented with 1 Â 10 4 MG-63 cells in serum-free medium. After incubation for 24 h at 37 C, the cells in the upper side were removed using a wet cotton swab. MG-63 cells migrated to the lower side was fixed with methanol, and stained with 0.5% crystal violet for 20 min. The OD values were read at 570 nm by using the microplate reader (Bio-Rad).

Cell cycle assay
After transfection, MG-63 cells were rinsed twice with PBS and fixed in 70% ethanol at 4 C overnight. Afterward, 50 lg/mL propidium iodide (PI) solution (Roche, Basel, Switzerland) was added to the six-well culture plate to stain cells for 30 min at room temperature. Four thousand events for each sample were acquired by using a FACS scan (Becton Dickson, San Jose, CA). Cells in G0/G1, S and G2/M phases of the cell cycle were analyzed by using the ModFit software (Verity Software House, Topsham, ME).
The transfected cells were cross-linked in 1% formaldehyde for 10 min at room temperature and rinsed twice with PBS. After this, MG-63 cells were collected and lysed with SDS lysis buffer (Upstate Biotechnology, Lake Placid, NY). The lysates sonicated in ultrasonic bath (Bioruptor, Diagenode) and then centrifuged at 15,000g for 10 min. The supernatants were diluted with ChIP dilution buffer (Upstate Biotechnology) and were immunoprecipitated with rabbit anti-H4K12ac (ab46983; Abcam, Cambridge, UK) overnight at 4 C. The anti-IgG antibody (2 lg, ab171870; Abcam) served as a control group. The beads were washed according to the previously described [21]. The immunoprecipitated DNA and serial dilutions of the 10% input DNA were analyzed by using RT-qPCR.

Western blot analysis
After 48 h transfection, MG-63 cells were collected and were washed once with PBS. Cell lysates were prepared by using RIPA lysis buffer (Beyotime, Shanghai, China). A BCA protein assay kit (Pierce, Appleton, WI) was executed to quantify the protein concentrations. The protein samples were then separated by SDS-PAGE and transferred to nitrocellulose membranes (Bio-Rad). After blocking in BSA solution, the membranes were incubated with the primary antibodies overnight at 4 C. Subsequently, HRP-conjugated secondary antibody (ab205718; Abcam) was added and incubated with the membranes for 1 h at room temperature. The Western blots were developed by using Western Blotting Luminol Reagent (Santa Cruz Biotechnology, Santa Cruz, CA) as recommended by the reagent manufacturer, meanwhile, the intensity of the bands was quantified by using ImageJ software (Bio-Rad).

Statistical analysis
All data are presented as the mean ± standard deviation (SD). Statistical analyses were performed using SPSS 19.0 statistical software (SPSS, Inc., Chicago, IL). The statistical analyses of multiple groups were performed using One-way analysis of variance (ANOVA) followed by the Duncan posthoc test. P-values of <.05 represented a statistically significant result.
H4K12ac regulated the carcinogenic effect of Ras-ERK1/ 2 in MG-63 cells To investigate whether H4K12ac participates in mediating the functions of Ras-ERK1/2 signalling pathway in the proliferation and migration of OS cells, we constructed the H4K12Q plasmid to mimic the acetylation state of H4K12ac. MG-63 cells were subsequently co-transfected with three group plasmids, which included pEGFP-N1 and H4 (GFP þ H4), pEGFP-
After transfection with H4K12Q, cell viability was obviously inhibited, as compared to Ras þ H4 group (p < .01 or p < .001, Figure 2(A)). Additionally, colony formation and transwell assays were performed to measure colonies and cell migration. As shown in Figure 2(B,C), the number of colonies and cell migration were both promoted in Ras þ H4 group as compared to GFP þ H4 group (p < .001), but were impeded in Ras þ H4K12Q group as compared to Ras þ H4 group (p < .001, Figure 2(B,C)). All the aforementioned results indicated that H4K12ac participated in regulating the carcinogenic effect of Ras-ERK1/2, which could directly affect proliferation and migration in MG-63 cells.

H4K12ac mediated the transcription of ERK1/2 downstream genes
To explore the regulatory effect of H4K12ac on the signal transduction process of Ras-ERK1/2, the downstream genes of ERK1/2 (CYR61, IGFBP3, WNT16B, NT5E and GDF15) were determined. The three plasmid groups of GFP þ H4, Ras þ H4 and Ras þ H4K12Q were constructed and transfected into MG-63 cells. RT-qPCR analytical results revealed that the relative mRNA levels of CYR61, WNT16B, GDF15 and NT5E were all upregulated, but the relative mRNA level of IGFBP3 was downregulated in Ras þ H4 group as relative to GFP þ H4 group (p < .001, Figure 3(A)). However, in Ras þ H4K12Q group, the effect of Ras G12V/T35S on the expression levels of these downstream genes of ERK1/2 were all reversed (p < .001, Figure  3(A)). Subsequently, ChIP assay was performed to further investigate the effect of H4K12ac on the transcription of ERK1/2 downstream genes. As shown in Figure 3(B), the input level of H4K12ac were significantly declined at the promoter regions of CYR61, IGFBP3, WNT16B, NT5E and GDF15 following Ras-ERK1/ 2 activation (p < .001). The data testified that H4K12ac participated in regulating the transcription of ERK1/2 downstream genes.
HDAC6 silence recovered H4K12ac expression and inhibited the carcinogenic role of Ras-ERK1/2 in MG-63 cells HDAC6 is an important deacetylase, which is involved in regulating numerous biological and pathologic processes [22]. In the next experiments, we explored the regulatory relationship between HDAC6 and H4K12ac as well as investigated the effect of HDAC6 on Ras-ERK1/2-mediated proliferation and migration

R E T R A C T E D
in MG-63 cells. RT-qPCR assay was performed the expression level of HDAC6 in MG-63 cells at transcriptional level after transfection with si-HDAC6-1 and si-HDAC6-2. The results presented that the expression level of HDAC6 was apparently decreased in si-HDAC6-1 and si-HDAC6-2 transfected cells (Figure 4(A)). Western blot results showed that the down-regulation of H4K12ac induced by Ras G12V/T35S was notably reversed in si-HDAC6-1 and si-HDAC6-2 transfected cells ( Figure  4(A)). Additionally, the results in Figure 4(B,C) showed that cell viability and proliferation increased by Ras G12V/T35S were significantly inhibited by HDAC6 silence (p < .001). The percentage of cells at S phase of cell cycle was increased by Ras G12V/T35S in MG-63 cells, however, after transfection with si-HDAC6-1 and si-HDAC6-2, the percentage of cells at S phase of cell cycle was obviously decreased (Figure 4(D)). Besides, we observed that HDAC6 silence reversed the regulatory effect of Ras G12V/T35S on the expression levels of CYR61, IGFBP3, WNT16B, NT5E and GDF15 (p < .001, Figure 4(E)). All the above results indicated that HDAC6 silence recovered H4K12ac expression and inhibited the carcinogenic role of Ras-ERK1/2 in MG-63 cells.

H4K12ac was down-regulated by Ras-ERK1/2 activation through degradation of HAT1
To investigate the regulatory relationship between HAT1 and Ras-ERK1/2 signalling, RT-qPCR and Western blot assays were  Figure 5(A-C) revealed that the expression level of HAT1 at transcriptional level was unaffected by Ras-ERK1/2 activation; however, the protein level of HAT1 was evidently downregulated by Ras-ERK1/2 activation. There was no effect of Ras-ERK activation on HDAC6 at transcriptional and post-transcriptional levels ( Figure 5(A,B)). In Figure 5(D), the results showed that the protein levels of HAT1 and H4K12ac were simultaneously decreased by Ras-ERK1/2 activation. Additionally, ChIP analytical results showed that the Ras-ERK1/2 activation significantly hindered the enrichment of HAT1 at the promoter regions of CYR61, IGFBP3, WNT16B, NT5E and GDF15 genes (p < .001, Figure 5(E)), suggesting that HAT1 was involved in mediating the transcription of ERK1/2 downstream genes. Subsequently, the MG132, a proteasome inhibitor, was used to treat MG-63 cells after transfection with Ras G12V/T35S and HAT1-HA plasmids. Results in Figure 5(F) demonstrated that the protein level of HAT1 was recovered in Ras G12V/T35S -transfected cells after treatment with MG132. Moreover, we found that the protein level of H4K12ac was notably declined by Ras G12V/T35S cells after 48 h transfection without MG132 stimulation ( Figure 5(G)). However, after administration with 25 lM MG132, the protein level of H4K12ac was recovered in a time-dependent manner in MG-63 cells ( Figure 5(H)). All above results indicated that down-regulation of H4K12ac induced by Ras-ERK1/2 activation might through regulating HAT1 degradation.

Ras-ERK1/2 activation degraded HAT1 by regulation of MDM2
To explore whether MDM2 is involved in mediating the degradation of HAT1 induced by activation of ERK1/2 pathway, Ras G12V/T35S and HAT1-HA following with 0.5, 1 and 2 lg MDM2-His were transfected into MG-63 cells. The plasmids of

R E T R A C T E D
Ras G12V/T35S and HAT1-HA without MDM2-His served as a control group. The results showed that the expression of MDM2 was notably decreased, while the expression level of HAT1 was increased in MG-63 cells after transfection with Ras G12V/T35S and HAT1-HA following with 0.5, 1 and 2 lg MDM2-His ( Figure 6(A,B)). Interestingly, after transfection with MDM2-MU plasmid, we observed that the down-regulated HAT1 protein level was notably recovered (Figure 6(C,D)), indicating that the degradation of HAT1 was mediated by MDM2. Additionally, in Figure 6(E), the results exhibited that activation of Ras-ERK1/2 signalling facilitated the protein level of MDM2, as well as declined H4K12ac protein level in MG-63 cells. The results hinted that the degradation of HAT1 induced by MDM2 was associated with Ras-ERK1/2 activationrepressed H4K12ac. MG-63 cells were then transfected with si-MDM2 and si-NC, and the transfection efficiency was presented in Figure 6(F). We finally observed that MDM2 silence prevented Ras-ERK1/2-repressed H4K12ac in MG-63 cells ( Figure 6(G)). All the aforementioned results indicated that Ras-ERK1/2 activation down-regulated H4K12ac possibly via MDM2-mediated HAT1 degradation.

Discussion
In the present study, we found that the Ras-ERK1/2 signalling pathway specifically down-regulated H4K12ac expression in MG-63 cells. Moreover, H4K12ac observably inhibited the carcinogenic effect, synchronously regulated the transcription of ERK1/2 downstream genes in MG-63 cells. It was intriguing that HDAC6 silence significantly recovered Ras-ERK1/2repressed H4K12ac expression. Moreover, the carcinogenic effect of Ras-ERK1/2 was also suppressed by HDAC6 silence in MG-63 cells. Furthermore, Ras-ERK1/2 impeded H4K12ac

R E T R A C T E D
expression was observed to be associated with MDM2-regulated the degradation of HAT1. Increasing pieces of evidence testified that Ras mutations contribute to boost the occurrence of various cancers [23,24]. Ras-ERK1/2 is an important signalling pathway in mitogenactivated protein kinases (MAPKs), which can be activated by various mechanisms, such as chromosome ectopic, cytokine mutations and overexpression of wild or mutant receptors [25]. Ras-ERK1/2 activation has been validated to exhibit critical roles in proliferation, apoptosis, migration and cell differentiation [26]. Recent study authenticates that the transcription and protein expression of Ras gene are regulated by histone acetylation modification [27]. Study from Liu et al. [21] found that Ras-PI3K signalling targeted the specific epigenetic modification of histone H3 acetylation at lysine 56 (H3K56ac), thereby mediating tumor cell activity in breast cancer. In this study, we found that activation of Ras-ERK1/2 specifically down-regulated H4K12ac expression in MG-63 cells. Therefore, we conjectured that H4K12ac might be involved in regulating the carcinogenic effect of Ras-ERK1/2 in OS cells.
Important evidence displays that histone H4 can be reversibly acetylated at lysine residues 5, 8, 12 and 16 [28]. One study reported that histone H4 acetylation at lysine 16 (H4K16ac) expression level was increased by Ras-ERK1/2 activation, which was implicated in regulating cell proliferation and migration in breast cancer cells [29]. With regard to H4K12ac, high expression of H4K12ac has been found in follicular adenomas [30], moreover, its aberrant expression might affect early embryonic development [31]. However, the regulatory effect of H4K12ac on Ras-ERK1/2-induced cell proliferation and migration in OS cells is still unclear. In our study, we constructed H4K12Q to mimic the acetylation state of H4K12ac. The results showed that cell viability, colonies and migration were promoted by Ras-ERK1/2 activation. However, after transfection with H4K12Q, the promoting effects were obviously reversed in MG-63 cells. Further, we also observed that H4K12Q suppressed the transcription of ERK1/2 downstream genes following Ras-ERK1/2 activation. These data indicated that H4K12ac could suppress the carcinogenic effect of Ras-ERK1/2 in OS cells.
HDAC6 is a class IIb HDAC, which is a key regulator in diversified biological and pathological processes [22]. HDAC6 has been proven to enhance the activity of ERK1/2 through deacetylation, as well as to regulate the oncogenic activity and nuclear localization of mutant K-Ras [22,32]. In terms of cancers, HDAC6 has been regarded as a crucial target of drug development for the treatment of cancers due to its main contribution in oncogenic cell transformation, which is also associated with tumorigenesis and improved survival of cancers [33]. However, the effect of HDAC6 on the pathogenesis of OS is still unknown. In our study, we observed that HDAC6 silence recovered the down-regulation level of H4K12ac expression induced by Ras-ERK1/2 activation. Additionally, the promoting effects of Ras-ERK1/2 on MG-63 cell proliferation, migration and the mRNA expression of CYR61, IGFBP3, WNT16B, NT5E and GDF15 were all reversed by HDAC6 silence. These findings indicated that HDAC6 might be an important regulator in the pathogenesis of OS. HAT1 is the earliest identified histone acetyltransferase, and the sole known B-type histone acetyltransferase [34]. RNAi screening confirmed that HAT1 might be a potential drug target in esophageal squamous cell carcinoma [35]. Evidence from Jin et al. [36] demonstrated that HAT1 could promote cell proliferation and induce cisplatin resistance in hepatocellular carcinoma. Additionally, Agudelo et al. [37] reported that HAT1 could catalyze the acetylation of histone H4 at lysines 5 and 12. However, whether HAT1 participates in regulating the expression H4K12ac following Ras-ERK1/2 activation is still worth studying. The results in the present study displayed that the down-regulation of H4K12ac induced by Ras-ERK1/2 activation might through degrading HAT1. More importantly, we further explored whether MDM2 was involved in mediating the degradation of HAT1 based on the previous study demonstrated the importance of MDM2 in histone acetyltransferases degradation [38]. As we expected, the results revealed that the degradation of HAT1 was mediated by MDM2. All the above findings indicated that Ras-ERK1/2 impeded H4K12ac expression might through degrading HAT1 mediated by MDM2.
Taken together, these data demonstrated that Ras-ERK1/2 signalling promoted the development of OS through downregulation of H4K12ac. Ras-ERK1/2-repressed H4K12ac was associated with the degradation of HAT1 mediated by MDM2. These findings might provide a novel treatment strategy for OS. This study did not detect the affinity of HAT1 and HDAC6 with substrates after activation of the Ras-ERK1/2 signalling pathway. Moreover, whether the localization of HAT1 is abnormal after activation of the ERK1/2 pathway, which results in the decrease of substrate-bound HAT1 enzyme and the decrease of H4K12ac level is worth explored. Further studies are still needed to explore the broader effects of Ras-ERK1/2-H4K12ac axis in OS.

Disclosure statement
No potential conflict of interest was reported by the authors.