The CDKN2B-AS1/ MIR497/TXNIP axis regulated macrophages

The polarization of macrophages plays a critical role in the pathophysiology of rheumatoid arthritis. The polarization states include pro-in ammatory M1 polarization and various alternative anti-in ammatory M2 polarization. Our preliminary results showed that CDKN2B-AS1/MIR497/TXNIP axis might play a role in macrophages extracted from rheumatoid arthritis patients. Therefore, we hypothesized that this axis regulated the polarization of rheumatoid macrophages. Flow cytometry was used to determine the surface polarization markers in M1 or M2 macrophages from healthy donors and rheumatoid arthritis patients. QPCR and western blotting were used to compare the expression of the CDKN2BAS1/MIR497/TXNIP axis in these macrophages. We interfered with the expression and function of the axis from upstream to downstream in the macrophage cell line MD to test its roles in macrophage polarization. Compared to cells from healthy donors, cells from rheumatoid arthritis patients expressed a higher level of CD40 and CD80 and a lower level of CD16, CD163, CD206, and CD200R after polarization, they also expressed higher CDKN2B-AS1, lower MIR497, and higher TXNIP. In macrophages from healthy donors, there was no correlation among CDKN2B-AS1, MIR497, and TXNIP. But in macrophages patients, they showed signi cant correlations. The CDKN2B-AS1 knockdown, MIR497 mimics suppressed M1 polarization but promoted M2 polarization in MD cells, while MIR497 knockdown and TXNIP overexpression did the opposite. This study demonstrated that elevated CDKN2B-AS1 in macrophages promotes M1 polarization and inhibits M2 polarization of macrophage by negatively regulating MIR497, thereby upregulated the expression of TXNIP.


Introduction
Rheumatoid arthritis is the most common long-term in ammatory joint disease that results in joints stiffness, swelling, and pain [1] . As an autoimmune disease, rheumatoid arthritis results from the selfattacking of the immune system [2] . Although rheumatoid arthritis is not curable, clinical treatments for rheumatoid arthritis can decrease the in ammatory activities in the joints to relieve pain and inhibit joint damage [3] . Much as many studies have shown that the disease involves several complex interactivities between the human body and environmental factors [4] , the exact mechanisms that drive rheumatoid arthritis remain obscure. Hence, clinical treatment of rheumatoid arthritis required further understanding of the pathogenesis of this disease.
Macrophages are one of the most critical cells in the pathophysiology of rheumatoid arthritis [5] .
Macrophages contribute to the secretion of many pro-in ammatory cytokines and chemokines such as TNF and IL-1β [6] . These pro-in ammatory factors activate many downstream cells and result in joint pain and joint damage [7] . On the other hand, macrophages also contribute to the recruitment of antiin ammatory cytokines such as IL-10 [8] . These anti-in ammatory factors can negatively regulate autoimmune activities and protect joint tissue [9] . A key activity of macrophages in rheumatoid arthritis is their plasticity, also call polarization [10] . The polarization of macrophages is activated by many environmental factors [11] . The polarized macrophages can acquire different abilities for speci c functions. These different polarization states include pro-in ammatory M1 polarization and various alternative antiin ammatory M2 polarization [12] . Although speci c macrophage polarization surface markers in rheumatoid arthritis have not been identi ed [13,14] , several studies have reported the in ammatory regulation role of macrophages in rheumatoid arthritis [15] .
Endogenous non-coding RNAs have been found to play critical roles in bone cells [16] . Micro RNAs (miRNAs) are a class of non-coding RNAs that can bind to the 3′ -UTR of their regulated genes to decrease the expression of their target [17] . A miRNA, miR-497 (MIR497), has been reported to participate in macrophages [18] , but its exact role has not been identi ed. Long non-coding RNAs (lncRNAs) are another class of non-coding RNAs that can regulate gene expression by binding to their target RNAs [19] .
TXNIP has been identi ed to be a critical regulator for in ammation in macrophages [22][23][24][25] , and its potential role in rheumatoid arthritis has previously been reported [26] . In addition, a study showed that TXNIP is regulated by MIR497 and CDKN2B-AS1 [20] .
Our preliminary results showed that the CDKN2B-AS1/miR497(MIR497)/TXNIP axis might play a role in macrophages extracted from rheumatoid arthritis patients. Therefore, we hypothesized that the CDKN2B-AS1/MIR497/TXNIP axis regulated the polarization of rheumatoid macrophages. This study explored the role of the CDKN2B-AS1/MIR497/TXNIP axis in macrophages collected from healthy donors and rheumatoid arthritis patients. Then con rmed the results using CDKN2B-AS1/MIR497/TXNIP axis knockdown and overexpression experiments. Our results reveal a novel role of the CDKN2B-AS1/MIR497/TXNIP in the polarization of rheumatoid macrophages. This study sheds light on the molecular mechanisms of rheumatoid macrophage polarization and provides potential pharmacological targets for rheumatoid arthritis therapy.  [27] . All donors were older than 18 years old. Patients with a rheumatoid arthritis disease duration of 1-3 years and a disease activity score are between 3-10 were included. The Blood samples were collected before the rst treatment. Patients and healthy donors included in this study have been informed and consented to the use of the samples. The clinical information of the patients was provided in the supplementary. The study has been approved by the Ethics Committee of First A liated Hospital of Zhengzhou University. A positive magnetic sorting technique with anti-CD14 antibody was used to sorted CD14 + monocytes out from peripheral blood mononuclear cells. The monocytes obtained were identi ed by ow cytometry using a CD14-FITC antibody.

Flow cytometry
We measured the surface polarization markers using ow cytometry. The methods were described previously [29,30] . Brie y, after blocking, the cells were stained with corresponding antibodies, then cells were xed and permeabilized. Flow cytometry analysis of cells was conducted on a Miltenyi MACSQuant 10 system. FlowJo V10 software was used to analyze the results.

Western Blotting
The expression of TXNIP protein was analyzed using a western blotting assay. The method was described previously [32] . Brie y, samples were lysed in lysing buffer (Pierce, Rockford, IL, USA) with protease inhibitors (Roche, Indianapolis, IN, USA). The proteins were separated in 10-12% sodium dodecyl sulfate-polyacrylamide gels followed by the transferring of the protein to nitrocellulose membranes (Pall Corporation, Port Washington, NY, USA). The membrane then was blocked in 5% skimmed milk in PBS. After the blocking, the membranes were incubated with the primary antibodies overnight at 4°C. Then, the membrane was incubated with secondary antibodies. ECL solution was used to visualize the protein on the membrane.

Statistical signi cance analysis
Data were present in means and standard deviations in the bar charts. A T-test or ANOVA was used to assess the signi cance (p < 0.01). Dunnett's post hoc tests were used to test the difference between groups. GraphPad Prism (version 8) was used to calculate statistics.

Polarization of macrophages contributed to the pathogenesis of rheumatoid arthritis.
This study aimed to explore mechanisms underlying the polarization of rheumatoid macrophages. Firstly, we determined the six surface polarization markers in macrophages from healthy donors (HD) and rheumatoid arthritis patients (RA). These surface polarization markers included CD40, CD80, CD16, CD163, CD206, and CD200R. The macrophages were induced by LPS plus IFNγ, IL10, or IL4 into M1 proin ammatory macrophages, IL10 M2 anti-in ammatory, and IL4 M2 anti-in ammatory macrophages respectively. The surface markers of cells were analyzed by ow cytometry. Compared to cells from healthy donors, cells from rheumatoid arthritis patients expressed a higher level of CD40 and CD80 ( Fig. 1AB), yet, they had a lower level of CD16, CD163, CD206, and CD200R ( Fig. 1C-F). Besides, we con rmed the indication of these biomarkers for polarization states using a commercially available macrophages cell line MD (Fig. 1G).

CDKN2B-AS1/ MIR497/TXNIP axis was involved in the polarization of rheumatoid macrophages
To explore the role of the CDKN2B-AS1/MIR497/TXNIP axis in rheumatoid macrophages, we determined the expression of CDKN2B-AS1, MIR497, and TXNIP mRNA in macrophages using QPCR assay. Results showed that compared to macrophages from healthy donors, macrophages from patients expressed higher CDKN2B-AS1, lower MIR497, and higher TXNIP mRNA ( Fig. 2A-C), indicating that they might play roles in the pathogenesis of rheumatoid arthritis. To analyze the association of the CDKN2B-AS1, MIR497, and TXNIP mRNA, we plotted (1) Correlation between CDKN2B-AS1 level and MIR497 level, (2) Correlation between CDKN2B-AS1 level and TXNIP mRNA level, (3) Correlation between MIR497 level and TXNIP mRNA level in monocytes from healthy donors and rheumatoid arthritis patients respectively.
Results showed that in macrophages from healthy donors there was no correlation among CDKN2B-AS1, MIR497, and TXNIP mRNA. However, in macrophages from rheumatoid arthritis patients, CDKN2B-AS1 was negatively correlated to MIR497, CDKN2B-AS1 was positively correlated to TXNIP, and MIR497 was negatively correlated to TXNIP (Fig. 2D-I).

TXNIP protein expression was correlated to its mRNA level
To con rm the potential involvement of TXNIP protein in rheumatoid arthritis, we used a western blotting assay to analyze TXNIP protein expression. Results showed that compared to macrophages from healthy donors, macrophages from patients expressed a higher level of TXNIP protein (Fig. 3AB). Then we plotted the correlation of TXNIP mRNA level and TXNIP protein level in both healthy donor and patient cells. Results showed that in both healthy donor and patient cells, the TXNIP mRNA level was signi cantly correlated to the TXNIP protein level (Fig. 3CD). This indicated that the TXNIP mRNA level can directly affect its protein level.

CDKN2B-AS1 knockdown suppressed M1 but promoted M2 polarization in MD cells
To validate the regulation of the CDKN2B-AS1/MIR497/TXNIP axis in the polarization of macrophages, we designed knockdown and overexpression experiments of CDKN2B-AS1/MIR497/TXNIP axis in MD cells. In this study, we have con rmed that the biomarkers we tested can indicate the M1 and M2 polarization in our inducted MD cell model (Fig. 1G). Thus, to facilitate the comparison, we normalized the data by multiplying the MD cells without transfection. We started from the upstream to the downstream of the CDKN2B-AS1/MIR497/TXNIP axis. Firstly, we transfected CDKN2B-AS1 shRNA to knock down the CDKN2B-AS1 level. Results showed that CDKN2B-AS1 was successfully knocked down, and this also resulted in an increase of MIR497 and a decrease of TXNIP mRNA and protein (Fig. 4 AB).
After the induction, we found that the knockdown of CDKN2B-AS1 suppressed M1 polarization but promoted M2 (both IL10 and IL4 induced) polarization in MD cells (Fig. 4C-F). These results suggested CDKN2B-AS1 might suppress the in ammatory activities of macrophages.
3.5. MIR497 positively regulated M1 but negatively regulated M2 polarization in MD cells Then we transfected MIR497 mimics to boost the effect of MIR497 and transfected MIR siRNA to knockdown MIR497. Results showed that MIR497 mimics were successfully detected by PCR, and this also resulted in a decrease of TXNIP mRNA and protein (Fig. 5AB). In the MIR497 knockdown experiment, the siRNA successfully knocked down the level of MIR497. Both the mimics and siRNA failed to affect its upstream target CDKN2B-AS1. After the induction, we found that the MIR497 mimics suppressed M1 polarization but facilitated M2 (both IL10 and IL4 induced) polarization in MD cells, while MIR497 siRNA remarkably promoted M1 polarization but signi cantly inhibited M2 (both IL10 and IL4 induced) polarization in MD cells (Fig. 5C-F). These results suggested that MIR497 is the downstream target of CDKN2B-AS1 and the upstream target of TXNIP mRNA. Additionally, MIR497might positively regulate the in ammatory activities of macrophages.

TXNIP suppressed M1 but promoted M2 polarization in MD cells
To validate the regulation of TXNIP in the polarization of macrophages, we overexpressed TXNIP in the cells and measure their polarization. After the transfection of the overexpression vectors, the TXNIP level was signi cantly increased by about 6 times, but this transfection did not affect the levels of CDKN2B-AS1 and MIR497 (Fig. 6 AB). After the induction, we found that the overexpression of TXNIP signi cantly increased the M1 polarization but reduced M2 (both IL10 and IL4 induced) polarization in MD cells (Fig. 4C-F). These results con rmed that TXNIP might promote the in ammatory activities of macrophages.

Discussion
In this study, we determined the polarization of macrophages using six surface polarization markers, including CD40, CD80, CD16, CD163, CD206, and CD200R. CD40 and CD80 were considered biomarkers for M1 macrophages as they are highly expressed in M1 macrophages but not in M2 macrophages [28] .
CD16 and CD163 were regarded as IL10 M2 biomarkers [29] , while CD206 and CD200R were have been wildly detected as IL4 M2 biomarkers [33] . Compared to cells from healthy donors, cells from rheumatoid arthritis patients expressed a higher level of CD40 and CD80, yet, they had a lower level of CD16, CD163, CD206, and CD200R. Our study demonstrated that compared to macrophages from healthy donors, macrophages from rheumatoid arthritis patients have a higher level of M1 polarization, but had a lower level of M2 polarization. We suggested the higher level of pro-in ammatory M1 polarization of cells is associated with the in ammatory in rheumatoid arthritis patients, on the other hand, the lower level of anti-in ammatory M2 biomarkers in rheumatoid arthritis patient cells indicated that the patient cells had a lower in ammatory.
We also explored the role of the CDKN2B-AS1/MIR497/TXNIP axis in rheumatoid macrophages. Surprisingly, in macrophages from healthy donors, there was no correlation among CDKN2B-AS1, MIR497, and TXNIP mRNA levels. These results suggested that CDKN2B-AS1/MIR497/TXNIP axis might not play role in healthy donors. However, in macrophages from rheumatoid arthritis patients, CDKN2B-AS1 was signi cantly negatively correlated to MIR497, CDKN2B-AS1 was signi cantly positively correlated to TXNIP, and MIR497 was signi cantly negatively correlated to TXNIP. These results indicated that CDKN2B-AS1/MIR497/TXNIP axis was activated only in rheumatoid macrophages but not in macrophages from healthy donors. We suggested that when the CDKN2B-AS1, MIR497, and TXNIP at a normal level, they are not able to signi cantly interacted with each other. Only when they are at an abnormal level can they regulate each other. This might also involve some unidenti ed feedback mechanisms that can be explored in the future.
To validate the regulation of the CDKN2B-AS1/MIR497/TXNIP axis in the polarization of macrophages, we interfered with the expression and function of the axis from upstream to downstream. The role of the CDKN2B-AS1/MIR497/TXNIP axis in macrophages was summarized in Fig. 7A. Our study demonstrated that CDKN2B-AS1 regulated TXNIP through MIR497. The complementary binding sites of the CDKN2B-AS1/MIR497/TXNIP axis were shown in Fig. 7B. Notably, because the complementary binding sites only have seven complementary base pairs, the CDKN2B-AS1/MIR497/TXNIP axis might only have relatively weak interaction, which might account for the deregulation of this axis in health donor cells. The sponging of MIR497 by CDKN2B-AS1 can inhibit the level of endogenous MIR497. Most of the miRNAs play roles in cells by their inhibition toward their target mRNA [17] . In this study, as MIR497 can directly suppress the translation of TXNIP mRNA to TXNIP protein, the CDKN2B-AS1/MIR497/TXNIP axis regulated the level of TXNIP. Our study also revealed that TXNIP had a novel regulation role in the polarization of macrophages. TXNIP can facilitate M1 pro-in ammatory polarization but inhibited M2 anti-in ammatory polarization in macrophages, thereby positively regulated the in ammation in rheumatoid arthritis.
However, the monocytes in this study were in vitro differentiated, which were different from the macrophages obtained from patients that might be affected by the internal environment of the patient. Yet, this study is focusing on the gene expression and the roles of the CDKN2B-AS1/MIR497/TXNIP axis. We think the effect of the genotype of the cells was more critical than the internal environment that induced the cell differentiation. We presumed that the cells from patients have a genotype that results in the onset of rheumatoid arthritis and the genotype would not be affected by the internal environment of the patient. Thus, we only studied in vitro differentiated cells. In addition, the mechanism of TXNIP regulating the polarization of macrophages has not been studied in this study. One potential mechanism involves might be the ion channel regulation in macrophages. Ion channels have been found to be critical Page 9/20 in many cells [34][35][36][37] , some macrophages expressed ion channels for their functions [38] . TXNIP has been found to be involved in some ion channel regulations [39] . Regarding the clinical treatment of rheumatoid arthritis, many anesthetics have been applied in therapy to relieve rheumatoid arthritis pain [40,41] . These drugs usually have multiple ion channel targets [42,43] . Many anesthetics can also have multiple ion channel targets [42] and regulate the immune system [44,45] that might interfere with rheumatoid arthritis. Whether the use of these anesthetics interacts with macrophages in rheumatoid arthritis is not clear. In addition, the potential role of the CDKN2B-AS1/MIR497/TXNIP axis in other bone diseases is also interesting. Studies have found that TXNIP is closely associated with bone metabolism [46][47][48] .
Rheumatoid arthritis and osteoporosis have been two of the most pervasive bone diseases [49,50] . For example, many pathways have been proposed to regulated osteoporosis [51][52][53][54] , but whether osteoporosis can be regulated by the CDKN2B-AS1/MIR497/TXNIP axis can also be explored in the future.

Conclusion
Generally, through studying macrophages from human blood samples and a cultured macrophage cell line, we demonstrate that elevated CDKN2B-AS1 in macrophages promotes M1 polarization and inhibits M2 polarization of macrophage by negatively regulating MIR497 for the upregulation of TXNIP. Our results revealed a novel role of the CDKN2B-AS1/MIR497/TXNIP axis in the polarization of rheumatoid macrophages and shed light on molecular mechanisms of rheumatoid macrophage polarization. This study also provided potential pharmacological targets for rheumatoid arthritis therapy.

Con ict of Interest Statement
There is no con ict of interest.

Availability of data
The data that support the ndings of this study are available from the corresponding author upon reasonable request. between negative and overexpressed groups was tested using a t-test, *p-value < 0.01.)