*2.5. E*ff*ecst of PADI2 Knockout on the Protein and mRNA Expression of Adhesion-Related Genes*

In the U937 cells and the differentiated macrophages, we found that the protein expression levels of phospho-FAK, FAK, phospho-paxillin, paxillin, and PAK1 were all decreased in the PADI2 knockout group compared with in the control group (Figures 5 and 6). For the gene expression of the adhesion-related proteins, we found that there were no statistically significant differences in the mRNA expression of *FAK*, *paxillin*, or *PAK1* between the PADI2 knockout group and the control group in the U937 cells and the differentiated macrophages.

**Figure 5.** Effects of PADI2 knockout in the protein and mRNA expression of focal adhesion kinase (*FAK*), *paxillin*, and *PAK1* in the U937 cells: (**A**) comparison of the relative protein expression levels of FAK and phospho-FAK in the PADI2 knockout group and the control group; (**B**) comparison of the relative protein expression levels of paxillin and phospho-paxillin in the PADI2 knockout group and the control group; (**C**) comparison of the relative protein expression levels of PAK1 in the PADI2 knockout group and the control group; and (**D**) Representative images showing the relative protein expression levels of FAK, paxillin, and PAK1 in the PADI2 knockout group and the control group; (**E**) the relative mRNA expression levels of *FAK*, *paxillin*, and *PAK1* in the U937 cells. In the U937 cells, the protein expression levels of the cell adhesion-related proteins, including FAK, phospho-FAK, paxillin, phospho-paxillin, and PAK1, were performed by Western blotting in the PADI2 knockout group and the control group. The relative protein expression was defined as the ratio of specific protein/hsp90 band intensity.

**Figure 6.** Effects of PADI2 knockout in the protein and mRNA expression of *FAK, paxillin*, and *PAK1* in the differentiated macrophages: (**A**) comparison of the relative protein expression levels of FAK and phospho-FAK in the PADI2 knockout group and the control group; (**B**) comparison of the relative protein expression levels of paxillin and phospho-paxillin in the PADI2 knockout group and the control group; (**C**) comparison of the relative protein expression levels of PAK1 in the PADI2 knockout group and the control group; and (**D**) Representative images showing the relative protein expression levels of FAK, paxillin, and PAK1 in the PADI2 knockout group and the control group; (**E**) the relative mRNA expression levels of *FAK*, *paxillin*, and *PAK1*. In differentiated macrophages, the protein expression levels of the cell adhesion-related proteins including FAK, phospho-FAK, paxillin, phospho-paxillin, and PAK1 were performed by Western blotting in the PADI2 knockout group and the control group. The relative protein expression was defined as the ratio of a specific protein/hsp90 band intensity.

#### **3. Discussion**

Vossenaar et al. [14], Hojo-Nakashima et al. [15], and our previous study [7] have shown that PADI2 protein levels increased during the macrophage differentiation, resulting in increasing protein citrullination. In the current study, we demonstrated that PADI2 is essential for macrophage proinflammatory cytokine secretion, cell adhesion, and apoptosis using CRISPR/Cas9-mediated knockout of PADI2. As for the target proteins of PADI2, we found the histone H3 citrullination did not change in the U937 cells, the differentiated macrophages, or the macrophages stimulated with LPS in the PADI2 knockout group compared with in the control group. Darrah et al. reported that histone H3 was prone to be citrullinated by PADI4, which might explain our finding [16]. Different antibodies used in diffident studies might detect different epitopes, which could affect the results of histone H3 citrullination. However, we demonstrated that the protein level of citrullinated NF-κB p65 was decreased in the PADI2 knockout group compared with in the controls. Sun et al. showed that the citrullination of NF-κB p65 could enhance *IL-1*β and *TNF-*α expression [11]. Our result also showed that the expression levels of *IL-1*β and *TNF-*α were indeed decreased in the U937 cells after PADI2 knockout. The differentiation of macrophage and further stimulation with LPS could decrease the *IL-1*β and *IL-6* expression in the PADI2 knockout group compared with in the controls. Most importantly, we found that the IL-1β, IL-6, and TNF-α concentrations in the culture soup were dramatically decreased in the PADI2 knockout group compared with in the controls. In addition, Sun et al. showed citrullinated NF-κB p65 was mediated by PADI4 using HeLa cells and neutrophils, which expressed high levels of PADI4 compared with those of monocytes and lymphocytes [6]. In the current study, we used U937 cells, a representative cell line for human monocytes. We demonstrated that PADI2 was also required to citrullinated NF-κB p65 in macrophages upon LPS stimulation. Lee et al. showed that PADI2 could interact with an inhibitor of nuclear factor kappa-B kinase subunit gamma (IKKγ) to suppress NF-κB activity using a murine cell line [17]. However, our results suggested that NF-κB activity decreased in PADI2 knockout U937 cells from a decreased expression of proinflammatory cytokines. Different cell lines used in different studies might explain these variations. Mishra et al. demonstrated that PADI2 and PADI4 activity in macrophages were required for inflammasome assembly and IL-1β release in a murine model [18], which is consistent with our findings. Since proinflammatory cytokines including IL-1β, IL-6, and TNF-α play a critical role in the immunopathogenesis of RA [19,20] and increased PADIs activities have been documented in patients with RA [21], targeting PADIs would be a novel strategy for RA treatment.

We found that PADI2 knockout significantly decreased the apoptotic rates in the U937 cells, the differentiated macrophages, and the macrophages stimulated with LPS. Liu et al. demonstrated the overexpression of PADI4-induced cell apoptosis in human leukemia (HL)-60 cells and human acute T leukemia Jurkat cells [22]. These findings suggested that both PADI2 and PADI4 could promote apoptosis. For the detail mechanism of cell apoptosis, Liu et al. found that PADI4 induced apoptosis mainly through cell cycle arrest and a mitochondria-mediated pathway [22]. Our study showed that PADI2 could activate caspase-2 and -9, leading to the activation of capases-3. Further studies are needed to clarify their molecular mechanisms.

We noted that the inhibition of PADI2 impaired the protein expression of phospho-FAK, FAK, phospho-paxillin, paxillin, and PAK1, resulting in impaired cell adhesion ability in macrophages. We could not detect the phosphorylation of PAK1, which was also not detected in macrophages in a previous study [23]. PAK1 belongs to one of the members of the PAKs family, which plays an important role in cell motility [24], and the phosphorylation of PAK1 is critical for cell migration instead of adhesion [25]. We noted that the protein levels, but not mRNA expression levels of FAK, paxillin, and PAK1, were decreased in the PADI2 knockout group compared with in the controls. We speculated that the citrullination of proteins might accelerate their degradation by changing their binding affinity to proteasome [7,26] or altered protein structure [27], and further studies are needed.

In conclusion, our study showed that PADI2 is essential for the multiple functions of macrophages in enhancing inflammatory cytokines production through the citrullination of NF-κB p65, promoting cell apoptosis with capase-2, -3, and -9 activation, and facilitating cell adhesion via FAK, paxillin, and PAK1.Targeting PADI2 could be a novel strategy for controlling inflammation triggers by macrophages.

#### **4. Material and Methods**

### *4.1. Purification of Anticitrullinated Protein Antibodies (ACPA) from Pooled ACPA(*+*) Sera in Patients with RA*

The study protocol was approved by the institutional review board of Dalin Tzu Chi Hospital, Buddhist Tzu Chi Medical Foundation (No. B10902001, 1 April 2020). The study was performed in accordance with the Declaration of Helsinki. Serum samples from ACPA-positive RA patients with high concentration of ACPAs (>340 IU/mL) and aged 20 years and above, which fulfilled the

2010 American College of Rheumatology (ACR)/European League Against Rheumatism (EULAR) criteria [28], were detected using an ELISA kit (Pharmacia Diagnostics AB, Uppsala, Sweden), collected and pooled. ACPAs were purified according to the method described previously [29]. In brief, the pooled sera containing high concentration of ACPAs from patients with RA were purified by affinity chromatography using an ÄKTA purifier 10 (GE Healthcare, Little Chalfont, UK) with UV detection at 280 nm for the collection of the desired fractions.
