*2.2. Upregulation of miR-941 in CD14*+ *Monocytes from PsA Patients by qRT-PCR, with Support Vector Machine Learning, Identified miR-941 as an Early Predictor for PsA*

In order to screen specific miRNA expressions in patients with PsA, we first profiled the miRNA expression from two pooled samples from HCs (from two males and one female, respectively), two pooled samples from PsA patients (from two males and one female, respectively) and one pooled sample from PsO patients (from two males). We collected approximately 32.0 million raw reads in total, and each sample yielded 6.4 million reads on average. Twelve miRNAs with transcripts per million (TPM) values higher than 1000 in at least one sample and with average expression ratios relative to HCs > 1.5 were identified. The overall expression levels of these 12 miRNAs from PsO, PsA, and HC monocytes are presented in Figure 1A,B. Notably, expression of miR-941 and miR-146b-5p were 1.5-fold greater in the PsA group compared to that in both HC and PsO group. To validate the results revealed by NGS, the expression of miR-941 and miR-146b-5p were measured by quantitative real-time reverse transcription polymerase chain reaction (qRT-PCR) in CD14+ monocytes from 40 PsA, 40 PsO, and 40 HC subjects. While expression level of miR-146b-5p was similar among groups (Figure 1C), miR-941 expression was significantly greater in PsA patients than that in both PsO patients and HCs (Figure 1D). MiR-941 expression was significantly higher in CD14+ monocytes derived from PsA patients compared to PsO (ORs: 2.87 (95% CI: 1.62–5.06), *p* < 0.001) or HCs (ORs: 3.85 (95% CI: 2.12–6.98), *p* < 0.001). Therefore, miR-941 could be a potential biomarker able to identify PsA patients from PsO or HCs. To address this, we used support vector machine (SVM) learning [15,16] to calculate the discrimination power of miR-941 with an SVM model. The result showed the area under the receiver operating curve (auROC) was 0.79 (Figure 1E), confirming that miR-941 expression alone is able to distinguish PsA patients from PsO or HCs.

**Figure 1.** High expression of miR-941 in CD14+ monocytes from PsA patients. (**A**) The RNA samples were analyzed from (1) the healthy control group (HC), including two pooled samples from two males and one female, respectively; (2) the psoriasis group (PsO), including one pooled sample from two males; and (3) the psoriasis arthritis group (PsA), including two samples from two males and one female, respectively. MicroRNAs (miRNAs) with transcript per million (TPM) values more than 1000 in at least one sample and with average expression ratio (normal to disease or disease to normal) more than 1.5 are plotted on the heat map. (**B**) The microRNA expression level was presented in the unit of TPM. (**C**,**D**) The expression levels of miR-146b-5p and miR-941 were measured in HCs (*n* = 40), PsO patients (*n* = 40) and PsA patients (*n* = 40) by quantitative real-time reverse transcription polymerase chain reaction (qRT-PCR). Patients with PsA showed increased expression of miR-941 in CD14+ monocytes. (**E**) We set training model from miR-941 expression of CD14+ monocytes from patients with PsA, PsO, and HCs using a Support Vector Machine (SVM) learning algorithm. To distinguish PsA from PsO and HC, the value of auROC was 0.79, indicating that miR-941 expression can distinguish PsA patients from PsO and HCs with high accuracy. \*\*\* *p* < 0.001.

#### *2.3. Enhanced Osteoclast Activation and Bone Resorption in PsA Patients with Increased miR-941 Expression*

Given that miR-941 expression is elevated in CD14+ monocytes from PsA patients, we first examined whether these cells exhibited enhanced capacity to differentiate into osteoclasts and demonstrated enhanced resorption activity. The result showed that osteoclast formation was higher in CD14+ monocytes from PsA patients than that from PsO patients or HCs (12.7 ± 1.2, 6.3 ± 0.8, and 5.5 ± 0.9/per HPF from 10 PsA patients, 10 PsO patients, and 10 HCs, respectively, *p* < 0.001; Figure 2A,C). Regarding the bone resorption, the average percent area of resorption pits in dentine slices was also significantly higher in the differentiated osteoclasts from PsA patients than that from PsO patients or HCs (7.2 ± 2.0%, 2.9 ± 2.1%, 1.8 ± 1.8%/per HPF from average of 10 PsA samples, 10 PsO samples and 10 HC samples, respectively, *p* < 0.001; Figure 2B,D), indicating the enhanced osteoclast formation and active resorption activity from monocytes in PsA patients. To examine if miR-941 is

important for osteoclastogenesis in PsA, we measured the expression of miR-941 in osteoclast from CD14+ monocytes of PsA, PsO patients and HCs. The expression of miR-941 in monocytes and osteoclasts from PsA patients are both higher than that from PsO patients (*p* < 0.05) or HCs (*p* < 0.05) (Figure 2E). These results suggest that miR-941 upregulation is in parallel with the elevated osteoclast differentiation potential in PsA patients.

**Figure 2.** PsA patients with higher miR-941 expression demonstrated increased osteoclast differentiation potential and resorption activity derived from CD14+ monocytes. CD14+ monocytes from PsA patients, PsO patients, and HCs were cultured with macrophage colony-stimulating factor (M-CSF) 20 ng/mL for three days, followed by nine days in 100 ng/mL receptor activator of nuclear factor-κB ligand (RANKL) and 100 ng/mL tumor necrosis factor-α (TNF-α) for osteoclast differentiation. For evaluation of resorption activity, the osteoclasts were cultured on dentine slices. (**A**) At day 13, cells were stained with tartrate-resistant acid phosphatase (TRAP) to calculate the number of CD14+ monocyte-derived osteoclasts from PsA patients and NCs. Scale bar: 50 μm. (**B**) At day 13, resorption pits made by osteoclasts from psoriatic patients and NCs were recorded under a bright field microscope. Scale bar: 50 μm. (**C**) Osteoclast formation from CD14+ monocytes was compared between 10 HCs, 10 PsO patients, and 10 PsA patients. The number of osteoclasts was quantified among the three groups. (**D**) The eroded surface area on the dentine slices from PsA patients (*n* = 10), PsO patients (*n* = 10), and HCs (*n* = 10) were quantified using ImageJ and expressed as % of total dentine slice area. (**E**) The expression of miR-941 in monocytes and osteoclasts from PsA patients are higher than that from HCs and PsO patients using qRT-PCR. \* *p* < 0.05; \*\* *p* < 0.01 and \*\*\* *p* < 0.001.

#### *2.4. miR-941 Inhibition Abolished the Osteoclast Activation and Functional Resportion in PsA Patients*

To address the direct involvement of miR-941 in the enhanced activation and functional resorption of osteoclasts derived from PsA patients, we examined whether miR-941 inhibition could reverse these effects. Peripheral CD14+ monocytes were obtained from 10 PsA patients and treated with M-CSF for 72 h. Cells were then divided into three treatment groups—one transfected with a control microRNA inhibitor (mock transfection control), one with a miR-941 inhibitor, and one without transfection. After transfection, cultures were treated with TNF-α and RANKL every three days for nine days to induce osteoclast formation. At day 13, the number of osteoclasts formed was significantly lower in the miR-941 inhibitor group (3.0 ± 1.1/HPF) than that in the non-transfected and mock transfected control groups (12.1 ± 3.4/HPF and 11.4 ± 2.2/HPF, respectively; *p* < 0.01) (Figure 3A,B). Similarly, resorption activity was significantly lower in the group transfected with miR-941 inhibitor (2.8 ± 1.2%) than that in the non-transfected and mock transfected control groups (7.2% ± 2.0% and 7.4% ± 2.0%, respectively; *p* < 0.01) (Figure 3A,C). The miR-941 inhibition abrogated the osteoclast development in PsA. The result also showed that microRNA-941 expression was reduced to 50% of baseline in the miR-941 inhibitor group as compared to that in the non-transfected and mock-transfected control miRNA groups (Figure 3B,C). These results indicated that the miR-941 inhibitor in CD14+ monocytes may reverse the active osteoclastogenesis and resorption activity in PsA patients.

**Figure 3.** miR-941 inhibitor reversed osteoclast differentiation and bone resorption activity in patients with PsA. The CD14+ monocytes were obtained from eight PsA patients and were subsequently activated into osteoclasts with treatment of M-CSF for 72 h. Cells were either nontransfected or transfected with the control-microRNA inhibitor or miR-941 inhibitor followed by TNF-α and RANKL very three days for nine days to study their ability to form osteoclasts and their activity of bone resorption. At day 13, the number of osteoclast formations and percentage of resorption pits were measured. (**A**) The cells were stained with TRAP to calculate the number of osteoclasts among non-transection, negative control miRNA and miR-941 inhibitor transfection groups, Scale bar: 50 μm. For the evaluation of resorption activity, the resorption pits were recorded by a bright field microscope, Scale bar: 50 μm. (**B**) The number of osteoclasts was quantified among the three groups (blue). The relative expression of miR-941, as standardized to the level of non-transfected osteoclasts, was depicted as orange color. (**C**) The eroded surface areas on the dentine slice were quantified using ImageJ software as percentage of expressions in total areas (blue line). The expressions of miR-941 in the osteoclasts were measured by qRT-PCR to evaluate the effect of miR-941 inhibitors transfection. The relative expression of miR-941, as standardized to the level of non-transfected osteoclasts, was depicted as orange color. \*\* *p* < 0.01.
