Serum Uric Acid as a Potential Biomarker in Patients With Rheumatoid Arthritis-associated Interstitial Pneumonia (RA-ILD)

Background: Uric acid (UA) has been reported participate in various inammatory and autoimmune diseases. Increasing evidence has shown that UA also plays an important role in lung inammation and brosis. We aimed to investigate the correlation between UA and rheumatoid arthritis (RA), especially rheumatoid arthritis-associated interstitial lung disease (RA-ILD). Methods: 266 RA patients and 138 healthy individuals were recruited in this study. RA was identied according to ACR/EULAR 2010 criteria. UA in serum and bronchoalveolar lavage uid (BALF), as well as clinic and laboratory Indexes were collected from participants enrolled in the study. Serum KL-6 was measured via ELISA and then Spearman correlation analysis was used to analysing their association. Subsequent the receiver operating characteristic (ROC) curve and area under the curve (AUC) were used to estimate the diagnostic value of UA in RA-ILD and UIP pattern of RA-ILD. HRCT and pulmonary function tests (PFT) were performed to evaluated the characteristics and pulmonary ndings of RA patients. Furtherly, we validated the effect of UA on the EMT and the expression of cytokines in A549 cells. Results: Compared with healthy controls (HC), UA levels in serum was signicantly higher in RA group (262.25±91.94 vs. 234.69±54.78µmol/L, P<0.01). Spearman correlation analysis revealed associations of SUA with risk factors of RA, including RF, anti-CCP, CRP and ESR (r=0.37, p=0.02; r=0.48, p<0.01; r=0.26, p<0.01; r=0.27, p<0.01). And higher UA was measured both in serum (291.81±102.42 vs. 252.38±6.15µmol/L, P<0.01) and BALF (393.3±222.6 vs. 204.0±120.3µmol/L, p<0.01) of RA-ILD patients, particularly those with UIP pattern (475.58±249.28 vs. 262.86±103.45µmol/L, p<0.01; 393.3±222.6 vs. 204.0±120.3µmol/L, p<0.01). Meanwhile, the correlation between the level of UA in serum and BALF and serum KL-6 concentration in RA were also signicant (r=0.59, p<0.01; r=0.43, p<0.01). Also, the negative correlations of UA level, both in serum and BALF, with lung function parameters including FEV1/FVC and FEV1% predicted were measured (r=-0.78; r=-0.87, p<0.01). In the ROC curve analysis of AUC, the AUC value of UA was 0.76 (95%CI=0.66-0.87, p<0.01). The sensitivity and specicity were 60% and 92%, respectively. The vitro experiment showed that UA stimulated the EMT in A549 cells, as well as induced the expression of cytokines, such as IL-1, IL-6 and TGF-β, in lung epithelial cell. Conclusions: This study suggests that UA is correlated with the ILD in RA. Particularly, the higher UA levels may be related to UIP, a pattern with worse prognosis, in patients with ILD. Therefore, UA may be an important contributing factor to the pathogenesis of RA-ILD. RF: rheumatoid factor; anti-CCP: anti-cyclic peptide containing citrulline; ESR: erythrocyte sedimentation rate; CRP: C-reactive protein; UA: uric acid; KL-6: Krebs von den Lungen-6; NSAIDs: Nonsteroidal Antiinammatory Drugs; DMARDs: disease-modifying anti-rheumatic drugs; DAS 28: 28-joint Disease Activity Score.


Introduction
Uric acid (UA) is a nal catabolite of purine found in organisms. In recent years, interest in uric acid as a regulator of in ammation and immune responses has grown [1,2]. Previous publications suggested that UA has the ability to stimulate the expression of C-reactive protein (CRP), brinogen and complement C3 in a dose-dependent manner [3]. Growing evidence indicates that serum uric acid is closely associated with circulating levels of interleukin-6 (IL-6), CRP and tumor necrosis factor-α (TNF-α), which implies that UA may play a critical role in systemic in ammation and in ammatory-related diseases, such as rheumatoid arthritis (RA) [4]. Moreover, a series of studies have reported that uric acid crystals are not just innocent bystanders but may be active players in the proliferation and differentiation of T cells that participate in the pathology of in ammatory disorders and autoin ammatory diseases. The experimental data demonstrated that UA coordinates with in ammasome-dependent cytokines to contribute to the differentiation of Th17 cells both in vitro and in vivo [5]. The correlation between Treg cell numbers and UA in systemic lupus erythematosus (SLE) has also been reported [6]. In addition, UA has been con rmed to be a considerable initiator and ampli er of Th2 cell function during immunity and allergic in ammation and to be an in ammatory mediator in allergic asthma [7]; additionally, it has been proposed that the local accumulation of UA, as an endogenous danger signal, can activate the Nod-like receptor protein 3 (NALP3) in ammasome, promoting the production of IL-1β, thereby contributing to lung in ammation and brosis [8]. Of note, increasing evidence has shown that a high level of UA, which is associated with both the presence of pulmonary hypertension (PH) and higher mortality, has prognostic value in ILD [9]. Furthermore, recent studies veri ed that uric acid crystals formed at the lung injury site might represent a key danger signal for activating the in ammasome to release in ammatory cytokines, thereby causing in ammatory lung pathologies [8]. Moreover, increased UA serum concentrations in patients with lung disease of different etiologies correlate with the prognosis and severity of the disease [10,11].
Rheumatoid arthritis (RA) is a chronic systemic in ammatory disease characterized by progressive and erosive destruction by arthritis accompanied by damage to multiple organs [12]. Interstitial lung disease (ILD), the most common extra-articular complication of RA, is associated with signi cant mortality and is a key factor affecting the prognosis of RA patients [13,14]. Since RA-associated ILD (RA-ILD) has a high mortality rate, clarifying its pathogenesis and treatment strategies is necessary. Recently, some documents con rmed a novel role for the danger signal uric acid in driving proin ammatory Th17 differentiation, and signi cant relationships of systemic sclerosis (SSc) with pulmonary brosis and elevated UA levels have been reported [15]. Although the potential function of UA in lung disease has increasingly appreciated, there is little information available on the clinical signi cance of UA in RA patients with ILD. Therefore, we hypothesize that UA might be engaged in the process of RA and is associated with the development of rheumatoid arthritis-associated interstitial lung disease (RA-ILD), and an assessment of UA may have clinical implications for identifying ILD with UIP patterns in RA patients. Herein, we analyzed the level of uric acid in RA-ILD patients and investigated the relationship between UA and other related laboratory indicators in RA-ILD patients. The data indicated that serum UA levels were correlated with the disease severity of RA-ILD patients, especially those with a UIP pattern.

Study population
In our study, 266 patients (> 18 years old) were recruited according to the 2010 ACR/EULAR criteria for de nite RA at The Second A liated Hospital of Nanjing Medical University between January 2018 and December 2020. We excluded patients suffering from other autoimmune diseases, such as systemic lupus erythematosus (SLE) and dermatomyositis (DM), acute infectious disease, gout, and severe hepatic and renal dysfunction, or those taking uric acid-lowering drugs. Current or former smokers were also excluded. General condition, past medical history, smoking status and duration of RA were obtained from medical clinical records. The ILD of all patients collected in this study was de ned according to HRCT evidence. Meanwhile, the patients were evaluated with pulmonary function tests (PFTs) and bronchoalveolar lavage (BAL). Additionally, 138 individuals matched by age and sex were also included as healthy controls (HCs).
All subjects who participated in this study voluntarily and provided written informed consent. The study protocol was approved by the medical ethics committee of The Second A liated Hospital of Nanjing Medical University.

Clinical and laboratory assessments
We registered the patients' general condition, including age, sex, duration of RA and comorbidities. Laboratory indexes such as C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), presence and plasma level of rheumatoid factor (RF) and anti-cyclic peptide containing citrulline (anti-CCP) were collected. Serum samples of patients were analyzed for UA level and the concentration of Krebs von den Lungen-6 (KL-6). All basic demographic and clinical characteristics of the participants are shown in Table. Bronchoalveolar lavage (BAL) Based on current evidence, BAL cellular analysis might be a supplementary means for the diagnostic evaluation of ILD. All RA patients included in this study underwent BAL, bronchoalveolar lavage uid (BALF) was collected and then UA concentrations were measured as described previously [16].

Measurement of Krebs von den Lungen-6 (KL-6)
The concentrations of KL-6 in serum were measured by an enzyme-linked immunosorbent assay (ELISA) kit according to previously described procedures [17].

Radiographic classi cation
According to the American Thoracic Society/European Respiratory Society (ATS/ERS) classi cation of idiopathic interstitial pneumonias (IIPs) [18,19], the HRCT patterns of RA-ILD were categorized into two major subgroups, namely, serum UAl interstitial pneumonia (UIP) and nonspeci c interstitial pneumonia (NSIP). Separately, the UIP pattern was characterized by traction bronchiectasis/bronchiectasis, reticular opacity and honeycombing, while the HRCT characteristic of NSIP was extensive ground glass opacity [20]. HRCT scans of each patient were independently reviewed by two trained thoracic radiologists who were blinded to all clinical information such as demographics and general state. Disagreements over the diagnosis were resolved by consensus.

Pulmonary function tests (PFTs)
Patients with RA-ILD who underwent PFTs ful lled the American Thoracic Society (ATS)/European Respiratory Society (ERS) guidelines [21][22][23], including forced vital capacity (FVC) and forced expiratory volume in 1 s (FEV1). All the data are presented as percentages of the normal predicted values. Moreover, we calculated the values of FEV1/FVC and FEV1% predicted.
Cell culture, treatment and analysis A549 cell line was used to resemble type II alveolar epithelial cells [24]. A549 cells were cultured in vitro.
Before UA treatment, A549 cells were maintained in serum-free media for 12 h prior to stimulation and then changed to complete media with increasing concentrations of uric acid (UA, 0, 200 and 400µmol/L) for 24 h. Supernatants were then collected and IL-1, IL-6, and TGF-β concentrations were determined by ELISA.

Real time Reverse transcription semiquantitative polymerase chain reaction (Real-Time RT-PCR).
Total RNA was extracted from A549 cells by TRIzol® RNA isolation (Gibco; Thermo Fisher Scienti c, Inc.) and puri ed with DNase I (Invitrogen; Thermo Fisher Scienti c, Inc.). according to the manufacturer's protocol. The primer sequences are available upon request. RNAs were reverse-transcribed into cDNA using SuperScriptTMII (Invitrogen Life Technology). Real-time quantitative PCR was performed by uorescent dye SYBR Green methodology using SYBR Green PCR Master Mix (Applied Biosystems) and the ABI Prism 7000 apparatus (Perkin-Elmer, Foster City, CA). Gene expression was normalized to the corresponding β-actin level and is presented as the fold change relative to that of the control.
A549 cells were seeded in 6-well plates. After the indicated treatment, the cells were xed and incubated with primary antibodies against α-SMA (Santa Cruz Biotechnology, Inc.) and E-cadherin (Santa Cruz Biotechnology, Inc.) overnight at 4°C, after which the cells were washed three times with PBS. The cells were stained as described in the online Supplementary Information [25].

Statistical analysis
All statistical analyses were carried out using GraphPad Prism version 8.0.2. Differences between groups were analyzed by the Student's t-test. Comparisons of categorical variables were conducted using Pearson chi-square tests. For nonparametric data, the results were expressed as median (range) values, and the differences between groups were analyzed by the Mann-Whitney U test. Spearman correlation analysis was performed to analyze the association of UA with clinical and laboratory indexes in RA patients. Univariate logistic regression analysis was performed to determine the factors associated with the presence of ILD. Multivariate logistic regression analysis was performed by including the confounding factors that were found to be signi cantly associated with the univariate analyses. The receiver operating characteristic (ROC) curve and area under the curve (AUC) were used to assess the validity and value of UA for RA-ILD. Any difference with a p value < 0.05 was regarded as statistically signi cant.

Results
Relationship of serum UA level with laboratory indexes of RA patients The demographic characteristics, the UA level in serum and other laboratory values of RA patients are presented in Table 1. To investigate the relationship between the serum levels of UA and RA, we compared the demographics and serum UA level of the RA and control groups. The serum UA level in the RA patients was signi cantly higher than that in the healthy controls (262.25 ± 91.94 vs. 234.69 ± 54.78 µmol/L, P < 0.01) (Fig. 1A). However, there was no signi cant difference in sex or age between the two groups. As shown in the Spearman correlation analysis, serum UA levels were positively correlated with RF and anti-CCP in patients with RA (r = 0.37, p = 0.02; r = 0.48, p < 0.01) ( Fig. 2A, B). Next, we evaluated the strength of the association between serum UA and CRP and ESR. The results showed that the serum UA level was apparently correlated with CRP and ESR (r = 0.26, p < 0.01; r = 0.27, p < 0.01).
The UA level was elevated, especially in RA patients with ILD, and correlated with the KL-6 level Based on the above data of the RA and healthy groups, as well as the Spearman correlation analysis results, we concluded that UA may be an important predictor of RA. Furthermore, we explored whether the UA level was associated with RA-ILD in RA patients. The level of serum UA was signi cantly higher in patients with RA-ILD than in those without ILD (291.81 ± 102.42 vs. 252.38 ± 6.15 µmol/L, p < 0.01) (Fig.  1B). In addition, more UA was detected in the BALF of the RA-ILD group than in the BALF of the non-ILD group (393.3 ± 222.6 vs. 204.0 ± 120.3 µmol/L, p < 0.01) (Fig. 1D). To further verify that UA levels were associated with lung injury markers in RA-ILD patients, we subsequently measured the concentration of KL-6, a biomarker of ILDs[26], in the serum of RA patients and examined the relationship between the levels of UA and KL-6 by correlation analysis. The results showed that serum UA levels positively correlated with KL-6 in patients with RA (r = 0.59, p < 0.01) (Fig. 2C). Similarly, the UA level in BALF also had a powerful correlation with serum KL-6 in RA patients (r = 0.43, p < 0.01) (Fig. 2D).
High serum and BALF UA levels were correlated with the UIP pattern and the progression of RA-ILD The above data showed that the UA level was higher in RA patients with ILD than in patients without ILD, which meant that high UA levels may be implicated in the onset of ILD in RA. Clinically, RA-ILD manifests many well-recognized phenotypes, in which the usual interstitial pneumonia (UIP) pattern is the most common forms of RA-ILD with a worse prognosis [14]. Thus, the distinction of UIP in RA-ILD patients is of important clinical value for guiding the treatment of RA-ILD patients. In view of the aforementioned ndings of UA levels in RA-ILD, we hypothesized that UA levels may be associated with the UIP pattern and the progression of RA-ILD. Based on the HRCT scans, 40 subjects with UIP patterns and 64 subjects with NSIP patterns were identi ed among the RA-ILD patients. Compared with the NSIP group, the level of serum UA in RA-ILD patients with the UIP pattern was markedly higher (475.58 ± 249.28 vs. 262.86 ± 103.45 µmol/L, p < 0.01) (Fig. 1C). Interestingly, the UA level in BALF was signi cantly higher in RA-UIP patients than in RA-ILD patients with the NSIP pattern (393.3 ± 222.6 vs. 204.0 ± 120.3 µmol/L, p < 0.01) (Fig. 1E). To investigate whether the UA level was correlated with a decline in lung function in RA-UIP patients, we evaluated the lung function of RA-ILD patients with the UIP and NSIP patterns. The results showed that the UA levels in serum and BALF were negatively correlated with FEV1/FVC (r=-0.78, P < 0.01; r=-0.46, P < 0.01) (Fig. 3A, C) and FEV1% predicted (r=-0.87, p < 0.01; r=-0.47, p < 0.01) (Fig. 3B, D). These data suggest that UA may be a signi cant contributor to the pathogenesis of RA-ILD.

Serum UA levels can be a disease marker for the detection of RA-ILD with UIP
To evaluate the diagnostic value of the serum UA level in clinical settings, we established an ROC curve and examined the sensitivity and speci city of the circulating factors in identi ed RA patients with UIP.
The area under the curve (AUC) was 0.661 (95% CI: 0.542-0.780; p < 0.001), with a sensitivity and speci city of 51.7% and 79.3%, respectively. (Fig. 3E). Thus, these results suggest that serum UA may be a positive biomarker for distinguishing the UIP pattern in RA-ILD patients, meaning that it can be a signi cant biomarker candidate for appraising the disease progression and severity of ILD in RA patients.

UA induces epithelial-to-mesenchymal transition (EMT) in A549 cells
Emerging evidence has demonstrated that EMT phenomena are potential contributors to ILD, that alveolar epithelial cells from interstitial lung diseases undergo epithelial-to-mesenchymal transitions after lung injury, and that these alveolar epithelial cells undergo EMT, which includes loss of their epithelial biomarkers and acquisition of mesenchymal ( broblast-like) cell biomarkers [27]. To assess whether UA could induce EMT in A549 cells, we performed immuno uorescence staining of EMT markers, including E-cadherin and α-SMA, after A549 cells were incubated with UA (0, 200 and 400 µM) for 24 h. We found that UA decreased the expression of the epithelial cell marker E-cadherin in a dose-dependent manner (Fig.  4A, B). In contrast, UA augmented the expression of the mesenchymal cell marker α-SMA in a dosedependent manner (Fig. 4C, D). These ndings revealed that high UA was shown to induce EMT in A549 cells.
The production of cytokines was induced by high UA in A549 cells It has been demonstrated that these proin ammatory cytokines exert direct proin ammatory effects on alveolar and bronchial epithelial cells, which are well-known lung pathology promoters [28,29]. Next, we explored whether UA is capable of stimulating the release of cytokines from A549 cells. A549 cells were treated with increasing concentrations of UA (0, 200 and 400 µM), and supernatants of A549 cells stimulated with UA were harvested for ELISA analysis. As shown in Fig. 5A.B.C, the production of IL-1, IL-6 and transforming growth factor-β (TGF-β) by A549 alveolar epithelial cells was markedly increased by UA in a dose-dependent manner compared with that of the control group, both at 24 h after treatment. A549 cells treated with increasing concentrations of UA were also harvested for quantitative RT-PCR analysis.
As shown in Fig. 5D.E.F, the mRNA expression of IL-1, IL-6 and TGF-β was upregulated by UA in a dosedependent manner compared with that of the control group, both at 24 h after treatment. These results revealed that high UA could induce the expression of cytokines in A549 cells.

Discussion
This is the rst report on the strong association between serum UA levels and the severity and progression of UIP in RA patients with ILD relative to RA patients without ILD. Furthermore, serum UA levels were signi cantly increased in the UIP pattern of RA-ILD patients. Intriguingly, the increased serum UA level in RA-ILD patients was positively correlated with the HRCT-UIP score (Fig. 2E) and rheumatoid factor (RF). ROC curves also showed that the serum UA level was an independent biomarker for the identi cation of disease severity and UIP in RA patients. These data suggest that the circulating UA concentration may be a potential biomarker for identifying UIP and evaluating the progression and severity of ILD in RA patients.
UA is an antioxidant that is remarkable in the airway mucous membrane, and its actual role, especially in the regulation of lung function, has attracted attention in various diseases [16,30]. The present study showed that the level of UA was signi cantly associated with RA and the presence of ILD in RA, and a higher UA level was measured in the serum of individuals with RA. Signi cant correlations between serum UA levels and other related laboratory parameters in RA, including RF and anti-CCP, were observed. Additionally, compared with RA patients without ILD, the levels of serum and BALF UA were robustly higher in patients with RA-ILD, suggesting that UA may have speci c correlations with ILD in RA patients. Furthermore, the serum UA level was elevated in the UIP pattern of RA-ILD patients compared to the NSIP pattern of RA-ILD patients, which indicated the apparent correlation between UA and the severity of ILD in RA. Thus, we believe that UA plays an important role in the pathogenesis of RA-ILD.
Regarding the underlying work, we have con rmed that serum UA levels have an obvious association with RA-ILD, and circulating UA levels have a strong correlation with ILD structural damage (HRCT score) and a signi cant correlation with pulmonary function impairment (FEV1/FVC, FEV1). Most notably, the circulating UA level has a remarkable association with the clinically severest RA-ILD disease phenotype, i.e., the UIP pattern. The overall function of UA in terms of its association with ILD is apparently correlated with the currently available marker KL-6. The good predictability of the serum UA level implied a potential role as a new biomarker for RA-ILD.
Some experimental and clinical reports have con rmed that UA might be closely related to in ammatory responses, and serum UA levels are positively associated with TNF-α, IL-6 and CRP [16,31]. Interestingly, the results of an in vitro experimental study revealed that monosodium urate crystals can activate human broblast-like synoviocytes (FLSs) from RA patients with a remarkable increase in the release of the in ammatory cytokine IL-6, which provided insight into the role of UA in provoking in ammation in RA [32]. In our study, the level of serum UA showed signi cant positive correlations with several in ammatory risk factors for RA, including CRP and ESR, suggesting that serum UA levels may be used as an effective indicator for predicting the in ammatory activity of RA.
Interstitial lung diseases are characterized by destruction of lung architecture due to excessive deposition of extracellular matrix proteins by myo broblast activation, leading to aggravated dyspnea and loss of lung function [33,34]. Epithelial-mesenchymal transition (EMT) plays a critical role in the accumulation of myo broblasts and the resultant deposition of extracellular matrix proteins, which is a crucial step in the progression of ILD [35]. Both EMT and aberrant phenotypic changes in alveolar epithelial cells are prominent for the progression of ILD [33,36]. To further verify that UA is involved in the pathogenesis of ILD, we investigated whether UA contributed to EMT of alveolar epithelial cells. Our data revealed that UA could signi cantly induce EMT in alveolar epithelial cells (A549 cells), including elevation of mesenchymal markers and decline of epithelial cell markers. In a recent study, soluble UA was found to contribute to lung interstitial brosis closely related to the NLRP3 in ammatory pathway [8]. In contrast, the consensus among the previous works showed that some proin ammatory cytokines, such as IL-1, IL-6 and TGF-β, secreted by alveolar epithelial cells are closely related to the development of ILD [37]. Thus, we detected the expression of IL-1, IL-6, and TGF-β in A549 cells treated with different concentrations of UA, and the results indicated that IL-1, IL-6, and TGF-β expression was elevated by UA in a dosedependent manner, suggesting the considerable value of UA in ILD pathological processes.

Conclusions
In conclusion, serum UA levels were signi cantly increased in the UIP pattern of RA-ILD patients. Intriguingly, the increased serum UA level in RA-ILD patients with UIP was positively correlated with the HRCT-score, RF and anti-CCP. ROC curves also showed that the serum UA level was an independent biomarker for the identi cation of disease severity and UIP in RA patients. Moreover, the vitro experiment further showed that UA induced epithelial-to-mesenchymal transition (EMT) and production of IL-1, IL-6 and TGF-β in A549 cells. Thus, this study established that UA is strongly associated with RA-ILD, especially UIP pattern, which can be an important indicator for predicting the onset and severity of RA-ILD. UA may also hold promise as a useful biomarker for RA-ILD. CRP: Creactive protein; IL-1: Interleukin-1; IL-6: Interleukin-6; IL-8: Interleukin-8; ILD: Interstitial lung disease; RF: rheumatoid factor; anti-CCP: anti-cyclic peptide containing citrulline; ESR: erythrocyte sedimentation rate; UA: uric acid; KL-6: Krebs von den Lungen-6; NSAIDs: Nonsteroidal Antiin ammatory Drugs; DMARDs: disease-modifying anti-rheumatic drugs; DAS 28: 28-joint Disease Activity Score.

Declarations
Ethics approval and consent to participate All patients gave written informed consent before any study-related procedures were performed. The study protocol was approved by the local institutional ethics committee of Nanjing Medical University.  Figure 1 The level of UA was measured in RA-ILD patients. The level of UA in serum of RA patients and healthy control (HC). Statistical difference was detected between RA patients and HC (p<0.01) (A). Comparison of serum UA levels between no-ILD and RA-ILD, as well as NSIP and UIP. The higher serum UA levels were observed in RA-ILD relative to RA without ILD (p<0.01) (B). Statistical difference of serum UA was also detected between UIP and NSIP pattern of RA-ILD (p<0.01) (C). The UA level in BALF of RA patients was also measured. Compared with RA patients without ILD, the level of UA in RA-ILD patients was higher signi cantly (p<0.01) (D). And a marked increase of the level of UA was observed in BALF in RA-UIP patients relative to RA patients with NSIP pattern (p<0.01) (E).

Figure 2
Correlation of UA with clinic and laboratory Indexes in RA patients. RF and anti-CCP were positively correlated with the serum UA in patients with RA (A-B). A positive correlation between serum UA and KL-6 was detected by Spearman test (C). UA in BALF was positively associated with KL-6 in RA (D). Serum UA also showed a positive correlation with the mean HRCT score (E).