S100A9/CD163 expression profiles in classical monocytes as biomarkers to discriminate idiopathic pulmonary fibrosis from idiopathic nonspecific interstitial pneumonia

Circulating monocytes have pathogenic relevance in idiopathic pulmonary fibrosis (IPF). Here, we determined whether the cell surface levels of two markers, pro-inflammatory-related S100A9 and anti-inflammatory-related CD163, expressed on CD14strongCD16− classical monocytes by flow cytometry could discriminate IPF from idiopathic nonspecific interstitial pneumonia (iNSIP). Twenty-five patients with IPF, 25 with iNSIP, and 20 healthy volunteers were prospectively enrolled in this study. The S100A9+CD163− cell percentages in classical monocytes showed a pronounced decrease on monocytes in iNSIP compared to that in IPF. In contrast, the percentages of S100A9−CD163+ cells were significantly higher in iNSIP patients than in IPF patients and healthy volunteers. In IPF patients, there was a trend toward a correlation between the percentage of S100A9+CD163− monocytes and the surfactant protein-D (SP-D) serum levels (r = 0.4158, [95% confidence interval (CI) − 0.02042–0.7191], p = 0.051). The individual percentages of S100A9+CD163− and S100A9−CD163+ cells were also independently associated with IPF through multivariate regression analysis. The unadjusted area under the receiver operating characteristic curve (ROC-AUC) to discriminate IPF from iNSIP was (ROC-AUC 0.802, 95% CI [0.687–0.928]), suggesting that these are better biomarkers than serum SP-D (p < 0.05). This preliminary study reports the first comparative characterization of monocyte phenotypes between IPF and iNSIP.


Patients and methods
Subject selection. Subjects with idiopathic pulmonary fibrosis (IPF) (n = 25) and idiopathic nonspecific interstitial pneumonia (iNSIP) (n = 25) and healthy volunteers (n = 20) were prospectively recruited at Iwate Medical University Hospital from April 2016 until March 2019. Healthy volunteers were consecutively enrolled from among subjects over 50 years of age. IPF and iNSIP who were newly diagnosed in accordance with the criteria outlined by the 2011 consensus statements of the American Thoracic Society, European Respiratory Society, Japanese Respiratory Society, and Latin American Thoracic Association, were consecutively enrolled in our study 22 . Patients receiving corticosteroids, immunosuppressants, and/or anti-fibrotic agents were excluded. Age, sex, smoking status, and pulmonary function tests were assumed as potential confounders in the present study.
Flow cytometry analysis. PBMCs were incubated with mixtures of fluorochrome-conjugated antibodies, and identified by flow cytometry. Data acquisition and analysis were performed using a BD CANTO II Flow Cytometer and BD FACS DIVA software (BD Biosciences, Flanklin Lakes, New Jersey). Dead cells were distinguished by 7-amino-actinomycin D (7-AAD, BD Biosciences) staining. Cells were not permeabilized for intra-staining. To identify monocyte lineage cells, the surface markers CD14-PE (Biolegend, San Diego, California, clone HCD14) and CD16-APC/Cy7 (Biolegend, clone 3G8) were used. Pro-inflammatory and anti-inflammatory phenotypes were characterized using S100A9-FITC (BioRad, Hercules, California, clone MAC387) and CD163-APC (Biolegend, clone RM3/1) antibodies (Supplementary Table). S100A9 and CD163 expression levels were determined using a positive dataset for each antibody, identified using a matched concentration of mouse IgG1 kappa isotype control for fluorochrome color (Biolegend, clone MOPC21).
Pulmonary function tests. Pulmonary function tests were performed within a month prior to flow cytometry analysis. The values of the forced vital capacity (FVC) and forced expiratory volume in 1 s were expressed as percentages of the predicted normal values, calculated according to sex, weight, and age 23 .

Measurement of KL-6 and SP-D.
Krebs von den Lungen-6 (KL-6) concentration was measured with a commercially available chemiluminescent enzyme immunoassay kit (Picolumi KL-6, Eisai, Tokyo, Japan) according to the manufacturer's instructions. Serum surfactant protein-D (SP-D) concentration was also measured with a commercially available enzyme immunoassay kit (Yamasa-EIA II, Kyowa-Medex, Tokyo, Japan). Statistical analysis. The patients' baseline characteristics are provided as mean ± standard error of the mean. Data are reported for the full cohort. The normality of distribution was estimated using the Kolmogorov-Smirnov test. The statistical significance of differences between the three groups was evaluated by one-way

Results
Patient characteristics are shown in Table 1. The smoking status, represented by the pack-year index, showed no difference among the three groups analyzed by one-way ANOVA. We identified three subsets of monocytes, as described above (Supplementary Fig. S1). The ratio of dot plot number of classical monocytes to that of all monocytes did not differ across the three groups (data not shown). We explored the intensity of S100A9 and CD163 expression in classical and non-classical monocytes. Although individuals of delta mean fluorescent intensity (M.F.I) of CD163 and S100A9 showed similar trends between classical and non-classical monocytes, the individual delta M.F.I of S100A9 and CD163 in classical monocytes but not that in non-classical monocytes showed a significant increase in iNSIP as compared to IPF patients (p < 0.05 and p < 0.05). In addition, the delta M.F.I of CD163 was pronouncedly decreased in non-classical monocytes relative to classical monocytes, while the delta M.F.I of S100A9 was almost equivalent between classical and non-classical monocytes ( Supplementary  Fig. S2). Furthermore, we analyzed dot plots of S100A9 and CD163 in classical monocytes. Flow cytometry dot plots demonstrated that classical monocytes were divided into four distinct categories: S100A9 + CD163 − (Gate 1), S100A9 + CD163 + (Gate 2), S100A9 − CD163 − (Gate 3), and S100A9 − CD163 + (Gate 4) ( Supplementary Fig. S1). The percentages of S100A9 + cells (Gate 1 + 2) and S100A9 + CD163 − (Gate 1) showed comparative characteristics among the three groups, and were significantly higher in IPF patients than in iNSIP patients (p < 0.01); the percentages of S100A9 + CD163 − (Gate 1) were represented using s log scale (Fig. 1). In contrast, the percentages of Table 1. Patient characteristics. IPF idiopathic pulmonary fibrosis, iNSIP idiopathic nonspecific interstitial pneumonia, RF rheumatoid factor, ANA anti-nuclear antibody, FVC forced viral capacity, FEV 1.0 forced expiratory volume in 1 s, LDH lactate dehydrogenase, KL-6 Krebs von den Lungen-6, SP-D surfactant protein-D, -not determined, n.s no significance. Data are provided as mean ± standard deviation. *p < 0.05, **p < 0.01, ***p < 0.001. www.nature.com/scientificreports/ S100A9 − CD163 + cells (Gate 4) were significantly increased in iNSIP patients compared to those in IPF patients and healthy volunteers, respectively (p < 0.01 and p < 0.05, respectively). The percentages of S100A9 + CD163 + cells (Gate 2) and the S100A9 + CD163 − cells (Gate 1)/S100A9 − CD163 + cells (Gate 4) ratio showed no significant differences among the three groups. There was no difference between cellular and fibrotic iNSIP (Supplementary Fig. S3). In non-classical monocytes, the percentages of each gate showed no significant differences among the three groups (data not shown). In IPF patients, the percentages of S100A9 + CD163 − monocytes showed a trend toward a moderate correlation with the serum levels of surfactant protein-D (SP-D) (r = 0.4158 [95% CI − 0.02042-0.7191], p = 0.051, Table 2). Univariate regression analysis showed that the percentages of S100A9 + cells (Gate 1 + 2), S100A9 + CD163 − cells (Gate 1), and S100A9 − CD163 + cells (Gate 4) had statistically significant Percentages of S100A9 − CD163 + monocytes (Gate 4). (Lower right) S100A9 + /CD163 + monocyte ratio (Gate 1/Gate 4). *p < 0.05, **p < 0.01, and ***p < 0.001 by one-way analysis of variance. www.nature.com/scientificreports/ association with IPF, discriminating it from iNSIP among the four parameters (Table 3). Multivariate logistic regression analysis revealed that the percentage of S100A9 + CD163 − cells (Gate 1) and S100A9 − CD163 + cells (Gate 4) was a diagnostic factor for discriminating IPF from iNSIP, independent from age, sex, smoking status, and %FVC (Table 3). Furthermore, we determined the diagnostic value of each parameter for differentiating IPF from iNSIP. According to the unadjusted ROC-AUC analysis, the percentage of S100A9 + CD163 − cells (Gate 1) showed the best diagnostic value for IPF (ROC-AUC 0.802, 95% CI [0.687-0.928]) and was a significantly better biomarker to discriminate IPF from iNSIP than serum SP-D (0.616, 95% CI [0.446-0.786]) (p < 0.05, Fig. 2). There were no significant differences between the adjusted ROC of each parameter of classical monocytes and the serum levels of KL-6 and SP-D.

Discussion
In the present study, we found characteristic differences in the expression of the classical monocyte cell surface markers S100A9 and CD163 between patients with IPF and iNSIP. Multiple group comparisons and multivariate logistic regression analysis indicated that IPF and iNSIP show distinct expression profiles of S100A9 and CD163 expression on the surface of classical monocytes. Macrophages in IPF have been immunohistochemically associated with fewer CD163 + cells than those in iNSIP, which is consistent with the findings of the present study 24,25 .
To our knowledge, this is the first report indicating that the cellular phenotypes of classical monocytes, which form a major population of monocyte lineages, are associated with IPF. These results may encourage further investigations to identify novel clinical biomarkers using monocyte lineages in the context of IPF and provide new insight into the pathogenic roles of circulating monocytes in both IPF and iNSIP. A gradually increasing body of literature is becoming available regarding the association between interstitial lung diseases and monocyte phenotypes. Trombetta et al. reported that increased rates of circulating  www.nature.com/scientificreports/ pro-inflammatory M1/anti-inflammatory M2 mixed CD14 + monocytes were more associated with systemic sclerosis with interstitial lung diseases than with systemic sclerosis without interstitial lung diseases 13 . Consistently, in the present study, different trends were observed between the percentages of CD163 + S100A9 + and CD163 + S100A9 − cells (Gate 2 + 4) and those of CD163 + S100A9 − cells (Gate 4). Taken together, these findings highlight the importance of the simultaneous determination of pro-inflammatory and anti-inflammatory phenotypes in the estimation of circulating monocytes. Greiffo et al. showed that circulating non-classical monocytes were significantly decreased in ILDs, including nonspecific interstitial pneumonia, hypersensitivity pneumonitis, and collagen vascular disease-associated interstitial pneumonia compared with those in healthy volunteers 14 . In contrast, non-classical monocyte counts and the expression of CX3CR1 in monocytes were higher in lung. In addition, CD163 expression on non-classical monocytes showed a significant increase in ILDs patients than in healthy volunteers. In the present study, CD163 expression on non-classical monocytes did not show significant differences among the three groups. In contrast, the percentages of S100A9 − CD163 + cells on classical monocytes were significantly increased in iNSIP patients than in both healthy volunteers and IPF. Of note, the levels of delta M.F.I with respect to the expression of CD163 on classical monocytes appeared to be higher than those on nonclassical monocytes, consistent with the results reported by Greiffo and coworkers. These findings suggested the importance of CD163 signals on classical monocytes rather than non-classical monocytes. Very recently, the expression of S100A9 and CD163 on circulating monocytes has been associated with age, smoking, and COPD 18 . In the present study, multivariate logistic regression analysis revealed that the percentage of S100A9 + CD163 − cells (Gate 1) and S100A9 − CD163 + cells (Gate 4) was a diagnostic factor for discriminating IPF from iNSIP, independently from the smoking status, in addition to age, sex, and %FVC. Moreover, the percentage of S100A9 + CD163 − cells (gate 1) in IPF tended to correlate moderately with the SP-D serum levels. In contrast, no parameters were correlated with disease activity and/or severity of iNSIP. Taken together, these results imply that circulating classical monocytes play multiple pathogenic roles in heterogeneous fibrotic conditions.
ROS have been reported to contribute to tissue damage in IPF, and S100A9 homodimers play pro-inflammatory roles and the production of ROS, while CD163 receptors on macrophages scavenge hemoglobin via heme-oxygaenase-1 pathways, and inhibit the production of ROS 17 . We dare to speculate that S100A9-dominant classical monocytes might perpetuate tissue damage from lung injury in IPF as precursors of S100A9 + macrophages on alveolar spaces, while CD163-dominant monocytes rather play a protective role with reversibility against lung injury in iNSIP, but not profibrosis that is classically characterized as a major role of M2 polarized macrophages. However, we cannot exclude the possibility that cellular phenotypes of classical monocytes are an epiphenomenon caused by the underlying pathologies.
The number of potential diagnostic biomarkers for IPF is growing: for instance, the serum levels of matrix metalloproteinase (MMP)-7, MMP-28, SP-D, and S100A9 have been explored as IPF biomarkers [26][27][28] . The serum levels of S100A9 and MMP28 are promising biomarkers for discriminating between IPF and other ILDs. In particular, Hara et al. reported a ROC-AUC value of 0.92 for serum S100A9 in the discrimination IPF from iNSIP 29 . In contrast, Bennett et al. reported no difference in the serum levels of S100A9 homo-dimer levels between IPF and fibrosing iNSIP patients, which is inconsistent with our results 27 . However, while S100A9 homo-dimers are rigid in the plasma membrane, they are unstable in the serum, which might explain the discrepancy between our results and those of Bennett et al. 30 . In the present study, multivariate logistic regression analysis suggested that the increased percentages of S100A9 + CD163 − cells (Gate 1) and S100A9 − CD163 + cells (Gate 4) contributed significantly (as per the parameters in the adjusted ROC-AUC analysis) towards the discrimination between IPF and iNSIP. In addition, the unadjusted ROC values of S100A9 + CD163 − cells (Gate 1) were significantly higher than those of serum SP-D. Our study therefore indicates that S100A9 and CD163 in classical monocytes could be potentially useful diagnostic markers for IPF.
This study has some limitations. First, the sample size is relatively small. Additionally, missing data were removed. Thus, further studies on a larger cohort of patients are needed to validate these preliminary results. Second, we could not determine the properties of S100A9 + CD163 + cells (Gate 2). When the three groups were compared, the results for Gate 2 were similar to those of Gate 1 but not to those of Gate 4. A similar trend was also found in the previous study regarding COPD 18 . We speculate that the properties of S100A9 + CD163 + cells are entirely different from those of S100A9 − CD163 + cells. This should be validated by the comparison of gene expression profiles. Third, we used only S100A9 and CD163 in circulating monocytes as pro-inflammatory and ant-inflammatory biomarkers, respectively, and found pro-inflammatory phenotypes in classical monocytes of IPF patients, and anti-inflammatory phenotypes in those of iNSIP patients. However, it remains to be elucidated whether other pro-inflammatory and anti-inflammatory surface markers can confirm these phenotypes. Forth, some papers have recently reported that circulating monocyte counts could predict the prognoses of IPF. However, we did not determine the total circulating monocyte counts in the present study; we only focused on the percentage of S100A9 + and CD163 + cells. Therefore, further studies are needed to validate our data within circulating classical monocytes.

Conclusion
In conclusion, we used a clinical cohort to identify differences in the cell-surface levels of S100A9 and CD163 in classical monocytes by flow cytometry. We found monocyte phenotypes that were specific to patients with IPF and those with iNSIP that had potentially better discriminatory ability for these clinical groups than other existing biomarkers for IPF. An effective diagnostic biomarker for the early detection of IPF would allow for timeous medical interventions that could improve patient outcomes for this important disease.