Low SUV39H1 expression was associated with high blood neutrophil counts
The enrollment of study population is presented in Figure 1. The levels of SUV39H1 protein in the peripheral blood mononuclear cells (PBMCs) of COPD patients (n=30) and normal control (n=13) were measured by Western blotting (Supplementary Figure 1). The characteristics of the study subjects with normal lungs or COPD evaluated with immunoblot assays are indicated in Supplementary Table 1. The lung function index was significantly decreased in the COPD patients compared with the normal controls, including smokers and non-smokers. Moreover, lung function was reduced in the more severe COPD patients compared with the mild COPD patients (GOLD Stage III/IV vs. I/II). As SUV39H1 controls genes encoding Th1 cytokines and non-specific inflammatory mediators (10), including IL-8, we tested whether low expression of SUV39H1 was associated with increased neutrophil counts. To this end, we divided the patients into low SUV39H1 expression (<0.5-fold average of the normal subjects) and high SUV39H1 expression (≥0.5-fold average of the normal subjects) groups (Figure 2). Additionally, characteristics of the COPD patients were compared between these two groups (Table 1). We found that the low SUV39H1 expression group had a significantly higher percentage of neutrophils in the blood (65.33% vs. 56.53%, p=0.015). There was no difference in the total leukocyte count between the groups (7,933/μL vs. 6,773/μL, p=0.110).
We also examined the levels of blood eosinophils, which are Th2 downstream effector cells. Although we found a trend towards an increase in the percentage of blood eosinophils in the high SUV39H1 group, the difference was not statistically significant (Figure 3, 2.01% vs. 3.29%, p=0.125). Interestingly, in contrast to the high SUV39H1 group, which had a wide variation in the percentage of blood eosinophils, the low SUV39H1 group had consistently low eosinophil percentages. The ratio of eosinophils/neutrophils showed a similar trend (3.3% vs. 6%, p=0.071).
Characteristics of the extended COPD cohort
Measuring SUV39H1 levels in PBMCs requires an adequate amount of blood; therefore, it was difficult to recruit more subjects to correlate SUV39H1 levels with additional clinical outcomes. We thus used an extended cohort to study blood cell counts instead. Medical records from 218 patients in our COPD registry cohort collected from March 2014 to December 2017 were reviewed; two patients were excluded for having liquid tumours, and three were excluded for missing data. Of the 213 enrolled patients, 112 (52.6%) were included in GOLD group A, 39 (18.3%) were included in group B, 21 (9.9%) were included in group C, and 41 (19.2%) were included in group D (Table 2). A total of 193 patients (90.6%) were male, the average age was 73.1 ± 8.4 years, and the average BMI was 23.57 ± 4.11. Eighty patients (37.5%) were current smokers, 113 patients (53.1%) were ex-smokers, and 20 patients (9.4%) were never smokers. Nineteen patients (8.9%) had a history of asthma or met the diagnostic criteria of ACO, which was defined in the joint project of GOLD and GINA in 2015.
The most common COPD-related comorbidities were coronary artery disease (n=54, 25.35%), malnutrition (n=38, 17.8%), diabetes mellitus (n=34, 16.0%), pulmonary hypertension (n=28, 13.2%), normocytic anaemia (n=26, 12.2%), heart failure (n=17, 8.0%), lung fibrosis (n=14, 6.6%), osteoporosis (n=12, 5.6%), anxiety/depression (n=9, 4.2%), obstructive sleep apnoea (n=8, 3.8%), and lung cancer (n=3, 1.4%). The Charlson Comorbidity Index score was 1.1 in average.
Regarding medications, 76 patients (35.7%) received dual bronchodilators (LABA + LAMA), 66 (31.0%) received triple therapy (LABA + LAMA + ICS), 36 (16.9%) received LAMA only, 18 (8.4%) received LABA + ICS, 12 (5.6%) received LABA only, 1 (0.5%) received LAMA + ICS, and 4 (1.9%) did not receive any inhaled treatment.
Neutrophilia was correlated with COPD comorbidities but not the frequency of moderate to severe exacerbations
To identify the associations of systemic inflammation with comorbidities, we grouped patients with COPD into two groups according to the number of comorbidities (low: 0-1 comorbidities; high: ≥ 2 comorbidities) (Table 3). We found that the high comorbidity group had a lower BMI (22.78 ± 5.41 vs 23.9 ± 3.41, p=0.026), a smaller percentage of Group A COPD patients (35.5% vs. 59.6%, p=0.0014), higher percentages of GOLD 4 (22.6% vs. 7.3%, p=0.002) and Group D COPD patients (32.3% vs. 13.9%, p=0.002), higher total leukocyte counts (9,187/μL vs. 7,983/μL, p=0.012), a higher neutrophil percentage (69.8% vs. 60.7%, p<0.001), a lower eosinophil percentage (2.3% vs. 3.3%, p=0.037), and a lower eosinophil/neutrophil ratio (4.1% vs. 6.03%, p<0.001) (Figure 4). The neutrophil ratio seemed to have the most significant difference. Next, we examined whether specific comorbidities related to neutrophilia. However, post hoc analysis using Dunn’s multiple comparison tests did not show statistical significance for any single comorbidity (Supplementary Figure 2). For the comparison of comorbidities, the average number of comorbidities in the high comorbidity group was 2.58 (vs. 0.55 in the low comorbidity group, p<0.001). The high comorbidity group had significantly more incidences of all comorbidities except lung cancer (3.2% vs. 0.7%, p=0.149). Further analyses revealed that the neutrophil percentage was more positively correlated with the number of comorbidities (Spearman's rank correlation coefficient r=0.388, p<0.001) than the Charlson Comorbidity Index (CCI) scores (Spearman’s r=0.171, p=0.013) (Figure 5).
In this study, we found that the high comorbidity group also had more moderate to severe exacerbations per year (1.5 vs. 0.9, p=0.005). To confirm that the actual role of neutrophils is related to the number of comorbidities or frequency of exacerbations, we tested the associations of blood cell counts with moderate to severe exacerbations of COPD (Figure 6) by dividing the patients into two groups: patients with non-frequent exacerbations (annual exacerbation rates of 0-1 exacerbations/year) and patients with frequent exacerbations (more than 1 exacerbation/year). There were no significant differences between these two groups in the total white blood cell count (p=0.078), neutrophil percentage (p=0.061), eosinophil percentage (p=0.570), or eosinophil/neutrophil ratio (p=0.397).
Neutrophilia is increased in groups with high COPD-related comorbidities
The number of comorbidities in our study was divided into few comorbidities (0-1) and high comorbidities (more than 2), which we showed in this study to be equivalent to the results for the CCI, and a relatively high correlation with COPD-related comorbidities was revealed. Thus, the neutrophil percentage in the peripheral blood of COPD patients with different numbers of comorbidities and CCI were further examined. Our data showed that the neutrophil percentage was significantly increased in the high COPD-related comorbidity groups (Cor 2-3 and Cor ³4) compared with the low comorbidity group (Cor 0-1) (Figure 7A). The neutrophil percentages in the high CCI subgroups (CCI 3-4 and ³5) were not significantly increased compared with that in the low CCI subgroup (Figure 7B). Our results are consistent with previous findings (1). Collectively, our results showed that the neutrophil percentage in COPD patients was markedly increased in those patients with a high number of comorbidities but not in those patients with a high CCI score (Figure 7).
Reduced SUV39H1 expression was associated with COPD comorbidities
As reduced SUV39H1 expression was related to neutrophilia, we next asked whether SUV39H1 is associated with comorbidities. By reanalysing the densitometry Western blotting data, we found a modest reduction in the SUV39H1 level in patients with low comorbidities (Figure 8). Moreover, lower levels of SUV39H1 in PBMCs were observed in COPD patients with high comorbidities than in those with low comorbidities. These data together suggest that impaired SUV39H1 expression leads to neutrophilia and thus comorbidities.
Collectively, reduced SUV39H1 expression in COPD patients is associated with neutrophilia and thus comorbidities. We suggest that preserving the expression of SUV39H1 may control the Th1 response and maintain the balance between Th1 and Th2 responses (Figure 9). Patients in this condition may have milder or eosinophilic inflammation once they also have asthma or other Th2-related conditions. This inference might indicate relatively good outcomes for eosinophilic COPD. In more severe COPD patients, the nearly depleted expression of SUV39H1 skews Th1 polarization, causing more dominant neutrophilia and COPD comorbidities.