Relationships of Haptoglobin Phenotypes with Systemic Inflammation and the Severity of Chronic Obstructive Pulmonary Disease

Chronic obstructive pulmonary disease (COPD) is caused by chronic inflammation. Many inflammatory mediators induce the low grade systemic inflammation of COPD. Haptoglobin (Hp) is synthesized in the liver and by lung epithelial and alveolar macrophage cells. However, associations of the serum concentration and phenotype of Hp with COPD are unclear. Therefore, we explored the association of the Hp concentration and Hp phenotype with the inflammatory response and COPD disease severity. We included healthy subjects and COPD patients. The Hp phenotype was categorized by SDS native-PAGE, and concentrations were determined by ELISA. In this trial Hp concentrations in COPD groups were significantly higher than those in healthy controls. There was a significant negative correlation between the Hp concentration and FEV1(%) (p < 0.001), while IL-6 and 8-isoprostane were positively correlated with the Hp concentration. As to the Hp phenotype, there were significant negative correlations between the FEV1 and both Hp2-1 and Hp2-2; IL-6 and 8-isoprostane were significantly positively correlated with Hp2-1 and Hp2-2. The ROC curve analysis of the Hp concentration was significantly higher than CRP. Hp concentrations and phenotype were positively correlated with the severity of COPD, especially Hp2-2. In the future, Hp can be considered a novel biomarker for identifying COPD.

It responds to the release of cytokines such as IL-6, IL-1, and TNF-α by the human liver, and by lung epithelial cells and alveolar macrophages, and increases in inflammation and infections 5,9 . The serum concentration of this protein is usually lower than normal in tissues. Changes in serum levels of this protein in asthmatic patients after allergen skin test trials were reported in children 10 .
The lungs are a major site of the extrahepatic synthesis of Hp. This protein protects the lungs against inflammatory agents, and its presence prevents lung injury 11,12 . Its deficiency leads to lung tissue destruction caused by natural bacteriostatic neutrophil function, leading to initiation of emphysema and COPD. A protein deficiency leads to early-onset emphysema, and usually the incidence of disease is exacerbated by smoking and in some cases progresses to asthma 13 . This protein exists in humans as three main phenotypes of Hp1-1, Hp2-2, and Hp2-1, which are determined by two alleles: Hp1 and Hp2 14 . Hp2-2 is a weaker antioxidant than Hp1-1 and Hp2-1 15 . There are functional differences between different Hp phenotypes. In this study, we investigated relationships of Hp phenotypes with systemic inflammation and disease severity in patients with COPD.

Results
In total, 126 participants, 35 in the control group and 91 in the COPD groups (38 COPD GOLD 1 + 2 group and 53 COPD GOLD3 + 4 group), were included in the study. Table 1 summarizes baseline characteristics of these participants. Several baseline characteristics demonstrated significant between-group differences. A significantly higher proportion of participants in the COPD group were older, men, smokers, steroid treatment, had lower pulmonary function tests compared to participants in the control group (all p < 0.05). Serum concentrations and phenotype levels of Hp were significantly higher in the COPD group compared to the control group (all p < 0.05). There was a significant difference in the proportion of participants with different Hp phenotypes, and a much higher proportion of COPD groups with high Hp2-1 and Hp2-2, but not Hp1-1 (Fig. 1A), concentrations was found compared to the healthy group (Fig. 1B). As to systemic inflammatory cytokines, we found that IL-6, 8-isoprostane, and CRP concentrations were significantly higher in the COPD groups compared to the control group (all p < 0.05, Fig. 2), which shows that systemic inflammatory cytokines were related to higher Hp2-1 and Hp2-2 levels in the COPD groups than in the healthy control group, but the Hp1-1 phenotype did not significantly differ. Table 2 shows correlations of baseline characteristics of FEV 1 , IL-6, and 8-isoprostane serum levels with phenotype concentrations of Hp. The serum concentration of Hp was significantly negatively correlated with gender and FEV 1 (both p < 0.01), and was positively correlated with IL-6 and 8-isoprostane (both p < 0.01). In the three Hp phenotype groups, in participants with a high Hp1-1 concentration, no significant correlations were detected among all variables. The Hp2-1 concentration was significantly negatively correlated with FEV 1 (p < 0.05) and was positively correlated with IL-6 and 8-isoprostane (both p < 0.01). The Hp2-2 concentration was significantly negatively correlated with gender and FEV 1 (both p < 0.05), and was positively correlated with age and IL-6 (both p < 0.05). The other results showed that BMI, steroid treatment and CRP were not significantly correlated with total Hp concentration and three Hp phenotypes. Table 3 shows the multiple linear regression analysis of the three Hp phenotypes with gender, smoking status, FEV 1 , IL-6, and 8-isoprostane. We adjusted for all variables that could possibly influence the concentration of Hp (i.e., age, gender, smoking status, the body-mass index, and FEV 1 ). After adjusting for all of those variables, the presence of Hp phenotypes still significantly differed between healthy and COPD participants. Results showed that gender, FEV 1 , IL-6, and 8-isoprostane were independent factors for different Hp phenotypes. Gender was negatively associated with the Hp2-2 concentration (p < 0.01) (Hp phenotype levels were adjusted for age, gender, and smoking status). A low FEV 1 was associated with high concentrations of the three Hp phenotypes (all p < 0.05). High IL-6 was associated with high concentrations of the three Hp phenotypes (all p < 0.05). High 8-isoprostane was associated with high Hp1-1 and Hp2-1 concentrations (both p < 0.05), but was not associated with Hp2-2. Figure 3 shows the ROC curve of sensitivity and specificity of CRP and serum levels of Hp for discriminating between the healthy and COPD groups. The area under the ROC curve (AUC) was 0.8131 (0.731~0.896, p < 0.01) for CRP concentrations in participants (Fig. 3A). The AUC was 0.8938 (0.831~0.957, p < 0.01) for serum levels of Hp in participants (Fig. 3B). The serum Hp cutoff value determined by the Youden index for participants was 187.0 mg/dl. The population of participants with Hp1-1 was too small to conclude whether it can be used to discriminate between healthy and COPD phenotypes. Further, the ROC curve of Hp2-1 and Hp2-2 levels for discriminating between the healthy and COPD groups. The AUC was 0.810 (0.674~0.946, p < 0.01) for the Hp2-1 concentration in participants (Fig. 4A).The AUC was 0.985 (0.959~1.000, p < 0.001) for the Hp2-2 level in participants (Fig. 4B). The Hp2-2 cutoff value determined by the Youden index for participants was 192 mg/dl.

Discussion
This study showed four major results: (1) distributions of Hp phenotypes differed between healthy participants and COPD patients; (2) concentrations of Hp phenotypes were positively significantly related to systemic inflammation and oxidative stress in COPD patients; (3) concentrations of Hp phenotypes were negatively related to the FEV 1 in participants; and (4) serum concentrations of Hp phenotypes were significantly higher in COPD patients than in healthy participants.  As to distributions of Hp phenotypes among various populations in the world 15 , the majority of Asians have very low HP 1F values compared to those on other continents, with Indians averaging about 0.05, with its virtual absence in Japanese, Koreans, and Taiwanese 16 . The distribution of the Hp1 allele is higher in Blacks than in Caucasians and even more so than in populations of Southeast Asia 15 . There was a decreased Hp2-1 frequency in those with a family history of bronchial asthma, and a low Hp2-2 frequency in those with adenocarcinoma of the lungs 15 . Different Hp phenotypes exhibited equal susceptibilities to pulmonary tuberculosis 17 . Tuberculosis patients with the Hp2-2 phenotype had a higher risk of mortality 17 . As to functional differences between Hp phenotypes, Hp2-2 may provide less protection against hemoglobin iron-driven peroxidation leading to lower A.
B.   concentrations of ascorbic acid compared to those carrying Hp1-1 and Hp2-1 18,19 . This study found a high Hp2-2 frequency in the most severe COPD GOLD3 + 4 group, and particularly Hp2-2 phenotype concentrations may be most capable of predicting systemic inflammation, according to the ROC analysis. Therefore, Hp may be used to differentiate severity in COPD patients, and it can be used to identify COPD severity.
Hp is an acute-phase protein, and its plasma concentration increases in response to a variety of stimuli 5 . A past study noted that plasma Hp increased in patients with an abdominal aortic aneurysm 20 and coronary artery disease (CAD) 21 . Hp may also play a proatherogenic role in which Hp acts as a chemoattractant to pre-B lymphocytes and monocytes in inflammatory adipose tissue 22 . Circulating concentrations of Hp were found to increase in parallel with COPD severity in patients 4 . Our study showed that the serum Hp concentration significantly increased and was positively related to systemic inflammation and oxidative stress in COPD patients, and this indicator significantly predicted systemic inflammation and oxidative stress, but not significantly correlated with steroid treatment for total Hp and Hp phenotype concentration.  Table 3. Linear regression analysis of haptoglobin phenotypes in the participants. Abbreviations: FEV 1 , Forced expiratory volume in first second; IL-6, interleukin-6; CRP, C-Reactive Protein.
A.  Total Hp concentrations vary with phenotype 23 , in Chinese subjects, Hp2-2 concentrations were highest in males 19 , and our study had similar results. An increase in Hp levels may generate a feedback dampening of the severity of cytokine release and may protect against endotoxin-induced effects 5,24 . As a nitric oxide (NO) scavenger the Hp-Hb complex has a role in regulating NO bioavailability and vascular homeostasis; Hp inhibits prostaglandin synthesis and so has an anti-inflammatory action 25 . This study found that Hp concentrations were significantly positively correlated with IL-6 and 8-isoprostane (both p < 0.01); IL-6 was associated with all three Hp phenotype concentrations (all p < 0.05). 8-Isoprostane was associated with subtype Hp1-1 and Hp2-1 concentrations (both p < 0.05), but was not associated with Hp2-2.

ROC curve for CRP
Our study has several limitations. (1) This study was relatively small in size, and study participants were recruited from a single hospital. Further, larger sample sizes from different centers are needed to confirm our results. (2) Furthermore, several differences in baseline characteristics also existed among study participants in the three groups. These cohort characteristics and differences may have introduced bias and affected the results of our study to some extent. Most of the COPD patients referred to this center were male; hence, we could not assess the effect of gender. Further samples are needed to increase female participation. (3) The severity of COPD was evaluated by FEV 1 alone. However, we know that the GOLD evaluation of COPD severity can be more complex and comprehensive. (4) Many methods can be used to identify Hp phenotypes, but we only used electrophoretic techniques. In the future, polymerase chain reaction-based methods may be developed to enable the identification of Hp allele types in subjects.

Conclusions
Hp concentrations were positively correlated with the severity of COPD, and IL-6 and 8-isoprostane serum concentrations. The Hp phenotype also influenced COPD severity and IL-6, especially the Hp2-2 phenotype, for which the disease severity and IL-6 levels were significantly higher than those of Hp1-1 and Hp2-1. In addition, the area under ROC curve of Hp concentration was significantly higher than that of CRP. In the future, Hp may be considered a novel biomarker for identifying COPD severity.  approved the study protocol. All subjects received written and oral information prior to inclusion and provided written informed consent. All methods were performed in accordance with the relevant guidelines and regulations.
Study participants. All study participants provided written consent. We enrolled 91 patients with COPD and 35 subjects as healthy controls from Shuang Ho Hospital (New Taipei City, Taiwan) between March 2014 and April 2016. Patients with COPD received a diagnosis from a physician and exhibited a post-bronchodilator forced expiratory volume in the first second (FEV 1 )/forced vital capacity (FVC) ratio of ≤70%. The classification of COPD severity followed the GOLD guidelines 26 . Healthy control subjects exhibited an FEV 1 /FVC ratio of ≥75% and an FEV 1 of ≥80% of the predicted value. All subjects were aged 40~80 years at the time of inclusion. Subjects with a known malignant tumor, active inflammatory disease (such as asthma, bronchiectasis, or other non-COPD-related disease), or exacerbation during the 3 months prior to the study were not included.
Hp phenotyping. The phenotypes of serum samples were determined by Hb-supplemented 6% native polyacrylamide gel electrophoresis (PAGE) and peroxidase chromogenic staining to detect Hb peroxidase activity, as described by Cheng et al. (Supplementary Fig. S1) 14 . Briefly, a sample containing 7 μL of serum was premixed with 5 μL of 8 mg/mL hemoglobin and equilibrated with 3 μL of sample buffer (0.625 mol/L Tris-base at pH 6.8, 50% glycerol, and 0.125 mg/L bromophenol blue). The mixture was run on a 6% native polyacrylamide gel (pH 8.8), with 4% polyacrylamide used as a top stacking gel (pH 6.8). Electrophoresis was performed at an initial voltage of 120 V, which was increased to 150 V when the dye front reached the separating gel. After electrophoresis, the Hp-was visualized by shaking the gel in freshly prepared peroxidase substrate (0.05% 3,3′-diaminobenzidine and 0.07% hydrogen peroxide in phosphate-buffered saline;). Results were confirmed by Western blotting using an α-chain-specific monoclonal antibody (mAb) 27 .
Hp purification. Serum of each Hp phenotype was chromatographed on an mAb-based affinity column followed by a high-performance liquid chromatographic (HPLC) procedure as described by Cheng et al. 27 .

Measurement of the human free Hp serum level.
Human free serum concentrations of Hp were measured using a phenotype-matched standard sandwich enzyme-linked immunosorbent assay (ELISA) as described in a previous publication 28 .

Statistical analysis.
Patients' demographic and clinical characteristics were summarized as the mean ± standard deviation for continuous data and the number (percent) for categorical data. For comparisons among multiple values, a one-way analysis of variance (ANOVA) with Turkey's post-hoc test was used. A Pearson correlation analysis was performed to identify correlations of Hp concentrations with age, anthropometric indicators, and CRP, IL-6, and 8-isoporstane concentrations. A linear regression analysis was performed to identify associations between Hp phenotypes and participants' characteristics. The optimal serum Hp concentration cutoff point for predicting COPD was determined based on maximization of the Youden index using a receiver operating characteristic (ROC) curve analysis. All statistical assessments were considered significant at p < 0.05. Statistical analyses were performed using SPSS 20.0 statistics software (SPSS, Chicago, IL, USA) and Graphpad Prism 5 software (GraphPad Software, San Diego, CA, USA).