Abstract
Background
Preserved ratio impaired spirometry (PRISm) is a spirometry pattern of interest regarding incident airflow obstruction and higher mortality risk. We applied a proteomic approach to gain more insight into the biological mechanisms associated with PRISm.
Methods
From the population-based Swedish Cardiopulmonary Bioimage Study (SCAPIS), participants in the Main (n=4835) and Pilot (n=1054) studies, were included as discovery and replication cohorts. The lower limit of normal (LLN) of post-bronchodilator forced expiratory volume in 1 s (FEV1), forced vital capacity (FVC) and FEV1/FVC was defined as the fifth percentile in healthy, never-smoking SCAPIS participants. Participants were subdivided into five groups: reference: FEV1/FVC≥LLN and FEV1≥LLN and FVC≥LLN (n=4084)); mild chronic airflow limitation (CAL): FEV1/FVC<LLN and FEV1≥LLN (n=278); moderate–severe CAL: FEV1/FVC<LLN and FEV1<LLN (n=170); restrictive spirometric pattern (RSP): FEV1/FVC ≥LLN and FVC<LLN (n=238); and PRISm: FEV1/FVC≥LLN and FEV1<LLN (n=238). Proximity extension assays were used to measure 168 proteins. The associations of each standardised protein were assessed with each study group by multiple linear regression models, adjusting for age, sex, body mass index (BMI), smoking, physical activity and study centres, and corrected for multiple testing to control for a false discovery rate of 5%.
Results
Eight proteins were associated with PRISm and replicated: tumour necrosis factor receptor superfamily member 10A, interleukin-6, paraoxonase 3 (negative association), renin, urokinase plasminogen activator surface receptor (U-PAR), E-selectin, matrix metalloproteinase 7 and chitinase-3-like protein 1. Interleukin-6 and U-PAR were also associated with moderate–severe CAL and E-selectin with RSP. In addition, five other proteins were associated with moderate–severe CAL and three with RSP.
Conclusion
Protein profile in PRISm differs from other spirometric patterns suggesting specific disease mechanisms.
Shareable abstract
This first study of proteomics of preserved ratio impaired spirometry (PRISm) identified a protein profile that discriminates PRISm form both severe airflow limitation and restrictive spirometric pattern https://bit.ly/3Wc4uqa
Introduction
Spirometry is a widely available tool to support the diagnosis and staging of obstructive lung diseases that are defined as a reduction in forced expiratory volume in 1 s (FEV1) relative to forced vital capacity (FVC). Spirometric lung function impairment in form of decreased FVC in presence of a preserved FEV1/FVC ratio is defined as restrictive spirometry pattern (RSP) [1]. If the FEV1/FVC ratio is preserved and FEV1 is decreased, “nonspecific” or “unclassified” patterns have previously been used to describe this combination. More recently, the term preserved ratio impaired spirometry (PRISm) was introduced to include subjects with impaired FEV1 in combination with a normal ratio. The prevalence of PRISm in population-based cohorts is high, ranging from 5.1% to 20.4% [2–5]. Some of this variation also depends on availability of pre- or post-bronchodilator measurements or if studies include subjects with preserved ratio with impaired FEV1 and/or FVC in the definition [3]. In 2023, PRISm was introduced in the Global Initiative of Obstructive Lung Disease's (GOLD) document (www.goldcopd.org) and defined as a post-bronchodilator preserved FEV1/FVC ratio of ≥0.7 and an impaired FEV1 of <80% of predicted values. The GOLD report also stated that “Despite an increasing body of literature on PRISm, significant knowledge gaps remain in relation to its pathogenesis and treatment”.
PRISm is characterised by an increased respiratory symptom burden [2, 6] and a higher risk of both all-cause, respiratory-related and cardiovascular (CV) mortality [5]. Similar findings were reported from a UK Biobank cohort study, in which subjects with PRISm displayed an increased risk of major adverse CV events and CV disease mortality, compared with those with normal lung function [4]. In longitudinal studies, PRISm has also been characterised by increased rates of transitions to both normal and obstructive spirometry over time [3, 7].
Protein biomarkers related to PRISm can offer insights into the underlying pathogenesis of PRISm. Proteomics detects and quantifies a high number of proteins simultaneously, enabling the discovery of new and clinically relevant biomarkers that can indicate pathways involved in disease development [8].
The overall aim of the present study was to apply a proteomic approach to identify proteins associated with different types of lung function impairment according to spirometry, with a special focus on PRISm. We hypothesised that PRISm is a more heterogeneous condition that could be associated with systemic protein biomarkers that differ from both obstructive and restrictive pattern of lung function impairment.
Methods
Population
This was a cross-sectional analysis of the population-based Swedish Cardiopulmonary Bioimage Study (SCAPIS), a collaborative research project between six Swedish universities (Gothenburg, Lund/Malmö, Linköping, Stockholm, Uppsala and Umeå). SCAPIS is an observational study including 30 154 participants (15 508 women), aged 50–64 years [9]. Before the start of SCAPIS, the core protocol was evaluated in a Pilot study at one of the test centres (Gothenburg) including 1111 participants [10]. Previous results on prevalence and comorbidities associated with RSP and PRISm in SCAPIS have been recently published [2].
In the present study, we used a discovery and replication approach with the SCAPIS Main cohort and SCAPIS Pilot study as the discovery and replication cohorts, respectively. Proteomics was analysed in plasma with the Olink CVD II and III arrays in 5073 randomly selected participants from the Main study and in 1068 individuals from the SCAPIS Pilot study. To achieve an even geographical distribution of participants from the Main study, there was an oversampling of subjects from Umeå. Of the participants, 4835 and 1054, respectively had complete data on spirometry and were included in the present study (figure 1).
FIGURE 1.
Flow chart. SCAPIS: Swedish Cardiopulmonary Bioimage Study; FEV1: forced expiratory volume in 1 s; FVC: forced vital capacity; QC: quality control.
Lung function
Dynamic spirometry was performed using a Jaeger MasterScreen PFT (Vyaire, Mettawa, IL, USA), 15 min after supervised inhalation of 400 μg of salbutamol with the subject in the sitting position and wearing a nose clip [11].
When interpreting the spirometry, the lower limit of normal (LLN) was defined as the fifth percentile of healthy SCAPIS participants [11]. Individuals were subdivided into five groups based on postbronchodilator spirometry:
1) Reference (normal spirometry): FEV1/FVC≥LLN and FEV1≥LLN and FVC≥LLN
2) Mild chronic airflow limitation (CAL): FEV1/FVC<LLN and FEV1≥LLN
3) Moderate–severe CAL: FEV1/FVC<LLN and FEV1<LLN
4) Restrictive spirometric pattern (RSP): FEV1/FVC≥LLN and FVC<LLN
5) Preserved ratio impaired spirometry (PRISm): FEV1/FVC≥LLN and FEV1<LLN.
Proteomics
Blood samples were drawn for proteomic analyses and plasma was stored at −70°C until analysis. Plasma samples were analysed at the Clinical Biomarkers Facility, Science for Life Laboratory, Uppsala University, using proximity extension assay (PEA) technology [8]. Two panels, Olink Multiplex CVD II and III 96×96 panels each measuring 92 proteins, were analysed. 16 proteins were excluded due to more than 15% of their values being below the corresponding limits of detection (LOD). All assay characteristics are available from the manufacturer's webpage (https://olink.com/products/compare and https://olink.com/knowledge/documents).
Assessments of covariates
Height and weight were measured and body mass index (BMI) was calculated as weight/height (kg·m−2). Questionnaire data included smoking status (never, former, present or unknown), pack-years and also level of physical activity during leisure time that was analysed by four questions adopted from the Public Health Agency of Sweden [12].
Statistics
For the discovery cohort, a total of 4835 subjects and 168 proteins on 58 plates were analysed. The proteins were used as the independent variables and spirometry categories were used as dependent variables. For each protein, values below the corresponding LODs were imputed with LOD/√2. All protein values were previously adjusted for storage time and plates, and then standardised as mean 0 and sd 1. The standardised residuals were used in further statistical analysis.
Multiple linear regressions were carried out, adjusting for age, sex, BMI, smoking (category and pack-years), physical activity (three categories) and study centre. The results were summarised as the estimated coefficients (β coefficients) with 95% confidence intervals.
The p-values were adjusted for multiple testing using the Benjamini–Hochberg method [13] to control for a false discovery rate at a 5% level. Proteins significantly associated with the spirometry categories were tested in the replication cohort using an identical approach. In the replication study, a nominal significance threshold of 0.05 was selected. All statistical analysis were performed using R [14].
Ethics
The SCAPIS multicentre study and the present study were both approved by the regional ethical review board in Umeå (Dnr 2010-228-31M and Dnr 2017-183-31). All participants gave their written informed consent.
Results
Characteristics of the study sample are presented in table 1. The study sample was representative of the full SCAPIS cohort in all included variables. Prevalence of the different lung function groups was also of similar magnitude both in the SCAPIS Main cohort and in the SCAPIS Pilot study.
TABLE 1.
Characteristics of the study sample.
| SCAPIS Main study | SCAPIS Pilot study | ||
|---|---|---|---|
| Discovery cohort, n=4835 |
Remaining cohort, n=25 319 |
Replication cohort, n=1054 |
|
| Age, years | 57.6±4.3 | 57.5±4.4 | 57.7±4.4 |
| BMI, kg·m−2 | 26.8±4.2 | 27.0±4.5 | 27.3±4.5 |
| Female | 2448 (50.6) | 13 060 (51.6) | 519 (49.2) |
| Smoking | |||
| Never | 2398 (49.6) | 12 333 (50.1) | 447 (42.7) |
| Former | 1762 (36.4) | 8840 (35.9) | 409 (39.0) |
| Present | 609 (12.6) | 3119 (12.7) | 156 (14.9) |
| Unknown | 66 (1.4) | 304 (1.2) | 36 (3.4) |
| Pack-years | 0 (0–12.0) | 0 (0–12.3) | 2.5 (0–17.0) |
| Physical activity | |||
| Low | 550 (11.4) | 2887 (11.4) | 170 (16.4) |
| Medium | 3666 (75.8) | 18 139 (71.6) | 708 (68.3) |
| High | 519 (10.7) | 2802 (11.1) | 158 (15.3) |
| Study centre | |||
| Gothenburg | 847 (17.5) | 5418 (21.4) | 1054 (100) |
| Lund/Malmö | 702 (14.5) | 5549 (21.9) | |
| Linköping | 820 (17.0) | 4218 (16.7) | |
| Stockholm | 839 (17.4) | 4218 (16.7) | |
| Uppsala | 841 (17.4) | 4195 (16.6) | |
| Umeå | 786 (16.3) | 1721 (6.8) | |
| FEV1/FVC<0.7 | 426 (8.8) | 2403 (9.6) | 107 (10.2) |
| New lung function groups according to the definitions in the present study: | |||
| Reference# | 4084 (84.5) | 20 602 (81.4) | 855 (81.1) |
| Mild CAL | 278 (5.7) | 1545 (6.1) | 56 (5.3) |
| Moderate–severe CAL | 170 (3.5) | 929 (3.7) | 57 (5.4) |
| RSP | 238 (4.9) | 1427 (5.6) | 64 (6.0) |
| PRISm | 238 (4.9) | 1358 (5.4) | 65 (6.2) |
Data are presented as mean±sd, n (%) or median (interquartile range). BMI: body mass index; FEV1: forced expiratory volume in 1 s; FVC: forced vital capacity; CAL: chronic airflow limitation; RSP: restrictive spirometric pattern; PRISm: preserved ratio, impaired spirometry; LLN: lower limit of normal. #: Reference (normal spirometry): FEV1/FVC≥LLN (fifth percentile) and FEV1≥LLN and FVC≥LLN.
Characteristics of the different lung function groups in SCAPIS Main cohort are presented in table 2.
TABLE 2.
Characteristics of the different lung function impairment groups in the SCAPIS Main cohort
| Normal spirometry, n=4084 | Moderate–severe CAL, n=170 | Mild CAL, n=278 | PRISm, n=238 | RSP, n=238 | |
|---|---|---|---|---|---|
| Definition | FEV1/FVC≥LLN FEV1≥LLN FVC≥LLN | FEV1/FVC<LLN FEV1<LLN | FEV1/FVC<LLN FEV1≥LLN | FEV1/FVC≥LLN FEV1<LLN | FEV1/FVC≥LLN FVC<LLN |
| Age, years | 57.6±4.3 | 58.1±4.3 | 57.4±4.4 | 57.8±4.3 | 57.6±4.3 |
| BMI, kg·m−2 | 26.7±4.1 | 27.4±5.1 | 25.9±3.8 | 29.0±5.5 | 28.6±5.5 |
| Female | 2074 (50.8) | 77 (45.3) | 156 (56.1) | 114 (47.9) | 110 (46.2) |
| Smoking | |||||
| Never | 2112 (51.7) | 40 (23.5) | 95 (34.2) | 112 (47.1) | 126 (52.9) |
| Former | 1491 (36.5) | 68 (40.0) | 108 (38.8) | 78 (32.8) | 72 (30.3) |
| Present | 421 (10.3) | 61 (35.9) | 73 (26.3) | 45 (18.9) | 37 (15.5) |
| Unknown | 60 (1.5) | 1 (0.6) | 2 (0.7) | 3 (1.2) | 3 (1.3) |
| Pack-years | 0.0 (0.0–10.8) | 17.5 (0.3–36) | 6 (0.0–21.1) | 0.6 (0.0–18.8) | 0.0 (0.0–13.8) |
| Physical activity | |||||
| Low | 419 (10.5) | 40 (24.4) | 36 (13.1) | 45 (19.6) | 46 (19.9) |
| Medium | 3111 (77.7) | 119 (72.6) | 216 (78.8) | 172 (74.7) | 172 (74.1) |
| High | 473 (11.8) | 5 (3.0) | 22 (8.0) | 13 (5.7) | 14 (6.0) |
| Study centre | |||||
| Gothenburg | 687 (16.8) | 42 (24.7) | 52 (18.7) | 51 (21.4) | 49 (20.6) |
| Malmö | 597 (14.6) | 26 (15.3) | 19 (6.8) | 48 (20.2) | 48 (20.2) |
| Linköping | 695 (17.0) | 26 (15.3) | 63 (22.7) | 26 (10.9) | 28 (11.8) |
| Stockholm | 703 (17.2) | 22 (12.9) | 62 (22.3) | 45 (18.9) | 39 (16.4) |
| Uppsala | 710 (17.5) | 35 (20.6) | 49 (17.6) | 36 (15.2) | 40 (16.8) |
| Umeå | 692 (16.9) | 19 (11.2) | 33 (11.9) | 32 (13.4) | 34 (14.2) |
| Post-bronchodilator spirometry | |||||
| FEV1 (% pred) | 101.9 (10.7) | 70.7 (8.2) | 91.9 (8.3) | 75.0 (5.0) | 76.6 (6.6) |
| FVC (% pred) | 101.4 (10.6) | 89.2 (10.4) | 108.4 (9.7) | 76.6 (6.8) | 74.8 (5.3) |
| FEV1/FVC (% pred) | 100.2 (5.5) | 79.2 (7.8) | 84.4 (3.8) | 97.9 (6.3) | 102.0 (6.6) |
Data are presented as mean±sd, n (%) or median (interquartile range). CAL: chronic airflow limitation; PRISm: preserved ratio, impaired spirometry; RSP: restrictive spirometric pattern; FEV1: forced expiratory volume in 1 s; FVC: forced vital capacity; LLN: lower limit of normal; BMI: body mass index.
An overlap between PRISm and RSP was seen in both the SCAPIS Main cohort and in the Pilot cohort as participants with a normal ratio (FEV1/FVC≥LLN) in combination with both impaired FEV1 (FEV1<LLN) and impaired FVC (FVC<LLN) fulfilled the criteria of both PRISm and RSP (supplementary figure S1).
Associations between all studied proteins and spirometry categories are presented in supplementary table S1. No proteins were associated with the group with mild CAL. In the SCAPIS Main cohort, we found 18 proteins that associated with the group with moderate–severe CAL after adjusting for age, sex, BMI, smoking, physical activity and centre, and after correction for multiple testing. Seven of these were replicated: growth/differentiation factor 15 (GDF-15), interleukin (IL)-1 receptor antagonist (IL-1RA), trefoil factor 3 (TFF3), IL-6, lectin-like oxidised LDL receptor 1 (LOX-1), urokinase plasminogen activator surface receptor (U-PAR) and adrenomedullin (ADM) (table 3 and figure 2a). 38 proteins were associated with RSP in the SCAPIS Main cohort and 4 of them were replicated: LOX-1, Receptor for advanced glycosylation end products (RAGE), angiotensin-converting enzyme 2 (ACE2) and E-selectin (SELE) (table 4 and figure 2b). All associations were in the positive direction except for RAGE that was negative. 16 proteins were significantly associated with PRISm in the adjusted model and eight of these replicated: tumour necrosis factor receptor superfamily member 10A (TNFRSF10A), paraoxonase 3 (PON3), IL-6, renin (REN), U-PAR, SELE, matrix metalloproteinase 7 (MMP7) and chitinase-3-like protein 1 (CHI3L1) (table 5 and figure 2c). Except for a negative association between PON3 and PRISm, all protein associations with moderate–severe CAL and PRISm were in the positive direction.
TABLE 3.
Associations between plasma proteins and moderate to severe chronic airflow limitation, sorted by significance in the replication cohort
| Protein | SCAPIS Main cohort (Discovery cohort) |
SCAPIS Pilot cohort (Replication cohort) |
||
|---|---|---|---|---|
| β-coefficient# (95% CI) | Adj. p-value¶ | β-coefficient# (95% CI) | Adj. p-value+ | |
| Growth/differentiation factor 15 (GDF-15) | 0.33 (0.19–0.47) | 00001 | 0.41 (0.16–0.65) | 0.0011 |
| Interleukin-1 receptor antagonist (IL-1RA) | 0.27 (0.14–0.41) | 0.0016 | 0.36 (0.13–0.60) | 0.0028 |
| Trefoil factor 3 (TFF3) | 0.24 (0.08–0.40) | 0.0292 | 0.39 (0.13–0.66) | 0.0041 |
| Interleukin-6 (IL-6) | 0.32 (0.17–0.47) | 0.0005 | 0.37 (0.11–0.63) | 0.0052 |
| Lectin-like oxidised LDL receptor 1 (LOX-1) | 0.34 (0.18–0.49) | 0.0005 | 0.29 (0.03–0.56) | 0.0300 |
| Urokinase plasminogen activator surface receptor (U-PAR) | 0.43 (0.28–0.57) | 1.67×10−6 | 0.26 (0.01–0.51) | 0.0381 |
| Adrenomedullin (ADM) | 0.33 (0.19–0.48) | 0.0001 | 0.25 (0.01–0.50) | 0.0441 |
SCAPIS: Swedish Cardiopulmonary Bioimage Study; CI: confidence interval. #: All proteins are previously adjusted for storage time and plate ID and standardised with mean 0 and sd 1. ¶: Adjusted for age, sex, body mass index (BMI), smoking (category and pack-years), physical activity, centre and corrected for multiple testing. +: Adjusted for age, sex, BMI, smoking (category and pack-years) and physical activity.
FIGURE 2.
Estimated association between plasma proteins (left panels) and a) moderate–severe chronic airflow limitation (CAL), b) restrictive spirometric pattern (RSP) and c) preserved ratio impaired spirometry (PRISm) of the SCAPIS Main cohort, adjusted for age, sex, body mass index (BMI), smoking (category and pack-years), physical activity and centre, and corrected for multiple testing. The red dots are those with adjusted p<0.05 by Benjamini–Hochberg method. Validation results in the Pilot cohort, adjusted for age, sex, BMI, smoking (category and pack-years) and physical activity (right panels). The red dots are those with p<0.05. All proteins were previously adjusted according to storage time and plate ID, and standardised with mean 0 and sd 1.
TABLE 4.
Associations between plasma proteins and restrictive spirometric pattern (RSP), sorted by significance in the replication cohort
| Protein | SCAPIS Main cohort (Discovery cohort) | SCAPIS Pilot cohort (Replication cohort) | ||
|---|---|---|---|---|
| β-coefficient# (95% CI) | Adj. p-value¶ | β-coefficient# (95% CI) | Adj. p-value+ | |
| Lectin-like oxidised LDL receptor 1 (LOX-1) | 0.22 (0.09–0.35) | 0.0072 | 0.40 (0.15–0.65) | 0.0019 |
| Receptor for advanced glycosylation end products (RAGE) | −0.21 (−0.34– −0.09) | 0.0091 | −0.37 (−0.63– −0.12) | 0.0041 |
| Angiotensin-con-verting enzyme 2 (ACE2) | 0.21 (0.09–0.33) | 0.0072 | 0.25 (0.01–0.49) | 0.0374 |
| E-selectin (SELE) | 0.19 (0.06–0.32) | 0.0240 | 0.26 (0.01–0.51) | 0.0435 |
SCAPIS: Swedish Cardiopulmonary Bioimage Study; CI: confidence interval. #: All proteins are previously adjusted for storage time and plate ID and standardised with mean 0 and sd 1. ¶: Adjusted for age, sex, body mass index (BMI), smoking (category and pack-years), physical activity, centre and corrected for multiple testing. +: Adjusted for age, sex, BMI, smoking (category and pack-years) and physical activity.
TABLE 5.
Associations between plasma proteins and preserved ratio impaired spirometry (PRISm), sorted by significance in the replication cohort
| Protein | SCAPIS Main cohort (Discovery cohort) | SCAPIS Pilot cohort (Replication cohort) | ||
|---|---|---|---|---|
| β-coefficient# (95% CI) | Adj. p-value¶ | β-coefficient# (95% CI) | Adj. p-value+ | |
| Tumour necrosis factor receptor superfamily member 10A (TNFRSF10A) | 0.25 (0.12–0.38) | 0.0034 | 0.36 (0.12–0.60) | 0.0033 |
| Paraoxonase 3 (PON3) | −0.24 (−0.36– −0.12) | 0.0018 | −0.32 (−0.55– −0.09) | 0.0057 |
| Interleukin-6 (IL-6) | 0.25 (0.12–0.38) | 0.0024 | 0.33 (0.09–0.58) | 0.0081 |
| Renin (REN) | 0.27 (0.14–0.40) | 0.0016 | 0.30 (0.05–0.55) | 0.0179 |
| Urokinase plasminogen activator surface receptor (U-PAR) | 0.22 (0.09–0.34) | 0.0101 | 0.27 (0.04–0.51) | 0.0236 |
| E-selectin (SELE) | 0.19 (0.06–0.32) | 0.0417 | 0.28 (0.03–0.53) | 0.0267 |
| Matrix metalloproteinase-7 (MMP7) | 0.19 (0.06–0.32) | 0.0495 | 0.27 (0.02–0.52) | 0.0353 |
| Chitinase-3-like protein 1 (CHI3L1) | 0.23 (0.10–0.36) | 0.0081 | 0.26 (0.01–0.51) | 0.0410 |
SCAPIS: Swedish Cardiopulmonary Bioimage Study; CI: confidence interval. #: All proteins are previously adjusted for storage time and plate ID and standardised with mean 0 and sd 1. ¶: Adjusted for age, sex, body mass index (BMI), smoking (category and pack-years), physical activity, centre and corrected for multiple testing. +: Adjusted for age, sex, BMI, smoking (category and pack-years) and physical activity.
The number of replicated proteins within each lung function impairment group and overlaps are illustrated in figure 3. U-PAR and IL-6 associated with both PRISm and moderate–severe CAL, while SELE associated with both PRISm and RSP. In addition, LOX-1 was associated with both moderate–severe CAL and RSP. The remaining protein associations were unique for only one lung function impairment group; five proteins for PRISm, four for moderate–severe CAL and two for RSP.
FIGURE 3.
Proteins associated with each lung function impairment group. (−) indicates a negative association. CAL: chronic airflow limitation; PRISm: preserved ratio impaired spirometry; RSP: restrictive spirometric pattern. GDF-15: growth/differentiation factor 15; ADM: adrenomedullin; IL-1RA: interleukin-1 receptor antagonist; TFF3: trefoil factor 3; LOX-1: lectin-like oxidised LDL receptor 1; U-PAR: urokinase plasminogen activator surface receptor; IL-6: interleukin-6; CHI3L1: chitinase-3-like protein 1; TNFRSF10A: tumour necrosis factor receptor superfamily member 10A; PON3: paraoxonase 3; REN: renin; MMP7: matrix metalloproteinase 7; SELE: E-selectin; ACE-2: angiotensin-converting enzyme 2; RAGE: receptor for advanced glycosylation end products.
Discussion
PRISm has emerged as a clinically significant spirometric pattern, associated with several adverse outcomes. We found a specific protein profile associated with PRISm, compared with both CAL and RSP in this, to our knowledge, first study of proteomics of PRISm. We identified eight proteins significantly associated with PRISm. Among these, IL-6 and U-PAR were also associated with moderate–severe CAL, while SELE was associated with RSP. Five proteins (CHI3L1, TNFRSF10A, PON3, REN and MMP7) were associated with PRISm in both the discovery and replication cohorts, with no association with any of the other groups. These proteins might offer insights into the underlying biological mechanisms behind PRISm. While the cross-sectional design is an obstacle to conclusions on clinical implications, the proteins identified in relation to PRISm here could be used in future prospective studies to evaluate their possible role as prognostic biomarkers for disease progression such as transition into COPD with subsequent exacerbations and/or FEV1 decline. Furthermore, although PRISm is associated with an increased respiratory burden and adverse outcomes, there is also a high rate of transition to normal spirometry over time [3, 7]. If any of the proteins recognised here could be useful in identifying these individuals, it would be a valuable prognostic tool, of importance both for the individual and the healthcare system.
As far as we know, there are no previous data on the proteins identified here in relation to PRISm. However, several of them have been identified in different lung diseases. CHI3L1, also known as YKL40, is a glycoprotein with an unclear biological function that is expressed and secreted by various cell types, including macrophages. Previous studies have indicated that YKL40 is related to both obstructive and interstitial lung diseases. Increased serum levels of YKL40 in both asthma–COPD overlap (ACO) [15] and interstitial lung disease (ILD) [16], imply this marker to be of importance in diagnosis and prognosis of both these diagnoses. Furthermore, YKL-40 has been related to faster lung function decline [17] and is therefore believed to be a risk marker for susceptibility to the long-term effects of cigarette smoking.
Signalling mediated by TNFRSF10A is essential for immune regulation, cell proliferation, cell death and morphogenesis [18]. Although the understanding of TNFRSF signalling is still incomplete, it has been established as an important pathway of human regulatory T-cells [19].
PON3 has antioxidant properties and seems to act as an important guardian against cellular damage from toxic agents [20]. The inverse relationship between PON3 and PRISm found here might reflect impaired antioxidant defences in PRISm and is in line with a previous community-based study, where higher PON3 related to both lower FEV1 and FVC [21].
REN is a highly specific endopeptidase within the renin–angiotensin–aldosterone system (RAAS) to stimulate aldosterone synthesis and the activation of RAAS may play a central role in the pathogenesis and progression of idiopathic pulmonary fibrosis [22]. The end-product of the RAAS pathway, angiotensin II, binds to angiotensin II type 1 (AT1R) or type 2 (AT2R) receptors. Activated AT1R, in contrast with AT2R, mediates pro-inflammatory and pro-hyperresponsive actions. In COPD lungs, a five-fold increase in the AT1R to AT2R ratio was associated with reduced lung function [23]. Renin has also been identified as a biomarker that is elevated in patients with the combination of coronary artery disease and COPD [24].
MMP7 is produced primarily by the epithelium of many organs, including the lungs. The production of MMP7 is highly upregulated by injury or exposure to bacteria, stimulating cell migration and coordinating inflammatory responses [25]. Both in COPD [26] and in community-based samples [21], MMP7 levels are inversely associated to both FEV1 and FVC. MMP7 has also been suggested as a biomarker of subclinical interstitial lung disease, linked to lower FVC. Furthermore, increased MMP7 levels at baseline have been associated with increased odds of exertional dyspnoea, interstitial lung abnormalities and all-cause mortality at the 10-year follow-up [27].
IL-6 and U-PAR were associated with both PRISm and moderate–severe CAL. IL-6 is produced by inflammatory cells and primary lung epithelial cells in response to a variety of different stimuli and is implied as a key modulator in the overall immune response as well as in the function of nonimmune cells. IL-6 could thereby play an active role in mediating specific inflammatory aspects in both asthma and other lung diseases [28]. In previous population-based samples, an inverse relationship was shown between IL-6 and both FEV1 and FVC [21, 29] and IL-6 levels also predicted a more rapid 5-year lung function decline [30]. U-PAR is a membrane-bound receptor implicated to play a role in neutrophil migration to sites of inflammation. U-PAR is increased in stable COPD patients and is further increased during COPD exacerbations, followed by a decline after treatment response [31].
Our findings support the notion that IL-6 and U-PAR may contribute to the inflammatory processes underlying both PRISm and CAL. The elevated levels of these biomarkers in PRISm might also reflect the substantial phenotype of PRISm that over time translates to CAL.
Despite a substantial overlap between PRISm and RSP, SELE was the only protein associated with both groups. SELE is a cell adhesion molecule expressed only on endothelial cells and is activated by inflammatory cytokines such as IL-1 and tumour necrosis factor-α. It plays an important part in inflammation and mainly mediates the adherence of neutrophils to inflammatory tissues. Increased levels of SELE have been related to COPD [32], but also to interstitial lung diseases [33] and may reflect the presence of endothelial injury.
Four proteins (GDF-15, ADM, IL-1RA and TFF3) were increased only in moderate–severe CAL (i.e. participants with the most severe airway obstruction) and all have previously been related to obstructive lung diseases.
Both GDF-15 [34] and ADM [35] are found to be related to exacerbation and higher mortality in COPD. Trefoil factors are secretory products of mucin producing cells and TFF3 is the most abundant of the trefoil factor peptides in the airways, present in goblet cells, submucosal glands and club cells. COPD patients display substantial higher concentrations of TFF3 in both sputum samples and serum [36].
The positive association between IL-1RA and moderate–severe CAL is in accordance with the findings of Rydell et al. [21], but is somewhat contradictory to some previous studies, where results from Mendelian randomisation suggested increased IL-1RA to play a role in preventing airway obstruction [37]. Thus, this issue needs to be further evaluated.
LOX-1 was the only protein associated with both moderate–severe CAL and RSP. Activation of LOX-1 by oxidised low-density lipoprotein is implicated in the progression of CV disease and LOX-1 may be a promising target for early diagnosis of CV disease and CV risk prediction along with being considered as a treatment target [38], but previous studies on LOX-1 in human lung disease are scarce or lacking.
The two proteins ACE2 and RAGE, here associated with RSP, have previously been related to lung fibrosis. In the respiratory system, ACE2 protects the lung from injury by neutralising the effect of accumulated angiotensin II [39]. RAGE, which was negatively associated with RSP here, is under-expressed in idiopathic pulmonary fibrosis lungs and data indicate that plasma soluble RAGE levels may reflect disease severity [40]. Also, in a previous proteomic study, both % predicted FEV1 and FVC were positively associated with RAGE [21].
In the present study, the recently published SCAPIS post-bronchodilator spirometry reference values were used [11]. Compared with pre-bronchodilator LLN, the post-bronchodilator values of LLN were higher for both FEV1/FVC ratio and FEV1 which could have impacted on group categorisation of some individuals.
Strengths of the present study include the large general population cohort from six different Swedish centres. Participants in both the external and replication cohorts were investigated using identical protocols for spirometry and proteomics measurements, thereby minimising variability and risk for potential bias. Furthermore, proteomics was analysed using the precise PEA-based technology, a validated and sensitive method, and with a broad spectrum of protein biomarkers. Another important strength includes the use of a reliable post-bronchodilator spirometry, thereby avoiding subjects with reversible airway obstruction being classified as PRISm.
The limitations include the fact that all participants were within the age range of 50–64 years, and the results cannot be generalised to other age groups. Also, as discussed above the cross-sectional nature of the study is an important limitation, as causality and clinical usefulness cannot be addressed. While adjusting for age, sex, BMI, smoking and physical activity, there could still be a risk of residual confounding.
To summarise, the identified protein biomarkers in moderate–severe CAL and RSP have previously been identified as related to obstructive and restrictive lung diseases, respectively. Interestingly, the protein profile associated with PRISm differed from both obstructive and RSPs and constitute a more heterogenous group with findings of several novel possible biomarker in PRISm. The mechanism by which PRISm might affect the circulating levels of several proteins is unclear and conclusions on causal effects are impeded by the cross-sectional design. However, the identified protein biomarkers in PRISm may be suitable for the assessment of progression and transition over time.
We conclude that the proteomic signatures discriminate PRISm from both severe airflow limitation and RSP. From a clinical point of view, the results of the present study strengthen the view that PRISm is a separate entity. Further research is warranted to validate these findings and elucidate the underlying mechanisms linking these protein biomarkers to lung function impairment. Finally, future studies are needed to better understand the potential role of these mechanisms for therapeutic interventions.
Acknowledgement
We would like to acknowledge the help of Biobank Sweden and the local biobank facilities for their services in handling of biological samples and biobanking.
Footnotes
Provenance: Submitted article, peer reviewed.
Ethics statement: The SCAPIS multicentre study and the present study were both approved by the regional ethical review board in Umeå (Dnr 2010-228-31M and Dnr 2017-183-31). All participants gave their written informed consent.
Author contributions: E. Lindberg, A. Malinovschi and A. Blomberg supervised and led the conception and design of the work and the writing of the statistical analysis plan together with X. Zhou. All authors critically revised the statistical analysis plan. X. Zhou performed the statistical analyses under the supervision of E. Lindberg. All authors contributed to the interpretation of the data. E. Lindberg, X. Zhou, A. Malinovschi and A. Blomberg drafted the manuscript. All authors critically revised the manuscript, gave final approval and agreed to be accountable for all aspects of the work. The corresponding author, E. Lindberg, is the manuscript's guarantor and attests that all listed authors meet authorship criteria and take full responsibility for the overall content and that no others meeting the criteria have been omitted. The SCAPIS project follows the International Committee of Medical Journal Editors recommendations for author contributions. All listed authors fulfil these recommendations.
Conflict of interest: A. Egesten reports consultancy fees from Arcede Pharma, Lund, Sweden. J.E. Engvall reports support for attending meetings from Acta Pediatrica. A. Lindberg reports consultancy fees from Swedish Medical Products Agency, SBU – Swedish Agency for Health Technology Assessment and Assessment of Social Service, Ongoing; payment or honoraria for lectures, presentations, manuscript writing or educational events from Bonnier Healthcare Sweden AB and AstraZeneca; participation on a data safety monitoring board or advisory board with AstraZeneca, Novartis, Boehringer Ingelheim and GSK; and a leadership role with Research Council for the Swedish Heart and Lung foundation 2019–2023. A-C. Olin reports consultancy fees from Chiesi AB; patents planned, issued or pending (WO2009045163, “Collection and measurement of exhaled particles”; WO2013117747, “A device and method for noninvasive analysis of particles during medical ventilation”); a leadership role with PExA AB; and stock (or stock options) with PExA AB. S. Söderberg reports payment or honoraria for lectures, presentations, manuscript writing or educational events, and support for attending meetings from, and participation on a data safety monitoring board or advisory board with Johnson & Johnson. The remaining authors have nothing to disclose.
Support statement: The main funding body of SCAPIS is the Swedish Heart and Lung Foundation. The study is also funded by the Knut and Alice Wallenberg Foundation, the Swedish Research Council, VINNOVA (Sweden's Innovation agency), the University of Gothenburg and Sahlgrenska University Hospital, Karolinska Institutet and Region Stockholm, Linköping University and University Hospital, Lund University and Skåne University Hospital, Umeå University and University Hospital, and Uppsala University and University Hospital. Funding information for this article has been deposited with the Crossref Funder Registry.
Supplementary material
Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.
Figure S1
01045-2024.SUPPLEMENT
Table S1
01045-2024.SUPPLEMENT2
References
- 1.Godfrey MS, Jankowich MD. The vital capacity is vital: epidemiology and clinical significance of the restrictive spirometry pattern. Chest 2016; 149: 238–251. doi: 10.1378/chest.15-1045. [DOI] [PubMed] [Google Scholar]
- 2.Torén K, Blomberg A, Schiöler L, et al. Restrictive spirometric pattern and preserved ratio impaired spirometry in a population 50-64 years. Ann Am Thoracic Soc 2024; 21: 1524–1532. doi: 10.1513/AnnalsATS.202403-242OC. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.He D, Sun Y, Gao M, et al. Different risks of mortality and longitudinal transition trajectories in new potential subtypes of the preserved ratio impaired spirometry: evidence from the English longitudinal study of aging. Front Med (Lausanne) 2021; 8: 755855. doi: 10.3389/fmed.2021.755855. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Zheng J, Zhou R, Zhang Y, et al. Preserved ratio impaired spirometry in relationship to cardiovascular outcomes: a large prospective cohort study. Chest 2023; 163: 610–623. doi: 10.1016/j.chest.2022.11.003. [DOI] [PubMed] [Google Scholar]
- 5.Yang S, Liao G, Tse LA. Association of preserved ratio impaired spirometry with mortality: a systematic review and meta-analysis. Eur Respir Rev 2023; 32: 230135. doi: 10.1183/16000617.0135-2023. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Kogo M, Sato S, Muro S, et al. Longitudinal changes and Association of Respiratory Symptoms with Preserved Ratio Impaired Spirometry (PRISm): the Nagahama study. Ann Am Thorac Soc 2023; 20: 1578–1786. doi: 10.1513/AnnalsATS.202301-050OC. [DOI] [PubMed] [Google Scholar]
- 7.Wijnant SRA, De Roos E, Kavousi M, et al. Trajectory and mortality of preserved ratio impaired spirometry: the Rotterdam Study. Eur Respir J 2020; 55: 1901217. doi: 10.1183/13993003.01217-2019. [DOI] [PubMed] [Google Scholar]
- 8.Assarsson E, Lundberg M, Holmquist G, et al. Homogenous 96-plex PEA immunoassay exhibiting high sensitivity, specificity, and excellent scalability. PLoS ONE 2014; 9: e95192. doi: 10.1371/journal.pone.0095192. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Bergström G, Persson M, Adiels M, et al. Prevalence of subclinical coronary artery atherosclerosis in the general population. Circulation 2021; 144: 916–929. doi: 10.1161/CIRCULATIONAHA.121.055340. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Bergström G, Berglund G, Blomberg A, et al. The Swedish CArdioPulmonary BioImage Study: objectives and design. J Intern Med 2015; 278: 645–659. doi: 10.1111/joim.12384. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Malinovschi A, Zhou X, Andersson A, et al. Consequences of using post- or prebronchodilator reference values in interpreting spirometry. Am J Respir Crit Care Med 2023; 208: 461–471. doi: 10.1164/rccm.202212-2341OC. [DOI] [PubMed] [Google Scholar]
- 12.Public Health Agency of Sweden. Hälsa på lika villkor . Date accessed: 4 December 2018. Date last updated: 2012.. https://www.folkhalsomyndigheten.se/ [Google Scholar]
- 13.Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc B Methodol 1995; 57: 289–300. doi: 10.1111/j.2517-6161.1995.tb02031.x. [DOI] [Google Scholar]
- 14.R Foundation for Statistical Computing. R: A Language and Environment for Statistical Computing. 2024.. https://www.R-project.org/ [Google Scholar]
- 15.Peng J, Wang M, Wu Y, et al. Chen L. clinical indicators for asthma-COPD overlap: a systematic review and meta-analysis. Int J Chron Obstruct Pulmon Dis 2022; 17: 2567–2575. doi: 10.2147/COPD.S374079. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Tong X, Ma Y, Liu T, et al. Can YKL-40 be used as a biomarker for interstitial lung disease?: A systematic review and meta-analysis. Medicine (Baltimore) 2021; 100: e25631. doi: 10.1097/MD.0000000000025631. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Guerra S, Halonen M, Sherrill DL, et al. The relation of circulating YKL-40 to levels and decline of lung function in adult life. Respir Med 2013; 107: 1923–1930. doi: 10.1016/j.rmed.2013.07.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Vanamee ÉS, Faustman DL. Structural principles of tumor necrosis factor superfamily signaling. Sci Signal 2018; 11: eaao4910. doi: 10.1126/scisignal.aao4910. [DOI] [PubMed] [Google Scholar]
- 19.Okubo Y, Mera T, Wang L, et al. Homogeneous expansion of human T-regulatory cells via tumor necrosis factor receptor 2. Sci Rep 2013; 3: 3153. doi: 10.1038/srep03153. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Rull A, García R, Fernández-Sender L, et al. Serum paraoxonase-3 concentration is associated with insulin sensitivity in peripheral artery disease and with inflammation in coronary artery disease. Atherosclerosis 2012; 220: 545–551. doi: 10.1016/j.atherosclerosis.2011.11.021. [DOI] [PubMed] [Google Scholar]
- 21.Rydell A, Nerpin E, Zhou X, et al. Cardiovascular disease-linked plasma proteins are mainly associated with lung volume. ERJ Open Res 2023; 9: 00321-2022. doi: 10.1183/23120541.00321-2022. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Gupta D, Kumar A, Mandloi A, et al. Renin angiotensin aldosterone system in pulmonary fibrosis: pathogenesis to therapeutic possibilities. Pharmacol Res 2021; 174: 105924. doi: 10.1016/j.phrs.2021.105924. [DOI] [PubMed] [Google Scholar]
- 23.Bullock GR, Steyaert I, Bilbe G, et al. Distribution of type-1 and type-2 angiotensin receptors in the normal human lung and in lungs from patients with chronic obstructive pulmonary disease. Histochem Cell Biol 2001; 115: 117–124. doi: 10.1007/s004180000235. [DOI] [PubMed] [Google Scholar]
- 24.Casselbrant A, Fedorowski A, Frantz S, et al. Common physiologic and proteomic biomarkers in pulmonary and coronary artery disease. PLoS ONE 2022; 17: e0264376. doi: 10.1371/journal.pone.0264376. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Manicone AM, Huizar I, McGuire JK. Matrilysin (matrix metalloproteinase-7) regulates anti-inflammatory and antifibrotic pulmonary dendritic cells that express CD103 (α(E)β(7)-integrin). Am J Pathol 2009; 175: 2319–2331. doi: 10.2353/ajpath.2009.090101. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Montaño M, Sansores RH, Becerril C, et al. FEV1 inversely correlates with metalloproteinases 1, 7, 9 and CRP in COPD by biomass smoke exposure. Respir Res 2014; 15: 74. doi: 10.1186/1465-9921-15-74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Armstrong HF, Podolanczuk AJ, Barr RG, et al. Serum matrix metalloproteinase-7, respiratory symptoms, and mortality in community-dwelling adults. MESA (Multi-Ethnic Study of Atherosclerosis). Am J Respir Crit Care Med 2017; 196: 1311–1317. doi: 10.1164/rccm.201701-0254OC. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Rincon M, Irvin CG. Role of IL-6 in asthma and other inflammatory pulmonary diseases. Int J Biol Sci 2012; 8: 1281–1290. doi: 10.7150/ijbs.4874. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Gimeno D, Delclos GL, Ferrie JE, et al. Association of CRP and IL-6 with lung function in a middle-aged population initially free from self-reported respiratory problems: the Whitehall II study. Eur J Epidemiol 2011; 26: 135–144. doi: 10.1007/s10654-010-9526-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Rydell A, Nowak C, Janson C, et al. Plasma proteomics and lung function in four community-based cohorts. Respir Med 2021; 176: 106282. doi: 10.1016/j.rmed.2020.106282. [DOI] [PubMed] [Google Scholar]
- 31.Gumus A, Altintas N, Cinarka H, et al. Soluble urokinase-type plasminogen activator receptor is a novel biomarker predicting acute exacerbation in COPD. Int J Chron Obstruct Pulmon Dis 2015; 10: 357–365. doi: 10.2147/COPD.S77654 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Takahashi T, Kobayashi S, Fujino N, et al. Annual FEV1 changes and numbers of circulating endothelial microparticles in patients with COPD: a prospective study. BMJ Open 2014; 4: e004571. doi: 10.1136/bmjopen-2013-004571 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Kumánovics G, Minier T, Radics J, et al. Comprehensive investigation of novel serum markers of pulmonary fibrosis associated with systemic sclerosis and dermato/polymyositis. Clin Exp Rheumatol 2008; 26: 414–420. [PubMed] [Google Scholar]
- 34.Husebø GR, Grønseth R, Lerner L, et al. Growth differentiation factor-15 is a predictor of important disease outcomes in patients with COPD. Eur Respir J 2017; 49: 1601298. doi: 10.1183/13993003.01298-2016. [DOI] [PubMed] [Google Scholar]
- 35.Stolz D, Kostikas K, Blasi F, et al. Adrenomedullin refines mortality prediction by the BODE index in COPD: the “BODE-A” index. Eur Respir J 2014; 43: 397–408. doi: 10.1183/09031936.00058713. [DOI] [PubMed] [Google Scholar]
- 36.Viby NE, Pedersen L, Lund TK, et al. Trefoil factor peptides in serum and sputum from subjects with asthma and COPD. Clin Respir J 2015; 9: 322–329. doi: 10.1111/crj.12146. [DOI] [PubMed] [Google Scholar]
- 37.Yang Z, Schooling CM, Kwok MK. Genetic evidence on the association of interleukin (IL)-1-mediated chronic inflammation with airflow obstruction: a Mendelian randomization study. COPD 2021; 18: 432–442. doi: 10.1080/15412555.2021.1955848. [DOI] [PubMed] [Google Scholar]
- 38.Barreto J, Karathanasis SK, Remaley A, et al. Role of LOX-1 (lectin-like oxidized low-density lipoprotein receptor 1) as a cardiovascular risk predictor: mechanistic insight and potential clinical use. Arterioscler Thromb Vasc Biol 2021; 41: 153–166. doi: 10.1161/ATVBAHA.120.315421. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Marčetić D, Samaržija M, Vukić Dugac A, et al. Angiotensin-converting enzyme 2 (ACE2) as a potential diagnostic and prognostic biomarker for chronic inflammatory lung diseases. Genes (Basel) 2021; 12: 1054. doi: 10.3390/genes12071054. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Manichaikul A, Sun L, Borczuk AC, et al. Plasma soluble receptor for advanced glycation end products in idiopathic pulmonary fibrosis. Ann Am Thorac Soc 2017; 14: 628–635. doi: 10.1513/AnnalsATS.201606-485OC [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.
Figure S1
01045-2024.SUPPLEMENT
Table S1
01045-2024.SUPPLEMENT2