Levels of Urinary Mercapturic Acids of Acrolein, Methacrolein, Crotonaldehyde, and Methyl Vinyl Ketone in Relationship to Chronic Obstructive Pulmonary Disease in Cigarette Smokers of the SubPopulations and InterRmediate Outcome Measures in COPD Study (SPIROMICS) Study

Abstract

Cigarette smoking is an established cause of chronic obstructive pulmonary disease (COPD). Numerous studies implicate acrolein, which occurs in relatively high concentrations in cigarette smoke and reacts readily with proteins, as one causative factor for COPD in smokers. Far less is known about the possible roles in COPD of the related α,β-unsaturated carbonyl compounds of cigarette smoke crotonaldehyde, methacrolein, and methyl vinyl ketone. In the study reported here, we analyzed mercapturic acids of these α,β-unsaturated compounds in the urine of 413 confirmed cigarette smokers in the Subpopulations and Intermediate Outcome Measures in COPD Study (SPIROMICS) – 202 with COPD and 211 without COPD. The mercapturic acids analyzed were 3-hydroxypropyl mercapturic acid (3-HPMA) from acrolein, 3-hydroxy-1-methylpropyl mercapturic acid (HMPMA-1) from crotonaldehyde, 3-hydroxy-2-methylpropyl mercapturic acid (HMPMA-2) from methacrolein, and 3-hydroxy-3-methylpropyl mercapturic acid (HMPMA-3) from methyl vinyl ketone. In models adjusting for age, sex, race, pack years of tobacco use and BMI, all four mercapturic acids were increased in individuals with COPD, but not significantly. Stratifying by GOLD status, there were increased levels of the metabolites associated with GOLD 3-4 compared to GOLD 0, with the methacrolein metabolite HMPMA-2 reaching statistical significance (adjusted odds ratio 1.23 [95%CI: 1.00 – 1.53]). These results highlight the possible role of methacrolein, which has previously received little attention in this regard, as a causative factor in COPD in cigarette smokers.

Graphical Abstract

INTRODUCTION

Chronic obstructive pulmonary disease (COPD) is currently the sixth leading cause of death in the world. The Global Burden of Disease Study reported 175 million cases of COPD in 2015; 3.2 million people died of COPD.¹ In the U.S. in 2020, 12.5 million people, or 5.0% of adults, reported a diagnosis of COPD.² COPD is often associated with tobacco smoking and the risk for COPD is 3-4 times greater in current smokers compared to non-smokers.^3–5 Approximately 40% of patients continue to smoke after a diagnosis of COPD.⁶ Nevertheless, it remains unknown why only 20% of adults with substantial tobacco smoke exposure develop clinically significant COPD. Biomarkers of tobacco smoke constituents involved in COPD etiology might be useful in identifying those smokers at highest risk, potentially leading to effective smoking cessation interventions to prevent the disease.

Numerous studies implicate acrolein (1), one of the most toxic compounds in cigarette smoke⁷ occurring in mainstream smoke at levels of 30.8 – 82.6 μg per cigarette (ISO smoking regimen) and 139-213 μg per cigarette (Canadian Intense regimen),⁸ as a main cause of COPD in cigarette smokers.⁹ The ease of reaction of acrolein with certain proteins of respiratory cells underlies its mechanisms of toxicity including oxidative stress, chronic inflammation, cell death, tissue remodeling and destruction, loss of lung elasticity and enlarged air spaces.¹⁰ The structurally related compounds crotonaldehyde (2, 10.1 μg per cigarette¹¹), methacrolein (3, 2-13 μg per cigarette¹²), and methyl vinyl ketone (4, 31.5 μg per cigarette¹¹) (Scheme 1) have reactive properties in common with acrolein but are less studied with respect to their possible role in COPD. While the cited levels per cigarette are average amounts, there is a wide range of exposures to these reactive compounds in cigarette smokers, resulting from the fact that their levels in smoke vary greatly because they are combustion products formed at high temperatures from multiple precursors in tobacco. One study demonstrated that mainstream smoke levels of acrolein vary up to 100-fold and those of crotonaldehyde nearly 30-fold depending on cigarette brand, filter ventilation and the machine smoking protocol used.¹³

Scheme 1.

Formation of the mercapturic acids analyzed in this study from α,β-unsaturated carbonyl compounds 1-4, considered as possible causative factors for COPD. ^aa, ɣ-glutamyl transpeptidase; b, cysteinyl glycinase; c, N-acetyl transferase; d, reductases

Biomarkers provide a potentially useful approach to assess individual exposure to specific compounds in cigarette smoke and other sources. After inhalation in cigarette smoke, compounds 1-4 all have the common property of reaction with cellular glutathione and other sulfhydryl compounds. The initially formed conjugates from the glutathione reactions are enzymatically processed and eventually excreted in the urine as the mercapturic acids 3-HPMA from acrolein, HMPMA-1 from crotonaldehyde, HMPMA-2 from methacrolein, and HMPMA-3 from methyl vinyl ketone (Scheme 1) which are accepted biomarkers of exposure to these compounds in cigarette smokers.^14–18 We have recently developed a method which separates HMPMA-1, 2, and 3, which were previously all attributed to products arising from crotonaldehyde (e.g. HMPMA-1).¹⁴ We applied the method in this study to test the hypothesis that differences in the levels of these urinary biomarkers as well as 3-HPMA could be related to the incidence and/or severity of COPD in cigarette smokers. Four hundred thirteen urine samples, all from cigarette smokers in the Subpopulations and Intermediate Outcome Measures in COPD Study (SPIROMICS) cohort, an observational study of COPD, were analyzed; the samples were divided between subjects who did or did not have COPD.

EXPERIMENTAL SECTION

Study Population:

Cases and controls were participants of the SPIROMICS cohort.¹⁹ Cases (n=202) were individuals with COPD, defined as FEV1 < 80% predicted and FEV1/FVC < 70 who were currently smoking tobacco and had a positive urinary cotinine test. Controls (n=211) were individuals without COPD and normal spirometry who were currently smoking tobacco and had a positive cotinine test.²⁰ Controls were matched for age (±5 years), sex and pack years of smoking (Table 1). Participants in the SPIROMICS study provided informed consent for sample collection and storage. All samples and participant data were de-identified. The current study was approved by the University of Minnesota Institutional Review Board.

Table 1.

Characteristics and urinary biomarkers of the study subjects

	Case (N=202)	Control (N=211)	Total (N=413)	p value
Gender				0.736
Male	101 (50.0%)	109 (51.7%)	210 (50.8%)
Female	101 (50.0%)	102 (48.3%)	203 (49.2%)
Race				0.999
White	109 (54.0%)	114 (54.0%)	223 (54.0%)
Black	91 (45.0%)	95 (45.0%)	186 (45.0%)
Mixed	2 (1.0%)	2 (0.9%)	4 (1.0%)
GOLD				< 0.001
0	0 (0.0%)	211 (100.0%)	211 (51.1%)
2	105 (52.0%)	0 (0.0%)	105 (25.4%)
3	74 (36.6%)	0 (0.0%)	74 (17.9%)
4	23 (11.4%)	0 (0.0%)	23 (5.6%)
BMI				< 0.001
Mean (SD)	25.45 (5.74)	28.45 (5.63)	26.98 (5.87)
Median (Range)	24.20 (13.80, 39.80)	28.00 (14.50, 40.50)	26.20 (13.80, 40.50)
Age				0.158
Mean (SD)	59.74 (7.58)	58.67 (7.84)	59.19 (7.72)
Median (Range)	60.00 (41.00, 79.00)	58.00 (43.00, 79.00)	59.00 (41.00, 79.00)
FEV1/FVC (%)				< 0.001
Mean (SD)	47.50 (12.02)	77.36 (4.64)	62.75 (17.46)
Median (Range)	47.51 (21.23, 69.43)	76.89 (70.16, 96.67)	70.41 (21.23, 96.67)
Pack-years				0.729
Mean (SD)	43.28 (16.68)	42.69 (17.74)	42.98 (17.21)
Median (Range)	40 (20, 116)	39 (20, 120)	40.00 (20.00, 120.00)
Cotinine (ng/mL)				0.174
Mean (SD)	1069 (815.20)	1177 (793.57)	1124 (805.05)
Median (Range)	865 (20, 4586)	1024 (22, 4100)	952 (20, 4586)
3′-Hydroxycotinine (ng/mL)				0.369
Mean (SD)	4202 (4308)	4600 (4668)	4406 (4495)
Median (Range)	2721 (79, 23531)	3030 (25, 33200)	2870 (25, 33200)
3′-Hydroxycotinine/cotinine				0.476
Mean (SD)	4.06 (3.34)	3.83 (3.24)	3.94 (3.29)
Median (Range)	3.15 (0.29, 23.91)	2.83 (0.17, 18.90)	2.95 (0.17, 23.91)
Total Nicotine Equivalents (nmol/ml)				0.486
Number missing	4	0	4
Mean (SD)	34.84 (30.62)	36.98 (31.41)	35.95 (31.01)
Median (Range)	25.08 (0.81, 166.02)	28.64 (0.60, 197.73)	26.99 (0.60, 197.73)

Urine Sample Collection and Storage.

At study enrollment, a spot urine sample was collected, aliquoted and stored at −80 °C prior to transfer and processing.

Analysis of Urine Samples.

3-HPMA, HMPMA-1, HMPMA-2, and HMPMA-3 were analyzed essentially as described previously using 0.200 mL of urine.¹⁴ The method was modified to include 3-HPMA by including the internal standard [CD₃]3-HPMA (100 ng; from Toronto Research Chemicals). For HMPMA-1, HMPMA-2, and HMPMA-3, the internal standards [N-acetyl-D₃]HMPMA-1 (50 ng; Toronto Research Chemicals), [¹³C₃¹⁵N]HMPMA-2 (50 ng¹⁴), and [¹³C₃¹⁵N]HMPMA-3 (50 ng¹⁴) were added to 200 μL urine. The urine samples were partially purified by solid phase extraction as previously described,¹⁴ concentrated to dryness on a SpeedVac, and stored at −20 °C until LC-ESI-MS/MS analysis with a TSQ Vantage Instrument (ThermoElectron) coupled with a Dionex Ultimate 3000 UHPLC system (NCS-3500RS pump and WPS-3000PL autosampler) equipped with a 150 x 2.1mm, 3μm particle size Force Biphenyl HPLC column. The column was kept at 5 °C and eluted with solvents A (0.1% HCO₂H in H₂O) and B (80% CH₃CN, 20% CH₃OH, and 0.1% HCO₂H) with a flow rate of 0.1 mL/min. The following solvent program was used: 97% A and 3% B increased to 10% B in 10 min and held for 6 min, then increased to 18% B in 4 min, and finally to 50% B in 1 min and held for 2 min. MS/MS was performed in the negative ESI mode, with the following transitions: unlabeled HMPMA, m/z 234 → m/z 105 [(M-1) → HOCH₂CH(S)CH₃]; [N-acetyl-D₃]HMPMA-1, m/z 237 → m/z 105; [¹³C₃¹⁵N]HMPMA-2 and [¹³C₃¹⁵N]HMPMA-3, m/z 238 → m/z 105 and unlabeled 3-HPMA, m/z 220 → m/z 91 [(M-1) → HOCH₂CH₂S]; [N-acetyl-D₃]3-HPMA, m/z 223 → m/z 91 . The retention times were as follows: HMPMA-1, 25.6 min, HMPMA-2, 26.3 min, HMPMA-3, 23.3 min and 3-HPMA, 17.7 min. Further details of the mass spectrometer settings and calibration curves for the analysis have been described.¹⁴

Statistical Analysis.

Demographics and baseline characteristics were summarized by COPD status and overall using descriptive statistics and compared using t-tests for continuous variables and chi-square for categorical variables. Logistic regression models were performed to investigate the impact of the four biomarkers on COPD status. The 4 biomarkers (3-HPMA from acrolein, HMPMA-1 from crotonaldehyde, HMPMA-2 from methacrolein, and HMPMA-3 from methyl vinyl ketone) were each used individually in a model to avoid bias due to multi-collinearity. BMI, gender, race, age, and pack-years of smoking were included in the models as covariates. Biomarkers were normalized per μg/ml urine of creatinine and pmol/ml urine of total nicotine equivalents (the sum of nicotine, cotinine, 3’-hydroxycotinine and their glucuroides) and log transformed due to skewness. Subgroup analyses were conducted for race, gender, and GOLD level (GOLD 2, GOLD 3 & 4, and GOLD 4).²¹ Additional analyses using log biomarkers without normalization and using creatinine, cotinine, 3’-hydroxycotinine, and nicotine levels as covariates, log biomarkers normalized by creatinine and using total nicotine equivalents as a covariate, and log biomarkers normalized by creatinine and using cotinine, 3’-hydroxycotinine, and nicotine as covariates were conducted. All reported p-values are two sided and a significance level of 0.05 was used. Statistical analyses were performed in R version 4.2.1 (“Funny Looking Kid”).

RESULTS

Study participant characteristics

Table 1 summarizes the demographics of the individuals in our study. This cohort consisted of 50.8 % males with an average age of 59 years; 45% of the subjects identified as Black and 54% White. All were currently smoking tobacco with an average cotinine level of 1069 ng/ml in cases and 1177 ng/ml in controls. Lung function (FEV1) was normal in controls (GOLD 0) and ranged from 52.0% predicted GOLD 2, 36.6% predicted GOLD 3 and 11.4% predicted GOLD 4 in cases. The diffusing capacity of the lungs for carbon monoxide (DLCO) was not available from this cohort. The average BMI for those with COPD was lower (mean = 25.45) than subjects without COPD (mean = 28.45, p < 0.001).

Metabolite levels related to COPD

To be included in this study, all participants were actively smoking as confirmed by a positive urine cotinine test and the analyses summarized in Table 1. Urinary biomarkers of nicotine exposure did not differ between cases and controls. These included cotinine (p = 0.174), 3’-hydroxycotinine (p = 0.369), ratio of 3’-hydroxycotinine/cotinine (p = 0.476) and total nicotine equivalents (p = 0.486) (Table 1).

Mean levels of the urinary metabolites of the α,β-unsaturated carbonyl compounds (pmol/ml urine ± S.D.) in all subjects were: 3-HPMA (from acrolein) 8221 ± 15470; HMPMA-1 (from crotonaldehyde) 4538 ± 10,210; HMPMA-2 (from methacrolein) 741 ± 1141; and HMPMA-3 (from methyl vinyl ketone) 2925 ± 3663. The amounts of the metabolites were normalized to account for urinary concentration (per μg/ml urine of creatinine) and tobacco exposure (pmol/ml urine of total nicotine equivalents) which were then log transformed due to skewness.

There was no consistent relationship between metabolite levels and age, sex, or race. There was an inverse relationship with body mass index (BMI). In models adjusting for age, sex, race, pack years of tobacco use and BMI, all four metabolites - 3-HPMA from acrolein (aOR 1.17 [95%CI: 0.98-1.41]), HMPMA-1 from crotonaldehyde (aOR 1.12 [95%CI: 0.93-1.36]), HMPMA-2 from methacrolein (aOR 1.16 [95%CI: 0.99-1.37]), and HMPMA-3 from methyl vinyl ketone (aOR 1.14 [95%CI: 0.94-1.37]) - were increased in individuals with COPD that spanned all GOLD levels (Figure 1). However, none reached statistical significance. When stratified by GOLD status, higher levels of three of the metabolites (of acrolein, crotonaldehyde and methacrolein) were associated with worsening lung function (GOLD 3-4) (Figure 2). Specifically, there was a 23% increased odds ratio of the acrolein metabolite 3-HPMA being associated with GOLD 3-4 compared to GOLD 0 (aOR 1.23 [95%CI: 0.98-1.55), 19% for the crotonaldehyde metabolite HMPMA-1 (aOR 1.19 [95%CI: 0.94-1.51), and 23% for the methacrolein metabolite HMPMA-2 (aOR 1.23 [95%CI: 1.00 -1.52). Only the methacrolein metabolite HMPMA-2 met statistical significance (p=0.047). The molar sum of the metabolites was not significantly related to COPD at any GOLD status or in the comparison of GOLD 3-4 to GOLD 0. Further comparisons are summarized in the Supporting Information, Table 1.

Figure 1.

Relationship of subject characteristics and biomarker levels of methacrolein (HMPMA-2), acrolein (3-HPMA), crotonaldehyde (HMPMA-1), and methyl vinyl ketone (HMPMA-3) to presence of COPD (odds ratio, 95% CI) in 413 confirmed cigarette smokers of the SPIROMICS study.

Figure 2.

Relationship of urinary biomarkers of acrolein (3-HPMA), crotonaldehyde (HMPMA-1), methacrolein (HMPMA-2), and methyl vinyl ketone (HMPMA-3) to COPD (odds ratio, 95% CI) among 413 cigarette smokers in the SPIROMICS study, stratified by GOLD status.

DISCUSSION

In this cohort of current smokers with and without COPD, our study is the first demonstrating an association of methacrolein exposure with COPD. We found a statistically significant association of urinary levels of HMPMA-2, the mercapturic acid of methacrolein, with a 23% increased odds of higher severity COPD (GOLD 3-4) in smokers. Increased risks of COPD and for GOLD 3-4 severity COPD were also observed for the metabolites of acrolein, crotonaldehyde, and methyl vinyl ketone, but they did not reach statistical significance.

Methacrolein is a potent sensory irritant that causes a concentration dependent decrease in respiratory rate in mice.²² It has been suggested that a chemical reaction with methacrolein, presumably Michael addition of cellular thiols or other nucleophiles, is involved in the activation of the sensory irritant receptor, and this occurs less readily than in the case of acrolein due to the electronic effects of the methyl substituent on its double bond,²² but further mechanistic studies have not been reported. Differences in the processing of glutathione conjugates of methacrolein versus the other α,β-unsaturated carbonyl compounds considered here could also have affected our results. In contrast, multiple studies summarized here have implicated acrolein in COPD, with a particular focus on its reactivity with nucleophilic sites in peptides and proteins as potentially important etiological factors for COPD in cigarette smokers, and similar effects would be expected for methacrolein although it is less reactive. Yasuo et al recently reported increased amounts of acrolein in the plasma of smokers who were COPD patients compared to smokers without COPD, although it was not clear why there would be free acrolein in plasma, or whether acrolein may have been released from proteins during analysis.²³ Earlier studies reviewed by Kehrer and Biswal summarized acrolein’s reactions with glutathione, its effects on glutathione depletion, and the consequences of acrolein oxidant toxicity.²⁴ Acrolein can also form protein adducts specifically targeting lysine²⁵ and cysteine.^26,27 Cai et al investigated reactions of acrolein with proteins and peptides; they concluded that cysteine was the most reactive amino acid and that the initial products reacted further with protein amino groups.²⁸ Reaction of acrolein with Cys-34 of albumin by Michael addition was reported by Columbo et al and others; reactions with lysine and histidine were also observed.^29–31 Reactions of acrolein with hemoglobin and albumin have recently been reviewed.³² Protein adducts of acrolein can also be generated during lipid peroxidation.³³ In the lung, acrolein disrupts structural and functional features of surfactant protein A by its reactions with cysteine, lysine, and histidine.³⁴ Acrolein is also associated with a pro-inflammatory state. Fachinnetti and coworkers demonstrated that acrolein reproduced effects related to COPD as observed in bronchiolar lavage samples of COPD patients. These effects, which included secretion of inflammatory mediators, were inhibited by glutathione and N-acetyl cysteine.³⁵ Moretto et al reported that acrolein as well as cigarette smoke caused release of interleukin 8 (IL-8) from cultured normal human lung fibroblasts and small airway epithelial cells. IL-8 activates adhesion molecules, neutrophils, T-cells, and monocytes; these studies implicated acrolein in chronic airway inflammation and COPD.³⁶ Cigarette smoke associated acrolein increases mononuclear cells, and eosinophil and neutrophil chemoattractants in murine lung tissue. Whereas depletion of aldehydes in smoke reduced IL-8 production in BEAS-2b cells, it was restored by acrolein.³⁷ Exposure models of differentiated human primary bronchial epithelial cells exposed to either cigarette smoke or a mixture of acetaldehyde, formaldehyde, and acrolein demonstrated increased IL-8; however this was only partially recapitulated by the aldehydes. This indicates that compounds in addition to these aldehydes contribute to the effects of cigarette smoke’s effects on airway epithelial cells.³⁸ The role of methacrolein was not considered in that study.

In animal models, treatment of rats with acrolein by i.p. injection resulted in multiple toxic effects including emphysematous lung tissue remodeling, enlargement of alveolar air spaces, apoptosis, endoplasmic reticulum stress, and emphysema; accumulation of acrolein protein adducts was observed in lung tissue from smokers with COPD.³⁹ Horinouchi et al reported that acrolein and methyl vinyl ketone activate the PKC/NOX/ROS signaling pathway leading to plasma membrane damage and apoptosis,⁴⁰ representing a possible mechanism in the development of emphysema. This was further highlighted when exposed primary nasal epithelial cell cultures to acrolein induced cell death.⁴¹ Consistent with all of this data, acrolein is ranked as the major non-cancer respiratory hazard of cigarette smoke.^7,42

Although the toxic effects and association of COPD with acrolein are clearly elucidated in the literature, our results indicate that further research on the effects of methacrolein on risk of COPD in smokers is necessary. While methacrolein is less reactive than acrolein and is found in substantially lower concentrations in cigarette smoke than acrolein, it may be detoxified less readily by the glutathione-S-transferase pathway, leading to its potentially increased persistence in the lung after inhalation of cigarette smoke. Although our study did not find a significant association of acrolein exposure with COPD, as measured by urinary 3-HPMA, the relationship that was observed is consistent with the reactive properties of α,β-unsaturated aldehydes, of which acrolein is the prime example. We also note that the mean concentration of 3-HPMA in urine was 10 times greater than that of the methacrolein metabolite HMPMA-1, which is consistent with the higher levels of acrolein than methacrolein in cigarette smoke. 3-HPMA is an accepted biomarker of acrolein exposure, but its levels are also affected by endogenous processes such as lipid peroxidation¹⁷ and it may therefore lack adequate specificity to probe the relationship between acrolein exposure by inhalation and COPD. Oral cell DNA adducts of acrolein may be a better measure of exposure, but further studies are required.⁴³

We did not observe consistent sex differences in the levels of the mercapturic acids quantified in this study. Some studies have found that women smokers tended to have more severe disease than men and a higher mortality rate.⁴⁴ It is not clear whether sex affects the course of COPD or if women have a higher susceptibility to toxicants in cigarette smoke. These are complex issues which require further study.

A clear strength of this study is its basis in the SPIROMICS study, a longitudinal prospective cohort study funded by the National Heart, Lung, and Blood Institute. The SPIROMICS study is designed specifically to identify biomarkers associated with COPD and has a collection of urine samples from cigarette smokers and non-smoking controls with COPD, as well as non-smoking controls without COPD. We analyzed urine samples from 413 cigarette smokers, 202 with COPD and 211 without COPD, in this unique biomarker study. A limitation of this study was the biomarkers themselves, which, while clearly derived from acrolein, methacrolein, crotonaldehyde, and methyl vinyl ketone, may have some endogenous sources unrelated to uptake of these compounds from cigarette smoke thus possibly decreasing their specificity.

In summary, we present evidence based on quantified urinary biomarkers that methacrolein exposure in cigarette smokers is significantly related to the incidence of COPD. These are the first results directly implicating this toxic α,β-unsaturated aldehyde as a potential cause of COPD in smokers. Biomarkers of exposure to methacrolein and related compounds may provide new mechanistic insights related to COPD etiology and supplement available approaches to identify those smokers at highest risk of COPD who could be targeted for intensive cessation strategies.

Supplementary Material

Data summary table (XLSX)