Association Between Serum Cathepsin S and Mortality in Older Adults FREE

Cathepsin S is a cysteine protease involved in major histocompatibility complex class II antigen presentation¹ and in intracellular and extracellular proteolysis.² Moreover, a causal interplay between cathepsin S activity and inflammatory activity has been proposed.^2- 9 Cathepsin S activity is regulated at the cellular level by its endogenous inhibitor cystatin C.¹⁰

Current data regarding factors that influence circulating levels of cathepsin S are scarce. Previous studies have reported higher circulating levels of cathepsin S in patients with obesity,^11- 12 diabetes, or cardiovascular disease.¹³ It also has been reported that weight loss reduces circulating levels of cathepsin S.^11- 12 Furthermore, higher circulating levels of cathepsin S have been shown to be associated with increased inflammatory activity.^11,14

Experimental studies suggest that cathepsin S activity is involved in the development of cardiovascular disease via promotion of atherosclerotic plaques and destabilization of advanced plaques.^15- 21 Moreover, cathepsin S activity has been implicated in the development of cancer via stimulation of cancer cell migration and tumor angiogenesis.^22- 27

Based on previous experimental reports suggesting a role for cathepsin S activity in the development of cardiovascular disease and cancer, we hypothesized that circulating levels of cathepsin S would be a marker for an increased mortality risk. Accordingly, we investigated the association between serum levels of cathepsin S and the risk of total mortality in a community-based sample of elderly men with prespecified analyses on the association between cathepsin S levels and cause-specific deaths from cardiovascular disease and cancer. As a second step, we investigated the association between cathepsin S and total mortality in an independent community-based cohort of elderly men and women.

The Uppsala Longitudinal Study of Adult Men (ULSAM) was initiated in 1970. All 50-year-old men born between 1920 and 1924 and living in Uppsala, Sweden, were invited to participate in a health survey aimed at identifying cardiovascular risk factors²⁸ (described in detail at http://www.pubcare.uu.se/ULSAM). All 70-year-old individuals living in Uppsala, Sweden, between 2001 and 2004 were eligible for the Prospective Investigation of the Vasculature in Uppsala Seniors (PIVUS) study²⁹ (described in detail at http://www.medsci.uu.se/pivus/pivus.htm). All participants gave written informed consent and the ethics committee of Uppsala University approved the study protocols.

The investigations in the ULSAM and PIVUS studies were performed using the same standardized methods, including anthropometrical measurements, blood pressure, blood sampling, and questionnaires regarding socioeconomic status, medical history, smoking habits, medication use, and physical activity level.^28- 29 Venous blood samples were drawn in the morning after an overnight fast and stored at –70°C.

Serum levels of cathepsin S were measured by an enzyme-linked immunosorbent assay (human cathepsin S [total] DY1183; R&D Systems, Minneapolis, Minnesota). In the ULSAM study, blood samples were frozen for 14.6 years (range, 12.9-16.7 years) until the analysis of cathepsin S was performed in April 2008. In the PIVUS study, blood samples were frozen for 5.5 years (range, 3.9-7.2 years) until the analysis of cathepsin S was performed in December 2009. The intraassay coefficient of variation was 7%. To date, there are no established reference values for circulating cathepsin S in humans.

Inflammatory markers and cystatin C were measured as previously described for the ULSAM study³⁰ and the PIVUS study.³¹ Diabetes mellitus was diagnosed using a fasting plasma glucose level of 126 mg/dL or higher (to convert to mmol/L, multiply by 0.0555) or confirmed if the participant was taking medication for diabetes. Prevalent cardiovascular disease at baseline was defined as a history of ischemic heart disease or cerebrovascular disease, Q or QS complexes, or left bundle-branch block on baseline electrocardiogram.

Leisure-time physical activity was assessed using the following questions: (1) Do you spend most of your time reading, watching television, going to the cinema, or engaging in other, mostly sedentary activities? (2) Do you often go walking or cycling for pleasure? (3) Do you engage in any active recreational sports or heavy gardening for at least 3 hours every week? (4) Do you regularly engage in hard physical training or competitive sport? The highest positive physical activity response level was used for each participant. Education level was stratified as low (elementary school, 6-7 years), medium (high school), or high (college studies).

The Swedish cause of death register was used to define total mortality, cardiovascular mortality (death from ischemic heart disease or cerebrovascular disease; International Classification of Diseases, Ninth Revision [ICD-9] codes 410-414, 430-438 or International Statistical Classification of Diseases, 10th Revision [ICD-10] codes I20-I25, I60-I69/G45), and cancer mortality (ICD-9 codes 140-239 and ICD-10 codes C00-D48). Data on cause-specific mortality were not available in the PIVUS cohort.

ULSAM Study. The associations of serum level of cathepsin S with total, cardiovascular, and cancer mortality were investigated using Cox proportional hazard regression and 3 multivariable models. Model A was age-adjusted, model B was for lifestyle factors (age, body mass index [calculated as weight in kilograms divided by height in meters squared], smoking status, leisure-time physical activity, and education level), and model C was for cardiovascular risk factors (age, systolic blood pressure, diabetes, smoking status, body mass index, total cholesterol, high-density lipoprotein cholesterol, antihypertensive treatment, lipid-lowering treatment, and prevalent cardiovascular disease).

Because cathepsin S activity is modulated by inflammation and cystatin C, we performed secondary analyses in which inflammatory markers (C-reactive protein, interleukin 6 [IL-6]) or cystatin C were added to model C. Because effects of long-term freezing on cathepsin S levels are uncertain, we also added freezer time as a covariate in separate models. Proportional hazards assumptions were confirmed using the Schoenfeld test.

In our primary analyses, we modeled cathepsin S as a continuous variable (expressed as 1-unit increase of serum cathepsin S). We also performed prespecified multicategory models comparing risk in cathepsin S quintiles 2, 3, 4, and 5 with that in quintile 1, and threshold models (quintile 5 vs quintiles 1-4). To gain additional insights into the potential nonlinearity of the associations, we examined the Cox regression models using penalized splines.

We performed stratified analyses in participants with and without prevalent cardiovascular disease at baseline (n = 249 and n = 760, respectively). Additionally, we performed tests for effect modification by prevalent cardiovascular disease and by including a multiplicative interaction term in model C. We also investigated the association between serum cathepsin S and cancer mortality after exclusion of participants with cancer at baseline or during the first 2 years of follow-up (n = 930) to limit the possibility of reverse causation as an explanation of our findings. In secondary analyses, multiple imputation methods were used to account for the potential influence of missing data.

PIVUS Study. We investigated the association between serum cathepsin S and total mortality in the PIVUS cohort using a similar protocol for the statistical analyses with the following major exceptions: (1) sex was included in all multivariable models (models A-C); (2) sex-stratified analyses were performed; and (3) an evaluation was performed to determine whether the association between cathepsin S and mortality was mediated by additional inflammatory pathways (IL-1β, IL-8, IL-10, tumor necrosis factor, and interferon γ).

A 2-sided P value of less than .05 was regarded as significant in all analyses. For all analyses, we used the STATA statistical software package version 10.0 (StataCorp, College Station, Texas).

Our analyses were based on the third examination cycle of ULSAM when participants were approximately 71 years old (baseline period: 1991-1995). Of the 1681 men who were invited, 1221 agreed to participate (73%). Of these 1221 men, 212 were excluded due to missing data on cathepsin S (n = 98) or other covariates (n = 114), leaving 1009 participants as the study sample. The incidence rate for mortality for those not included in the analyses (nonparticipants or individuals with missing data at baseline) was 5.70/100 person-years at risk.

Baseline characteristics are shown in Table 1. During follow-up (median, 12.6 years; range, 0.1-15.3 years; end of follow-up: 2006), 413 participants died (incidence rate: 3.59/100 person-years at risk). The incidence rate per 1-unit increase of cathepsin S was 0.44/100 person-years at risk. A total of 131 deaths were due to cardiovascular disease (incidence rate: 1.14/100 person-years at risk) and 148 deaths were due to cancer (incidence rate: 1.29/100 person-years at risk). The incidence rates in cathepsin S quintiles are shown in Table 2.

A 1-unit increase of cathepsin S was associated with a 1.04 higher hazard ratio (HR) for total mortality after adjustment for age (model A), lifestyle factors (model B), and established cardiovascular risk factors (model C; Table 3). In multicategory and threshold models, participants in the highest quintile of cathepsin S were at higher risk for total mortality compared with participants in quintile 1 and quintiles 1 through 4, respectively (models A-C; Table 3). The association between cathepsin S and mortality remained similar after the addition of cystatin C (HR per 1-unit increase of cathepsin S, 1.03 [95% CI, 1.01-1.06]; P = .02), IL-6 (HR, 1.04 [95% CI, 1.01-1.06]; P = .008), and freezer time of the sample (HR, 1.03 [95% CI, 1.01-1.06]; P = .02) to multivariable model C. This association was no longer statistically significant after adding serum C-reactive protein to model C (HR per 1-unit increase, 1.03 [95% CI, 1.00-1.05]; P = .06).

The cumulative incidence of total mortality in quintile 5 vs quintiles 1 through 4 is shown in Figure 1. Examination of regression splines suggests a linear increase in the HR for mortality with increasing levels of cathepsin S (Figure 2). The association between serum levels of cathepsin S and mortality was essentially unaltered when missing data were imputed (eTable 1 and eTable 2).

In continuous models, a 1-unit increase of cathepsin S was associated with a 1.05 higher HR for cardiovascular mortality in models A and B (Table 3). In threshold models, participants in quintile 5 had HRs that were between 1.62 and 1.82 higher for cardiovascular mortality compared with participants in quintiles 1 through 4 (models A-C; Table 3).

The multiplicative interaction term for cardiovascular disease at baseline was not statistically significant (P = .55). Nonetheless, we performed stratified analyses in participants with and without prevalent cardiovascular disease at baseline because previous experimental and clinical data indicate an important role for cathepsin S in the destabilization of advanced atherosclerotic plaques in patients with prevalent cardiovascular disease. In these stratified analyses, the association between serum cathepsin S and cardiovascular disease mortality appeared stronger in participants with prevalent cardiovascular disease, although none of these stratified analyses was statistically significant (with cardiovascular disease: cathepsin S quintile 5 vs quintiles 1-4, HR, 1.80 [95% CI, 0.98-3.34]; P = .06, model C; without cardiovascular disease: HR, 1.35 [95% CI, 0.82-2.22]; P = .23).

The association between cathepsin S and cardiovascular mortality remained similar after adjusting for C-reactive protein, IL-6, and cystatin C (P < .03 for all; eTable 3).

A 1-unit increase of cathepsin S was associated with HRs that were 1.05 to 1.06 higher for cancer mortality (models A-C; Table 3). Moreover, in multicategory models, participants in the highest quintile of cathepsin S were at higher risk for cancer mortality compared with participants in quintile 1 (models A-C; Table 3). The results were similar after exclusion of participants with a history of cancer at baseline and those who developed cancer during the first 2 years of follow-up (model C; HR for 1-unit increase of cathepsin S, 1.06 [95% CI, 1.01-1.11], P = .01) and after the addition of C-reactive protein, IL-6, or cystatin C to multivariable model C (P < .05 for all; eTable 4).

Of the 2025 individuals invited to participate, 1014 agreed to participate (50%). Of these 1014 individuals, 27 were excluded due to missing data on cathepsin S (n = 13) or other covariates (n = 14), leaving 987 participants as the study sample (baseline period: 2001-2004).

The baseline characteristics are shown in Table 1. The mean (SD) serum level of cathepsin S was 19.5 (0.6) μg/L in men and 19.0 (0.6) μg/L in women. During follow-up (median: 7.9 years; range, 0.3-9.8 years; end of follow-up: 2010), 100 participants died (incidence rate: 1.32/100 person-years at risk). The incidence rate per 1-unit increase of cathepsin S was 0.15/100 person-years at risk. The incidence rates in cathepsin S quintiles are shown in Table 2.

A 1-unit increase of cathepsin S was associated with a 1.03 higher HR for total mortality after adjustment for age, sex, lifestyle factors, and established cardiovascular risk factors (models B and C; Table 4). Moreover, in multicategory models, participants in quintile 5 had a HR of between 2.10 and 2.25 compared with participants in quintile 1 (models A-C; Table 4). The risk estimates were similar after adjustment for various inflammatory markers and cystatin C, although some associations were not statistically significant (eTable 5). The cumulative incidences of total mortality in quintile 5 vs quintiles 1 through 4 are shown in Figure 1. Examination of regression splines suggests a linear increase in mortality with increasing levels of cathepsin S (Figure 2).

There was no statistically significant effect modification for sex (P = .15). Nonetheless, we performed sex-stratified analyses. In these analyses, the association between cathepsin S and mortality appeared stronger in women than in men in model A (women: HR for 1-unit increase of cathepsin S, 1.06 [95% CI, 1.01-1.11], P = .03; men: HR for 1-unit increase of cathepsin S, 1.01 [95% CI, 0.97-1.06], P = .58) and in model B (women: HR for 1-unit increase of cathepsin S, 1.05 [95% CI, 1.00-1.11], P = .04; men: HR for 1-unit increase of cathepsin S, 1.02 [95% CI, 0.97-1.07], P = .42); however, neither analysis was significant in model C (eTable 6).

In 2 independent cohorts of elderly individuals (age ≥70 years), higher serum levels of cathepsin S were associated with increased mortality risk even after adjustment for age, lifestyle factors, and cardiovascular risk factors. Moreover, in the ULSAM cohort, serum cathepsin S was independently associated with cause-specific mortality from cardiovascular disease and cancer. These results were similar after further adjustment for inflammatory markers and cystatin C and after excluding those with a diagnosis of cancer at baseline or up to 2 years after. In the PIVUS cohort, the association between cathepsin S and mortality appeared stronger in women. However, this finding should be taken with caution given the modest sample size of the different strata. Our community-based data confirm and extend previous experimental research that suggest that cathepsin S activity may be involved in the pathological processes leading to cardiovascular disease, cancer, and death.

In patients with peripheral artery disease, cathepsin S levels have not been associated with increased mortality.³² The conflicting results between our study and this prior study³² may be explained by differences in the sample size, number of events, follow-up time, and clinical characteristics of the study samples. To our knowledge, no prior study has reported the association between circulating levels of cathepsin S and mortality risk in the general population.

Although we cannot establish causality in our observational study, there are several potential pathways that may explain the association between cathepsin S and mortality. First, multiple lines of evidence suggest a role for cathepsin S at different stages in the atherosclerotic process (Figure 3A). Cathepsin S aggravates foam cell formation by degrading low-density lipoprotein cholesterol and reducing cholesterol efflux from macrophages.¹⁷ In knockout studies in mice, deficiency of cathepsin S reduces atherosclerosis by up to 60%.²⁰ In human atherosclerotic lesions, the immunoreactivity of cathepsin S is localized mainly in the fibrous caps and macrophage-rich shoulder regions where plaque ruptures usually occur.¹⁹ The fact that the association between cathepsin S and cardiovascular mortality appeared stronger in participants with prevalent cardiovascular disease may indicate that circulating levels of cathepsin S could be a marker of the plaque's vulnerability in patients with prevalent cardiovascular disease. Another possible interpretation could be that higher circulating levels of cathepsin S are a consequence of prevalent cardiovascular disease, which in turn leads to a higher mortality risk (ie, reverse causation). Additional studies are needed to elucidate this issue.

Second, cathepsin S is highly expressed in different malignant tissue²⁶ and is involved in the progression of cancer via tumor proliferation, angiogenesis, and cell migration (Figure 3B).^{22- 23,25,27,33} The fact that the association between serum level of cathepsin S and cancer mortality was robust even after exclusion of participants with prevalent cancer at baseline and of those who developed cancer during the first 2 years of follow-up suggests that cathepsin S activity may be involved early in the processes leading to clinically overt cancer. However, it is possible that some individuals had undiagnosed cancer at baseline.

Third, other potential pathways should be highlighted. Cystatin C, the endogenous inhibitor of cathepsin S activity,¹⁰ is a strong predictor for cardiovascular disease^30,34 and cancer.³⁵ Moreover, cathepsin S activity is involved in inflammatory processes^14,36- 40 and in conditions with metabolic disturbances such as obesity,^11- 12 diabetes,¹³ and dyslipidemia⁴¹; these factors are associated with an increased risk of mortality.⁴² The fact that the association between cathepsin S and mortality was similar after adjustments for cystatin C, inflammatory markers, and established cardiovascular risk factors suggests that these pathways are not the sole explanation of our findings. Still, we cannot exclude residual confounding by these pathways because circulating biomarkers may have limited usefulness as indicators of intracellular processes.

Given its putative role in atherogenesis and tumorigenesis, cathepsin S has been put forward as a possible target of pharmaceutical intervention and the development of selective cathepsin S inhibitors is ongoing.^4,18,43- 46 Some of these inhibitors are being evaluated in various phases clinical trials.⁴⁷ If these drugs are found to be effective, tools for identifying target groups and for monitoring treatment need to be developed. Our study suggests that it is possible to assess serum levels of cathepsin S in large populations. Future intervention trials are needed to evaluate whether cathepsin S inhibition is a safe and effective pharmacological target for these diseases.

The strengths of our investigation include the use of 2 independent community-based study samples with longitudinal data and the detailed characterization of the study participants. To our knowledge, the ULSAM and PIVUS cohorts are the largest cohorts that have analyzed circulating levels of cathepsin S.

Limitations include the unknown generalizability to other age and ethnic groups. Moreover, some of the risk estimates in the PIVUS cohort did not reach statistical significance. Thus, additional large-scale studies in other ethnicities and in other age groups are needed to properly validate our findings. Still, the consistency of the association between cathepsin S and mortality in the 2 cohorts would argue against the results as chance findings.

Higher serum levels of cathepsin S were associated with an increased risk for mortality in 2 independent community-based cohorts. Further research is needed to determine whether cathepsin S measurements have any utility in clinical settings.