A total of 52 subjects in the lifestyle group (LIFE) (28 males, 24 females) and 52 in the control group (CON) (27 males, 25 females) participated in this cross-sectional study (Table 1). The sex distribution did not differ between groups (Χ2 = 0.039, p = 0.844) and analyses were conducted for all study participants combined. A separate analysis (data not shown) showed that all group differences reported in this paper were significant when comparing males and females separately.
Table 1
Subject characteristics. Data expressed as mean ± standard deviation.
Variable | Lifestyle (n = 52) (28M, 24F) | Control (n = 52) (27M, 25F) |
Age (yrs) | 47.5 ± 12.2 | 51.1 ± 10.5 |
Education (yrs) | 16.3 ± 3.4 | 15.2 ± 3.1 |
Weight (kg) | 67.8 ± 11.8* | 101.5 ± 15.5 |
Height (cm) | 172 ± 9.0 | 171 ± 9.5 |
Waist circumference (cm) | 82.5 ± 7.5* | 112.2 ± 11.5 |
Sagittal abdominal diameter (SAD) (cm) | 17.5 ± 2.13* | 26.9 ± 3.4 |
BMI (body mass kg/height m2) | 22.9 ± 2.6* | 34.5 ± 4.0 |
Body fat (%) | 22.6 ± 7.3* | 40.8 ± 7.2 |
Fat mass index (fat mass kg/height m2) | 5.32 ± 1.90* | 14.2 ± 3.57 |
VO2max (ml.kg.−1min−1) | 37.7 ± 8.8* | 20.3 ± 8.4 |
* p < 0.001, group difference. |
Age, education level, and height did not differ significantly between LIFE and CON groups (Table 1). Marital status (LIFE, 71% married, 29% other; CON, 57% married, 43% other) did not differ significantly between groups (Χ2 = 6.20, p = 0.102). Race and ethnic backgrounds (LIFE, 88% white, 12% other; CON, 73% white, 27% other) did not differ significantly between groups (Χ2 = 6.86, p = 0.231). Group medical history comparisons for 41 past or current conditions showed significant differences for hypertension (LIFE 13%, CON 29%), arthritis (LIFE 4%, CON 21%), depression (LIFE 8%, CON 21%), sleep problems (LIFE 4%, CON 19%), and gallstones (LIFE 0%, CON 13%) (all p ≤ 0.05). Current medication use differed significantly between groups for beta-blockers/ace inhibitor hypertension drugs (LIFE 8%, CONs 23%), statin-based drugs (LIFE 8%, CONs 23%), depression drugs (LIFE 8%, CONs 21%), and metformin (LIFE 0%, CON 8%) (all p ≤ 0.05). None of the subjects in LIFE and only 2 in CON were current cigarette smokers, but 12% and 27% of LIFE and CON, respectively, reported smoking at least 100 cigarettes in their entire lifetimes (Χ2 = 3.96, p = 0.047). Body mass, waist circumference, body mass index (BMI), sagittal abdominal diameter (SAD), and body composition (body fat percentage and fat mass index (FMI)) differed significantly between the groups (p < 0.001) (Table 1).
Estimated VO2max was 86% higher in LIFE vs. CON (p < 0.001) (Table 2). Total physical activity at work, as part of house and yard work, to get from place to place, and during recreation, exercise, and sport was calculated as MET-min/week and was 136% higher in LIFE vs. CON (p < 0.001) (Table 2). Weight adjusted leg/back and handgrip dynamometer strength were significantly higher in LIFE vs. CON (p < 0.001) (Table 2). Resting heart rate (RHR) and diastolic blood pressure (dBP) but not systolic blood pressure (sBP) was significantly lower in LIFE vs. CON (p < 0.001) (Table 2).
Table 2
Lifestyle characteristics and related biomarkers. Data expressed as mean ± standard deviation. * p < 0.01, group difference.
Variable | Lifestyle (n = 52) | Control (n = 52) |
VO2max (ml.kg.−1min−1) | 37.7 ± 8.8* | 20.3 ± 8.4 |
Physical activity (MET-min/week) | 5463 ± 3723* | 2319 ± 2614 |
Leg/back strength (kg/kg body mass) | 1.62 ± 0.6* | 1.10 ± 0.5 |
Handgrip strength (sum right, left hands) (kg/kg body mass) | 1.49 ± 0.6* | 1.01 ± 0.5 |
Resting heart rate (RHR) (beats/min) | 61.2 ± 10.1* | 70.0 ± 10.5 |
Systolic blood pressure (sBP) (mm Hg) | 115 ± 14.5 | 117 ± 14.5 |
Diastolic blood pressure (dBP) (mm Hg) | 67.8 ± 10.0* | 73.0 ± 10.5 |
Total mood disturbance (TMD) | 90.4 ± 7.1* | 94.8 ± 8.1 |
Fruit & vegetable (servings/day) | 5.1 ± 2.0* | 2.9 ± 1.4 |
Red meat (servings/day) | 0.45 ± 0.64* | 1.21 ± 0.72 |
Food nutrient index (FNI) | 74.8 ± 15.8* | 60.9 ± 15.8 |
Serum C-reactive protein (CRP) (mg/L) | 0.77 ± 0.78* | 2.61 ± 2.63 |
Serum insulin (U/L) | 4.45 ± 2.7* | 16.6 ± 12.6 |
Serum glucose (mg/dl) | 86.8 ± 6.7* | 99.8 ± 19.8 |
Homeostatic model assessment of insulin resistance (HOMA-IR) | 0.97 ± 0.62* | 4.23 ± 3.31 |
White blood cell count (109/L) | 4.8 ± 1.1* | 5.8 ± 1.5 |
Serum cholesterol (mg/dl) | 188 ± 31.5 | 188 ± 34.1 |
Serum low density lipoprotein (LDL) cholesterol (mg/dl) | 106 ± 25.9 | 111 ± 30.1 |
Serum high density lipoprotein (HDL) cholesterol (mg/dl) | 68.0 ± 17.5* | 52.6 ± 17.9 |
Serum triglycerides (mg/dl) | 75.3 ± 30.4* | 141 ± 121 |
Serum bilirubin (mg/dl) | 0.63 ± 0.31* | 0.45 ± 0.20 |
Serum albumin (g/L) | 4.61 ± 0.24* | 4.41 ± 0.27 |
Serum carbon dioxide (mEq/L) | 24.3 ± 1.7* | 22.7 ± 2.6 |
Serum alkaline phosphatase (IU/L) | 64.5 ± 16.7* | 76.8 ± 27.3 |
Total mood disturbance (TMD) calculated from the Profile of Mood States questionnaire (POMS) was modestly but significantly lower in LIFE vs. CON (p = 0.004). Other questions related to stress and anxiety did reveal important group differences.
Fruit and vegetable intake was higher and red meat intake lower in LIFE vs. CON (both p < 0.001). The food nutrient index (FNI) calculated from 3-day food records and daily intake of eight micronutrients (vitamins A,C,D,E, folate, calcium, magnesium, potassium) was significantly higher in LIFE vs. CON (p < 0.001) (Table 2).
Biomarkers including serum C-reactive protein (CRP) (-70%), serum insulin (-73%), serum glucose (-13%), the homeostatic model assessment of insulin resistance (HOMA-IR) (-77%), and serum triglycerides (-47%), but not total serum cholesterol or LDL-cholesterol, were significantly lower in LIFE vs. CON (all p < 0.001) (Table 2). HDL-cholesterol was 29% higher in LIFE vs. CON (p < 0.001) (Table 2). Serum bilirubin, albumin, and carbon dioxide were significantly higher and serum alkaline phosphatase lower in LIFE vs. CON (all p < 0.007). Other serum chemistries including blood urea nitrogen (BUN), creatinine, glomerular filtration rate (GFR), sodium, calcium, protein, and aspartate aminotransferase (AST) did not differ significantly between groups (data not shown). Serum thyroid stimulating hormone (TSH) and calcitriol (1,25-OH vitamin D) did not differ between groups (data not shown).
The untargeted proteomics analysis identified 970 proteins in at least one DBS sample, and 725 proteins in every sample. Principal component analysis (PCA) revealed outlier data from one study participant, and these data were removed from the proteomics analysis. Two-sample t-tests (LIFE vs. CON) with permutation-based FDR correction (Q < 0.05 alpha level), showed that blood levels of 39 proteins were found to be significantly different between groups. When the data were expressed as LIFE/CON group ratios, a total of 18 blood proteins were lower and 21 proteins were higher in LIFE vs. CON. Protein descriptions and functions for the 39 proteins differing between the LIFE and CON groups are listed in Table 3.
Table 3
Descriptions and names of proteins that were lower (n = 18) and higher (n = 21) in the lifestyle versus control group (all 39 proteins, FDR correction, q < 0.05). Group fold differences (lifestyle/control) are log2 transformed with minus values representing proteins downregulated in the lifestyle vs. control group. “*” indicates the 15 proteins that were in common with the list of 20 proteins from multivariate modeling that predicted group membership. The 5 of 20 proteins included in the multivariate model that are not listed here were: P62877, Q9UNM6, Q13619, A4D1P6, Q9UBV8.
Group Fold Difference (log2) | Protein ID# | Gene | Protein names and short descriptions of functions. |
-0.5919 | *P02743 | APCS | Serum amyloid P-component; An acute phase protein involved in inflammation via complement-dependent pathways. |
-0.441927 | P00738 | HP | Haptoglobin; Captures hemoglobin; acts as an antimicrobial and antioxidant and modulates the acute phase response. |
-0.397979 | *P00751 | CFB | Complement factor B; Part of the alternate pathway of the complement system. |
-0.362202 | *P00740 | F9 | Coagulation factor IX; Participates in the intrinsic pathway of blood coagulation. |
-0.327764 | P0C0L5 | C4B | Complement C4-B; Non-enzymatic component of C3 and C5 convertases; propagates the complement pathway. |
-0.303047 | P02763 | ORM1 | Alpha-1-acid glycoprotein 1; Transport protein; modulates immune activity during the acute-phase reaction. |
-0.300494 | *P01024 | C3 | Complement C3; Plays a central role in the activation of the complement system. |
-0.286746 | P05156 | CFI | Complement factor I; Cleaves alpha-chains of C4b, C3b with cofactors C4-binding protein and factor H. |
-0.286538 | *P68104 | EEF1A1 | Elongation factor 1-alpha 1; Role in the positive regulation of interferon gamma transcription in T-helper 1 cells. |
-0.282922 | *P19652 | ORM2 | Alpha-1-acid glycoprotein 2; Transport protein; modulates immune activity during the acute-phase reaction. |
-0.275192 | P08603 | CFH | Complement factor H; Functions as a cofactor in the inactivation of C3b. |
-0.27405 | P08311 | CTSG | Cathepsin G; Cleaves complement C3 and has antibacterial activity. |
-0.264261 | *P43652 | AFM | Afamin; Vitamin E binding protein. |
-0.258812 | P04004 | VTN | Vitronectin; Cell adhesion factor for inhibition of the terminal cytolytic complement pathway. |
-0.244408 | P02751 | FN1 | Fibronectin 1; Involved in cell adhesion, cell motility, opsonization, wound healing, and maintenance of cell shape. |
-0.229351 | P05546 | SERPIND1 | Heparin cofactor 2; Thrombin inhibitor activated by the glycosaminoglycans, heparin or dermatan sulfate. |
-0.202753 | P36955 | SERPINF1 | Pigment epithelium-derived factor; Neurotrophic protein; induces neuronal differentiation; inhibits angiogenesis. |
-0.132368 | P01031 | C5 | Complement C5; Activation initiates the spontaneous assembly of the membrane attack complex. |
0.105871 | *Q15435 | PPP1R7 | Protein phosphatase 1 regulatory subunit 7; Regulatory subunit of protein phosphatase 1, and is required for completion of the mitotic cycle. |
0.129966 | Q86X55 | CARM1 | Histone-arginine methyltransferase CARM1; Methylates guanidino nitrogens of arginyl residues and acts on histones and the regulation of gene expression. |
0.155758 | P02647 | APOA1 | Apolipoprotein A-I; The major protein component of high density lipoprotein (HDL) in plasma. Participates in the reverse transport of cholesterol from tissues to the liver. |
0.160256 | P06396 | GSN | Gelsolin; Actin-modulating protein that is involved in filament assembly; influences cell processes including cells of the immune system. |
0.166127 | P14550 | AKR1A1 | Alcohol dehydrogenase [NADP(+)]; Catalyzes the NADPH-dependent reduction of aldehydes to alcohols. |
0.175456 | P01008 | SERPINC1 | Antithrombin-III; Most important serine protease inhibitor in plasma that regulates the blood coagulation cascade. |
0.188729 | P08397 | HMBS | Porphobilinogen deaminase; Produces an enzyme involved in heme production. |
0.194536 | O94903 | PLPBP | Pyridoxal phosphate homeostasis protein; Intracellular regulation of pyridoxal 5'-phosphate (PLP), vitamin B6. |
0.202922 | *P02766 | TTR | Thyroid hormone-binding protein; Transports thyroxine from the bloodstream to the brain. |
0.217191 | *P37840 | SNCA | Alpha-synuclein; Involved in the regulation of dopamine release and transport. |
0.228831 | *P05089 | ARG1 | Arginase-1; Key element of the urea cycle converting L-arginine to urea and L-ornithine. |
0.244906 | *P01023 | A2M | Alpha-2-macroglobulin; Inhibits all four classes of proteinases by a unique 'trapping' mechanism. Inhibits inflammatory cytokines and inflammatory cascades. |
0.26744 | *P02768 | ALB | Serum albumin; Main function is the regulation of the colloidal osmotic pressure of blood; acts as a carrier protein for many molecules. |
0.267692 | O95445 | APOM | Apolipoprotein M; Involved with lipid transport. |
0.269348 | P02654 | APOC1 | Apolipoprotein C-I; Plays a central role in HDL and very low density lipoprotein (VLDL) metabolism. |
0.273906 | *P54252 | ATXN3 | Ataxin-3; Protein homeostasis maintenance, transcription, cytoskeleton regulation, myogenesis, and degradation. |
0.279698 | P05090 | APOD | Apolipoprotein D; Encodes a component of HDL; involved in lipoprotein metabolism and the transport and binding of bilin (e.g., bilirubin). |
0.301995 | *P35858 | IGFALS | Insulin-like growth factor-binding protein complex acid labile subunit; Binds insulin-like growth factors, increasing their half-life; influences protein complexes, cell adhesion. |
0.405038 | P01871 | IGHM | Immunoglobulin heavy constant mu; Constant region of immunoglobulin heavy chains. |
0.411452 | P04278 | SHBG | Sex hormone-binding globulin; Androgen transport protein; regulates clearance rate of steroid hormones. |
0.415369 | Q15848 | ADIPOQ | Adiponectin; Adipokine that controls fat metabolism and insulin sensitivity, and anti-inflammatory activities. |
The 39 proteins were mapped onto STRING v11.5 to build protein–protein interaction (PPI) networks (http://string-db.org/) (Fig. 1). Figure 1 was developed with Cytoscape (Institute for Systems Biology. Cytoscape. 2023. Available from: https://www.cytoscape.org). A reactome pathway enrichment analysis showed that the primary pathways affected by group status were innate immune responses including complement activation and neutrophil degranulation (all FDR < 0.001) (Fig. 1). The pathways most influenced by group status for the blood proteins that were higher in LIFE vs. CON were plasma lipoprotein assembly, remodeling, and clearance, and HDL remodeling (all FDR < 0.025) (Fig. 1).
The normalized relative intensity data for the 18 blood proteins that were lower and the 21 proteins that were higher in the LIFE vs. CON groups were averaged separately and then correlated with the variables in Tables 1 and 2. The variables with the strongest correlations were included in stepwise regression analyses. For the 18 blood proteins that were lower in LIFE vs. CON, the model that emerged from stepwise regression included FMI and CRP [F = 75.5(2,97), p < 0.001] (r2 = 0.609) (r2 = 0.587 for FMI alone). Figure 2 depicts the scatterplot for FMI and the average normalized relative intensity for the 18 proteins that were lower in LIFE vs. CON. For the 21 blood proteins that were higher in LIFE vs. CON, the best fit model included SAD, total white blood cell count (WBC), HDL-cholesterol, HOMA-IR, and the number of servings/day for fruits and vegetables [F = 56.4(5,97), p < 0.001] (r2 = 0.744) (r2 = 0.52 for SAD alone).
A single group (LIFE and CON) discriminator was optimized over all 103 samples using logistic Lasso regression and a leave-one-out (LOO) iteration of the same was performed to get an estimate of how well a proteome-based discriminator might perform on new samples. From the logistic categorization scores obtained from these models, receiver-operator-characteristic (ROC) curves were plotted to illustrate the discrimination capabilities of these models. Results for both the single (over-fit) model and the LOO models are shown in Fig. 3. The area under the curves (AUC) for the single model and LOO models were 0.99 and 0.88, respectively, with p-values of 2.6e-18 and 1.3e-11. The discriminator trained on all samples produced a multivariate model based on 20 proteins, and 15 of these were in common with the list of 39 proteins in Table 3 (noted with an “*”). The LOO iteration yielded LIFE vs. CON group memberships that were 82% correct with a Fisher Exact Test p-value of 6.5e-11.