The study protocol was approved by the Research and Ethics Committee of the Hospital General Yanga from Córdoba, having Institutional Review Board approval number HGC 0012/17-1. A signed informed consent was obtained from all study participants before commencing the study. The study followed the Good Clinical Practice Guidelines and was conducted in accordance with the ethical principles defined by the Declaration of Helsinki 1975, revised in 2013. An estimation of the required sample size was performed by applying the equation N = 2[(Z_α + Z_β)]² × SD²/d² for studies on the means of groups, where N is the number of participants, Zα = Z at 95%, Zβ = Z at 80%, SD = 0.2 (variance of carbonyl groups that has the reference value) [18], and d = 0.4 (minimum value expected in the change of carbonyl groups). According to this equation, we were required to include 10 participants per study group. All participants were students recruited from the Academic Division of Basic Sciences at the Universidad Juárez Autónoma de Tabasco, México.

Since abnormal glucose metabolism and dyslipidemia have been related to an increased risk of adverse health outcomes (cardiovascular disease, type 2 diabetes, stroke, and metabolic syndrome) in male Mexican adults [1920], this study focused on enrolling men. Considering the possibility of a high drop-out rate of about 50%, 15 men were included per study group. Young adults with and without class II obesity were included because data on the effects of cocoa in these populations are scarce.

This is a before-and-after study where 2 groups of 15 men each were included in the study: normal weight and obese. The participants were classified according to their body mass index (BMI) which was calculated by dividing the body weight in kilograms (kg) by the height in square meters (m²). In accordance with the World Health Organization definitions, participants with a BMI between 18.5–24.9 kg/m² were considered as normoweight (NW), while the obese participants considered class II obesity (CIIO) had a BMI between 35–39.9 kg/m² [21]. Anthropometric measurements (weight, age, height, BMI, and blood pressure) were taken in each group. The inclusion criteria were voluntary participation, age between 19 and 25 yrs, self-reported good general health status, and BMI between 18 and 25 kg/m² or 35 and 40 kg/m². Exclusion criteria included not signing the informed consent, smokers, subjects receiving treatment with medications, consumption of antioxidants, alcohol, food supplements and/or vitamins during the last 6 months, and presence of an active infection.

Based on their BMI, participants were assigned to the 2 groups. The clinical history, anthropometric measurements, and the blood pressure of each participant were obtained. At baseline (Day 0), the participants attended the clinic following an overnight fast (×12 h) for the collection of blood samples to determine the biochemical parameters, establish the homeostasis model assessment of insulin resistance (HOMA-IR), redox status (biomarkers of oxidative damage and antioxidant defenses), and structural changes of serum insulin.

After 24 h, each participant commenced the daily consumption of high-flavanol cocoa (0.375 g/kg body mass in 300 mL water) for 7 days; the NW and CIIO groups received an average of 25 g (1,050 mg of flavonoids) and 39 g (1,638 mg of flavonoids) per day, respectively. These intervention doses were approximately 2 times higher than doses used for observing maximal cardiovascular effects (800 mg of flavonoids) [22] and a moderate benefit on performance in healthy individuals (688 mg/day) [23].

The cocoa powder used in this study was Hershey’s (100% cocoa) which has higher total polyphenols and flavanols. Furthermore, it also comprises the following ingredients: corn syrup solids, sunflower oil, and ≤ 2% food additives (salt, carrageenan, vanillin, aspartame, sodium caseinate, dipotassium phosphate, monoglycerides, acesulfame potassium, diglycerides, soy lecithin, and silicon dioxide) [24]. The nutritional content per 25 g is presented in Table 1. Participants were instructed to consume the appropriate amount of cocoa, not to change their dietary habits during the study, and avoid caffeine, tea, alcohol, red wine, dark chocolate, or dietary supplements of these foods throughout the intervention. On day 8, cocoa was not consumed, and all assessments performed at baseline were repeated.

A total of 10 mL blood was drawn from the antecubital vein of the arm using heparin as an anticoagulant; of this, 4 mL was separated and used to obtain plasma. The plasma was used for analyzing the biochemical parameters (2 mL) and to determine the biomarkers of oxidative stress damage, antioxidant defense, and total antioxidant capacity (TAC) (0.5 mL).

The remaining 6 mL blood was used in the assays to determine the structural changes of recombinant human insulin (see below). The same amount of blood was obtained on day 8, and all assessments performed at baseline were repeated.

All determinations were measured by colorimetric and kinetic kits using the MINDRAY BS-120 chemistry analyzer. Blood glucose was determined using the GOD-PAP method (Wiener lab/Glicemia enzimática AA/Rosario, Argentina), and insulin levels were determined by an ELISA method (Insulina 6001011; Mex-Lab, Mexico City, Mexico).

HOMA was calculated as follows: [fasting glucose (mg/dL)/fasting insulin (lU/mL)/22.5] [25].

The lipid profile was evaluated by measuring the total cholesterol (Wiener lab/Colestat enzimático AA/Rosario, Argentina) and triglycerides (Wiener lab/TG Color GPO-PAP AA/Rosario, Argentina) using the CHOD-POD and GPO-POD methods, respectively.

High-density lipoprotein (c-HDL) (HDL Colesterol monofase AA; Wiener Lab, Rosario, Argentina) and low-density lipoprotein cholesterol (c-LDL) (LDL Colesterol monofase AA; Wiener Lab) were measured by enzymatic clearance.

C-reactive protein was evaluated by immunoturbidimetric tests (Turbitest AA; Wiener Lab).

The separated plasma was used to determine the biomarkers of oxidative damage. Lipid damage was assessed by the quantification of lipoperoxidation products malondialdehyde (MDA) and 4-hydroxynonenal (4-HNE) [26].

MDA was determined using 200 µL of plasma diluted in H₂O (1:1) mixed with 650 µL of a solution of 1-methyl-2-phenylindole in a mixture of acetonitrile/methanol (3:1). The reaction was then started by adding 150 µL of 37% hydrochloric acid. Absorbance was measured at 586 nm after incubating the reaction mixture at 45°C for 40 min. A sample blank was included, in which the plasma was replaced by water.

4-HNE was determined following the same procedure of MDA, replacing hydrochloric acid by methanesulfonic acid containing 34 µM Fe⁺³ and incubated at 45°C for 30 min. The final absorbance at 586 nm was a linear function of the concentrations of both MDA and 4-HNE. Therefore, the 4-HNE concentration was assessed after subtracting the MDA concentration, and the results were expressed as micromolar concentration.

Quantification of the carbonyl groups was established as oxidative damage to proteins. Free carbonyl groups were determined in 50 µL plasma mixed with 500 µL 2,4-dinitrophenylhydrazine (dinitrophenylhydrazine [DNPH] 10 mM in 2.5 M HCl). All samples had their corresponding blank (50 µL plasma in 500 µL 2.5 M HCl). Samples and blanks were incubated for 1 h at room temperature, protected from light and with agitation every 15 min. Subsequently, 1 mL of 5% trichloroacetic acid (TCA) was added to the reaction mix and the solution was centrifuged at 3000 rpm for 10 min. After discarding the supernatant, the precipitate was dissolved in 1 mL 2.5% TCA, centrifuged at 3,000 rpm for 10 min, and the supernatant was discarded. The precipitate was dissolved in 2 mL ethanol/ethyl acetate (1:1, v/v) and centrifuged again (3,000 rpm for 10 min). The supernatant was eliminated and the precipitate was dried for 5 min, dissolved in 1 mL of 6 M guanidine containing 20 mM potassium phosphate at pH 2.3, and incubated for 10 min at 37°C. Absorbance was determined at 370 nm the DNPH formation was determined. The molar extinction coefficient for DNPH (E = 22,000/M−1/cm−1) was used to calculate the concentration of carbonyls [27]. Results are expressed in nmol of osazones/mg of protein. Total protein was measured as a reference parameter according to Lowry’s method [28].

Thiol concentration and TAC are considered as a reflection of the antioxidant defense capability. Total thiol concentration was determined in 100 µL of plasma using Ellman’s reagent. The concentration of the thiol groups was measured with a photometric assay at 412 nm, which was calculated from a standard curve constructed using cysteine [29].

TAC in the plasma was assessed by a colorimetric assay kit (Cayman Chemical Company, Ann Arbor, MI, USA) according to the manufacturer’s instructions.

Structural changes of insulin were determined using recombinant human insulin (IRH; Lily Laboratories, Mexico City, Mexico) and exposing it to the blood obtained from all study participants, according to a method previously described [530]. Briefly, 30 IU (1.5 mg) of native insulin was placed on a 5 cm portion of membrane dialysis tubing with a cut-off of 3,500 Da (Spectrum Laboratories, Inc., Rancho Dominguez, CA. USA). The dialysis tubes were incubated at 37ºC in 6 mL blood (obtained from NW or COII subjects) for 3 h. After incubation, the bags containing the insulin were washed with distilled water. Five microliters (90 µg) of insulin were used to determine the effects of blood on the insulin structure using nondenaturing polyacrylamide gels. Coomassie blue staining of the polyacrylamide gels was performed to assess any changes in the electrophoretic pattern.

Data are expressed as the mean ± SE. Student’s t-test was employed to determine intergroup differences. The paired t-test was used to determine intragroup differences before and after cocoa consumption. P-values ≤ 0.05 are considered to be statistically significant. All tests were performed using the Prism 5 software (GraphPad, San Diego, CA, USA).

Two groups of participants (NW and CIIO) were included in the study. Each group included 15 participants. Data in Table 2 shows that the anthropometric features of both groups are similar. All participants were homogeneous in age and height. As expected, the obese group had higher weights and the calculation of BMI classified them as CIIO.

Since several studies indicate a relationship between obesity, inflammation, and oxidative damage, we decided to analyze the redox state of the participants. Oxidative damage was established by quantification of the respective biomarkers. A major degree of lipid damage was observed in the CIIO group as determined by the high MDA values, compared to the NW group (Table 2). No statistically significant differences were observed in the levels of 4-HNE or protein damage biomarker (carbonyl groups). Considering the antioxidant status, the CIIO group had a lower TAC in comparison with the NW group, but no differences were observed in the concentration of thiol groups (Table 2).

At baseline, the clinical biochemical parameters of both groups are presented in Table 3. Interestingly, the value of LDL-cholesterol (LDL-C) was higher than the accepted normal values in both groups. In addition, the CIIO group had higher levels of insulin and HOMA-IR as compared with the NW group, indicating the presence of IR (Table 3, baseline values before intervention).

Analysis of the inflammation markers revealed higher levels in the CIIO group as compared to the NW group (Table 3). These data suggest the presence of an inflammatory state in the CIIO group.

The establishment of IR in the CIIO group and the previous evidence that indicates the capacity of IR blood to modify insulin suggests that blood from these participants have the capacity to modify the hormone insulin. This was determined by analyzing the changes generated in the electrophoretic pattern of the hormone after its incubation with the blood obtained from study participants. After incubation, the hormone was loaded onto 15% nondenaturing polyacrylamide gels and separated by electrophoresis.

Coomassie blue staining of the polyacrylamide gel was performed to assess the electrophoretic pattern of the hormone. Fig. 1A shows the electrophoretic mobility of the hormone at the baseline stage. In all samples, a band of approximately 6 kDa (molecular mass of the insulin) was observed. Importantly, a band of 25 kDa was observed in all the insulin incubated in blood from CIIO participants (Fig. 1A). These data demonstrate the generation of structural changes in the insulin that modifies its electrophoretic mobility.

To investigate the effect of cocoa consumption on biochemical, inflammation, and oxidative damage biomarkers, NW and CIIO groups consumed an average of 25 and 39 g per day of cocoa for 7 days, respectively. The effects of this intervention are presented in Table 3. Interestingly, amelioration of the LDL-C level was observed in both groups. Furthermore, the improved insulin levels in the CIIO group generated a decrease in the HOMA and in IR (Table 3, values after intervention). Conversely, consumption of cocoa had no effect on CRP and ESR. However, a notable decrease was observed in the arginase activity of the CIIO group; this enzyme plays a critical role in the inflammatory process associated with obesity (Table 3, value after intervention).

We also investigated the effects of the intervention on biomarkers of oxidative damage. As shown in Table 4, consumption of cocoa resulted in reduced oxidative damage biomarkers (MDA and carbonyl groups) and generated an increase of 4-HNE in both groups. These changes were accompanied by an increase in the TAC and decrease in thiol groups (Table 4).

It is important to note that the consumption of cocoa also had a positive impact on the molecular damage observed in insulin. Before cocoa consumption, the incubation of insulin with CIIO blood generated a band of 25 kDa in the electrophoretic pattern, which was considered a modification of the hormone (Fig. 1A). Interestingly, this band disappeared in the electrophoretic pattern generated after cocoa consumption. Fig. 1B is a representative polyacrylamide gel showing the electrophoretic pattern obtained before and after cocoa consumption: a band of approximately 6 kDa was present in all samples (insulin); however, the modified insulin band (25 kDa, Fig. 1B, lane 2 and 4) disappeared after cocoa consumption (lane 3 and 5). These data demonstrate the protective effect generated by the consumption of this functional food for insulin.