Measurement of sweat lactate levels in exercise and non-exercise activities using capillary electrophoresis system with contactless conductivity detection and cyclodextrin-modified buffer

2.1 Chemicals

Two cyclodextrin additives (Fine chemical grade) were obtained from Cyclolab (Budapest, Hungary), viz.: carboxymethyl-β-cyclodextrin sodium salt and heptakis(2,3,6-tri-O-benzoyl)-β-cyclodextrin. Sodium chloride, potassium nitrate, sodium sulfate, potassium oxalate, maleic acid, acetic acid, lactic acid, and potassium dihydrogen phosphate were of analytical grade (AR) and were purchased from Sigma-Aldrich (St. Louis, MO, USA). 2-(N-morpholino)ethanesulfonic acid (MES, 99% assay) was obtained from Acros Organics (USA). 3-(N-morpholino)propanesulfonic acid (MOPS, 98.5% assay, ultrapure grade), and 4-(2-hydroxyethyl)-1-piperazineethane sulfonic acid (HEPES, 99% assay) were acquired from Affymetrix Inc. (Ohio, USA). L-Histidine (L-His, 99.5% assay, BioUltra grade), cetyltrimethylammonium bromide (CTAB, biology grade), and boric acid (analytical grade) were supplied by Merck (Darmstadt, Germany). Orthophosphoric acid (85% assay, UNIVAR® Analytical Reagents) was obtained from Ajax Finechem Pty Ltd. (Australia).

2.2 Preparation of standard and buffer solutions

Standard working solutions of chloride, nitrate, sulfate, oxalate, maleate, acetate, lactate, and phosphate were prepared in ultrapure water by diluting individual 100.0 mM aqueous standard solutions to concentrations of 0.1, 0.3, 0.5, 1.0, 3.0, and 5.0 mM. The MES/L-His buffer was freshly prepared by diluting stock solutions of MES and L-His (150 mM each) to a concentration of 40.0 mM MES and 40.0 mM L-His (pH 6.0) with 0.05 mM CTAB. CTAB was added to reverse the EOF and allow for the migration reversal of anions. This buffer exhibits low conductivity and is commonly employed for CE-C⁴D works (Kubáň and Hauser 2004). The running buffer also contained two different CD additives, TRIME-β-CD and CM-β-CD, at concentrations of 0.05, 0.1, and 0.5 mM. Before use, the buffer solution was degassed by sonication.

2.3 Collection and preparation of sweat samples

The collection of sweat samples from volunteers was conducted with the approval of the Mahidol University Central Institutional Review Board (MU-CIRB 2020/116.1805). Six healthy volunteers, comprising 3 females and 3 males, were recruited for the study and provided written informed consent. Detailed information about the participants, including their gender, age, height, and weight, can be found in Table S1 of Supplementary Material A.

Two activities were performed to collect the sweat samples: a non-exercise activity after a 30-minute sauna session at 60 °C and 70% humidity, and an exercise activity after a 30-minute treadmill run covering a distance of 3.0 km. Before each activity, the volunteers were instructed to clean their skin using soap. Approximately 0.3–0.4 mL aliquots of sweat were directly collected from the face area and transferred into 0.5 mL microcentrifuge tubes. The collected sweat samples were then stored at a temperature of 4 °C until further analysis. Each sweat sample was prepared by diluting it with ultrapure water and adding an internal standard (1.0 mM maleate). For the non-exercise activity samples, a 25-fold dilution was performed, while a 50-fold dilution was applied to the exercise activity samples.

2.4 Capillary electrophoresis with Capacitively Coupled contactless conductivity detector (C⁴D)

The CE system was assembled in-house and connected to a contactless conductivity detector (C⁴D, eDAQ ET 120, Denistone East, NSW, Australia). The CE unit was placed in a Plexiglas box and connected to a micro switch that controlled the ON/OFF function of the high-voltage power supply unit (Spellman CZE1000R, Hauppauge, USA). The signal from the detector, measured in millivolts, was recorded using an E-corder (ED410, Denistone East, NSW, Australia) and analyzed using the eDAQ Chart software (eDAQ, Denistone East, NSW, Australia). A fused-silica capillary (360.0 µm o.d. and 50.0 µm i.d.) was provided by Polymicro Technologies (Phoenix, AZ, USA). A total length of 30.0 cm with an effective length of 18.5 cm was used for the capillary. Before analysis, the capillary was conditioned by sequentially flushing it with 0.1 M NaOH, ultrapure water, and the MES/L-His buffer for 5 min using a manual syringe pump. The hydrodynamic injection was performed by raising the inlet end of the capillary 7.5 cm higher than the outlet end for 3.0 s. A potential of −10.0 kV (a field strength of –333 V cm⁻¹) was applied for the analysis.

3.1 Application of cyclodextrin-modified buffers

Our study investigates the potential of TRIME-β-CD and CM-β-CD as new additives for anion separation (see Table 2). These CDs were chosen based on their commercial availability. While they have been extensively studied for enantiomeric separation, their application in anion separation remains unexplored. TRIME-β-CD is a non-ionic CD that has not been employed previously as a buffer modifier for anion separation. In contrast, CM-β-CD is an anionic CD type that has found application in various contexts involving anionic chiral compounds (Chankvetadze et al., 1995, Organero et al., 2007). Using the latter CD for anion binding might initially seem counterintuitive because of the intrinsic repulsion between similar charges. However, we hypothesized that using the CM-β-CD modifier buffer for anion separation is based on the degree of inclusion complexation strength. The anions with variation of distinct properties such as charge and size, exhibit diverse binding affinities. Despite the electrostatic repulsion, these anions can access the hydrophobic cavity of the CD, forming a host–guest complex. The resulting strength and stability of these complexes influence the anion migration.

The eight anions (chloride, nitrate, sulfate, oxalate, maleate, acetate, lactate, and phosphate anions) are investigated for their migration affecting binding behavior with these two CDs. Table 3 presents their structures. As listed, TRIME-β-CD (carboxymethylated) and CM-β-CD (trimethylated) have distinct electrostatic properties due to their substituents. Fig. 1 shows the electropherograms of the separation of eight anions using CM-β-CD and TRIME-β-CD additives, see Fig. 1(A) and (B), respectively. The migration patterns of the anions exhibited varying separation characteristics. Slight changes in anion migration were observed at all concentrations of TRIME-β-CD (0.05, 0.1, 0.5 mM) as shown in Fig. 1B(b)–(d). However, regarding CM-β-CD, smaller changes in the migration times were observed for peaks 1–4 (chloride, nitrate, sulfate, and oxalate), as shown in Fig. 1A(b)–(d). On the other hand, significantly longer anion migration times were observed for peaks 5–8 (maleate, acetate, lactate, and phosphate) for CM-β-CD addition, Fig. 1A(b)–(d). These findings show that CM-β-CD has a pronounced effect on the migration of these latter anions. Broadening of peaks 7–8 (lactate and phosphate) was obtained when the amounts were more than 0.05 mM. The binding affinity of TRIME-β-CD with all the anions was lower compared to that of CM-β-CD. This translated into slight changes in the migration times of all anions when TRIME-β-CD was employed. To provide a clearer observation of these changes see Fig. 1C and 1D. The figures offer a clearer view of the changes in anion migration times in response to the concentrations of the CD.

Table 3 Name, abbreviation, and chemical structure of cyclodextrin additives.

	CM-β-CD	TRIME-β-CD
Name (Abbreviation)	Carboxymethyl-β-cyclodextrin sodium salt, (CM-β-CD)	Heptakis(2,3,6-tri-O-methyl)-β-cyclodextrin, (TRIME-β-CD)
Structure

Abbreviation: CD; cyclodextrin, DS; the average number of substituents per one CD molecule.

Close

Fig. 1

Electropherograms of eight anions by capillary electrophoresis with C⁴D using buffer modified with CD additive: (A) CM-β-CD and (B) TRIME-β-CD at (a) 0 mM, (b) 0.05 mM, (c) 0.1 mM and (d) 0.5 mM. The migration times of the eight anions (chloride, nitrate, sulfate, oxalate, maleate, acetate, lactate, and phosphate) are plotted in response to the concentration of CD modification: (C) CM-β-CD and (D) TRIME-β-CD. Note: Lines connecting the points in the graph are included for visual clarity and do not represent any specific trends. The running buffer is 40.0 mM MES/L-His buffer (pH 6.0) containing 0.05 mM CTAB. Peak identification: 1. Chloride (0.3 mM), 2. Nitrate (0.3 mM), 3. Sulfate (0.3 mM), 4. Oxalate (0.4 mM), 5. Maleate (1.0 mM, IS), 6. Acetate (0.4 mM), 7. Lactate (1.0 mM), 8. Phosphate (1.0 mM), *unidentified peaks. (See Section 2.4. for CE conditions).

Export to PPT

Based on the results, a concentration of 0.1 mM of TRIME-β-CD was chosen for modifying the buffers for anion separation. This modifier composition offers improved performance, as evidenced by the high separation efficiency (N/m) ranging from 2.5 to 5.8 × 10⁴ m⁻¹ and resolution (R_s) ranging from 0.88 to 4.69, achieved for the eight analyzed anions. The inclusion of TRIME-β-CD resulted in narrow peaks with high resolution. It is essential to achieve effective separation of lactate and phosphate peaks, whichare anions contained in sweat (phosphate, chloride, and sulfate) (Maughan 1991). Hence selectivity in the separation of these anions is critical.

3.2 Measurements of lactate in human sweat

The lactate measurement method was validated following the US-FDA guidelines for the validation of bioanalytical methods (US-FDA Guidelines, 2018). The CE analysis with conductivity detection employed a running buffer consisting of 40.0 mM MES/L-His buffer (pH 6.0), which contained 0.05 mM CTAB and 0.1 mM TRIME-β-CD. The evaluated parameters included linearity, intra and inter-day precision, limit of detection (LOD), specificity, accuracy, and repeatability with the latter based on relative migration time (RMT), as detailed in subsequent sections.

3.2.2

3.2.2 Measurement of lactate contents in sweat: Exercise activity (after treadmill running) and non-exercise activity (after sauna)

Sweat samples were collected from six volunteers who participated in two activities: a non-exercise activity and an exercise activity. The electropherograms of the sweat samples from the two activities are shown in Fig. 2(b) and 2(c), respectively, and were compared to the standard electropherograms of lactate, chloride, and maleate (IS) peaks (Fig. 2(a)). The lactate, chloride, and IS peaks were well separated in approximately 2.5 min. Spiking with a standard solution of lactate (1 mM) confirmed the lactate peak from the increase in peak height. The chloride peak in the sweat samples was also identified and analyzed using maleate as the IS. The mean chloride content of the samples was found to be 65 mM (n = 3). Moreover, peak specificity was established for lactate and phosphate to confirm that the identified lactate peak was not overlapped by any interfering substances.

Close

Fig. 2

Electropherograms of (a) a standard mixture of eight anions; (b) diluted sweat (25-fold) from a non-exercise activity (after sauna); (c) diluted sweat (50-fold) from an exercise activity (after treadmill running). The samples were spiked with standard lactate (0.5 mM) and maleate (1.0 mM, IS). Peak identifications: 1. Chloride (0.3 mM), 2. Nitrate (0.3 mM), 3. Sulfate (0.3 mM), 4. Oxalate (0.4 mM), 5. Maleate (1.0 mM, IS), 6. Acetate (0.4 mM), 7. Lactate (1.0 mM), 8. Phosphate (1.0 mM). The running buffer condition is 40.0 mM MES/L-His buffer (pH 6.0) containing 0.05 mM CTAB and 0.1 mM TRIME-β-CD. (See Section 2.4. for CE conditions).

Export to PPT

Lactate measurement and recovery study: Table 5 lists the measured lactate concentrations from the sweat samples collected from the non-exercise and exercise activities. The lactate values in the sweat samples collected from the non-exercise activity ranged from 15 to 36 mM (4% RSD), while those collected from the exercise activity ranged from 26 to 136 mM (4% RSD). Calculation of percent recoveries of spiked samples was carried out to determine the accuracy of the method. The percent recovery of the spiked diluted sample was calculated using the formula:% recovery = [[S₁ - S₂] / S₀] × 100, where S₀ is the peak area ratio of the spiked standard lactate solution, S₁ is the peak area ratio of the spiked diluted sample, and S₂ is the peak area ratio of the original diluted sample. Table 5 lists the percent lactate recoveries for all samples, which ranged from 80 ± 1 – 100 ± 2% (<1% RSD) for the non-exercise activity (n = 6) and from 100 ± 1 – 105 ± 1% (<2% RSD) for the exercise activity (n = 6). The mean heart rates were 80 ± 0 bmp (0% RSD, n = 6) and 137 ± 9 bmp (6% RSD, n = 6) for the non-exercise and exercise activities, respectively (refer to Fig. 3A, right y-axis). The bar plot in Fig. 3A overlays the lactate levels (left y-axis) obtained from the same volunteer for the two types of sweat formation activities. Table S1 in Supplementary Material A presents the heart rates of all volunteers measured during each activity. According to the bar graph, the lactate levels observed in all samples collected during the exercise activity were higher, ranging from 26 to 136 mM. The mean lactate value was 94 ± 48 mM (51% RSD, n = 6). In contrast, during the non-exercise activity, the lactate levels ranged from 15 to 36 mM, with a mean value of 24 ± 7 mM (30% RSD, n = 6) for the same group of volunteers.

Table 5 Measured lactate concentrations in sweat samples collected from volunteers who participated in two activities: a non-exercise (after sauna) and an exercise (after treadmill running), relative migration time (RMT), recoveries of spiked diluted samples, and lactate content in samples.

Sample	Lactate concentration (mM)			RMT ± SD	Percent recovery ± SD	Lactate in sample ± SD, n = 3 (mM)
	Diluted sample ± SD, n = 3	Spiked concentration	Spiked diluted sample ± SD, n = 3	(% RSD), n = 3
Non-exercise (after sauna)
V1	1.43 ± 0.07	1	2.40 ± 0.06	1.25 ± 0.02 (1.40%)	96 ± 2	36 ± 2
V2	0.88 ± 0.09	1	1.88 ± 0.09	1.24 ± 0.02 (1.20%)	100 ± 1	24 ± 2
V3	0.88 ± 0.07	1	1.87 ± 0.08	1.20 ± 0.02 (1.40%)	100 ± 2	23 ± 1
V4	0.92 ± 0.06	1	1.87 ± 0.05	1.20 ± 0.01 (0.95 %)	95 ± 2	23 ± 2
V5	0.80 ± 0.02	1	1.78 ± 0.03	1.25 ± 0.01 (0.96%)	98 ± 1	20 ± 1
V6	0.60 ± 0.02	1	1.44 ± 0.02	1.27 ± 0.02 (1.70%)	80 ± 1	15 ± 1
Exercise (after treadmill running)
V1	2.20 ± 0.02	0.5	2.70 ± 0.01	1.28 ± 0.02 (1.30%)	105 ± 1	110 ± 1
V2	2.70 ± 0.02	0.5	3.23 ± 0.02	1.24 ± 0.01 (1.00%)	100 ± 2	136 ± 1
V3	2.24 ± 0.03	0.5	2.75 ± 0.04	1.36 ± 0.01 (0.50%)	100 ± 2	112 ± 2
V4	2.70 ± 0.01	1	3.70 ± 0.01	1.32 ± 0.01 (1.05%)	100 ± 1	135 ± 1
V5	0.84 ± 0.01	1	1.88 ± 0.01	1.30 ± 0.00 (0.15%)	104 ± 1	42 ± 1
V6	1.05 ± 0.03	1	2.10 ± 0.02	1.32 ± 0.00 (0.30%)	104 ± 2	26 ± 1

Note: Dilution factors for all samples were 25-fold and 50-fold for non-exercise and exercise activities, respectively.

The mean migration time values were between 134.28 ± 1.27 (0.95% RSD) and 141.50 ± 0.43 (0.30% RSD) for IS (maleate peaks), and between 169.00 ± 1.90 (1.12% RSD) and 185.26 ± 1.18 (0.64% RSD) for lactate peaks across all tested sweat samples. Each sample was injected in triplicate.

Close

Fig. 3

(A) Bar graphs of the sweat lactate concentrations from V1 to V6 (volunteers who participated in two activities: a non-exercise (after sauna) and an exercise (after treadmill running)) and scatter plots of heart rates from both activities. Dash lines are the mean values of the heart rate for both activities, with values of mean ± SD (n = 6). (B) Scatter plot showing the measured lactate concentrations of samples V1 – V6 at 1, 2, 3, 4, and 24 weeks of storage at 4 °C. The samples were collected from volunteers who participated in the non-exercise activity. Dash lines are the mean ± SD (n = 5) with% RSD.

Export to PPT

Relative migration time: The RMT (n = 3 injections) of the lactate and IS peaks for all injected diluted sweat samples are listed in Table 5. The mean of RMT of the 12 samples was 1.27 ± 0.05 (4% RSD, n = 12), demonstrating the good repeatability of the CE system. The result could also be due to the conditioning of the capillary surface, particularly the rate of CTAB equilibration with the silica capillary surface which might be slow.

In summary, validation parameters included lactate linearity (0.01–5.0 mM, r² = 0.9999), intra- and inter-day precision (< 4% RSD for RMT and < 7% RSD for peak area ratio), limit of detection (LOD) of 0.042 mM, percent recoveries (80–105%), and system repeatability (evaluated based on RMT < 4% RSD. The values of these parameters not only demonstrate the method's accuracy and reliability but also emphasize its practical suitability for quantifying lactate in human sweat with peak specificity.

3.3 Comparison of lactate measurement methods for two activities (exercise and non-exercise) with a proposed decision threshold and potential applications

Table 1 provides an overview of previous methods utilized for the analysis of sweat lactate from both non-exercise and exercise activities. These methods employ various approaches and are characterized by the detection and levels of lactate measured, the number of analyzed samples, and the number of volunteers involved. The calibration range and limit of detection (LOD) for each method are also presented. Among these methods, Gutmann and Wahlefeld's method focuses on measuring LDH activity (Fellmann et al., 1983, Alvear-Ordenes et al., 2005), while the remaining methods primarily measure LOX activity (Mitsubayashi et al., 1994), see Table 1. Our method involves the utilization of capillary electrophoresis with buffer-modified TRIME-β‐CD for precise lactate measurement. Table 1 includes the calibration range and limit of detection (LOD) for our method, as well as for other methods, offering comprehensive information regarding their analytical performance. In addition to the comparison of methods, we evaluated the ratio of lactate levels measured between exercise and non-exercise conditions. The ratio values obtained ranged from approximately 1–5.2. Our developed method demonstrated a significant increase in mean lactate levels measured during exercise for the same group of volunteers, the mean lactate values were 94 ± 48 mM during exercise and 24 ± 7 mM during non-exercise activities, resulting in a four-fold increase. The results of our developed method demonstrate a significantly improved range in the measurement of sweat lactate levels. This method offers a broader linearity range (i.e., 0.1–5.0 mM) when compared to previous studies using CE-C⁴D (Law et al., 2005; Pormsila et al., 2011). This enables differentiation between the two types of activities, i.e., exercise and non-exercise, compared to previous methods (see Table 1 for details). These findings highlight the significant advancements and capabilities of various analytical methods in the analysis of sweat lactate. Nevertheless, the measurements of sweat lactate levels can be employed to indicate the intensity of exercise activity. This study suggests a decision level of 40.0 mM for sweat lactate, determined based on a cut-off level of 36.0 mM. This decision level is established by adding a guard band of + 1.645 × SD to the cut-off value, where SD is the standard deviation of the data at the cut-off value (WADA, 2019). Lactate levels in both sweat and blood are commonly associated with exercise intensity (McCaughan et al., 2000). With increased exercise intensity, lactate production rises, accumulating in the blood and subsequently becoming detectable in sweat. However, blood sampling is invasive with infection risks and requires a more complex, time-consuming lactate analysis method due to its matrix content. In contrast, a sweat sample offers straightforward markers of exercise physiology and performance evaluation. Our suggested decision threshold for sweat lactate is meaningful in evaluating exercise intensity levels. Further research involving a more diverse population and larger sample sizes as well as a broader range of exercise parameters, such as variations in exercise duration and levels of intensity would further strengthen the validity of this research.