Protocol and application of basal erythrocyte transketolase activity to improve assessment of thiamine status

Thiamine (vitamin B1) is an essential micronutrient required as a cofactor in many metabolic processes. Clinical symptoms of thiamine deficiency are poorly defined, hence biomarkers of thiamine status are important. The erythrocyte transketolase activity coefficient (ETKac) is a sensitive measure of thiamine status, but its interpretation may be confounded where the availability of the transketolase enzyme is limited. Basal ETK activity per gram of hemoglobin provides a complementary biomarker of thiamine status; however, its measurement and calculation are poorly described. Here, we describe in detail the assessment of basal ETK activity, including the calculation of path length in microplates and the molar absorption coefficient of NADH specific to the assay, and the measurement of hemoglobin in sample hemolysates. To illustrate the application of the methods, we present ETKac and basal ETK activity from women in The Gambia and UK. In conclusion, we present a clear protocol for the measurement of basal ETK activity that will permit the harmonization of methods to improve replication between laboratories.


INTRODUCTION
Thiamine is an essential micronutrient required as a cofactor for many metabolic enzymes. Clinical symptoms of thiamine deficiency affect the cardiovascular, muscular, and nervous systems. However, symptoms are nonspecific. 1 Blood biomarkers of thiamine status, therefore, provide a complementary approach to diagnose acute deficiency.
Furthermore, there is some evidence to suggest that chronic subclinical deficiency may have long-term consequences for cognition and gross motor skills, 2-4 underlining the need for a reliable and wellcharacterized biomarker that can provide biochemical evidence of thiamine insufficiency. 2 Assessment of the ETKac requires simultaneous or consecutive measurement of basal and stimulated transketolase activity, the lat- and more than 1.25 indicate a moderate or high risk of deficiency, respectively. 2,5,9 While ETKac is a useful biomarker of thiamine status and is able to discriminate between groups, there is concern that some clinical conditions, drugs, or long-term thiamine deficiency may reduce the levels of the apoenzyme (the inactive transketolase). 5,[10][11][12][13][14] In a scenario of low apoenzyme, less thiamine is required to stimulate the transketolase, and a ratio in the normal range may be obtained, regardless of the availability of thiamine. Therefore, the measurement of basal ETK activity provides an additional and complementary biomarker of thiamine status to aid the interpretation of other thiamine status indicators, providing context for the interpretation of ETKac. 5,15 Despite its potential utility, the derivation of absolute activity has not received detailed descriptions in the literature. The basal and stimulated activities measured in the calculation of the ETKac are used exclusively for the calculation and are not sufficient, alone, to derive basal ETK activity. Additional calculations are required to determine the basal activity in units of enzyme activity as well as to express the absolute activity per unit weight of hemoglobin (Hb).
Published reports of basal ETK tend to lack sufficient methodological details to allow replication [16][17][18][19] or use obsolete technology 20,21 rather than current 96-well microplate methods. Where methods are better described, assumptions used in the calculations are commonly not shown and descriptions assume a degree of understanding that may limit their reproducibility. 20,22 More recent reports of basal ETK activity typically lack detail on how the basal activity was measured and/or calculated. 16,18,23 This inhibits further advances in the field by impeding replication and initiatives to harmonize methods between laboratories.
We recently published a modern and detailed protocol describing the steps involved to measure ETKac. 24 Here, we expand that work and describe the equipment, analytical methods, and calculation required to determine basal ETK activity. The method described herein can be replicated with the aim to improve harmonization among laboratories. In addition, we provide example datasets from the UK and The Gambia encompassing a wide range of thiamine status.

Summary of ETK activity measurement
ETK activity is a function of the conversion of NADH to NAD+ in the reaction whereby dihydroxyacetone phosphate is converted to glycerol-3-phosphate (G-3-P) as part of the final steps in a set of ETKdependent reactions. In practice, the rate of oxidation of NADH is measured by monitoring the change in absorbance (ΔA) at 340 nm. 24 In the assay, the hemolysate is added to assay reagents without (basal) and with additional ThDP (stimulated). The activity coefficient is calculated as the ratio of stimulated to basal ETK activity. The experimental details for the measurement of basal ETK activity in the context of the activity coefficient were described previously. 24 In order to generate meaningful values for basal ETK activity, it is necessary to express activity in absolute units of enzyme activity relative to the Hb concentration of the diluted hemolysate used in the assay. Therefore, the following parameters are required: • Basal ETK activity

Equipment and reagents
Reagents were Sigma brand and purchased from Merck Life Science UK Ltd (Dorset, UK); details are provided in Table S1. Full details of the materials and methods for the measurement of ETKac were published previously. 24 Hemolysate Hb was quantified using Drabkin's reagent (Sigma-Aldrich, product code: D5941; Merck Life Science UK Limited).
Control materials for Hb were purchased from Bio-Rad (Watford, UK).

Sample type and preparation
Venous blood was collected into EDTA blood tubes using standard techniques. After centrifugation, plasma and buffy coat were removed and samples were frozen at −70 • C to lyse the cells, and for storage and transport. After thawing, deionized water (1:2 washed erythrocytes: deionized water) was added to ensure cell lysis. 24 Lithium heparin as an anticoagulant may also be used. 24 To ensure adequate volume for sampling handling, the recommended volume of washed erythrocytes for the analysis of ETKac and Hb is between 200 and 500 μl, however, the absolute minimum volume required is 90 μl.

Calculation of basal activity
Basal ETK activity per gram of Hb is expressed as μmol G-3-P produced per gram Hb per minute or U/gHb) and was calculated based on a modified version of the equation by Bayoumi and Rosalki 21 (Equation 1). Note that this equation includes values calculated in our laboratory that are further explained below: • ΔA is the change in absorbance per minute measured at 340 nm (this is the raw basal activity measured in the ETKac assay) • 245 is the total volume (μl) of assay reagents and hemolysate per well of the 96-well microplate • 30 is the volume (μl) of hemolysate within the assay volume (a division of total volume by assay volume allows this equation to express activity in the sample component within the total assay volume only) • 5.64 is the MAC of NADH at 340 nm in L −1 mmol −1 cm −1 calculated using our assay conditions (see step 5b below) • 0.67 cm is the calculated light path of liquid in the well (see step 5a below) • Multiplication by 1000 converts millimoles of G-3-P produced to micromoles. This allows expression as units (U) of transketolase activity per minute 25 • Hb concentration (g/L) is measured in each sample hemolysate; this concentration is divided by 10 to give the assay dilution Hb concentration (g/L). Multiplying the basal activity by 1/Hb (assay dilution) concentration gives basal activity in U/gHb or μmol G-3-P produced per gram enzyme activity per minute.
In practice, it is simpler if the basal activity (not corrected for Hb) is calculated first using part of Equation (1)

Absorbance, path length, and MAC
As stated by Beer's law, absorbance is equal to the MAC × path length (of the light path) × concentration. As detailed above, ΔA is measured at 340 nm in the ETK activity assay.
Published values for MACs are based on measurements under specific conditions using a spectrophotometer and cuvettes with a standard 1-cm path length. Therefore, it is inappropriate to use such published values where these standard experimental conditions are not met. An assay-specific MAC needs to be calculated according to the actual path length of the 96-well microplate and the assay conditions.

Path length
Calculation of path length through a 96-well microplate is challenging as it is dependent not only on the volume of liquid in the well but also on the shape of the well and the viscosity of the mixture that affects the meniscus. It may be possible to calculate path length using the geom-

MAC (ε)
The MAC of NADH is widely published as 6.22 × 10 3 M −1 cm −1 ; however, this value is dependent on other variables, including assay components, pH, and temperature. 26 It is, therefore, recommended that a laboratory-specific value be calculated.

Measurement and calculation of hemolysate Hb concentration using a 96-well microplate method
Hemolysates prepared for the ETKac assay were also used to mea- Hematology-16 Control "Low" and "Normal" were used as controls.
A 1-in-3 dilution in deionized water of the low control was prepared and used as an additional control. Reconstituted Drabkin's reagent (200 μl) was added to each well. The microplate was then transferred to the plate reader, shaken at medium speed for 5 s, and read at 540 nm.
The Thermo SkanIt software was used to produce a standard curve and to calculate Hb concentrations in the hemolysate samples and controls.

Final calculation of basal ETK activity
The measured Hb hemolysate concentration from above was divided by 10 to give the assay dilution Hb concentration in g/L. For information, summary statistics of the measured concentrations of Hb in the sample hemolysates and raw basal reaction rates are reported in Table S2.
Adding Hb concentration to the simplified Equation (3) gives Equation (4): As shown in Equation (4), multiplication of the basal activity by 1/Hb (in effect dividing the basal activity by Hb concentration) gives U/gHb.  [29][30][31][32] ETKac results were available from previous analyses in our laboratory. 29,31 Hb was measured and basal ETK activity was calculated in the same samples. Table 1 shows the ETKac and basal ETK activity for the two populations as well as the proportions of the population against the cut-offs described above. The Gambian population was at higher risk of thiamine deficiency and had a significantly higher mean ETKac and lower mean basal ETK activity compared with the UK sample (t-test, both ps < 0.0001).

RESULTS
ETKac country distributions overlapped ( Figure 1A), but the distribution was shifted right in the Gambian population sample, indicating a greater level of insufficiency (higher ETKac values) compared with the UK population sample. Distribution of basal ETK activity ( Figure 1B) showed a similar pattern between the two population samples, with lower basal activities in The Gambia.
Notwithstanding uncertainties over the cut-offs, relationships between ETKac and basal ETK activity are shown in Figure 2 together with the cut-offs for sufficiency and deficiency for ETKac 2 and reference limits proposed for basal ETK activity. 17 The Gambian population sample (Figure 2A)   Measurement and calculation of the basal ETK activity has been described previously. In early studies dating back four decades or more, sufficient detail was provided to follow the method and repeat the protocol. 20,21,33,34 More recently, typical reporting of the methods is scant or points to previously published manuscripts that may or may not provide detail or provide incomplete method details to allow replication of the method. [16][17][18][19]35 In addition, no previously published method provides details of the steps required to calculate basal ETK activity using the 96-well microplate method.
If a laboratory measures ETKac, then the additional calculation of basal ETK activity is relatively straightforward if the methods outlined here are followed. Experiments to calculate values for path length and MAC in a specific laboratory only need to be performed once. The assay to measure Hb in 96-well microplates is straightforward, inexpensive, and quick. Although the method to measure ETKac and basal ETK activity has multiple steps, it requires relatively little in the way of costly equipment and associated specialist, technical knowledge. For this reason, it is particularly suited to resource-limited countries where there is often evidence of thiamine deficiency.
A set of commonly used ETKac cut-offs to define thiamine deficiency are based on studies conducted in different patient groups. 27,36,37 In contrast, thiamine deficiency cut-offs for basal ETK activity are not well described. Soukaloun et al. used receiver operator curve analysis to determine a basal ETK activity cut-off of 0.59 U/gHb and reported good sensitivity and specificity to determine the risk of clinical infantile beriberi (75% [48-93%] and 85% [66-96%], F I G U R E 2 Relationship between ETKac and basal ETK activity (U/gHb) in The Gambia (left) and UK (right). Shaded areas indicate commonly used cut-offs for thiamine sufficiency (green), moderate risk of deficiency (yellow), high risk of deficiency (orange), and ETKac purportedly associated with beriberi (red). 2 The horizontal dashed line indicates a previously proposed but unconfirmed lower limit of the proposed reference range (0.50 U/gHb) for UK women aged 25-34 years. 17 Abbreviations: ETK, erythrocyte transketolase; ETKac, erythrocyte transketolase activity coefficient. respectively). This was in contrast to the use of ETKac cut-offs that showed no discrimination between cases and controls. 17 The cut-off value of 0.59 has since been used in other studies, 18,23,35 but remains unsubstantiated. Soukaloun et al. 17 also reported reference ranges of between 0.47 and 0.57 U/gHb for children and women aged between 4 and 49 years and derived from UK NDNS data. 17 However, the validity of these cut-offs and reference ranges and their suitability across different populations, age, and sex groups has not yet been tested.
Basal ETK activity is reported to be more closely related to ThDP than ETKac, 18 possibly because of the potential ambiguity in activity coefficients when apoenzyme levels are low, and was observed to better predict infantile beriberi than ETKac. 17 Other studies have reported that ETKac provides a more robust and sensitive marker of thiamine status. 8 The question of whether basal ETK activity or the ETKac better reflects thiamine status is not resolved and may depend on the population under study. What is evident is that basal ETK activity is useful for the interpretation of ETKac in the context of variable apoenzyme concentrations.
One of the limitations of the method overall, partly due to the higher concentration of thiamine in leukocytes, 5 is the requirement for washed red blood cells. Whole blood has been considered as a simple, alternative sample type since its use removes the need for saline washes immediately after sample collection and prior to sample freezing. Limited evidence to date suggests differences in transketolase activity but similar ETKac, 38,39 but further work is required to demonstrate the utility of whole blood and to better understand the effects of raised leukocyte count. 40 16,41 Animal studies have suggested that chronic thiamine deficiency may cause low apoenzyme levels. 41 Observation of a relationship between transketolase activity and age was observed in children 42 but not adults. 16 Different forms of transketolase may also influence the measurement of enzyme activity. 43 Although these studies provide some information on potential factors that may affect the transketolase enzyme, the evidence base is small and further work is necessary to confirm the findings.
Data from the UK and The Gambia show the contrasting thiamine status of these two populations. Values for Gambian ETKac data were published previously, 29 and, as may be expected based on ETKac, mean basal ETK activity was lower in The Gambia than in the UK. Notwithstanding the fact that the cut-off for basal ETK activity is based on limited data and may not be equally applicable in these populations; a large proportion of women judged to be thiamine sufficient on the basis of ETKac cut-offs had low basal ETK activity. The appropriateness of these cut-offs and the implications for health require further enquiry.
Previous studies from different countries have reported similar basal ETK activities to those reported here, with study means/medians generally ranging between ∼0.5 and ∼1 U/gHb. Mothers in Laos had a mean basal ETK activity of 0.58 U/gHb. 17 A number of other studies performed in Laos reported values in malaria-infected patients 18,19 and in infants with 17 and without 17,44 clinical signs of thiamine deficiency.
Other clinical populations studied include diabetics and alcoholics in Poland 16 and, in the UK, patients with chronic fatigue syndrome 45 and alcohol withdrawal syndrome. 23 Study designs vary and do not always include healthy control groups. Bailey et al. studied apparently healthy adolescents in the UK and reported a mean basal ETK activity of 0.90 U/gHb. 22 As discussed, comparison between studies is hindered by variability in methods and/or lack of method detail.

CONCLUSION
We present a method to measure and calculate basal ETK activity.
This will allow laboratories to introduce and/or harmonize protocols and review assay performance to facilitate collaboration between laboratories for the measurement of thiamine status.
Harmonization of methods will allow researchers to more readily combine datasets to better understand the utility of basal ETK activity as a complementary marker of thiamine status. Harmonization is a necessary precursor to bridging the gap between defining cut-offs for deficiency and understanding the nutritional (i.e., thiamine deficiency) and the broader etiology of thiamine-related disorders. Furthermore, the measurement of transketolase activities in erythrocytes and other tissues allows the exploration of the role of thiamine and the effectiveness of thiamine analogs as clinical treatments, for example, in renal disease 46