HPLC grade methanol was obtained from J. T. Baker (Phillipsburg, NJ, USA). Trichloroacetic acid (TCA), potassium ferricyanide (K₃Fe(CN)₆), thiamine hydrochloride (vitamin B1), thiamine pyrophosphate (TDP), thiamine monophosphate chloride (TMP) and hydrochloric acid were purchased by Sigma-Al- drich (St. Louis, MO, USA). Diethyl ether was obtained from Cromoline (Diadema, SP, Brazil), disodium hydrogen phosphate (Na₂HPO₄・7H₂O), sodium dihydrogen phosphate (NaH₂PO₄・H₂O), sodium hydroxide and triethylamine (TEA) were obtained from Reagen (Colombo, PN, Brazil). Ethylenediaminetetraacetic acid (EDTA) was purchased from Vetec (Duque de Caxias, RJ, Brazil). All reagents were of the highest purity available. For all solutions in which water had to be added, Milli-Q deionized water was used.

The HPLC system was a Shimadzu chromatograph (LC-10AD, Tokyo, Japan) with a fluorescence detector (FLD-Shimadzu spectrofluorometric detector RF- 551, Tokyo, Japan), a 200 µL loop (Rheodyne 7725-I, CA, USA) and a LC-10 AD PUMP. This system was equipped with a 5 µm particle size analytical column (Purospher Star RP-18 end capped −250 mm × 4.6 mm, ID-Merck, Darmstadt E.R., Germany) and a pre-packed column (Purospher Star RP-18 −4 × 4 mm- Merck, Darmstadt E.R., Germany). An integrator (Shimadzu C-R7Ae plus) was used to analyze the chromatographic data.

The mobile phase consisted of an equimolar buffer solution of 0.14 M NaH₂PO₄・H₂O and Na₂HPO₄・7H₂O and methanol (70:30, v/v), pH 7.0, plus 0.1% TEA. Before the chromatographic analyses, the phase was filtered through 0.45 µm membrane filters (Millipore Durapore) and vacuum degassed prior to use. Chromatographic analyses were performed at 25˚C ± 2˚C and the compounds were eluted isocratically over 15 min runtime at a flowrate of 1 mL/min. The fluorescence detector was set at an excitation wavelength of 367 nm and an emission wavelength of 435 nm, high sensitivity and range of 1.5. Both peak height and area are proportional to concentration.

One milligram per milliliter stock solutions of TDP, TMP and B1 were prepared in 0.1 M hydrochloric acid and aliquoted out for keeping at −20˚C. Standard solutions (100, 325, 550, 775, 1000 ng/mL for TDP; 50, 100, 175, 250 ng/mL for TMP and 2, 10, 50, 80, 100, 150 ng/mL for B1) were prepared daily by dilutions of the stock solutions, aliquoted out and stored at 4˚C until derivatization and analysis. 150 µL of standards were derivatized by addition of 150 µL of freshly prepared potassium ferricyanide in 15% (w/v) sodium hydroxide. A blank sample was used containing 150 µL of water and 150 µL of potassium ferricyanide solution.

Sixteen male Wistar rats weighing 250 - 300 g were collected from the Centro de Bioterismo da Universidade Federal de Minas Gerais (CEBIO-ICB-UFMG vivarium). The rats were housed in plastic cages in groups of four, maintained on a 12:12 h light-dark cycle and fed ad libitum. Animals were killed by decapitation in random order, blood samples were collected, their brains were removed and the tissues were separated and processed as described in Section 2.5. The present study was approved by the Ethics Committee for Care and Use of Laboratory Animals of the Universidade Federal de Minas Gerais (CEUA-UFMG protocol number: 107/2010) and the care and use of animals were done according to the National Institutes of Health Guide for Care and Use of Laboratory Animals [45] .

As mentioned above, the brains were rapidly removed from the cranium and prefrontal cortex, thalamus and hippocampus were dissected out on an ice-cold plate. The tissues were weighed and placed in 1.5 mL microcentrifuge tubes, and stored at −80˚C until analysis. Samples of 500 µL of blood were collected in tubes containing EDTA 6% (w/v) and stored at −20˚C until the day of the assay. All assays were done within one week. On the day of the assay the samples were thawed, homogenized and the thiamine and its phosphate esters extracted. The assays were then carried out in the supernatant extract within one minute after derivatization. All these methods are described below.

The brain tissue and blood samples were processed as described in [1] with few modifications in the quantities of some reagents as detailed below. The tissues were homogenized in 15 vol. of TCA 10% (85:15, v/v) and blood samples were homogenized in TCA 10% (1:1 v/v) for protein precipitation. Samples were vortexed vigorously for 1 minute, left standing on ice for 15 min and vortexed vigorously for 1 minute more. Immediately after, the samples were centrifuged (10,000 ×g for 15 min at 4˚C) and aliquoted until derivatization for thiamine and its phosphate esters analysis. Each supernatant was filtered (Millipore 0.45 μm, 13 mm) and transferred into a falcon tube and washed twice with 500 µL of diethyl ether saturated with water (5:1 v/w) to remove TCA. Aliquots of the solution (100 µL) were transferred to a corresponding tube and methanol (31 µL) was added. The samples were left standing on ice until derivatization. Briefly, 150 µL of supernatant mixture plus methanol were derivatized by adding 150 µL of freshly prepared potassium ferricyanide in 15% (w/v) sodium hydroxide. After the samples were vortexed and analyzed after 1min at room temperature. The derivatization of the thiamine and its phosphate esters in under alkaline conditions with potassium ferricyanide allows the formation of fluorescent thiochromes derivative [1] .

The assay validations were carried out with hippocampus samples of male Wistar rats and standards of thiamine and its phosphate esters.

Linearity and detection limit: The linearity of the detector response to standard solutions of TDP, TMP and B1 were determined. Calibration curves obtained from standard solutions of 100, 325, 550, 775 and 1000 ng/mL of TDP; 10, 50, 100, 175 and 250 ng/mL of TMP, and 10, 50, 80, 100 and 150 ng/mL of B1 were done.

Recovery: For evaluating the recovery of TDP, TMP and B1, a mixture of these compounds was added to samples of hippocampus at the moment of homogenization, at final concentrations of 100, 250, 400, 550 and 700 ng/mL of TDP; 25, 75, 125, 175 and 250 ng/mL of TMP and 20, 40, 60, 80 and 100 ng/mL of B1. Recovery was calculated as: [(final concentration-initial concentration)/ added concentration].

Stability: Experiments were done to determine the stability of thiamine and its phosphate esters in two kinds of samples/conditions: 1) hippocampus homogenate and 2) supernatant extract after derivatization process. In the first case, homogenized hippocampus samples were stored at −80˚C and assayed 7, 15 and 30 days after homogenization. In the latter, the stability of the thiochrome derivatives in the supernatant extract was assessed 1, 30, 60, 120 and 240 minutes after potassium ferricyanide oxidation. For the two conditions, the data obtained were expressed as nmol/g of tissue for each compound separately or by adding the three obtained values in a single representative total value (TDP + TMP + P = Total, nmol/g of tissue).

Experimental variation: Intraday reproducibility was calculated from 10 consecutive injections of hippocampus samples and interday reproducibility was obtained by comparing the means of 10 replicates over two consecutive days.

Hippocampus samples from male Wistar rat brains were used in the assays to validate the method. Mean elution times for TDP, TMP and B1 were 3.82, 4.28 and 11.67 min, respectively. Thus, in a total run time of less than 12 min the thiamine and its phosphorylated forms were eluted from the HPLC system. Figure 1(a) shows the chromatographic profile of standard solutions of TDP, TMP and B1, and Figure 1(b) shows the chromatographic profile for these parameters extracted from hippocampus sample. There was no peak in the chromatographic profile obtained with blank samples, indicating that the potassium ferricyanide oxidation does not interfere with the assay.

The linearity of the detector response to standard solutions of TDP, TMP and B1 were determined by calibration curves and the obtained data are shown in Figures 2(a)-(c), respectively. The result of each concentration is the average of triplicate assays. Linear correlation between the peak areas and concentrations was demonstrated for the three parameters, with r-values ranging from 0.993 to 0.996. Both peak height and area are proportional to concentration.

The accuracy of the method was estimated by recovery assays. Table 1 shows the recovery percentage for TDP, 101.7% ± 0.8%, TMP, 96.04% ± 2.7%, and B1, 97.3% ± 1.43% from hippocampus samples. The obtained accuracy indicates that the methods for sample processing and chromatographic analysis shown in the present study can be used for quantitative analyses of these compounds in the rat brain tissue. The values of reproducibility and accuracy obtained here are in accordance with the data described by other authors who used different chromatographic conditions [37] [46] .

Figure 3 shows the data obtained in the experiments carried out to investigate the stability of thiamine and its phosphate esters in two different conditions. (a) shows data of the stability of the compounds in hippocampus homogenate and (b) in supernatant extract after the derivatization process. As mentioned in the Methods section, in the first case (Figure 3(a)), homogenized hippocampus samples were stored at −80˚C and assayed 7, 15 and 30 days after homogenization. In the second condition (Figure 3(b)), the stability of the thiochrome derivatives in the supernatant extract was assessed 1, 30, 60, 120 and 240 minutes after potassium ferricyanide oxidation. The data obtained for the two conditions were expressed as nmol/g of tissue for each compound separately or by adding the three obtained values in a single representative total value (TDP + TMP + P = Total, nmol/g of tissue).

The data show that after the derivatization process the brain extract samples have to be analyzed in less than 30 min. In the present study, all samples were

injected into the chromatographic system within 1 min. Comparing to data obtained by other study, 30 min can be considered a lower stability [41] . They found that TDP, TMP and B1 kept in a cell culture after derivatization were stable until 3 h, at room temperature. After this period, they detected TDP hydrolysis to TMP. One possible explanation for the difference between these results could be not only the kind of samples, cells and tissues, but also light effects on the compound stability. The derivatized samples were stored in the dark at room temperature [41] . Further systematic studies have to be done to test this hypothesis. Despite the relative low stability it is important to consider that the use of potassium ferricyanide as a derivatizing agent has many advantages, among which are an adequate sensitivity and being safer than other derivatizing agents, like cyanogen bromide and mercuric chloride [44] . As described by other authors [47] , 5M phosphoric acid can also be added to derivatized samples before injecting into the HPLC. In some samples we detected that it can improve the resolution of the phosphate esters (data not shown).

The experimental variation of the analyses is shown in Table 2. Intraday reproducibility was calculated from 10 consecutive injections of derivatized supernatant obtained from the hippocampus, yielding relative standard deviations (R.S.D.s) of 2.66%, 4.50% and 7.43% for TDP, TMP and B1, respectively. Interday reproducibility was obtained by comparing the mean of 10 replicates over two consecutive days. The obtained interday R.S.D.s were 37.54%, 25.39% and 25.87% for TDP, TMP and B1, respectively. The data show that the intraday reproducibility is high with small R.S.D. for each of the three assessed parameters. On the other hand, the interday reproducibility was low showing high R.S.D.s. These data indicate that after derivatization the analyses have to be done on the same day.

Following the validation of the method using hippocampus sample, as described above, we also determined the levels of TDP, TMP and B1 in the thalamus, PFC and blood samples from male Wistar rats. Table 3 shows the concentrations of B1 and its phosphorylated forms expressed in nmol/g of tissue or in nmol/L of

blood. In addition, it also shows the retention time, in minutes, for each of the three assessed variables. The results are expressed as the average of two independent experiments. In each experiment the measurements were done in triplicate for hippocampus, PFC and blood. The measurements for thalamus were done in duplicate.

Extraction procedures of B1 and its phosphate esters and HPLC quantifications after derivatization under alkaline conditions to form fluorescent thiochrome derivatives were previously reported by other authors who used biological samples obtained from different kinds of tissues [42] [43] [44] [46] [48] [49] [50] [52] and cell culture [41] . However, these authors did not validate the method used for the specific brain samples assessed and also none of them measured these parameters in the hippocampus and prefrontal cortex as shown in the present study. The majority of them used homogenates of the whole brain, cortical hemispheres, cerebral cortex and cerebellum as well as samples from liver, kidney, heart and urine. Similarly, to the present study, the phosphate ester parameters were measured in the thalamus samples [43] , but they did not assess the vitamin B1 levels nor did they use a validated method for measurement of these compounds in brain tissue.

The TDP concentrations shown here for each of the three rat brain areas are within the same range, from 6 to 15 nmol/g of brain tissue, described by other authors [46] [48] , who used distinct kinds of samples compared to that used in the present study. The values (nmol/g of tissue) found here were 8.8, 9.6, and 11.2 in the hippocampus, thalamus and prefrontal cortex, respectively.

The methods described here are also useful for measuring levels of B1 and its phosphate esters in rat blood. The average concentration found for the total thiamine (TDP + TMP + B1) in rat blood was 203.3 ± 8.7 μg/L (n = 8). This value is similar to that described by others authors: 283.1 ± 0.75 μg/L (n = 7) [52] and 243.4 ± 12.1 μg/L (n = 7) [53] .

This study has limitations, including the stability of thiamine and its phosphate esters, and the amount of the sample.

Thiamine and its phosphate esters are temperature and light-sensitive and their stability is low. Therefore, samples should be storage in −80˚C for 1 week and the analysis should be carry out in the same day and until one minute after the derivatization.

Since the thalamic sample obtained is very small, it only allowed for a duplicate run rather than the standard triplicate.

Due to our difficulty in obtaining a standard for thiamine triphosphate (ThTP), we were unable to perform analysis for the compound.