Metabolic Brain Disease

, Volume 23, Issue 1, pp 115–122

Glucose loading precipitates focal lactic acidosis in the vulnerable medial thalamus of thiamine-deficient rats

  • Darren Navarro
  • Claudia Zwingmann
  • Nicolas Chatauret
  • Roger F. Butterworth

DOI: 10.1007/s11011-007-9076-z

Cite this article as:
Navarro, D., Zwingmann, C., Chatauret, N. et al. Metab Brain Dis (2008) 23: 115. doi:10.1007/s11011-007-9076-z


Glucose loading in thiamine-deficient patients is known to precipitate Wernicke’s Encephalopathy; however, the mechanisms responsible have not been fully elucidated. Lactate accumulation occurs in brains of thiamine-deficient rats. In order to determine whether glucose loading in thiamine-deficient rats causes selective lactic acidosis in vulnerable brain structures, cerebral pH was measured autoradiographically using 14-labeled 5,5-dimethyloxazolidine-2, 4-dione ([14C]DMO) in the medial thalamus, a vulnerable brain region, versus cerebral cortex, a brain region that is spared in thiamine deficiency. Following administration of a glucose load, regional lactate levels and de novo lactate synthesis measured by 1H-13C-NMR spectroscopy, increased significantly to 21.86 ± 3.04 μmol/g (wet weight) in the medial thalamus (p < 0.001) and pH in this brain region was decreased significantly from 7.08 ± 0.04 to 6.87 ± 0.05 (p < 0.001). No such changes were observed in cerebral cortex following a glucose load. These results demonstrate that the increased production and accumulation of brain lactate result in acidosis following glucose loading in thiamine deficiency. Alterations of brain pH could contribute to the pathogenesis of thalamic neuronal damage and consequent cerebral dysfunction in Wernicke’s Encephalopathy.


Thiamine deficiency Lactate Glucose loading Cerebral acidosis Wernicke’s Encephalopathy 


Thiamine deficiency continues to be a problem in some world communities leading to infantile beriberi (McGready et al.2001; Luxemburger et al.2003); and Wernicke’s Encephalopathy (WE) is a serious complication of chronic alcoholism and other disorders associated with grossly impaired nutritional status (Butterworth 1989;Butterworth et al.1991). The precise cause of the selective lesions in WE is still unknown.

Lactate accumulation and its increased de novo synthesis have been shown to occur in brains of rats made thiamine-deficient (TD) by the thiamine antagonist, pyrithiamine (McCandless 1982; Butterworth and Heroux 1989; Navarro et al.2005). However, the mechanisms by which lactate may bring about neuronal injury are not fully understood. Although it is well established that glucose loading of TD patients precipitates Wernicke’s Encephalopathy (Wallis et al.1978; Watson et al.1981), the mechanisms implicated have not been established. Lactic acidosis is one such mechanism.

The present study was undertaken to determine the extent to which glucose loading leads to brain lactate accumulation and using 13C-NMR spectroscopy, its de novo synthesis within the TD medial thalamus (a brain region vulnerable to thiamine deficiency), and whether this increase of lactate results in acidosis. Local cerebral pH (LCpH) was determined using an autoradiographic technique and carbon 14-labelled 5,5-dimethyloxazolidine-2, 4-dione ([14C]DMO) as a pH marker.

Materials and methods

Thiamine deficiency protocol

Adult male Sprague Dawley rats weighing 200–225 g obtained from Charles River (St. Constant, Quebec, Canada) were used in all experiments. Rats were housed individually under constant conditions of temperature, humidity and 12 h light/dark cycles and had free access to water at all times. Rats were allowed to adapt to their environment for 3 days prior to the initiation of TD protocol, which consisted of feeding a diet deficient in thiamine (MP Biochemicals, Solon, OH, USA) and daily administration of pyrithiamine hydrobromide (0.5 mg per kg body wt) intraperitoneally (i.p.) (Troncoso et al.1981) until onset of anorexia and ataxia (12 days), but prior to appearance of opisthotonus. This time-point is not associated with significant neuronal cell loss in this model of thiamine deficiency (Todd and Butterworth 1999). All animal treatment procedures were approved by the Animal Ethics Committee of Saint-Luc Hospital (CHUM) and were carried out in accordance with the National Institute of Health Guide for the Care and Use of Laboratory Animals (NIH Publications No. 80-23).

Sample preparation for nuclear magnetic resonance (NMR) spectroscopic studies

For NMR investigations, [1-13C]glucose (Cambridge Isotopes, Andover, MA, USA) was administered either as a labeling dose (200 mg/kg, i.p.) or a glucose load (500 mg/kg, i.p.) 60 min prior to sacrifice. Rats were killed by decapitation, brains were removed from the skull, immediately frozen in isopentane over dry ice, and dissected on ice to obtain medial thalamus (MT, vulnerable brain region) and frontal cortex (FC, non-vulnerable brain region). Tissue samples were ground over liquid nitrogen and homogenized in 12% perchloric acid (PCA) at 0°C using a motor-driven polished glass tube-Teflon homogenizer. The homogenate was centrifuged at 40,000×g for 15 min and this was repeated once. The supernatants were combined and neutralized on ice with KOH. The precipitated KClO4 was removed by centrifugation (4,000×g, 15 min). Arterial blood was immediately mixed with 20% PCA, centrifuged (4,000×g, 40 min), neutralized with KOH, and centrifuged again (40,000×g, 15 min) to precipitate KClO4. The supernatants from each sample were lyophilized.

NMR spectroscopy

The lyophilized PCA extracts of the blood and the brain tissue were dissolved in 0.6 ml deuterium oxide (D2O; Merck, Darmstadt, Germany) and centrifuged. Prior to NMR analysis, the pH was adjusted to 7.0 with DCl and NaOD. 1H- and 13C-NMR spectra were recorded on a Bruker DRX 600 spectrometer, operating at frequencies of 600 MHz for 1H-measurements, and 150.9 MHz for 13C-measurements.

1H-NMR spectra were recorded with a 5-mm H,C,N inverse triple resonance probe (5-mm HX probe), 400 accumulations, repetition time 15 s, spectral width 7,183 Hz (DRX 600) or 3,623 Hz (NB/WB 360). Chemical shifts were referenced to lactate at 1.33 ppm. 13C-NMR spectra were recorded with a 5-mm 1H/13C dual probe, 20,000 accumulations, repetition time 2 s, composite pulse decoupling with WALTZ-16, spectral width 47,619 Hz (DRX 600). Chemical shifts were referenced to the C-3 signal of lactate at 21.3 ppm.

Measurement of regional brain lactate concentrations

Lactate concentrations were determined enzymatically (adapted for measurement in 96-well microtiter ELISA plates) using a commercially available kit (Roche Molecular Biochemicals, Mannheim, Germany). Concentrations of unlabeled lactate and its pool size ([13C] + [12C]) were determined from fully relaxed 1H-NMR spectra of brain extracts, obtained after injection of [1-13C]glucose, using known lactate concentrations as internal standards.

Determination of local cerebral pH (LCpH)

The methods used to determine local cerebral pH are based on previous studies using [14C]DMO (Hakim 1984; Kobatake et al.1984), with minor modifications. After 12 days of thiamine deficiency, rats were anesthetized with isofluorane and the femoral artery and jugular vein cannulated, after which they were allowed to recover for at least 4 h. Once the serum pH was stable and within the range established for normal rats, 60 μCi of [14C]DMO (specific activity, 55 mCi/mmol American Radiolabeled Chemicals, St. Louis, USA) dissolved in saline solution at a concentration of 100 μCi/ml was injected into the jugular vein. In order to examine the effects of glucose loading on acidosis, animals were administered either a glucose load (500 mg/kg in saline, i.p.) or a saline (i.p.) 45 min before the [14C]DMO injection (2 h before decapitation). This 2 h timeline for the glucose/vehicle injection was employed in order to allow ample time for acidosis to occur while not interfering with the administration and of [14C]DMO. Arterial samples were taken at 30, 45, 60 and 75 min after [14C]DMO injection for determination of plasma [14C]DMO radioactivity and at 30 and 60 min for determination of plasma pH.

Seventy-five minutes following [14C]DMO injection, animals were sacrificed, the brain was rapidly removed, frozen in isopentane on dry ice and stored at −80°C. Brain sections, 20 μm thick, were then cut in a microtome cryostat (International Equipment Company, USA) at −22°C. Sections were collected on gelatin-coated slides and dried at room temperature within 60 min of slicing since [14C]DMO is somewhat unstable at room temperature and could be lost from tissue slices exposed for long periods to room temperature. Sections were then exposed to X-ray film (Hyperfilm MP, Amersham Pharmacia Biotech, Buckinghamshire, UK) for 8 days at 4°C. Local tissue concentrations of 14C were determined from the optical densities of the medial thalamus, and autoradiographic [14C] micro-scales (Amersham Biosciences, UK) were included in every audioradiograph. Films were developed and tissue concentrations of [14C]DMO were determined by quantitative densitometry analysis using an MCID computer-based densitometer and image-analysis system (Imaging Research, St. Catharines, Canada). The amount of ligand bound to various regions was calculated from the specific activity of the ligand.

Calculation of LCpH

Autoradiographic studies in which [14C]DMO concentrations are used to determine regional pH consider the brain to be composed of two functional compartments, one plasma (p) and the other tissue (t). The LCpH was calculated from the following formula:
$$\frac{{\left[ {{\text{DMO}}^{\text{ - }} } \right]_t }}{{\left[ {{\text{DMO}}^{\text{ - }} } \right]_p }}\; = \;\frac{{10^{{\text{pH}}t} - 6.13 + 1}}{{10^{{\text{pH}}p} - 6.13 + 1}}$$
Comprehensive derivations of this equation are well-documented in previous studies using [14C]DMO as a pH marker in brain (Hakim 1984; Kobatake et al.1984). This equation was used to calculate the regional pH in this study where pHp and pHt are plasma and tissue pH respectively, and [DMO-] is the ionized species concentration in plasma or tissue. The pK of DMO is 6.13. From the above equation, the concentration of DMO in brain is proportional to the pH of the tissue. In all animals the water contents of plasma and brain gray matter and white matter were assumed to be 93, 81 and 69%, respectively (Katzman and Pappius 1973). [DMO]t was obtained from the densitometric readings; pHp was the average of arterial pHs determined at 30 and 60 min following [14C]DMO injection; and [DMO]p was the average of the plasma concentration in the last 45 min of the experiment (30, 45, 60, 75 min). This average was used to avoid reliance on one determination and was made feasible by equilibrium studies showing that after the initial 30 min, [14C]DMO concentrations in plasma declined by only 4% per h (Kobatake et al.1984). The calculated LCpH values (pHt) are a composite of intracellular and extracellular pH, but are weighted more heavily by pH in the intracellular space due to its larger size.

Statistical methods

Studies were carried out on four rats per group. The relatively small “n” here is due to the prohibitively high cost of 1-13C-glucose. Data are expressed as mean ± SD values. Regional lactate concentrations and the LCpH values were compared using one-way analysis of variance (ANOVA) to detect overall variation. Differences were considered significant when p ≤ 0.05.


Regional brain lactate concentrations

[1-13C]glucose is converted to [3-13C]pyruvate via the glycolytic pathway, and subsequently to [3-13C]lactate via lactate dehydrogenase. 1H-13C-NMR spectroscopy, (the heteronuclear spin pattern of lactate in the high-field region of 1H-NMR spectra), allows for distinction between lactate synthesized after glycolytic transformation of [1-13C]glucose ([3-13C]lactate) which is evident in the spectra as an additional doublet splitting due to 1H-13C-coupling (J = 128 Hz), and the unlabeled isotopomer ([12C]lactate) synthesized from unlabeled pyruvate. Table 1 shows the total concentrations of lactate ([12C] + [13C]) calculated from the high-field region of 1H-NMR spectra in both cerebral cortex and medial thalamus, with or without a glucose load. Following administration of a glucose load, the concentration of lactate increased significantly from 17.80 ± 2.07 to 21.86 ± 2.26 μmol/g (wet weight) in the vulnerable medial thalamus (p < 0.05), while levels remained relatively unchanged in the non-vulnerable cerebral cortex.
Table 1

Effect of glucose loading on regional brain lactate concentrations and local cerebral pH in TD rats


No glucose load

Glucose load

Regional brain lactate concentration (μmol/g)

 Cerebral cortex

11.92 ± 1.04

12.79 ± 1.50

 Medial thalamus

17.80 ± 2.07

21.86 ± 2.26*

Local cerebral pH

 Cerebral cortex

7.09 ± 0.04

7.10 ± 0.05

 Medial thalamus

7.05 ± 0.04

6.87 ± 0.05**

Concentrations of lactate determined by 1H-NMR spectroscopy of brain extracts obtained after administration of either 200 mg/kg (labeling dose) or 500 mg/kg (glucose load) [1-13C]glucose. pH values were measured autoradiographically using [14C]DMO administered 75 min prior to sacrifice. Glucose load (or vehicle) was administered 1 h before sacrifice. Values represent means ± SD for n = 4 animals per group. Data between all experimental groups were analyzed using ANOVA, where *p < 0.05, **p < 0.001.

Autoradiographic determination of LCpH in TD rat brain

Immediately prior to administration of [14C]DMO to TD rats arterial pH was 7.37 ± 0.08, and rectal temperature was 35.6 ± 0.5°C. These parameters were not significantly different in TD rats administered a glucose load.

Figure 1 shows representative autoradiographic sections at the level of the medial thalamus made 75 min after [14C]DMO administration in TD rats with or without a glucose load. Acidosis, indicated by hypodensity, can be clearly seen in the medial thalamus of the glucose-loaded TD rat. Table 1 also shows the calculated values for LCpH in the cerebral cortex versus the medial thalamus, in the presence or absence of a glucose load. LCpH (pHt values) were decreased significantly from 7.05 ± 0.04 to 6.87 ± 0.03 (p < 0.001) in the medial thalamus following a glucose load, while no significant differences were observed in the cerebral cortex.
Fig. 1

Acidosis due to glucose loading in TD rat brain. Representative coronal sections from TD rats administered [14C]DMO following 11 days of thiamine deficiency, with or without administration of a glucose load (500 mg/kg). Note the appearance of thalamic hypodensity, implying acidosis following a glucose load


Lactate accumulation in brains of TD animals has been well-documented since the pioneering experiments of Peters in the early 1930s. More recently, the selective accumulation of lactate has been shown to occur within cerebral structures which are susceptible to histological damage with continued thiamine deficiency (Navarro et al.2005). This pattern of vulnerability closely resembles that encountered in WE in humans.

Thiamine deficiency does not result in early compromise of brain activities of the thiamine-dependent enzyme PDH (Butterworth et al.1985), but activity of a second thiamine-dependent dehydrogenase, αKGDH, on the other hand has been found to be substantially decreased, leading to impairment in TCA cycle flux (Gibson et al.1984; Butterworth et al.1986). Impaired pyruvate oxidation resulting from decreased entry of pyruvate into the TCA cycle resulting from loss of activity of αKGDH is accompanied by increased de novo synthesis of lactate, which has been found to be elevated in circulation and brain tissue of both experimental animals and patients with WE (Holowach et al.1968; McCandless et al.1968; McCandless 1982; Munujos et al.1996).

Cerebral acidosis has been shown to occur in the medial thalamus of TD rats following the onset of opisthotonus (Hakim 1984). Medial thalamus is known ultimately to manifest neuronal cell loss in thiamine deficiency. Results of the present study reveal that the concentration of lactate in the TD medial thalamus prior to administration of a glucose load is already elevated relative to that of the TD cerebral cortex, however this produces no observable differences in tissue pH between the two brain regions (Table 1). Thalamic pH values in the present study are comparable to those previously published using this and other techniques (Roos 1971; Kogure et al.1980; Hakim 1984; Kobatake et al.1984).

Administration of a glucose load has been shown to precipitate WE in TD patients exhibiting no prior neurological symptoms (Wallis et al.1978; Watson et al.1981); however, the mechanisms involved in this phenomenon remain unclear. In the present study, brain lactate concentrations increased to 21.86 ± 2.26 μmol/g (wet weight) in the TD medial thalamus following the administration of a glucose load. Such a concentration has the potential to adversely affect cellular function. For example, in periods of hypoxia and ischemia, recovery varies with the nutritional state of the animals, and results of previous studies suggest that tissue lactate concentrations exceeding 20 μmol/g (wet weight) are detrimental for recovery (Myers 1979; Rehncrona et al.1980, 1981). This focal accumulation of lactate was accompanied by a significant decrease in local cerebral pH; a characteristic of later stages of thiamine deficiency (Hakim 1984). In contrast, both lactate concentrations and tissue pH values within the non-vulnerable cerebral cortex remained unaffected following administration of a glucose load.

Previous studies reveal that an acidic environment may have detrimental effects on cellular metabolism (Rehncrona et al.1980; Van Nimmen et al.1986); and pH changes can affect vascular tone and may lead to deleterious hemodynamic changes (Kuschinsky et al.1972). Acidosis is also known to enhance the production of reactive oxygen species, and oxidative stress is a feature of TD neuropathology, as shown by the accumulation of heme oxygenase-1, ferritin, reactive iron and superoxide dismutase in microglia, nitrotyrosine and 4-hydroxynonenal in neurons, as well as induction of endothelial nitric oxide synthase within vulnerable brain regions such as MT (Calingasan et al.2000; Kruse et al.2004). Endothelial cells are believed to be targets of free radicals (Kontos 1985, 1989); and impairment of blood-brain barrier function has been demonstrated in TD rats (Zelaya et al.1995) and in WE patients (Schroth et al.1991). Thus the microvasculature may be a potential primary target of acidosis-mediated damage, suggesting that a compromise in the blood-brain barrier, as observed in thiamine deficiency, may occur secondarily to focal cerebral acidosis in brain regions vulnerable to thiamine deficiency.


Research funded by the Canadian Institutes for Health Research (CIHR). We thank Professor Dieter Leibfritz, University of Bremen, for the generous availability of the NMR laboratory.

Copyright information

© Springer Science+Business Media, LLC 2007

Authors and Affiliations

  • Darren Navarro
    • 1
  • Claudia Zwingmann
    • 1
  • Nicolas Chatauret
    • 1
  • Roger F. Butterworth
    • 1
  1. 1.Neuroscience Research Unit, CHUM (Campus Saint-Luc)University of MontrealMontrealCanada

Personalised recommendations