No evidence of an intracellular lactate shuttle in rat skeletal muscle

Abstract

The concerted view is that cytosolic pyruvate is transferred into mitochondria and after oxidative decarboxylation further metabolized in the tricarboxylic acid cycle. Recently this view has been challenged. Based on experimental evidence from rat skeletal muscle it has been concluded that mitochondria predominantly oxidize lactate in vivo and that this constitutes part of an ‘intracellular lactate shuttle’. This view appears to be gaining acceptance in the scientific community and due to its conceptual importance, confirmation by independent experiments is required. We have repeated the experiments in mitochondria isolated from rat soleus muscle. Contrary to the previously published findings we cannot find any mitochondrial respiration with lactate. Analysis of lactate dehydrogenase (LDH) by spectrophotometry demonstrated that the activity in the mitochondrial fraction was only 0.7 % of total activity. However, even when external LDH was added to mitochondria, there were no signs of respiration with lactate. In the presence of conditions where lactate is converted to pyruvate (external additions of both LDH and NAD⁺), mitochondrial oxygen consumption increased. Furthermore, we provide theoretical evidence that direct mitochondrial lactate oxidation is energetically unlikely. Based on the present data we conclude that direct mitochondrial lactate oxidation does not occur in skeletal muscle. The presence of an ‘intracellular lactate shuttle’ can therefore be questioned.

The current view of oxidative metabolism is that cytosolic pyruvate is transferred into mitochondria, and after oxidative decarboxylation to acetyl-CoA is further metabolized in the tricarboxylic acid (TCA) cycle. In a recent paper this view is challenged, and it is concluded that mitochondria predominantly oxidize lactate in vivo (Brooks et al. 1999). The evidence for this conclusion is (a) isolated mitochondria oxidize lactate with a similar rate to that of pyruvate, (b) lactate exceeds cytosolic pyruvate concentration by an order of magnitude and (c) gel electrophoresis and immunocytochemistry demonstrate the presence of lactate dehydrogenase (LDH) in isolated mitochondria.

The findings have been incorporated into a model termed ‘intracellular lactate shuttle’ by which lactate is transported into the mitochondrial matrix by the monocarboxylate transporter located in the inner mitochondrial membrane. According to the hypothesis, lactate is converted to pyruvate in the mitochondrial matrix and oxidized. The model is an extension of the lactate shuttle hypothesis where lactate is exchanged between muscle tissues and between muscle fibres. Brooks et al. (1999) suggest that the presence of an intracellular lactate shuttle could explain why fully oxygenated muscles produce lactate.

The idea of direct mitochondrial lactate oxidation appears to be gaining a wide acceptance (Brooks et al. 1999; Brooks, 2000; Gladden, 2001; Hood, 2001). If the idea is correct, the current view of mitochondrial fuel utilization would be altered. Owing to the conceptual importance of the findings, we attempted to confirm the results in experiments with mitochondria from rat soleus muscle. Soleus muscle is mainly composed of slow-oxidative (type I) fibres and would, according to the lactate shuttle hypothesis, be one of the main tissues for lactate oxidation (Brooks et al. 1999).

The purpose of this study was to test the hypothesis that mitochondria can oxidize lactate directly. Contrary to the findings presented by Brooks et al. (1999) we: (a) find no evidence that mitochondria can use lactate as a substrate without prior conversion to pyruvate in the cytosol and (b) find insignificant activities of LDH in the mitochondrial fraction when analysed by gel electrophoresis and quantitative spectrophotometry. Furthermore, we provide theoretical evidence that direct mitochondrial lactate oxidation is energetically unlikely.

METHODS

Animal care and feeding

Sprague-Dawley rats (250-300 g; BKI:SD) obtained from B & K Universal (Sollentuna, Sweden) were housed at 23 °C on a 12 h light-12 h dark cycle. Animals were given free access to a B & K Universal standard rat and mouse diet and tap water. Rats used in this study were cared for in accordance with the Principles for the Utilization and Care of Vertebrate Animals Used in Testing, Research and Training. The experimental protocol was approved by the local ethical committee on animal experiments.

Preparation of muscle samples

Animals were anaesthetized and killed with an overdose of pentobarbital sodium (100 mg kg⁻¹). Soleus muscles from both legs were dissected and divided into two parts. One part (10-15 mg) was quenched in liquid nitrogen, freeze-dried and stored at −70 °C. This part was used for analysis of protein content and enzyme activities. The remaining part (115-170 mg) was used for isolation of mitochondria.

Isolation of mitochondria

Mitochondria were isolated as previously described (Tonkonogi & Sahlin, 1997). Briefly, muscle specimens were cut with scissors, homogenized in the presence of bacterial proteinase (Nagarse EC 3.4.21.62) and fractionated by differential centrifugation. The final mitochondrial pellet was resuspended (0.4 μl mg⁻¹ initial muscle) in a medium consisting of (mmol l⁻¹): mannitol, 225; sucrose, 75; Tris, 10; EDTA, 0.1; pH 7.40 (7.7-12.4 mg mitochondrial protein ml⁻¹). Prior to measurements of respiration, the mitochondrial suspension was kept on ice. Citrate synthase (CS) is exclusively located in mitochondria and measurements of CS activity in muscle homogenate and the mitochondrial fraction (see below) were used to estimate the percentage of mitochondria freed from the muscle.

Polarography

The respiration rates of isolated mitochondria were measured with a Clark-type electrode (Hansatech DW1, Norfolk, UK) at 25 °C. The measurements were carried out in 0.5 ml reaction medium containing (mmol l⁻¹): mannitol, 225; sucrose, 75; Tris, 10; KCl, 10; K₂HPO₄, 10; EDTA, 0.1; malate, 2; pH 7.0. Substrate was either pyruvate or lactate or combinations of pyruvate and lactate. The solubility of oxygen in the medium was considered to be equal to 237.5 μmol l⁻¹.

Respiration was initiated by the addition of 7.5 μl of the mitochondrial suspension to the reaction medium and a conventional respiratory experiment with state 4-3-4 transitions was performed. State 3 was initiated by adding ADP (final concentration 300 μmol l⁻¹). Respiratory control index (RCI) was calculated as the ratio of the respiratory rate in state 3 to the rate of oxygen uptake after exhaustion of ADP (state 4).

Protein content and activities of LDH and CS

Muscle specimens were dissected from solid non-muscle constituents, powdered and homogenized (12.5 mg dry weight muscle ml⁻¹) in Triton solution containing Triton X-100 (0.05 % v/v), 50 mm KH₂PO₄ and 1 mm EDTA at pH 7.5. The homogenate was centrifuged for 30 s (1400 g) and the supernatant analysed for protein and the activity of CS and LDH. The activities of CS and LDH were also determined in the mitochondrial suspension. Mitochondria were disrupted by freeze thawing and dilution (×20) in ice-cold Triton buffer prior to assay of enzyme activities. Protein concentration was determined using a commercial kit (BCA Protein Assay Reagent Kit, Pierce, Rockford, IL, USA). The activity of citrate synthase (EC 4.1.3.7) was determined spectrophotometrically at 30 °C (Alp et al. 1976). The activity of LDH was determined according to Bergmeyer & Bernt (1974). Each sample was analysed in duplicate.

Analysis of LDH isoenzymes with electrophoresis

Mitochondrial suspension treated with Triton and sonicated (4 μl, 21-34 μg protein) or muscle homogenate (2 μl, 8 μg protein) was added to agarose (1 %) gels prepared in electrophoresis buffer (Sigma 705-1). After electrophoretic fractionation (30 min, 100 V), isoenzyme activity was visualized with a colorimetric procedure (Sigma, procedure no. 705-EP) and scanned.

RESULTS

The protein content of the mitochondrial suspension was 8.2 ± 1.1 (mg protein) ml⁻¹ (mean ± s.e.m.). The activity of CS in the mitochondrial fraction corresponded to 15.1 ± 1.9 % of the activity in whole muscle. Since CS is exclusively located in the mitochondria, the yield of mitochondria was also 15.1 %. The activity of LDH in whole muscle was 160 ± 8 mmol (kg wet wt)⁻¹ min⁻¹. The activity of LDH in the mitochondrial fraction varied between 0 and 2.1 mmol (kg wet wt)⁻¹ min⁻¹ (average 1.2), which corresponds to 0-1.2 % (average 0.7 %) of the muscle activity.

When pyruvate-malate was used as substrate, isolated mitochondria showed a high-ADP-stimulated respiration (state 3 or V_max) (Fig. 1A and Table 1) and were tightly coupled. In contrast, the respiratory rate with lactate- malate was very low (Fig. 1B) and similar to that with malate alone (Table 1). Addition of LDH to the oxygraph decreased the respiratory rate, but further additions of NAD⁺ and pyruvate stimulated respiration to 42 and 92 % of V_max (Fig. 1B and Table 1). The sensitivity of state 3 respiration to pyruvate was investigated by adding increasing concentrations of pyruvate to the oxygraph. Approximately 50 % of V_max was reached when [pyruvate] was 0.015 mm (Table 1). Addition of malate is necessary for an adequate function of the tricarboxylic acid cycle. When malate was omitted from the substrate there was no respiration with lactate, and only a submaximal respiration with pyruvate (data not shown).

Figure 1. Representative oxygraph traces of mitochondrial respiration with pyruvate or lactate.

Respiration medium contained malate (2 mm), 127 μg mitochondrial protein ml⁻¹ and 5 mm pyruvate (A) or 5 mm lactate (B). Final concentrations of added substances were: ADP 0.3 mm, LDH 3 U ml⁻¹, NAD⁺ 0.6 mm, pyruvate 5 mm. The numbers below traces denote the rate of oxygen consumption (μmol O₂ l⁻¹ min⁻¹).

Table 1.

Respiratory rates and respiratory control index (RCI) of isolated mitochondria of rat skeletal muscle

Substrate (mm)	n	State 4 (pre)	State 3	State 4 (post)	RCI
Py (5 mm) + Ma (2 mm)	4	30.4 (3.4)	154 (7)	30.3 (2.7)	5.2 (0.4)
La (5 mm) + Ma (2 mm)	4	10.4 (1.3)	15.8 (2.5)
+ LDH (3 U ml−1)	4		11.3 (0.9)
+ NAD+ (0.6 mm)	4		64.0 (5.2)
+ Py (5 mm)	3		141 (15)	37.3 (4.6)	3.8 (0.3)
Ma (2 mm)	4	10.0 (1.0)	15.0 (2.3)
+ Py (0.015 mm)	4		78.8 (8.7)
+ Py (0.1 mm)	4		137 (12)
+ Py (5 mm)	4		145 (11)	26.6 (3.2)	5.6 (0.4)

Respiratory rates in nmol O₂ (mg protein)^-1 min^-1. Py, pyruvate; Ma, malate; La, lactate. Concentrations of added substrates correspond to final concentration in oxygraph medium. Values are means (s.e.m.). State 3 was initiated by addition of 0.3 mm ADP (final concentration). RCI = state 3/state 4 (post).

The activity of LDH isoenzymes in muscle homogenate and mitochondria was determined with agarose gel electrophoresis. Soleus muscle had a high activity of isoenzymes with two or more of the four subunits in the H-form (LDH-1, LDH-2, LDH-3), whereas isoenzymes dominated by the M-form (LDH-4 and LDH-5) were visible but less abundant (Fig. 2). There were no clear traces of LDH in the mitochondrial fraction, despite application of three to four times more protein than for homogenates.

Figure 2. Agarose gel electrophoresis of LDH isoenzymes in muscle homogenate (MH) and muscle mitochondria (Mit).

Muscle homogenate (2 μl, 8 μg protein) was added to lanes b and d. Mitochondrial suspension (4 μl) was added to lanes a (21 μg protein) and c (34 μg protein). The signs (- and +) refer to the polarity of the electrodes.

DISCUSSION

The present results demonstrate that mitochondria isolated from skeletal muscle readily oxidize pyruvate but cannot oxidize lactate directly. Additions of both LDH and NAD⁺ in the presence of lactate increased respiration to almost 50 % of V_max. The amount of lactate converted to pyruvate under these conditions can be calculated from the LDH equilibrium (apparent equilibrium constant 2.76 × 10⁻⁵ at 25 °C and pH 7.0) (Hakala et al. 1956). During the prevailing conditions ([lactate] = 5 mm, [NAD⁺] = 0.6 mm, 25 °C and pH 7.0), [pyruvate] would be 0.015 mm. The observed respiration, when pyruvate was added to a concentration of 0.015 mm (Table 1), was close to 50 % of V_max. This is similar to the respiration with lactate in the presence of both LDH and NAD⁺, and demonstrates that the system responds adequately when lactate is converted to pyruvate. It is well documented that lactate can be released by one muscle and oxidized by another muscle (or tissue) depending on the muscle glycogen level and the energetic state (Ahlborg et al. 1975; Gollnick et al. 1981). However, our results indicate that oxidation of lactate can only occur after extramitochondrial conversion to pyruvate.

Impact of the redox state on the LDH reaction

LDH catalyses an equilibrium reaction, and the direction of the LDH reaction is dependent on the concentrations of substrates and products:

Mitochondrial pH (pH 7.24) is similar to that in the cytosol (pH 7.04) during state 3 respiration (Addanki et al. 1967). Spectrophotometric measurements of the redox level in isolated mitochondria, which are relevant for the present experimental conditions, demonstrate that the NAD⁺-NADH redox couple is nearly completely reduced in state 4 (i.e. in the presence of substrate) and reduced to 42-63 % in state 3 (Chance & Williams, 1955b; Jobsis & Duffield, 1967). Furthermore, in intact toad sartorius muscle, surface fluorescence measurements demonstrate that mitochondrial NAD⁺-NADH is highly reduced in both quiescent and contracting muscle (about 70 and 45 % of the anoxic level, respectively; Jobsis & Duffield, 1967). Mitochondrial NADH/NAD⁺ ratio has from quantitative analysis of NADH and NAD⁺ before and after CN poisoning (assuming negligible protein binding of NADH) been estimated to be 0.55 in soleus muscle (Sahlin & Katz, 1986). In contrast to mitochondria, the NAD⁺-NADH redox couple is highly oxidized in the cytosol. From the LDH equilibrium, cytosolic NADH/NAD⁺ has been estimated to be about 0.004 in quiescent soleus muscle (Sahlin & Katz, 1986).

The NAD⁺-NADH redox couple is thus several orders of magnitude more reduced in mitochondria than in the cytosol. The maintenance of a highly reduced state is a well-known and accepted feature of mitochondria and is a pre-requisite for the redox drive of the electron transport chain and the function of oxidative phosphorylation. Conversion of lactate to pyruvate in mitochondrial matrix is, due to the reduced redox state, energetically unlikely, and would violate the first law of thermodynamics. If we hypothetically assume that LDH is present in the mitochondrial matrix and that the mitochondrial membrane is permeable to both pyruvate and lactate, the difference in redox state between mitochondria and cytosol would create a futile cycle by which pyruvate is reduced to lactate in mitochondria and vice versa in the cytosol (Fig. 3). Mitochondrial NADH would be oxidized and the driving force for the electron transport chain would disappear. It can be concluded that operation of LDH in the mitochondrial matrix would be incompatible with oxidative phosphorylation.

Figure 3. Schematic model describing the hypothetical scenario if LDH was localized in both the mitochondrial matrix (LDHm) and the cytosol (LDHc).

The presence of a monocarboxylate transporter in the mitochondrial membrane will mediate transmembrane transport of both pyruvate and lactate. In contrast to the cytosol, the redox state of the NAD⁺-NADH redox couple is highly reduced in the mitochondrial matrix and the flux of the LDH reaction (if LDH is present) would be from pyruvate to lactate (the intracellular lactate shuttle hypothesis would require the converse flux). The system would abolish the redox gradient between mitochondria and cytosol, remove the redox drive and prevent oxidative phosphorylation. Presence of LDH in the mitochondrial matrix is thus bioenergetically highly unlikely.

The average activity of LDH in the mitochondrial fraction was less than 1 % of total muscle activity and close to the detection limit. The finding that LDH activity in the mitochondrial fraction is very low is consistent with previous reports in cardiac tissue (0.9 % of total activity, Brandt et al. 1987) and liver tissue (2.4 % of total activity, Brandt et al. 1987; and 1.6 % of total activity, Kline et al. 1986). Complete removal of cytosolic enzymes during the preparation of mitochondria is difficult and it seems likely that the remaining low activity corresponds to cytosolic contamination of the mitochondrial fraction, possibly due to adsorption. When extra LDH was added to tissues prior to homogenation there was a large increase in mitochondrial LDH (Brandt et al. 1987), and this supports the view of cytosolic contamination. Digotinin fractionation of mitochondria demonstrated that almost all of the mitochondrial LDH was present in the intermembrane space (i.e. between the inner and outer mitochondrial membrane) (Brandt et al. 1987). Brooks et al. (1999) reported that LDH-1 and LDH-5 were present in mitochondria from skeletal muscle when measured with agarose gel electrophoresis. However, we could not detect any clear trace of LDH in the mitochondrial fraction despite using the same method and the addition of 20-30 times more protein (21-34 μg) than that used by Brooks et al. (1999). The negligible activity of LDH in the mitochondrial matrix is consistent with the inability of mitochondria to oxidize lactate. Brooks et al. (1999) used immunocytochemistry to measure the intracellular localization of LDH, and reported that the frequency of LDH-1 labelling in soleus muscle mitochondria exceeded that of the surrounding areas. Our results demonstrate that LDH-1 and LDH-2 are the dominating isoenzymes in soleus muscle and that mitochondrial LDH activity is negligible. From this perspective the report of a higher frequency of LDH-1 in mitochondria than in surrounding areas is surprising. The results obtained from immunocytochemistry are dependent on the specificity of the primary antibodies used, which in this case remains to be shown.

The results from this study extend the findings by (Popinigis et al. 1991) where mitochondria from human skeletal muscle showed no coupled lactate oxidation. However, our results disagree in many ways from those of Brooks et al. (1999). In contrast to the results reported by Brooks et al. (1999) we do not find any mitochondrial respiration with lactate. Secondly, we find the activity of LDH in the mitochondrial fraction to be negligible. Possible reasons for this discrepancy will be discussed.

It could be argued that our system is unable to oxidize lactate due to system restrictions (impurities, disturbing substances, etc.). However, since mitochondria were able to oxidize lactate after additions of LDH and NAD⁺, it is clear that the system responds adequately.

Another possibility is that the mitochondria used by Brooks et al. (1999) contained impurities of LDH, or alternatively that LDH was lost in our preparation. We have used a slightly different method to isolate mitochondria than that used by Brooks et al. (1999), and it is possible that the impurities of cytosolic LDH were higher in their mitochondrial preparation. However, the finding that mitochondria can oxidize lactate (Brooks et al. 1999) cannot solely be explained by contamination with cytosolic LDH since our preparation of mitochondria could not oxidize lactate even when a high activity of LDH was added to the system.

Different concentrations of ADP were used to induce state 3 respiration. We used 300 μM [ADP], whereas Brooks et al. (1999) used 0.005 μM. The apparent K_m of respiration for ADP in isolated mitochondria is about 20 μM (Chance & Williams, 1955a). The respiration at 0.005 μM ADP would be far from V_max and could give rise to abnormal results. Indeed, under our conditions such a low concentration of ADP has non-detectable affects on respiration. Observed V_max during pyruvate oxidation was 293 nmol O₂ (mg protein)⁻¹ min⁻¹ at 37 °C (Brooks et al. 1999) and 154 nmol O₂ (mg protein)⁻¹ min⁻¹ in the present study at 25 °C. When the difference in temperature is taken into account V_max is similar. This is astonishing considering the vast difference in used [ADP] to elicit state 3 respiration. Although we have been unable to find any published errata in subsequent volumes of the journal where the article by Brooks et al. (1999) was published (Proceedings of the National Academy of Sciences of the USA), we are inclined to believe that the given [ADP] of 5 nm is a typographical error.

The experiments in this paper were performed on rat soleus muscle. The experiments by Brooks et al. (1999) were performed on mitochondria from rat hindlimb skeletal muscle, cardiac muscle and liver. Theoretically one could argue that soleus muscle is a unique type of skeletal muscle and does not express the ‘intracellular lactate shuttle’. However, this appears highly unlikely since soleus muscle contains mainly slow oxidative (type I fibres) which, according to the lactate shuttle hypothesis, would be especially adapted for lactate oxidation (Brooks et al. 1999). Furthermore, we have in preliminary experiments investigated mitochondria isolated from muscle biopsies taken from human vastus lateralis, which has a mixed fibre type composition. Mitochondria from human muscle were unable to oxidize lactate, but oxidized puruvate at a high rate (authors’ unpublished observations).

We conclude that (a) mitochondria are unable to oxidize lactate directly, (b) LDH activity in skeletal muscle mitochondria is negligible and (c) oxidation of lactate in the mitochondrial matrix is thermodynamically impossible due to the highly reduced state of the mitochondrial NAD⁺-NADH redox couple. Our results are in conflict with those presented by Brooks et al. (1999), where mitochondria from rat hindlimb were shown to readily oxidize lactate and to contain high activities of LDH. We have no plausible explanation for this discrepancy.