Reply from Arend Bonen, Hideo Hatta, Graham P. Holloway, Lawrence L. Spriet and Yuko Yoshida

Evidence for direct oxidation of lactate by cardiac and skeletal muscle mitochondria has only been shown in one study to date (Brooks et al. 1999b). Since then, this concept has been criticized on theoretical and experimental grounds (Sahlin et al. 2002) and on the basis that malate rather than lactate oxidation was measured (Ponsot et al. 2005). Moreover, evidence for direct mitochondrial oxidation of lactate has failed to garner support from at least five other laboratories (Popinigis et al. 1991; Rasmussen et al. 2002; Sahlin et al. 2002; Willis et al. 2003; Ponsot et al. 2005), as well as ours (Yoshida et al. 2007). Thus, for almost a decade this hypothesis has failed to meet the basic scientific test of independent verification.

The present critique of our work mirrors a previous critique (Brooks, 2002) of studies that also failed to observe an intracellular lactate shuttle (Rasmussen et al. 2002; Sahlin et al. 2002). The central point in all critiques (foregoing letter and Brooks, 2002; Hashimoto et al. 2005) is that muscle homogenization used by us, and others, disrupts mitochondria and results in the loss of intramitochondrial constituents, particularly LDH. Brooks feels that this does not apply to his slightly different muscle homogenization method. Despite the often repeated assertions about this matter (foregoing letter and Brooks, 2002; Hashimoto et al. 2005), LDH contamination may well account for the apparent presence of LDH in mitochondria (Brooks et al. 1999b). Indeed, using only a poorly expressed plasmalemmal protein (GLUT1) as a marker to infer no contamination of mitochondrial preparations (Butz et al. 2004) is not appropriate.

There are compelling data to indicate that Brooks' central methodological critique (i.e. loss of mitochondrial LDH) is without merit. In permeabilized muscles, which have completely intact mitochondria but which contain no cytosolic LDH, there was no evidence of direct mitochondrial lactate oxidation (Ponsot et al. 2005). Thus, mitochondrial LDH does not exist. Hence, there is no LDH to be lost from within mitochondria during their preparation.

Items (a), (b), (c), (d) and (e) in the foregoing letter are based on the assumption that intramitochondrial constituents were there to be lost. The work of Ponsot et al. (2005) illustrates this assumption to be incorrect. Indeed, LDH contamination may be at the root of the so-called intracellular lactate shuttle (see discussion section entitled ‘Is direct mitochondrial lactate oxidation due to mitochondrial contamination’ in Yoshida et al. 2007). Such LDH contamination has already led to one retraction (Popinigis et al. 1991).

Point (f), concerning Fig. 7C and D in Yoshida et al. (2007) is erroneous. Specifically:

(i)we disagree with the assertion that mitochondrial FAT/CD36 represent false positive results. We (Campbell et al. 2004; Bezaire et al. 2006; Holloway et al. 2006, 2007) and others (Schenk & Horowitz, 2006) have provided substantial novel evidence that FAT/CD36 is a key regulator, along with CPT1, of mitochondrial fatty acid oxidation.

(ii)the fact that our MCT data disagree with that of Brooks' observations may simply reflect the quality of the antibodies used, not their loss from mitochondria. Indeed, early reports from their laboratory have demonstrated considerable detection problems with both MCT1 and -4 (i.e. compare MCT1 and -4 fibre-type differences in McCullagh et al. 1996, 1997; Bonen et al. 2000a,b with those of McClelland & Brooks, 2002). Their recent work, particularly that with our MCT4 antibody (Hashimoto et al. 2005), now supports our previous observations.

We disagree with many statements in the foregoing letter.

(i)It is wrong to state (see foregoing letter), that ‘Yoshida et al. confirm that mitochondrial lactate dehydrogenase (mlDH) is essential for mitochondrial lactate oxidation’. Instead, we show that exogenous LDH provision to mitochondria, akin to cytosolic LDH, is required (see Fig. 9 in Yoshida et al. 2007), because there is no intramitochondrial LDH.

(ii)It is incorrect to state that we were unable to detect mitochondrial MCTs previously. We have long been able to do so, but chose not to publish this work until we had satisfied ourselves that mitochondrial contamination with other subcellular organelles (e.g. plasma membrane) was not a problem (see Benton et al. 2004). Mitochondrial contamination is, however, clearly evident in Brooks' studies (Brooks et al. 1999a; Dubouchaud et al. 2000).

(iii)There is no proof that co-localization of LDH, MCT1 and CD147 near mitochondria (Hashimoto et al. 2005) can be taken to mean that they are located within mitochondria. Moreover, a recent Nature report noted that (a) fluorescence microscopy is fraught with dangers and (b) ∼50% of the studies may have been performed improperly (Pearson, 2007). Despite Brooks' work (Brooks et al. 1999b), there remains great skepticism about intramitochondrial LDH (see (Yoshida et al. 2007) and references therein).

(iv)It is incorrect to compare Table 2 with another study (Stellingwerf et al. 2006). Lactate concentrations in Table 2 would have to be within mitochondria, but as we demonstrated, direct mitochondrial lactate oxidation does not occur. It is the residual extramitochondrial LDH contamination that enabled the minimal lactate oxidation, as discussed in Yoshida et al. (2007).

(v)It is incorrect to assert that Chatham et al. (2001) provide prima fascia evidence for the intracellular lactate shuttle. This study also fits another model (Gladden, 2004; and see Fig. 10 and the Discussion section entitled ‘Reconciling different views’ in Yoshida et al. 2007).

(vi)It is incorrect to state that Schurr (2006) supports the intracellular lactate shuttle. He reviews the debate surrounding the astrocyte–neuron lactate shuttle.

(vi)We did not overlook studies by Mootha et al. (Mootha et al. 2003) and Taylor et al. (Taylor et al. 2003). They only provided an index of proteins located in or near mitochondria. Only one reported some cytosolic LDH near mitochondria (Taylor et al. 2003).

In summary, Brooks has consistently mounted a highly vigorous defense of the intracellular lactate shuttle hypothesis. However, the central critique (i.e. loss of mitochondrial constituents) directed at all dissenting reports has not been supported by experimental data (Ponsot et al. 2005). It is doubtful that all laboratories, except Brooks' group, have ‘stumbled technically and conceptually’. Among six independent groups, not one has been able to observe direct mitochondrial lactate oxidation in muscle. Therefore, the robustness of the intracellular lactate hypothesis (Brooks et al. 1999b) is highly suspect (Gladden, 2007), and appears to have been based on unsuspected LDH contamination of mitochondria (Popinigis et al. 1991; Yoshida et al. 2007) and/or malate rather than lactate oxidation (Ponsot et al. 2005). Indeed, Gladden's commentary of our work states ‘In toto, these results along with previous congruent reports by other investigators may lay to rest the idea of intramitochondrial oxidation of La⁻’ (Gladden, 2007).