Malignant tumors inhabit a complex carbohydrate metabolism which differs from that of non-neoplastic cells with two main paradigms: 1) malignant cells produce large amounts of lactate even in the presence of sufficient oxygen for aerobic glycolysis [1
] 2) intermediates of the TCA are used for fatty, amino and nucleic acid synthesis [3
]. Thus, the extensive glucose uptake of cancer cells is needed not only for energy supply but also to provide the components for cellular growth and a high amount of reducing equivalents such as NADPH [3
]. To accomplish this, high levels of pyruvate are needed which can be introduced either into the TCA, converted into Acetyl-CoA or degraded to lactate by LDH [4
]. The latter results in an excess of lactate [3
]. In concordance with this theory of changes in tumor cell metabolism Koukourakis et al. showed that LDH5 was overexpressed in 36 out of 76 non-small lung cell cancers (47%) [7
]. In our cohort none of the 34 non-neoplastic tissue controls was positive for LDH5, while 89.43% of tumors were positive for LDH5. As this is highly statistically significant (p < 0.001) LDH5 can also be proposed as an immunohistochemical marker for neoplasias in the lung.
LDH is composed of two different subunits, LDH-H and LDH-M. Each LDH isoenzyme is composed of 4 subunits, thus recombination of these subunits leads to 5 different isoenzymes, LDH1 with 4 LDH-H subunits to LDH5 with 4 LDH-M subunits. These isoenzymes differ in their specific substrate binding which is highest for lactate in LDH1 and for pyruvate in LDH5, respectively [8
]. Maekawa et al. recently reported that the promoter for the LDHB gene is silenced in cancer cells via hypermethylation. This leads to restricted transcription of the LDH-M subunit eliminating transcription of the isoenzyms LDH1 to LDH4 [9
]. Upon this pathogenetical background our results with LDH5 being overexpressed in NSCLC are in agreement with the proposed tumor biology.
We observed decreasing LDH5 expression levels in differentiated carcinomas (SCC and AC) compared to LCs. No correlation between these two subgroups concerning the percentage of positive cancer cells could be detected. This suggests that the decrease in staining intensity could be due to the larger cellular volume in LCs in terms of a "dilution effect" for staining intensity.
Our cohort of lung cancers showed comparable data in terms of patients' age, disease stage and survival as published in the literature [11
]. As the patients in our cohort received primary surgery for curative treatment, small cell lung carcinomas (SCLC) were not included into the study due to the small number of available surgical specimens. At first glance, the frequency of large cell carcinomas is relatively high (27.1%). But taking into account that published data with proportions of around 8 to 10% for LCs usually include SCLC and rare entities [12
], adjusting these figures in analogy to our cohort by recalculating the percentages without SCLC gives proportions of around 22% for large cell carcinomas [12
]. This is in the range of our cohort.
Koukourakis et al. reported a link between LHD5 overexpression and survival in a set of 76 SCCs and 36 ACs [15
]. Taking the same parameters we could not find similar significant associations, even by omitting LCs and only investigating ACs and SCCs (figure ; p = 0.785 for all NSCLC, p = 0.551 for AC and SCC combined, p = 0.662 for AC, p = 0.145 for SCC). The diversity between the two studies can be explained not only by the fact that we used a different antibody but also by the investigated figures: the median percentage of stained tumor cells was calculated by 56.67% in our cohort, Koukourakis experienced a median of 80% of cytoplasmatic positive carcinoma cells. Thus, we were not able to reproduce the cited results in a large and well characterized cohort of NSCLC patients.
Figure 9 Kaplan-Meier survival curves comparing LDH5 expression according to Koukourakis et al in all NSCLC (p = 0.785) (a), AC combined with SCC (p = 0.551) (b), AC (p = 0.662) (c) and SCC (p = 0.145) (d).
In order to further analyze the impact of LDH5 expression on patient's survival a positive correlation only with advanced (mediastinal) lymphnode metastases (pN0 + pN1 vs pN2 + pN3) and a positive correlation between the expression of LDH5 and TKTL1 was apparent. The latter is overexpressed in a large portion of NSCLC [16
] and is an independent indicator for aggressive subtypes in NSCLC, as we have shown in another study. It could therefore be discussed if LDH5 overexpression does not appear as a linear function with a steady increase from little to highly aggressive cancers but rather in smaller steps which then reach a saturation level. Fantin et al. describe the necessity for upregulation of LDH in cancer cells for malignant transformation. Their results also stress the close link between lactate production and oxidative phosphorylation. This link is according to their results dependent on the upregulation of LDH but also regulated through metabolite concentrations [17
]. This also supports our data and hypotheses.
The metabolism of healthy and functionally active cells is optimized for productivity in terms of fulfilling their duty to synthesize, degrade, transport or contract. In contrast to this, malignant tumor cells do not meet these demands but only strive for cellular growth and mitosis. According to this theory high amounts of ATP as demanded within healthy cells are not only not required but act contra-productive for the synthesis of basic elements for cellular growth such as nucleic acids, proteins and fatty acids. By degradation of pyruvate to lactate the pool of reductive equivalents on the one hand and the availability of citric-acid cycle intermediates for fatty and amino acid synthesis on the other hand is raised [3
]. Our results which describe the overexpression of LDH5 in tumor cells and its correlation with TKTL1 further supports this theory of a glucose metabolism optimized for cellular growth within malignant tumors.