Targeting lactate-fueled respiration selectively kills hypoxic tumor cells in mice

Abstract

Tumors contain oxygenated and hypoxic regions, so the tumor cell population is heterogeneous. Hypoxic tumor cells primarily use glucose for glycolytic energy production and release lactic acid, creating a lactate gradient that mirrors the oxygen gradient in the tumor. By contrast, oxygenated tumor cells have been thought to primarily use glucose for oxidative energy production. Although lactate is generally considered a waste product, we now show that it is a prominent substrate that fuels the oxidative metabolism of oxygenated tumor cells. There is therefore a symbiosis in which glycolytic and oxidative tumor cells mutually regulate their access to energy metabolites. We identified monocarboxylate transporter 1 (MCT1) as the prominent path for lactate uptake by a human cervix squamous carcinoma cell line that preferentially utilized lactate for oxidative metabolism. Inhibiting MCT1 with alpha-cyano-4-hydroxycinnamate (CHC) or siRNA in these cells induced a switch from lactate-fueled respiration to glycolysis. A similar switch from lactate-fueled respiration to glycolysis by oxygenated tumor cells in both a mouse model of lung carcinoma and xenotransplanted human colorectal adenocarcinoma cells was observed after administration of CHC. This retarded tumor growth, as the hypoxic/glycolytic tumor cells died from glucose starvation, and rendered the remaining cells sensitive to irradiation. As MCT1 was found to be expressed by an array of primary human tumors, we suggest that MCT1 inhibition has clinical antitumor potential.

Figure 6. MCT1 inhibition delays tumor growth, induces tumor core necrosis, and decreases tumor hypoxia.

(A) MCT1 is expressed at the plasma membrane of mouse LLc cells. Representative pictures show fluorescent staining of MCT1 (red) and nuclei (blue) in cultured cells. (B) From day 0, LLc tumor growth was determined in groups of mice treated with daily CHC (25 μmol in 200 μl i.p.) or vehicle. n = 11–17. (C) Representative H&E staining of biopsies of size-matched tumors after treatments. Dashed lines delineate necrosis (n). (D) MCT1 is not expressed at the plasma membrane of mouse TLT cells. Representative pictures show fluorescent staining of MCT1 (red) and nuclei (blue) in cultured cells. (E) Similar analysis was performed as in B, but using TLT cells. n = 6. (F) Similar analysis was performed as in A, but using WiDr human colorectal adenocarcinoma cells. (G) Similar analysis was performed as in B, but using WiDr cells in athymic Balb/C mice. n = 9–15. (H) Representative histological pictures of pimonidazole staining of WiDr tumor biopsies at the end of the tumor growth delay assay are shown with H&E counterstaining. Top: Analysis of whole tumor sections revealed extensive necrosis at the core of tumors from an animal treated with the MCT1 inhibitor. Bottom: MCT1 inhibition decreased tumor hypoxia around arterioles and at the tumor-muscle interface at the tumor periphery. Scale bars: 20 μm (A, D, and F); 200 μm (C and H). Error bars represent SEM.

Figure 9. MCT1 is expressed in a variety of different human tumor cell lines and primary human tumor biopsies.

(A and B) MCT1 was detected by western blot (A) and confocal microscopy (B) in human tumor cell lines and control tissues. Note the plasma membrane expression of the lactate transporter in WiDr, FaDu, SiHa, and PC-3 cancer cells. (C) MCT1 (red) and nuclei (blue) were detected using immunofluorescence in biopsies of primary human colon, breast, and head and neck human cancers. (D) MCT1 and hypoxia were detected in cryoslices of a primary human lung cancer. The patient had received EF5 before tumor biopsy. Representative confocal microscopy pictures revealed that the staining of MCT1 (green) and of the hypoxia marker EF5 (red) did not overlap. Scale bars: 20 μm (B); 100 μm (C and D).

Figure 1. Metabolic characterization of oxidative SiHa and glycolytic WiDr tumor cell lines.

(A) pH of cell culture supernatants after reaching cell confluence on day 0 was measured with a pH meter. Inset shows typical dishes containing confluent WiDr and SiHa cells in phenol red–containing medium. **P = 0.0013 (2-way ANOVA; n = 3). (B) An equal amount (2 × 10⁷ cells/ml) of viable SiHa and WiDr cells were placed in a sealed tube containing an EPR oxygen sensor. EPR measurements revealed a significant difference (P < 0.0001) in the rate of oxygen consumption between SiHa and WiDr tumor cells (Student’s t test; n = 5–9). (C) At time 0, confluent cells received fresh medium containing glucose and FBS. Glucose utilization (solid lines, left y axis) and lactate concentration in the cell supernatant (dotted lines, right y axis) were determined enzymatically. Note the different scales of the left and right y axes. n = 4. Error bars represent the SEM and are sometimes smaller than symbols.

Figure 2. Lactate is a substrate for oxidative tumor cell metabolism.

(A and B) Enzymatic assays were used to determine glucose utilization (solid lines) and lactate concentration in the supernatant of confluent cells (dotted lines). Note the different scales of the left and right y axes in A. At time 0, cells received fresh medium containing glucose, FBS, and sodium lactate (A) or medium containing sodium lactate but no glucose and FBS (B). n = 4–5. (C and D) ρ0 SiHa and WiDr cells were produced by a chronic treatment with low-dose ethidium bromide. Then, cells were cultured in fresh medium containing glucose, FBS, and sodium lactate from time 0. Glucose utilization (solid lines, left y axes) and lactate concentration in the supernatant of confluent cells (dotted lines, right y axes) were assayed enzymatically to compare the metabolic activity of wild-type versus ρ0 SiHa cells (C) and wild-type versus ρ0 WiDr cells (D). Note the different scales of the left and right y axes. n = 3–5. (E and F) EPR measurements of tumor cells oxygen consumption by SiHa cells (E) and WiDr cells (F) in the indicated experimental culture media. Statistical analyses are presented in Table 1. Error bars represent the SEM and are sometimes smaller than symbols.

Figure 3. Oxidative tumor cells prominently express MCT1.

(A) qRT-PCR analysis showing higher MCT1 mRNA expression in oxidative SiHa tumor cells compared with glycolytic WiDr tumor cells. ***P = 0.0002 (Student’s t test; n = 8–13). (B) qRT-PCR analysis also showed that oxidative SiHa cells expressed higher levels of MCT1 compared with MCT4 mRNA. *P = 0.0116 (Student’s t test; n = 3–13). (C) MCT1, but not MCT4, is expressed at the plasma membrane of oxidative SiHa tumor cells. Representative confocal pictures show fluorescent staining of MCT1 (red), MCT4 (red), and nuclei (blue) in cultured cells. (D) SiHa and WiDr cells were loaded with the intracellular pH sensor C.SNARF1-AM. Intracellular pH was determined from fluorescence emission before and after addition of sodium lactate to the cell culture medium maintained at pH 7.3. Columns represent the difference (Δ) between the intracellular pH measured in the presence of exogenous lactate and the intracellular pH measured in the absence of exogenous lactate. ^#P = 0.0393 (Student’s t test; n = 4). (E and F) Immunohistological analyses of tumor biopsies revealed that MCT1 is expressed in both SiHa and WiDr tumors in vivo and that MCT1 expression and hypoxia are mutually exclusive. (E) Representative pictures of whole SiHa and WiDr tumor sections. (F) Representative pictures are shown with H&E counterstaining. Arrows indicate typical mutually exclusive MCT1 and pimonidazole stainings in WiDr tumors. Scale bars: 20 μm (C), 0.5 mm (E and F). Error bars represent SEM.

Figure 4. MCT1 inhibition blocks lactate-fueled tumor cell respiration.

(A and B) Enzymatic assays were used to determine glucose utilization (solid lines) and lactate concentration (dotted lines) in the supernatant of confluent cells. Note the different scales of the y axes in A. At time 0, cells received fresh medium containing glucose, FBS, and sodium lactate (A) or medium containing sodium lactate but no glucose and FBS (B). The MCT1 inhibitor CHC was added where indicated. n = 4–5. (C–E) EPR measurements of tumor cell oxygen consumption by SiHa (C and E) and WiDr cells (D) in the indicated experimental culture media. Statistical analyses of C and D are presented in Table 1. Error bars represent the SEM and are sometimes smaller than symbols.

Figure 5. MCT1 inhibition prevents lactate-fueled ATP production and the survival of oxidative tumor cells.

(A and B) ATP content was determined over time in SiHa (A) and WiDr (B) cells using a bioluminescence assay. Cells were cultured in the indicated media. *P < 0.05 compared with medium containing only glucose. ^#P < 0.05 compared with medium containing only lactate (Student’s t test; n = 4). (C) SiHa cells were transfected with a specific MCT1 shRNA or siRNA or a control vector, or were left untreated. After a 24-hour recovery period, cells were cultured in the indicated media from time 0. Cell death was determined over time using a NucleoCounter. Slope comparison: **P = 0.0014 and ***P = 0.0003 compared with wild-type cells; ^##P = 0.0013 and ^###P = 0.0003 compared with control vector transfection (Student’s t test; n = 3–7). Error bars represent the SEM and are sometimes smaller than symbols.

Figure 7. Model for therapeutic targeting of lactate-based metabolic symbiosis in tumors.

Hypoxic tumor cells depend on glucose and glycolysis to produce energy. Lactate, the end-product of glycolysis, diffuses along its concentration gradient toward blood vessels. By contrast, oxygenated tumor cells import lactate (a process mediated by MCT1 located at the cell plasma membrane) and oxidize it to produce energy. In the respiration process, lactate is a substrate preferred to glucose. As a consequence, glucose freely diffuses through the oxygenated tumor cell sheath to fuel glycolysis of distant, hypoxic tumor cells. This metabolic symbiosis can be disrupted by MCT1 inhibition. Upon MCT1 inhibition, oxidative tumor cells switch from lactate oxidation to glycolysis, thereby preventing adequate glucose delivery to glycolytic cells, which die from glucose starvation. This glycolytic switch is associated with a decrease in oxygen consumption of surviving tumor cells, which is responsible for increased tumor pO₂. MCT1 inhibition is thus a potent antitumor strategy that indirectly eradicates hypoxic/glycolytic tumor cells. GLUT, glucose transporter.

Figure 8. MCT1 inhibition radiosensitizes tumors.

LLc tumor–bearing mice were treated with CHC (daily doses of 25 μmol in 200 μl i.p.) or vehicle (200 μl) from day 0, local radiotherapy (1 dose of 6 Gy) on day 3, or combined treatments. (A) Tumor volume was determined daily. (B) Treatment efficacy was compared after determination of tumor growth delays. ***P < 0.001 compared with 6 Gy; ^###P < 0.001 compared with MCT1 inhibitor alone (1-way ANOVA; n = 6–12). Error bars represent the SEM and are sometimes smaller than symbols.