MCT down-regulation contributes to reduced conversion of hyperpolarized [1-13C]-pyruvate to [1-13C]-lactate in IDH1 mutant glioma cells
Pavithra Viswanath1, Jose Izquierdo-Garcia1, Chloe Najac1, Larry Cai1, Russell Pieper2, and Sabrina M Ronen1

1Radiology, University of California San Francisco, San Francisco, CA, United States, 2Neurological Surgery, University of California San Francisco, San Francisco, CA, United States

Synopsis

In this study we used hyperpolarized 13C-MRS to investigate pyruvate to lactate flux in IDH1 mutant cells. We found reduced hyperpolarized [1-13C]-lactate production from hyperpolarized [1-13C]-pyruvate in IDH1 mutant cells compared to wild-type. While there was no difference in lactate dehydrogenase A activity or NAD+/NADH, IDH1 mutant cells and patient samples showed reduced expression of monocarboxylate transporters MCT1 and MCT4. Comparison of hyperpolarized [1-13C]-lactate production between IDH1 wild-type and mutant lysates confirmed that reduced MCT expression was responsible for reduced hyperpolarized [1-13C]-lactate production. Thus, our study indicates that reduced MCT expression is a metabolic feature of the IDH1 mutation.

Introduction

Mutations in isocitrate dehydrogenase 1 (IDH1) are predominant driver mutations in low-grade gliomas, and are associated with a broad metabolic reprogramming1,2. In particular, we recently showed that PDH activity is down-regulated in IDH1 mutant glioma cells relative to wild-type cells, resulting in reduced pyruvate flux into the tricarboxylic acid (TCA) cycle3. Reduced PDH activity could also potentially increase pyruvate flux to lactate in IDH1 mutant cells. The goal of this study was, therefore, to use hyperpolarized (HP) 13C-MRS to probe the metabolism of HP [1-13C]-pyruvate to [1-13C]-lactate in IDH1 wild-type and mutant glioma cells.

Methods

We performed our studies on normal human astrocytes (NHAs) expressing IDH1 wild-type (NHAIDHwt) or mutant (NHAIDHmut) enzyme generated as described1. MRS studies were performed on a 500-MHz INOVA spectrometer (Agilent). [1-13C]-pyruvate containing 15mM OX063 trityl radical was hyperpolarized for ~1.5h using the Hypersense DNP polarizer (Oxford Instruments) and dissolved in 6mL of isotonic buffer (40mM Tris-HCl, 3mM EDTA, pH 7.8). For live cell studies we used an MR-compatible bioreactor system4. For studies of cell lysates, cells (~108) were lysed in 80mM Tris HCl, 200mM NaCl, 2mM NADH, pH 8 and lysates were placed in a 10mM NMR tube. In both cases HP pyruvate was injected to a final concentration of 5mM. Single-transient 13C spectra were acquired every 3s for 300s using 5ºpulses, 40k data points and spectral width of 20kHz. Data were analyzed using Mnova (Mestrelab) as follows. Spectra were summed and the integral for lactate normalized to the integral for pyruvate and to cell number. LDHA activity was determined using a spectrophotometric assay and Michaelis-Menten constant (Km) and maximal velocity (Vmax) calculated using Lineweaver-Burke analysis. NAD+/NADH ratio and NADH concentration were measured using a kit (Biovision). Protein expression was determined using Western blotting for MCT1 (Aviva) and MCT4 (Aviva). The Cancer Genome Atlas (TCGA) data analysis was performed as described previously5. All experiments were performed in triplicate and significance assessed using an unpaired t-Test.

Results and Discussion

A representative 13C-MR spectral array of HP [1-13C]-lactate production in the NHA model is shown in Fig. 1A. Surprisingly, our results indicated that HP [1-13C]-lactate production was significantly reduced, rather than increased, in live NHAIDHmut cells compared to NHAIDHwt (Fig. 1B&1C; 30.8±14.8%, p<0.05). Intracellular lactate concentration, which can affect hyperpolarized lactate production, was previously shown to be lower in NHAIDHmut cells compared to NHAIDHwt and could explain our findings1. Nonetheless, we wanted to determine whether any other factors could explain our data. We therefore investigated LDHA activity, NAD+/NADH ratio, NADH concentration, and monocarboxylate transporter (MCT) expression. We found no difference in LDHA activity between NHAIDHwt and NHAIDHmut cells as assessed by the Km for pyruvate (Fig. 2A) or the Vmax (Fig. 2B) for LDHA. We also detected no difference in the NAD+/NADH ratio (Fig. 2C) or cellular NADH concentration (Fig. 2D). We then examined MCT1 and MCT4 expression. Our results revealed that expression of both MCT1 (Fig. 3A&3B) and MCT4 (Fig 3C&3D) was significantly reduced in NHAIDHmut cells relative to NHAIDHwt (33.3±6.4%, p<0.05 for MCT1 and 30.4±2%, p<0.005 for MCT4). In order to confirm that reduced MCT expression was, indeed, affecting HP [1-13C]-lactate production in NHAIDHmut cells, we assessed HP [1-13C]-pyruvate metabolism in lysed NHAIDHwt and NHAIDHmut cells. We observed no difference in HP [1-13C]-lactate production between NHAIDHwt and NHAIDHmut lysates (Fig. 4A&4B), confirming the role of MCTs in explaining the observed reduction in HP [1-13C]-lactate production in NHAIDHmut cells. Finally, in order to determine whether the reduction in MCT expression in our genetically engineered model was clinically relevant, we compared MCT1 and MCT4 mRNA levels between IDH1 wild-type and mutant patient biopsy samples from the TCGA database. The mean normalized z-scores for MCT1 (Fig. 5A, 0.22 vs. -0.21, p<0.05) and MCT4 (Fig. 5B, 0.25 vs. -0.15, p<0.005) mRNA were significantly lower in mutant IDH1 glioma samples (n=218) compared to wild-type (n=68). These results indicated that our findings with regard to reduced MCT expression were clinically relevant. Previous work has also shown that in clinical biopsy samples LDHA expression is silenced (although no difference in LDHA expression was detected in our genetically engineered NHA model)6. Collectively, the findings from the current study with regard to MCT expression, and the previous findings regarding LDHA silencing, indicate that HP [1-13C]-lactate production is likely to be significantly lower in IDH1 mutant gliomas in the clinic when compared to wild-type IDH1 tumors. Furthermore, down-regulation of MCT expression is likely a significant metabolic consequence of the IDH1 mutation.

Acknowledgements

NIH R01CA172845, NIH R01CA154915, NIH R21CA161545 and the UCSF Brain Tumor Center Loglio Collective.

References

1) Izquierdo-Garcia et al., PLOS ONE, 10:e0118781, 2015. 2) Reitman et al., PNAS,108:3270-3275, 2011 3) Izquierdo-Garcia et al., Cancer Research, 75:1-11, 2015. 4) Brandes et al. Breast Cancer Research 12, R84, 2010. 5) Izquierdo-Garcia et al., PLOS ONE, 9(9): e108289, 2014. 6) Chesnelong et al., Neuro Oncol, 16: 686-95, 2014.

Figures

Fig. 1. (A) 13C-MR spectral array showing HP [1-13C]-lactate production from HP [1-13C]-pyruvate in live NHAIDHwt cells. (B) Representative summed-up spectra showing peaks for [1-13C]-lactate and [1-13C]-pyruvate in live NHAIDHwt and NHAIDHmut cells. (C) Ratio of HP [1-13C]-lactate to [1-13C]-pyruvate normalized to cell number for live NHAIDHwt and NHAIDHmut cells.

Fig. 2. (A) Km for pyruvate of LDHA from NHAIDHwt and NHAIDHmut cells (B) Vmax of LDHA from NHAIDHwt and NHAIDHmut cells. (C) NAD+/NADH ratio in NHAIDHwt and NHAIDHmut cells. (D) NADH concentration in the NHA model.

Fig. 3. Western blots for MCT1 (A) and MCT4 (C) in NHAIDHwt and NHAIDHmut cells. Quantification of MCT1 (B) and MCT4 (D) expression in NHAIDHwt and NHAIDHmut cells.

Fig. 4.(A) Representative summed up spectra showing peaks for lactate and pyruvate in lysate from NHAIDHwt and NHAIDHmut cells. (C) Quantification of HP [1-13C]-lactate to HP [1-13C]-pyruvate normalized to cell number for NHAIDHwt and NHAIDHmut lysate.

Fig. 5. Comparison of normalized expression z-scores for MCT1(A) and MCT4(B) mRNA in low-grade glioma patient samples. Boxes denote mean z-score and whiskers denote 5-95 percentile for tumors within the group (n=68 for IDH1 wild-type and n=218 for IDH1 mutant). A negative z-score represents expression value below the population mean.



Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)
4173