Celine Taglang1, Georgios Batsios1, Meryssa Tran1, Anne-Marie Gillespie1, Sabrina Ronen1, and Pavithra Viswanath1
1Radiology, University of California San Francisco, San Francisco, CA, United States
Synopsis
Elevated
choline phospholipid metabolism is a hallmark of cancer. Here, we show that
silencing choline kinase a (CKα), the rate-limiting enzyme in
choline phospholipid biosynthesis, abrogates total choline (tCho) production
from [2H9]-choline in patient-derived glioma models,
indicating that tCho production from [2H9]-choline
predominantly reflects CKα
activity. Importantly, we show that [2H9]-choline
metabolism to tCho serves to delineate tumor from normal brain in mice bearing
orthotopic patient-derived low-grade gliomas. Furthermore, [2H9]-choline
informs on response to therapy, at early timepoints when anatomical alterations
cannot be detected, pointing to the ability of [2H9]-choline
to assess pseudoprogression, which is a challenge in glioma imaging.
Introduction
Aberrant
choline phospholipid metabolism is a metabolic hallmark of cancer1,2.
Choline is a dietary nutrient that is metabolized to phosphocholine (PC) by
choline kinase α (CKα) and subsequently
incorporated into phosphatidylcholine, which is the primary membrane
phospholipid in mammalian cells1,2 (see
Fig. 1). Due to the high membrane
turnover associated with uncontrolled proliferation, tumor cells upregulate PC
and phosphatidylcholine biosynthesis1. This metabolic
phenotype has been leveraged for tumor imaging using 1H-magnetic
resonance spectroscopy (MRS)-based detection of total choline (tCho; composite
peak of choline, PC and glycerophosphocholine, a degradation product of
phosphatidylcholine, since these peaks cannot be resolved in vivo) in several
cancers including gliomas1,2. However,
1H-MRS detects steady-state metabolite levels and does not
interrogate metabolic pathway activity. Recent studies suggest that 2H-MRS
following administration of 2H-choline can be used to monitor production
of tCho in vivo3,4. The
goal of the current study was to determine whether tCho production from 2H-choline
reflects CKα
activity and establish the utility of 2H-choline for monitoring
glioma response to therapy in vivo.Methods
Cell
studies: Patient-derived high-grade glioblastoma (GBM1), low-grade
oligodendroglioma (BT88) and low-grade astrocytoma (BT257) cells were
maintained as previously described5-8. CKα expression was silenced by
RNA interference using a mix of Smartpool siRNAs against CKα9,10. Non-targeting
siRNA was used as control. Silencing was confirmed by measuring CKα (gene CHKA) mRNA
levels using quantitative RT-PCR11. Cells
were incubated in media containing 56mM choline
chloride-(trimethyl-d9) ([2H9]-choline) for 48h. 2H-MR
spectra were acquired from live cell suspensions on a Varian 14.1T scanner using
a 16mm 2H surface coil and a
pulse-acquire sequence (TR=260ms,
NA=2500, complex points=512, flip angle=64o, spectral width=2kHz). Data
was analyzed in Mnova. Peak integrals were corrected for saturation and normalized
to the natural abundance of semi-heavy water (HDO, 4.75 ppm) collected from a
vial containing saline.
In
vivo studies: We examined mice bearing
orthotopic BT257 tumors generated as described earlier8. Tumor-free
mice were used as healthy controls. Tumor volume was determined by T2-weighted
MRI using a 14.1T scanner equipped with a single-channel 1H volume coil
and a spin echo multi-slice sequence (TE=20ms,
TR=1200ms, field of view=30×30 mm2, matrix=256×256, slice
thickness=1mm, number of averages=4)12. For treatment
response assessment, BT257 tumor-bearing mice were treated with temozolomide
(50 mg/kg, 6 days/week, intraperitoneally). [2H9]-choline
metabolism was monitored before (day zero, D0) and after (day 7, D7)
treatment. Following injection of a bolus of [2H9]-choline
(200mg/kg) via a tail-vein catheter over 2.5 min, non-localized 2H-MR
spectra were acquired over 50min with a pulse-acquire sequence (TR=500ms,
averages=500, complex points=512, flip angle=64, spectral width=2kHz, temporal
resolution = 4min 10s). Normalized tCho levels were calculated by dividing tCho
peak integrals by pre-injection HDO.
Statistical analysis: All
results are expressed as mean±standard
deviation. Statistical significance was assessed using an unpaired two-tailed
Student’s t-test with p<0.05 considered significant.Results
[2H9]-choline
provides a readout of CKα
activity: Previous
studies have demonstrated tCho production in tumor regions in vivo
following administration of [2H9]-choline3,4.
However, since the spectral resolution of 2H-MRS does not allow
differentiation of choline, PC or glycerophosphocholine, it is not clear
whether the 2H-tCho peak reflects choline uptake or metabolism via CKα. We examined the effect of
silencing CKα on [2H9]-choline
metabolism in GBM1 and BT88 cells. First, we confirmed that CKα mRNA and activity were
significantly reduced in siCKα cells
relative to siCtrl (Fig. 2A-2B). Importantly, silencing CKα abrogated tCho production
from [2H9]-choline in both GBM1 and BT88 models (Fig. 3A-3B).
These results suggest that PC produced via CKα
activity is the predominant component of the 2H-tCho peak produced
from [2H9]-choline, consistent with previous studies
indicating that CKα is a
key rate-limiting enzyme in choline phospholipid biosynthesis in cancer1,2.
[2H9]-choline
can be used to non-invasively monitor response to therapy in patient-derived glioma
models: Next, we examined [2H9]-choline metabolism
in mice bearing orthotopic BT257 tumors and compared to tumor-free healthy
controls. As shown in the representative 2H-MR spectra and
quantification in Fig. 4A-4B, tCho labeling was significantly higher in BT257
tumors relative to normal brain. We then examined whether [2H9]-choline
metabolism is altered by treatment with temozolomide (TMZ), which is standard
of care for glioma patients13. TMZ induced tumor shrinkage as assessed by
T2-weighted MRI in BT257 tumor-bearing mice, an effect that was apparent by D15
(Fig. 5A). Importantly, [2H9]-choline flux to tCho was
reduced in BT257 tumor-bearing mice at D7 of TMZ treatment (Fig. 5B-5C), when
no difference in tumor volume was detectable (see Fig. 5A), pointing to the
potential ability of [2H9]-choline to assess early
response to therapy prior to the onset of anatomical alterations.Conclusions
To
date, non-invasive methods of assessing choline phospholipid biosynthetic
activity are missing. PET-based radiolabeled choline tracers interrogate
choline uptake while 1H-MRS monitors steady-state choline metabolism.
Our results showing that silencing CKα
abrogates tCho production from [2H9]-choline in multiple
patient-derived tumor models indicates that [2H9]-choline
tracks the activity of CKα,
which is a key rate-limiting enzyme in choline phospholipid biosynthesis. Importantly,
our studies show that [2H9]-choline provides an early
readout of response to therapy, prior to MRI-detectable volumetric alterations.
These results suggest that [2H9]-choline has the
potential to assess pseudoprogression, which is a major challenge in glioma
imaging10.Acknowledgements
This
study was supported by NIH R01CA239288, Department of Defense W81XWH201055315
and UCSF Brain Tumor Center Loglio and NICO initiatives.References
1 Cheng,
M., Bhujwalla, Z. M. & Glunde, K. Targeting Phospholipid Metabolism in
Cancer. Frontiers in oncology 6, 266, doi:10.3389/fonc.2016.00266
(2016).
2 Glunde,
K., Bhujwalla, Z. M. & Ronen, S. M. Choline metabolism in malignant
transformation. Nature reviews. Cancer
11, 835-848, doi:10.1038/nrc3162
(2011).
3 De
Feyter, H. T., Thomas, M., Ip, K., Behar, K. & de Graaf, R. Delayed mapping
of 2H-labeled choline using Deuterium Metabolic Imaging (DMI) reveals active
choline metabolism in rat glioblastoma. 2021
Annual Meeting of the International Society for Magnetic Resonance in Medicine
(2021).
4 Veltien,
A. et al. Simultaneous Recording of
the Uptake and Conversion of Glucose and Choline in Tumors by Deuterium
Metabolic Imaging. Cancers (Basel) 13, doi:10.3390/cancers13164034 (2021).
5 Mancini,
A. et al. Disruption of the beta1L
Isoform of GABP Reverses Glioblastoma Replicative Immortality in a TERT
Promoter Mutation-Dependent Manner. Cancer
cell 34, 513-528 e518 (2018).
6 Giannini,
C. et al. Patient tumor EGFR and
PDGFRA gene amplifications retained in an invasive intracranial xenograft model
of glioblastoma multiforme. Neuro-oncology
7, 164-176 (2005).
7 Kelly,
J. J. et al. Oligodendroglioma cell
lines containing t(1;19)(q10;p10). Neuro-oncology
12, 745-755,
doi:10.1093/neuonc/noq031 (2010).
8 Mazor,
T. et al. Clonal expansion and
epigenetic reprogramming following deletion or amplification of mutant IDH1. Proceedings of the National Academy of
Sciences of the United States of America 114, 10743-10748, doi:10.1073/pnas.1708914114 (2017).
9 Viswanath,
P. et al. Mutant IDH1 gliomas
downregulate phosphocholine and phosphoethanolamine synthesis in a
2-hydroxyglutarate-dependent manner. Cancer
& metabolism 6, 3, doi:10.1186/s40170-018-0178-3 (2018).
10 Viswanath,
P. et al. 2-hydroxyglutarate-mediated
autophagy of the endoplasmic reticulum leads to an unusual downregulation of
phospholipid biosynthesis in mutant IDH1 gliomas. Cancer research, doi:10.1158/0008-5472.can-17-2926 (2018).
11 Viswanath,
P. et al. Non-invasive assessment of
telomere maintenance mechanisms in brain tumors. Nature communications 12,
92, doi:10.1038/s41467-020-20312-y (2021).
12 Batsios,
G. et al. PI3K/mTOR inhibition of
IDH1 mutant glioma leads to reduced 2HG production that is associated with
increased survival. Sci Rep 9, 10521,
doi:10.1038/s41598-019-47021-x (2019).
13 Stupp,
R. et al. Radiotherapy plus
concomitant and adjuvant temozolomide for glioblastoma. The New England journal of medicine 352, 987-996, doi:10.1056/NEJMoa043330 (2005).