The authors have declared that no competing interests exist.
Conceived and designed the experiments: CSW SMR. Performed the experiments: CSW AHB PE JLIG. Analyzed the data: CSW PE JLIG. Wrote the paper: CSW SMR.
Histone deacetylase (HDAC) inhibitors have emerged as effective antineoplastic agents in the clinic. Studies from our lab and others have reported that magnetic resonance spectroscopy (MRS)-detectable phosphocholine (PC) is elevated following SAHA treatment, providing a potential noninvasive biomarker of response. Typically, elevated PC is associated with cancer while a decrease in PC accompanies response to antineoplastic treatment. The goal of this study was therefore to elucidate the underlying biochemical mechanism by which HDAC inhibition leads to elevated PC. We investigated the effect of SAHA on MCF-7 breast cancer cells using 13C MRS to monitor [1,2-13C] choline uptake and phosphorylation to PC. We found that PC synthesis was significantly higher in treated cells, representing 154±19% of control. This was within standard deviation of the increase in total PC levels detected by 31P MRS (129±7% of control). Furthermore, cellular choline kinase activity was elevated (177±31%), while cytidylyltransferase activity was unchanged. Expression of the intermediate-affinity choline transporter SLC44A1 and choline kinase α increased (144% and 161%, respectively) relative to control, as determined by mRNA microarray analysis with protein-level confirmation by Western blotting. Taken together, our findings indicate that the increase in PC levels following SAHA treatment results from its elevated synthesis. Additionally, the concentration of glycerophosphocholine (GPC) increased significantly with treatment to 210±45%. This is likely due to the upregulated expression of several phospholipase A2 (PLA2) isoforms, resulting in increased PLA2 activity (162±18%) in SAHA-treated cells. Importantly, the levels of total choline (tCho)-containing metabolites, comprised of choline, PC and GPC, are readily detectable clinically using 1H MRS. Our findings thus provide an important step in validating clinically translatable non-invasive imaging methods for follow-up diagnostics of HDAC inhibitor treatment.
The histone deacetylase (HDAC) enzymes catalyze the removal of acetyl groups from the epsilon-amine of lysine residues of histone tails. The opposing actions of HDACs and histone acetyltransferases (HATs) dictate the acetylation status of histone lysine residues, regulating gene transcription through chromatin packaging
Clinical use of therapeutic drugs is significantly enhanced when a non-invasive means of longitudinally monitoring treatment efficacy is available. Traditional imaging methods may not be adequate for rapid monitoring of molecular drug action and response, since in many cases molecular drug action results in tumor stasis rather than shrinkage
We have previously used MRS to develop and mechanistically validate biomarkers of response to emerging targeted therapies
MCF7 cells were obtained from ATCC. Unique DNA “fingerprint” identities (i.e., variable number tandem repeat PCR products) have been established for the cell line used in this study, and the identity of the cell line was confirmed in association with its use in the experiments described here.
Cells were cultured in DMEM supplemented with 10% heat-inactivated fetal bovine serum, 2 mM L-glutamine, 100 units/mL penicillin and 100 µg/mL streptomycin. For MRS labeling studies, custom DMEM without choline (UCSF Cell Culture Facility) was supplemented with [1,2-13C]choline chloride at a final concentration of 28 µM (the concentration normally present in DMEM). For all experiments, cells were harvested in their logarithmic phase of proliferation.
For HDAC inhibition, cell cultures were incubated with 10 µM SAHA (courtesy of Drs. W. Bornmann and A. Pal, University of Texas M.D. Anderson Cancer Center for initial studies, and Cyman Chemical Company for subsequent studies; findings were within experimental error), replenishing drug and medium every 24 h. The final concentration of DMSO was 1∶1000 in culture medium.
The effect of drug treatment on cell proliferation was determined using the WST-1 reagent assay (Roche). Cells were seeded in 96-well plates and treated for 4 to 48 h with 0, 5, 10, and 20 µM SAHA. After treatment, WST-1 reagent was incubated in wells for 1 h and cell viability was determined by quantification of absorbance at 440 nm using a spectrophotometer (Tecan).
The effect of SAHA on HDAC activity was determined using the Fluor de Lys fluorometric assay (Biomol), following the manufacturer’s instructions. Cells were seeded in 96-well plates and incubated with SAHA for 48 h. The Fluor de Lys substrate was added for 1.5 h, medium was transferred to adjacent wells, and cells were rinsed with PBS. Fluor de Lys developer was added to both the wells containing cells and medium. After 20 min, fluorescence was measured at 460 nm using a spectrofluorometer (Tecan).
Control and treated cells were harvested by trypsinization, washed in PBS and fixed in 95% ethanol. After fixation cells were washed and resuspended in PBS supplemented with 10 µg/ml RNase A (Sigma) and 20 µg/mL propidium iodide (Biotium). After 30 minutes of staining, cells were analyzed using a fluorescence-activated cell sorting (FACS) Calibur flow cytometer and CellQuest Pro software (BD Biosciences). Single cells were gated away from clumped cells by using forward light scattering on an FL2-width versus FL2-area dot plot. Percentages of cells in the G1, S, and G2/M phases were determined by plotting a histogram of FL2-A. Cell size distribution was determined by Coulter counter. Approximately 5×106 cells per sample were diluted into 10 mL of Isoton solution in a clean cuvette. Coulter counter readings provided cell number and cell size.
For initial MRS studies, MCF7 cells were treated for 48 h with 0, 5, 10 and 20 µM SAHA (n = 2). For mechanistic studies, MCF7 cells were treated for 48 h with 10 µM SAHA. Culture medium was replaced with [1,2-13C]choline-labeled medium 6 h prior to extraction. Cells (∼3×107) were extracted using the dual-phase extraction method, as previously described
Experiments were performed to assay the total cellular activity of choline kinase as previously described
Experiments were performed to measure the activity of phosphatidylcholine-specific CTP:phosphocholine cytidylyltransferase, as previously described
The activities of phospholipase A1 and A2 were determined using the EnzChek phospholipase A assay kits (Invitrogen) according to manufacturer’s instructions. The fluorescent cleavage product was measured (495 nm excitation, 515 nm emission) by an Infinite M200 fluorescence spectrophotometer (Tecan).
Control and treated cells were lysed using Cell Lysis Buffer (Cell Signaling). Lysates, were run on 4%–20% Tris-HCl Gels (Bio-Rad) by SDS-PAGE and electrotransferred onto PVDF membranes (Millipore). Blots were blocked and incubated with primary antibodies, anti-PLA2G4C (Abcam), anti-PLA2G10 (Abnova), anti-ChoK α (Sigma), anti-SLC44A1 (Abcam) or anti-β-Actin (Cell Signaling), then incubated with secondary antibody anti-IgG HRP-linked antibody (Cell Signaling). Immunocomplexes were visualized using ECL Western Blotting Substrate (Pierce). Quantification was performed using ImageJ (NIH) by measuring band intensities in scanned blots, and normalizing to loading control. Results are expressed as percentage of controls.
All results are expressed as mean ± SD and represent an average of three repeats unless otherwise stated. Two-tailed unpaired Student’s
Total cellular RNA was isolated from approximately 1×107 cells after 48 h of treatment with SAHA or vehicle (DMSO), using the RNeasy Mini Kit (Qiagen) according to manufacturer’s instructions. RNA quality was determined by Bioanalyzer (Agilent), considering RNA integrity number (RIN) values of ≥8.0 acceptable (most values were 9.5 or higher). Microarray hybridization was performed at the UCSF Genomics Core Laboratories using the Human Gene 1.0 ST assay (Affymetrix). Arrays were analyzed by fluorescence detection using the Agilent GeneArray Scanner (Agilent). Data acquisition was performed using the MicroArray Suite 5.0 software (Affymetrix). Microarray experiments were performed with 4 repeats of each condition. Microarray data is available through the ArrayExpress public repository [EMBL:E-MTAB-339] in compliance with standards of the Microarray Data Gene Expression Society.
(A) WST-1 assay showing anti-proliferative effects of 0, 5, 10 and 20 µM over a 48-hour treatment period; (B) Cellular PC and GPC levels following treatment with 0, 5, 10 and 20 µM SAHA at 48-hour time point.
(A) Representative 31P spectra of control (bottom) and SAHA-treated (top) MCF7 cell extracts, showing increases in PC and GPC after 48-hour treatment; (B) Representative 1H spectra (3.10–3.30 ppm) of choline-containing metabolites, highlighting increased tCho levels after 48-hour treatment.
In order to determined the underlying mechanism of this effect we chose to focus further studies on treatment with 10 µM SAHA for 48 h, a dose that led to a reduction in cell proliferation to 50±5% (P<0.001, n = 3) of control (
31P MRS demonstrated that PC increased to 129±7% of control (P = 0.006, n = 3;
To evaluate the contribution of
Representative 13C spectra of control (bottom) and SAHA-treated (top) MCF7 cell extracts labeled with [1,2-13C]choline, showing increased uptake and incorporation of labeled choline as [1,2-13C]PC.
Interestingly, there is no [1,2-13C]choline peak visible in the 13C spectrum. The concentration of labeled choline does not reach the detection limit, which suggests that choline is phosphorylated shortly after being transported into the cell. The relative rates of the two steps involved in PC synthesis can be estimated using the signal-to-noise ratio of the PC peaks (SNR values ≥20)
In order to explain the observed changes in choline-containing metabolite concentrations, we analyzed the activities of choline kinase and CTP: phosphocholine cytidylyltransferase, the enzymes directly involved in PC production and consumption respectively. The cellular activity of choline kinase was measured by 1H NMR assay in cell lysates.
(A) Simultaneous detection of PC and choline peaks in MCF7 cell cytosolic preparations over 1 hour; (B) Time courses of PC production in control and SAHA-treated cells in representative experiments, showing increased choline kinase activity with treatment.
The activity of CTP:phosphocholine cytidylyltransferase was measured in total cell lysates by 31P NMR assay. These assays revealed a rate of 10.0±2.2 fmol CDP-choline/cell per hr and no significant change in treated cells (P = 0.15).
(A) Buildup of CDP-choline peaks in MCF7 cell lysates over 1.5 hours; (B) Time courses of CDP-choline production in control and SAHA-treated cells, showing no significant change in cytidylyltransferase activity with treatment.
Fluorimetric assessment of PLA activities measured in total cell lysates revealed a 162±18% (P = 0.03) increase in PLA2, while no significant change was seen in PLA1 (115±12%; P = 0.13).
To verify that the increases in enzyme activities are due to elevated levels of enzymes, we determined the relative expressions of annotated genes associated with choline metabolism
Function | Gene Symbol | Gene Title | MA | % Control | FDR |
Choline Transporters | SLC5A7 | solute carrier family 5, member 7 | −0.13 | 91 | 0.3 |
SLC44A4 | solute carrier family 44, member 4 | 0.04 | 103 | 0.75 | |
SLC44A5 | solute carrier family 44, member 5 | 0.08 | 106 | 0.54 | |
SLC22A1 | solute carrier family 22, member 1 | 0.07 | 105 | 0.66 | |
SLC22A2 | solute carrier family 22, member 2 | 0.00 | 100 | 0.99 | |
SLC22A3 | solute carrier family 22, member 3 | −0.01 | 99 | 0.96 | |
Choline Kinases | |||||
Cytidylyltransferases | PCYT1A | phosphocholine | −0.01 | 99 | 0.96 |
cytidylyltransferase 1 alpha | |||||
PCYT1B | phosphcholine | 0.08 | 106 | 0.53 | |
cytidylyltransferase 1 beta | |||||
Choline Phosphotransferase | − |
||||
Phospholipase A | PLA1A | phospholipase A1 member A | 0.11 | 108 | 0.48 |
PLA2G1B | phospholipase A2, group IB | 0.24 | 118 | 0.074 | |
PLA2G2A | phospholipase A2, group IIA | −0.16 | 90 | 0.13 | |
PLA2G2C | phospholipase A2, group IIC | 0.05 | 104 | 0.8 | |
PLA2G2D | phospholipase A2, group IID | 0.03 | 102 | 0.83 | |
PLA2G2E | phospholipase A2, group IIE | −0.22 | 86 | 0.22 | |
PLA2G3 | phospholipase A2, group III | −0.06 | 96 | 0.74 | |
PLA2G4E | phospholipase A2, group IVE | −0.04 | 97 | 0.83 | |
PLA2G4F | phospholipase A2, group IVF | −0.22 | 86 | 0.058 | |
PLA2G5 | phospholipase A2, group V | −0.13 | 91 | 0.53 | |
PLA2G7 | phospholipase A2, group VII | −0.04 | 97 | 0.83 | |
PLA2G12A | phospholipase A2, group XIIA | 0.05 | 104 | 0.61 | |
PLA2G12B | phospholipase A2, group XIIB | 0.12 | 109 | 0.3 | |
Phospholipase D | |||||
PLD2 | phospholipase D family, member 2 | 0.11 | 108 | 0.52 | |
PLD4 | phospholipase D family, member 4 | 0.11 | 108 | 0.59 | |
PLD5 | phospholipase D family, member 5 | 0.26 | 120 | 0.054 | |
− |
|||||
Lysophospholipase | − |
||||
LYPLA2 | lysophospholipase II | 0.20 | 115 | 0.061 | |
Glycerophosphocholine | |||||
Phosphodiesterase | |||||
GDPD2 | glycerophosphodiester | −0.21 | 86 | 0.18 | |
phosphodiesterase domain 2 | |||||
− |
|||||
GDPD5 | glycerophosphodiester | 0.03 | 102 | 0.89 | |
phosphodiesterase domain 5 |
Genes for which log2<3 were considered as not expressed.
To further confirm the most significant microarray findings, Western blot analysis was used to monitor the protein levels of SLC44A1, ChoK α, PLA2G4C and PLA2G10. Our findings are illustrated in
β-actin served a loading control.
Novel therapeutic approaches are increasingly developed to target specific molecular genetic events associated with cancer. These advances will lead to more personalized cancer treatment and are expected to result in improved response and reduced toxicity. However, several challenges remain. Most significantly, many targeted therapies result in tumor stasis rather than shrinkage. Consequently, there is a critical need for non-invasive functional imaging biomarkers that confirm drug delivery and molecular drug activity at the tumor site.
Previous studies have reported increased PC levels in response to HDAC inhibition in cell lines and in tumors of several cancer types. We showed that PC levels were inversely correlated with HDAC activity in SAHA-treated PC3 cells
Our studies focused on treatment with 10 µM SAHA, a dose that led to a 50% inhibition in cell proliferation, and thus could be considered as mimicking response to treatment. This dose was consistent with a previous study on 8 breast cancer cell lines, in which the IC50 ranged from 0.81 to 18.6 µM
Drug-induced metabolic changes are the result of altered enzyme activities. Elevated PC levels in response to HDACi reflect the modulation of enzymes involved in the Kennedy pathway of phospholipid biosynthesis
Our results suggest that PC increases as the result of increased choline transport and phosphorylation. The use of 13C-labeled choline provided a means to characterize the incorporation of choline. After short-term choline incubation in SAHA-treated cells, an increase in [1,2-13C]PC was observed, indicating that the elevated levels of endogenous PC are the result of elevated PC synthesis via choline transport and choline phosphorylation. Choline can be transported into the cell by several classes of proteins: high-affinity choline transporters (SLC5A7), intermediate-affinity choline transporter-like proteins (SLC44A1, etc.), and low-affinity polyspecific organic-cation transporters (SLC22A1, etc.)
Choline phosphorylation is catalyzed by choline kinase. The activity of the enzyme depends on which subunits (ChoK α and ChoK β) form the dimer, as a recent study demonstrated that the activity of ChoK α is much greater than that of the β isoform
However, interpretation of the 13C MR results and previous studies suggest that elevated PC levels following treatment are more likely a result of increased choline uptake, at least in our system. Previous studies in MCF7 cells have reported that choline transport is the rate-limiting step in PC synthesis
Our findings with regard to ChoK α are in line with studies of belinostat that demonstrated that the increase in PC was associated with an induction of ChoK α expression, not only in cells but also in the
Increased expression of the choline transporter SLC44A1 was observed in MCF7 cells following treatment with the HSP90 inhibitor 17AAG, which led to an increase in PC both in cells and
An additional metabolic change was observed in the MR results – increased GPC. The increase in GPC following treatment could have been caused by the increased catabolic breakdown of PtdCho by PLA2 (and lysophospholipase) and/or the inhibition of GPC phosphodiesterase. Indeed, measurement of total PLA2 activity showed a significant increase, and expression of several PLA2 isoforms increased. Only a limited number of studies have looked at the modulation of PLA2 isoforms with oncogenesis or response to treatment. We have previously found that PLA2G10 was significantly elevated in MCF7 cells following HSP90 inhibition
Further studies are needed to assess our findings in other cancer types in both cell and tumor models. Nonetheless, this study was the first to systematically assess a range of enzymes that, collectively, lead to the increased PC and GPC levels associated with HDAC inhibition. Additional studies are also needed to fully understand the complex interplay between HDAC and choline metabolism. Comparison of results to other similar studies suggest that HSP90 may possibly be responsible for mediating some of the effects observed in our study, including the elevation in SLC44A1 and PLA2
In summary, this study is a first step in developing the necessary mechanistic understanding that can lead to establishing clinically applicable non-invasive imaging methods, which can help inform on target modulation and potential response to HDAC inhibitor treatment.
We thank Drs. Christopher Barker and Linda Ta of UCSF Genomic Core for expert technical assistance in performing gene arrays, and Dr. Alisha Holloway of the Biostatistics Core for performing data analysis of microarray results.