| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 281, Issue 42, 31222-31233, October 20, 2006
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
1
From the
Department of Biochemistry and Molecular Biology, Indiana University School of Medicine, Evansville, Indiana 47712, the Biochemistry and Molecular Biology, Indiana University School of Medicine, Indianapolis, Indiana 46202, the ¶North Shore Long Island Jewish Research Institute, Manhasset, New York 11030, the **Department of Microbiology and Immunology, Indiana University School of Medicine, Evansville, Indiana 47712, the AFLAC Center for Cancer and Blood Disorders, Emory University School of Medicine, Atlanta, Georgia 30322, and the ||Department of Surgery, State University of New York, Stony Brook, New York 11794
Received for publication, May 11, 2006 , and in revised form, August 23, 2006.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
Deciphering the mode of asparaginase action in anti-cancer therapy is complicated by the observation that asparaginase preparations also possess glutaminase activity, rapidly reducing circulating concentrations of glutamine in the plasma of patients (11). Glutamine deamination values are highly correlated with serum asparaginase activity (12). Several reports suggest that the cytotoxic effects of asparaginase are related to reductions in cellular glutamine (11, 13, 14). Preclinical testing of a novel, glutaminase-free form of asparaginase (isolated from the Vibrio succinogenes microbe, subsequently reclassified as Wolinella succinogenes) found this enzyme to retain anti-lymphoma properties while lacking hepatotoxicity and immunosuppressive actions (7, 15–18). These studies suggest that depletion of glutamine may be one reason for asparaginase toxicity.
Amino acid deprivation is known to inhibit growth and protein synthesis via increased phosphorylation of the 
subunit of the translation factor, eukaryotic initiation factor 2 (eIF2). In response to nutrient depletion, phosphorylation of eIF2 by the GCN2 kinase reduces global translation, allowing cells to conserve resources and initiate a reconfiguration of gene expression to either alleviate cellular conditions of stress or trigger programmed cell death (19, 20). Transcription factors whose expression increases in response to amino acid starvation includes asparagine synthetase and the CAAT/enhancer-binding protein homologous protein (CHOP, also known as growth arrest DNA damage-inducible gene 153, GADD153) (21–24). Microarray analysis of mRNA from leukemic cells treated with asparaginase report induction of both asparagine synthetase and CHOP. Expression of asparagine synthetase was considered experimentally for many years both a marker of asparaginase efficacy and developing drug resistance, but recent work has challenged the concept of using asparagine synthetase as a diagnostic indicator (25–27). There are no data examining asparagine synthetase or CHOP levels in non-tumor tissues of whole organisms treated with asparaginase.
Dietary amino acid deprivation also mediates translational control via the mammalian target of rapamycin (mTOR) kinase. These events include decreased phosphorylation of eukaryotic initiation factor 4E-binding protein (4E-BP1) and ribosomal protein S6 kinase (S6K1) that function to stimulate mRNA translation and regulate cell size (28) (917). Asparaginase has been suggested to inhibit S6K1 phosphorylation in leukemic cell lines (29). Yet, although the effects of rapamycin in the whole animal are well documented (30–32), there are no studies reporting the effects of asparaginase on mTOR signaling in vivo.
Although the inhibition of protein synthesis has been implicated as the basis for altered function of several non-target tissues following asparaginase treatment, mechanisms by which asparaginase reduces protein synthesis remain unknown. Herein we report comparative effects of two forms of asparaginase (differing in the ability to degrade glutamine) on the regulation of protein synthesis in tissues that are associated with secondary complications, namely the liver, pancreas, and spleen. This study is the first to show tissue-specific changes in protein synthesis, eIF2 phosphorylation, and mTOR signaling following asparaginase treatment in the whole animal. We also show that prevention of glutamine depletion by the use of a novel glutaminase-free enzyme prevents some but not all measured stress responses elicited by a contemporary asparaginase. Specifically, both asparagine synthetase and CHOP are induced by both enzymes, in an apparent eIF2 phosphorylation-independent manner. Finally, unlike dietary amino acid deprivation, phosphorylation of eIF2 following asparaginase is initially but not solely dependent on GCN2. These data provide important insight into the cellular stress mechanisms elicited by asparaginase and support testing of Wolinella asparaginase in chemotherapeutic regimens to treat ALL.
|
EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
Animals—The following study protocol was approved by the Institutional Care and Use Committee at the Indiana University School of Medicine-Evansville. Six to 8 week old male and female C57BL/6J mice (also referred to as GCN2+/+) and GCN2 null mice (also referred to as GCN2–/–; backcrossed onto the C57BL/6J genetic background 8–10 generations (35)) were maintained on a 12-h light:dark cycle and provided free access to commercial rodent chow (PMI International, Brentwood, MO) and tap water prior to the experiment. On the day of the experiment, mice were given a single injection of phosphate-buffered saline that contained an enzyme activity of 0, 1.5, or 3.0 IU of L-asparaginase (from E. coli or Wolinella) per gram body weight. It is reported that mice are resistant to L-asparaginase immunosuppression and hepatotoxicity up to 2.0 IU/g (36), so these doses were chosen to represent enzyme activities below and above this threshold. All mice were injected in the morning and allowed free access to food and water throughout the day. At the indicated times (usually 6 h later but 15 min and 1 h were also studied), mice were euthanized by decapitation, and the liver, pancreas, and spleen were dissected carefully, rinsed in ice-cold phosphate-buffered saline, weighed, and frozen immediately in liquid nitrogen. Trunk blood was collected to obtain serum for analysis of amino acid profiles. In a separate experiment, C57BL/6J mice were given a single intraperitoneal injection of 200 mg/kg ammonium chloride or 0 or 3 IU/g BW E. coli asparaginase and euthanized 30 min later. Blood was collected to measure serum ammonia concentrations, and the liver was examined for phosphorylation of eIF2 (see below for description of analyses).
Amino Acid Profiles—Serum was obtained by centrifugation of clotted blood and then snap-frozen and stored at –20 °C. Serum samples were sent to the Indiana University School of Medicine Quantitative Amino Acid Core Facility (under the direction of Dr. Edward Liechty) for the determination of amino acid profiles by the ninhydrin method, using standard ion-exchange chromatography with a Beckman 6300 automated amino acid analyzer.
Tissue Amino Acid Concentrations—Frozen powdered tissue was incubated in 3% perchloric acid on ice and then centrifuged for 10 min at 10,000 x g. The collected supernatant was applied onto a cation exchange column (Dowex AG 50W-X8 resin, 100–200 mesh hydrogen form, Bio-Rad), eluted with ammonium hydroxide, and dried to completion using a Savant evaporator. Free amino acid analysis was performed using a Waters 2690 high-performance liquid chromatographic separation module (Waters, Milford, MA) and 474 fluorescence detector with pre-column derivatization (ortho-phthalaldehyde/3-mercaptopropionic acid). The internal standard used was methionine sulfone. The separation column used was a Synergi 4µ MAX-pro, 250 x 4.6 mm (Phenomenex, Torrance, CA) heated at 40 °C.
Plasma Ammonia—Ammonia concentrations were measured using a commercial enzymatic method (Diagnostic Chemicals Limited, Oxford, CT). Duplicate 100-µl samples were added to 3.0 ml of a buffer mixture that contained (in mM) 2.2 ADP, 3.5 -ketoglutarate, and 0.2 NADPH. Subsequent addition of 20 µl of glutamate dehydrogenase (1200 IU/ml) catalyzes the formation of glutamate from glutamine and oxidation of NADPH to NADP+ at room temperature. Absorbance at 340 nm was measured before the addition of enzyme (t = 0) and following reaction completion (t = 6 min). The change in absorbance was used to calculate plasma ammonia concentration using ammonia standards and blanks assayed alongside samples.
Quantitative Reverse Transcription-PCR of Asparagine Synthetase—Total RNA was extracted from frozen tissue using TriReagent (Molecular Research Center, Inc., Cincinnati, OH) followed by DNase treatment (VersaGene DNase kit, Gentra Systems). The A260/280 ratio was between 1.8 and 2.0 following RNA clean-up. mRNA expression was determined by quantitative real-time PCR using TaqMan chemistry. The relative expression levels of asparagine synthetase mRNA was determined using the Eurogentec RTqPCR mastermix (Eurogentec, Belgium) and ABI PRISM 7700 Sequence Detection System. The PCR mix contained 1x master mix and 0.125 µl of Euroscript+RT and RNase Inhibitor (reverse transcription, 0.125 unit/µl and RNase inhibitor, 0.05 unit/µl). Asparagine synthetase and 
-actin primers and probes were added at final concentrations of 200 nM and 100 nM, respectively. The primers used for asparagine synthetase were: forward, 5'-GGAGAGGGGTCAGATGAACTT-3' and reverse, 5'-CTCCTCCTCGGCCTTCTC-3'. 1 µg of total RNA was used per reaction in a 25-µl reaction volume. All samples were run in duplicates. The thermal cycler conditions were 48 °C for 30 min, 95 °C for 10 min, and 45 cycles of 95 °C for 0.15 min and 60 °C for 1 min. Data were analyzed using Sequence Detection System software version 1.9.1. Results were obtained as Ct (threshold cycle) values. Ct is inversely proportional to the starting template copy number. -Actin was used as reference gene and normalizer. Relative expression of asparagine synthetase gene in five samples treated with 3 units of ASNase was calculated in comparison to untreated three control samples using the 
Ct method (User Bulletin no. 2, Applied Biosystems Inc.). Results were expressed as -fold change with respect to the experimental control. The data were analyzed by Student's t test with one-tail distribution and two-sample unequal variance. Statistical significance was set at p < 0.05.
Protein Synthesis—Fifteen minutes before they were euthanized, mice were intraperitoneally injected with a bolus solution of L-phenylalanine (250 mg/kg), comprising 60% unlabeled L-phenylalanine (Sigma-Aldrich) and 40% phenylalanine labeled with deuterium (L-[2H5]phenylalanine, Cambridge Isotopes, Andover, MA) for the measurement of tissue protein synthesis as previously described (37, 38). Each tissue sample was processed to determine the enrichment of labeled phenylalanine into liver protein as previously described (39). Tissue L-[2H5]phenylalanine enrichments were measured by monitoring the ions at m/z 336 and 341 of the tertiary butyldimethylsilyl derivative on a model MD800 GC-MS (Fisons Instruments) operated under electron impact (40). The intraperitoneal route of injection to ensure constant precursor enrichment has been previously validated for use in visceral tissues (35, 38, 41).
Tissue Preparation for Immunoblot Analysis—Tissues were homogenized as previously described (35) using a glass-on-glass homogenizer in 7 volumes of buffer A consisting of (in mM)20 N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (pH 7.4), 100 KCl, 0.2 EDTA, 2 EGTA, 1 dithiothreitol, 50 sodium fluoride, 50 -glycerophosphate, 0.1 phenylmethylsulfonyl fluoride, 1 benzamidine, and 0.5 sodium orthovanadate. The homogenates were immediately centrifuged at 10,000 x g for 10 min at 4 °C for analysis of protein expression and phosphorylation state as described below.
Phosphorylation of eIF2—Phosphorylation of eIF2 was assessed using an antibody that recognizes the protein only when it is phosphorylated at Ser-51 (Cell Signaling Technology, Inc., Beverly, MA). Results were normalized for total eIF2 with an antibody that recognizes the protein irrespective of phosphorylation state (Santa Cruz Biotechnology, Santa Cruz, CA).
PEK/PERK Activation—Frozen samples were homogenized in a detergent-containing buffer (50 mM Tris-HCl (pH 7.9), 150 mM NaCl, 1% Nonidet P-40, 0.1% SDS, 50 mM NaF, 37 mM -glycerolphosphate, and 1 mM dithiothreitol) supplemented with protease inhibitors (100 µM of phenylmethylsulfonyl fluoride, 0.15 µM aprotinin, 1 µM leupeptin, and 1 µM pepstatin) and centrifuged for 10 min at 10,000 x g. Resulting aliquots of equal protein concentration were added to SDS sample buffer and loaded onto polyacrylamide gels for resolution by SDS-PAGE. PEK activation was measured as a shift upwards in migration as detected by immunoblot analysis, using a polyclonal antibody prepared against PEK (42).
Phosphorylation of 4E-BP1—Phosphorylation of 4E-BP1 was measured as a change in migration during SDS-PAGE as detected by immunoblot analysis as described previously (43). Briefly, an aliquot of the 10,000 x g supernatant was boiled for 10 min and centrifuged at 10,000 x g for 30 min at 4 °C. The resultant supernatant was added to SDS sample buffer and subjected to protein immunoblot analysis using a polyclonal 4E-BP1 antibody (Santa Cruz Biotechnology, Santa Cruz, CA).
Phosphorylation of S6K1—Phosphorylation of S6K1 was measured as a decrease in mobility during SDS-PAGE as described previously (43). Briefly, an aliquot of the 10,000 x g supernatant was added to SDS sample buffer. Immunoblot analysis was then performed using a polyclonal S6K1 antibody (Santa Cruz Biotechnology).
Protein Expression of CHOP/GADD153—10,000 x g supernatant aliquots of equal protein concentration were added to SDS sample buffer and loaded onto 12.5% polyacrylamide gels for resolution by SDS-PAGE. Resolved proteins were electrotransferred onto polyvinylidene difluoride membranes, and immunoblot analysis was then performed using a monoclonal anti-CHOP/GADD153 antibody (Santa Cruz Biotechnology).
Statistics—All data were analyzed by the STATISTICA statistical software package for the Macintosh, volume II (StatSoft, Tulsa, OK). Data were analyzed using two-way analysis of variance to assess main effects, with drug and dose as the independent variables. When a significant overall effect was detected, differences among treatment groups were assessed with Duncan's Multiple Range post-hoc test. The level of significance was set at p < 0.05 for all statistical tests.
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
Glutamine Is Reduced in Liver and Spleen following a Single Injection of L-Asparaginase Derived from E. coli but Not W. succinogenes—As expected, both forms of asparaginase reduced the concentration of asparagine in the blood below the detectable levels of our instrument (below 5 µmol/liter) at both enzyme doses (Fig. 1A). As a result, the amount of aspartic acid increased substantially in the circulation of all asparaginase-treated mice irrespective of dose or form of enzyme (Fig. 1B). Asparaginase from E. coli harbors glutaminase activity, which is reported to catalyze the breakdown of glutamine at 3–5% the rate of asparagine hydrolysis (45). Consistent with this idea, treatment with E. coli asparaginase incrementally reduced circulating glutamine concentrations, resulting in 
3-fold and 6-fold increases in serum glutamic acid in the 1.5 and 3.0 IU/g groups, respectively (Fig. 1, C and D). In contrast, treatment with Wolinella asparaginase did not reduce serum glutamine, although a significant increase in circulating glutamate was noted at the higher dose of enzyme. An explanation for this result is that even though glutamine is not being enzymatically cleaved by Wolinella, its turnover in the body is increased in response to simple asparagine depletion. By way of the asparagine synthetase reaction, glutamine would provide the NH3 group to make more asparagine, leading to a greater production of glutamate in the blood. Although no changes in urea were noted across enzyme doses, circulating ammonia increased 2-fold following administration of the higher dose of E. coli asparaginase (Fig. 2, E and F). In contrast, treatment with Wolinella did not significantly alter ammonia concentrations. There were no changes in serum concentrations of any of the essential or other non-essential amino acids by either enzyme (data not shown).
To determine if the reduction in circulating glutamine by E. coli asparaginase was reflected at the tissue level, we examined intracellular concentrations of asparagine, glutamine, aspartate, and glutamate in the liver, pancreas, and spleen over time (Table 1). Intracellular concentrations of aspartate and glutamate did not change over time and remained similar to control values, which were (nanomoles/g tissue) as follows: aspartate, 354 ± 57 (liver), 1600 ± 201 (spleen), and 253 ± 70 (pancreas); and glutamate, 2077 ± 386 (liver), 2809 ± 484 (spleen), and 2295 ± 169 (pancreas). Asparagine concentrations fell over time in both liver and spleen p = 0.08), while glutamine concentrations were maintained or increased for a time before falling below control values at 6 h post-injection. Interestingly, both asparagine and glutamine concentrations remained stable in the pancreas.
|
|
Phosphorylation of eIF2 at serine 51 was calculated as a ratio of phospho-Ser-51 over the total amount of eIF2 detected in the sample. Following treatment with E. coli asparaginase, p-eIF2 in the liver increased slightly at the 1.5 IU/g dose and over 2-fold (p < 0.05) at the 3.0 IU/g dose (Fig. 3A). In spleen, p-eIF2 again increased stepwise, reaching a 2-fold increase at the 3.0 IU/g dose (Fig. 3C). In contrast, there were no changes in p-eIF2 in liver or spleen following administration of Wolinella asparaginase (Fig. 3, D and F) and neither enzyme altered the phosphorylation of eIF2 in pancreas (Fig. 3, B and E). Of the tissues examined from untreated mice, pancreas had the highest basal levels of p-eIF2. This is consistent with our work and that of others demonstrating a high abundance of the eIF2 kinase, PEK (also referred to as the PKR-like ER stress-induced kinase, or PERK) in pancreas, required for its endocrine and exocrine functions (42, 46, 47).
Increased Circulating Ammonia Does Not Increase eIF2 Phosphorylation—In addition to altering the availability of specific amino acids, current preparations of asparaginase are known to increase ammonia concentrations, resulting in complications related to hyperammonemia. Indeed, circulating ammonia was found to be significantly increased following injection of E. coli asparaginase. To determine the contribution of this factor to the stress response in liver, mice were injected with ammonium chloride at a neurotoxic dose and compared with mice treated with the highest dose of E. coli asparaginase. As shown in Fig. 4, whereas ammonium chloride treatment was effective at raising plasma ammonia above that of E. coli asparaginase, only asparaginase increased eIF2 phosphorylation in liver. Thus, while increased ammonia certainly causes some untoward events in vivo (for example, acute encephalopathy related to hyperammonemia is reported in patients (48)), the current data point to amino acid deprivation as a primary factor in inducing the reduction in liver protein synthesis by asparaginase.
|
kinase activated in response to asparaginase, we treated both GCN2+/+ and GCN2–/– mice with E. coli asparaginase and measured p-eIF2 in the liver at various times (Fig. 5, bottom two rows). In contrast to our findings with dietary amino acid deprivation, we found that only very early time points following asparaginase treatment demonstrated GCN2 dependence. Later time points demonstrated a secondary means to induce p-eIF2. An accumulation of misfolded protein in the endoplasmic reticulum due to environmental stress can activate PEK/PERK that together with other stress sensors, such as IRE1 and ATF6, activate the unfolded protein response (also known as ER stress response). A recent study reports both GCN2 and PEK/PERK contributes to eIF2 phosphorylation after activation of the unfolded protein response (49). To determine if the auxiliary eIF2 kinase PEK elicits translational control in response to the longer asparaginase treatments, we assayed for PEK activation in the liver of both GCN2+/+ and GCN2–/– mice following injection with E. coli asparaginase. Activation of PEK was noted in mouse embryonic fibroblasts treated with thapsigargin as an upward shift signifying retarded electrophoretic mobility due to increased PEK phosphorylation that accompanies its activation (50). In contrast, no PEK activation was noted in any liver samples at any time point.
A subsequent effort to examine activation of the third ubiquitously expressed eIF2 kinase PKR was unsuccessful, because we were unable to detect PKR in liver preparations (data not shown). PKR functions in cellular defense against viral infection. An important feature of this defense system is induced PKR transcriptional expression by interferon, and our inability to measure PKR may reflect the low levels of this eIF2 kinase in liver in the absence of interferon. Regulation of protein phosphatases has also been suggested to affect eIF2 phosphorylation, and such regulation may be important for translational control in response to asparaginase treatment. Thus, although the primary mechanism to induce p-eIF2 is identified as GCN2-dependent, one or more secondary mechanisms yet unidentified exist in response to asparaginase. This difference separates asparaginase treatment from that of dietary deprivation for an essential amino acid, such as leucine (35).
Asparagine Depletion Induces Expression of Asparagine Synthetase and CHOP/GADD153 Independent of p-eIF2—Both amino acid starvation and asparaginase treatment induce asparagine synthetase gene expression at the level of DNA transcription by inducing specific trans-acting factors to bind specific amino acid response elements in the asparagine synthetase promoter region (51). Increased basal expression or rapid induction of asparagine synthetase has been historically considered a marker of asparaginase resistance in leukemic cell lines in vitro, but recent reports using primary cells from patients have challenged this relationship in vivo. Interestingly, there is no published information regarding asparagine synthetase expression in normal tissues following asparaginase treatment. Thus, we were interested to determine not only if either form of asparaginase induced asparagine synthetase, but if expression of asparagine synthetase correlated with the phosphorylation state of eIF2. As seen in Fig. 6, both forms of asparaginase were equally effective at increasing asparagine synthetase gene expression in liver. Furthermore, asparagine synthetase mRNA expression was elevated by E. coli asparaginase in pancreas, a tissue that did not demonstrate increased p-eIF2. Thus, the asparagine synthetase response to asparaginase does not require eIF2 phosphorylation.
Another marker of cellular stress is increased expression of CHOP/GADD153. CHOP protein expression is induced in response to a range of cellular insults, including the build-up of misfolded proteins in the ER and also amino acid limitation (52, 53). Glutamine regulates the expression of CHOP (24, 54), and CHOP expression is increased following increased eIF2 phosphorylation (20), so it was hypothesized that expression of this protein might better reflect the differential stress response between the two asparaginase enzymes as compared with asparagine synthetase. Interestingly, CHOP was also induced by both enzymes, although induction by Wolinella occurred only at the higher dose (Fig. 6). Thus, although asparaginase induces both asparagine synthetase and CHOP, these responses are not ascribed to simply glutamine depletion nor p-eIF2.
|
|
-form) by the density of the total protein present in the sample. For S6K1, phosphorylation was calculated similarly but as a ratio of the density of the upper bands (+-forms) over the total density of all bands present. Following administration of 3.0 IU/g E. coli asparaginase, phosphorylation of 4E-BP1 and S6K1 were both decreased 41% in liver as compared with control (p < 0.05, Figs. 7A and 8A). In comparison, Wolinella asparaginase did not alter the phosphorylation of either protein in liver (Figs. 7D and 8D). In pancreas, phosphorylation of 4E-BP1 and S6K1 was reduced to a similar extent by 1.5 and 3.0 IU/g of E. coli asparaginase (p < 0.05, Figs. 7B and 8B). In contrast, Wolinella asparaginase treatment of pancreas resulted in significant hyperphosphorylation of both 4E-BP1 and S6K1 at the higher dose (Figs. 7E and 8E). There were no differences in 4E-BP1 and S6K1 phosphorylation in spleen following administration of either enzyme (Figs. 7C, 7F, 8C, and 8F). Thus, signaling downstream of mTOR in liver and pancreas is repressed by E. coli asparaginase and not by Wolinella. Finally, to determine if reduced phosphorylation of 4E-BP1 reflects decreased formation of the eIF4F mRNA cap-binding complex, we examined the binding of eIF4G to eIF4E in the liver and pancreas of mice treated with 0 versus 3 IU/g BW E. coli asparaginase. It was found that the binding of eIF4G to eIF4E decreased in the liver but not pancreas (supplemental Fig. 1S). These findings are consistent with the work of Anand and Gruppuso (55), which show that formation of the eIF4G·eIF4E complex during refeeding is rapamycin-resistant; i.e. is not dependent on signaling via mTOR. Thus, the 4E-BP1 phosphorylation pattern does not necessarily reflect formation of the mRNA cap-binding complex.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
|
asparagine + glutamate + AMP + PPi. An increase in asparagine synthetase gene expression alone cannot effectively produce asparagine under conditions of reduced glutamine. This may be why recent studies show poor correlation of asparagine synthetase gene expression with asparaginase efficacy or resistance (25–27). Based on our data and the data of others (56), the regulation of glutamine supply is an important factor to consider in understanding both the cytotoxic and anti-leukemic effects of asparaginase. Glutamine is produced from glutamate by way of the glutamine synthetase reaction. Aslanian and Kilberg (57) report glutamine synthetase activity to be induced in asparaginase-resistant cells, and a recent report shows that inhibiting glutamine synthetase kills leukemic cells resistant to asparaginase (58). However, glutamine depletion can also be dangerous for the patient, for Acinetobacter glutaminase-asparaginase, an enzyme that depletes glutamine to a greater extent than E. coli asparaginase, was evaluated in patients with leukemia during the 1980s and found to be highly toxic, halting clinical trials (59). Glutamine is classified as "conditionally essential," meaning it is required in the diet under conditions of hypermetabolic stress such as post-surgery or burn injury and also in patients nourished by total parenteral nutrition. Glutamine has vital roles in interorgan nitrogen handling, gastrointestinal health, and immune cell responsiveness, and is functionally indispensable for protein synthesis (60). This last point is particularly relevant to the action of asparaginase in liver, for treatment of rat hepatocytes with E. coli asparaginase reduces protein synthesis and cell growth only when cellular levels of glutamine fall (14, 61). Our current data demonstrating that intracellular levels of glutamine fall in the liver and spleen support this concept. A decrease in the synthesis of important plasma proteins by the liver is suggested as a root cause for thromboembolism and cerebrovascular events during asparaginase therapy (11). The concept of adding glutamine to the diet of patients treated with asparaginase to prevent these complications is unstudied. Alternatively, the use of a form of asparaginase that preserves glutamine status may serve to preserve healthy organ function while retaining tumor killing properties.
Durden, Distasio, and coworkers (18, 62) were the first to demonstrate that asparaginase from Wolinella (then classified as Vibrio) possesses anti-lymphoma activities while lacking hepatotoxicity and maintaining immune responses in the spleen. Other studies comparing Wolinella and E. coli asparaginases in human pancreatic carcinoma cells report that both enzymes reduce cell growth, whereas only E. coli asparaginase inhibits protein synthesis (63). Taken together with the current results, E. coli asparaginase facilitates a broad spectrum of cytotoxic effects due in part to a decrease in glutamine status. In this regard, Wolinella asparaginase may be a safer alternative to contemporary asparaginases, resulting in fewer secondary complications linked to glutamine depletion. It should be noted that Wolinella asparaginase is currently under development through the NIH/NCI-RAID Developmental Therapeutics program but has yet to be administered to patients. Because it remains unknown if glutaminase activity is required for antitumor activity in human ALL, the current results support further development and testing of Wolinella asparaginase in this regard.
|
A striking difference between asparaginase and dietary amino acid deprivation is the extent to which p-eIF2 is GCN2-dependent. Feeding mice a diet lacking in a single essential amino acid increases p-eIF2 in a strictly GCN2-dependent fashion (43, 67), and mice deleted for GCN2 are unable to adapt to a leucine-devoid diet, resulting in increased morbidity and mortality (35). In contrast to dietary leucine deprivation, asparaginase activates but does not require GCN2 for stimulating p-eIF2. Auxiliary means of initiating p-eIF2 by a kinase are activated during oxidative or ER stress, with GCN2 identified as the secondary kinase (49). Furthermore, in GCN2–/– mouse embryonic fibroblasts switched to leucine-free media, at 6-h incubation time there is GCN2-independent p-eIF2, followed by induction of ATF3 at 10 h, suggestive of the presence of a secondary kinase to phosphorylate eIF2 during amino acid depletion (68). Nevertheless, this does not appear to be the case in the current study with respect to asparaginase, for activation of PEK did not occur. Although activation of PKR cannot yet be ruled out, the most likely secondary means of increasing p-eIF2 by asparaginase involves repressing eIF2 phosphatase activity. Reduced eIF2 phosphatase activity is documented in livers perfused with histidinol (69). The mechanism for decreased eIF2 phosphatase activity is unclear but may involve the constitutive repressor of eIF2 phosphorylation (CReP), a regulatory subunit of the type 1 protein phosphatase complex. CreP knockdown by RNA interference activates p-eIF2 and protects cells against ER stress (70). It is unknown how this or other regulatory subunits contribute to the cellular response to nutrient deprivation.
Terminal cell stress directed toward apoptosis is reflected by induction of the transcription factor, CHOP/GADD153. CHOP is induced by multiple cell stressors, including amino acid deprivation (71). CHOP is also increased in human ALL cell lines treated with E. coli asparaginase (72). Increased promoter activity of CHOP is reportedly dependent on eIF2 kinase signaling, for in GCN2–/– mouse embryonic stem cells, p-eIF2 and induction of CHOP are completely blocked in response to leucine starvation (20). So it was a bit surprising to find that Wolinella asparaginase induced CHOP independent of p-eIF2. However, a close examination of the literature reveals that our results are remarkably similar to those of Fafournoux's group, which showed that in cells incubated in media containing low (35 and 70 µM) leucine concentrations, CHOP mRNA expression is increased despite no significant reduction in general protein synthesis (73). In that study, the authors conclude that the inhibition of protein synthesis is not responsible for induction of CHOP by amino acid limitation. This finding is important, for amino acid limitation versus complete removal from the media is a more physiological scenario and closer comparison to the whole animal. Furthermore, nuclear run-on assays and mRNA decay studies suggest that a primary mechanism leading to increased CHOP mRNA levels in response to amino acid deprivation is not transcriptional, but rather via marked stabilization of the mRNA (24, 73). It is unknown to what extent mRNA stability controls induction of asparagine synthetase in response to asparaginase in vivo, but this may be an important mechanism to consider as asparagine levels in somatic cells do not fall to zero as they do in cells in culture. Collectively, these data suggest that the regulation of CHOP and asparagine synthetase by asparaginase is more complex than simply transcriptional induction directed by p-eIF2.
|
Addition of asparaginase to human leukemic cells (B and T lineage) inhibits phosphorylation of 4E-BP1 and S6K1 but not other growth-related serine/threonine kinases such as Akt and cdk2 (29). In the current study, mTOR signaling was not repressed in spleen. Considering that the spleen is composed primarily of B and T cells undergoing maturation or differentiation, our results identify a mechanistic distinction between normal and malignant lymphocytes in response to asparaginase. On the other hand, E. coli asparaginase decreases phosphorylation of S6K1 and 4E-BP1 in liver and pancreas similar to rats fed an amino acid-devoid diet or treated with rapamycin (31, 43). Inhibitors of mTOR, including rapamycin, are currently under investigation as a single agent or in combination therapy for the treatment of hematological malignancies (75), an interesting trend considering that asparaginase functions on some level as an mTOR inhibitor. It is unknown how asparaginase represses mTOR or if combining asparaginase with rapamycin would enhance killing of leukemic lymphoblasts or only increase general cytotoxicity. Further exploration in this area is needed to resolve these questions.
One of the more common side effects to asparaginase therapy is acute pancreatitis. The pathogenesis of acute pancreatitis in patients receiving asparaginase is not understood. Rodent models of acute pancreatitis demonstrate p-eIF2 alongside global down-regulation of protein synthesis (76). In the current study, intracellular concentrations of both asparagine and glutamine remained stable over time, providing a reason why E. coli asparaginase did not alter protein synthesis or p-eIF2 in pancreas. In the pancreas, fasting and food intake does not alter p-eIF2 but instead regulates the mTOR pathway (77). Thus, the pancreas responds to E. coli asparaginase similar to that of nutrient deprivation. Interestingly, treatment with Wolinella asparaginase paradoxically increased phosphorylation of S6K1 and 4E-BP1. The reason for this difference is unknown, but we speculate that this effect may be due in part to food intake. Preliminary measurements indicate that E. coli asparaginase slightly inhibits food intake over 6–12 h, whereas Wolinella asparaginase does not (data not shown). Thus, mice treated with Wolinella may simply have ingested chow near the point of euthanasia. It is worth mentioning that, even though mice treated with E. coli asparaginase do not stop eating, any reduction in food intake as a result of enzyme treatment may serve to exacerbate effects on nutrient-responsive signaling pathways over time. This point is one we are further examining.
In conclusion, E. coli asparaginase reduces cellular growth and function by repressing multiple signaling pathways in a tissue-specific fashion. The depletion of glutamine in addition to asparaginase produces a severe cellular stress that is not evident by asparagine depletion alone. Further investigation into other potential contributing factors, such as differential clearance of the enzymes by the kidney, the impact of exogenous nutrition, and tissue differences in the proteolytic response, amino acid biosynthesis, and gene expression events, will further help explain the disparity between the two enzymes and shed light on how to improve the safety and efficacy of this class of chemotherapeutic agents.
|
FOOTNOTES |
|---|
The on-line version of this article (available at http://www.jbc.org) contains supplemental Fig. S1.
1 To whom correspondence should be addressed: Indiana University School of Medicine-Evansville, 8600 University Blvd., Evansville, IN 47712. Tel.: 812-465-1199; Fax: 812-465-1184; E-mail: tganthon{at}iupui.edu.
2 The abbreviations used are: ALL, acute lymphoblastic leukemia; mTOR, mammalian target of rapamycin; S6K1, 70-kDa ribosomal protein S6 kinase; 4E-BP1, eukaryotic initiation factor 4E-binding protein 1; eIF2, eukaryotic initiation factor 2; GCN2, general control non-derepressible 2; ER, endoplasmic reticulum; PEK/PERK, pancreatic eIF2 kinase/PKR-like endoplasmic reticulum resident kinase, PKR, double-stranded RNA-dependent protein kinase; CHOP/GADD153, CAAT/enhancer-binding protein homologous protein/growth arrest DNA damage-inducible gene 153; BW, body weight; ASNase, asparaginase.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
