Elevation of Homocysteine and Excitatory Amino Acid
`Neurotransmitters in the CSF of Children Who Receive
`Methotrexate for the Treatment of Cancer
`
`By Charles T. Quinn, James C. Griener, Teodoro Bottiglieri, Keith Hyland, Arleen Farrow, and Barton A. Kamen
`
`Purpose: Folate deficiency, either by diet or drug, in-
`creases plsma homocysteine (Hcy). Hcy damages cerebro-
`vascular endothelium, and hyperhomocysteinem
`ia is a risk
`factor for stroke. Hcy is metabolized to excitatory amino acid
`(EAA) neurotransmitters, such as homocysteic acid (HCA) and
`cysteine sulfinic acid (CSA), which may cause seizures and
`excitotoxic neuronal death. We postulated that excess Hcy
`and EAA neurotransmitters may partly mediate methotrexate
`(MTX)-associated neurotoxicity.
`Patients and Methods:
`In this retrospective analysis, we
`used high-performance
`liquid chromatography
`(HPLC) to
`measure Hcy, HCA, and CSA in CSF from two groups of chil-
`dren: (1) a control group of patients with no MIX exposure,
`and (2) a treatment group of patients whoe had received MTX
`no more than 7 days before a scheduled lumbar puncture.
`
`Results: The treatment group had a significantly (P =
`.0255) greater concentration of Hcy in CSF (0.814 Lmol/L +
`0.215 [mean + SEM], n = 23) than the control group (0.210
`pmol/L + 0.028, n = 34). HCA and CSA were not detected
`in CSF from control patients (n = 29); however, MTX caused
`marked accumulation of CSF HCA (119.1 jcmol/L + 32.0, n
`= 16) and CSA (28.4 cmol/L + 7.7, n = 16) in the treatment
`group. Patients with neurologic toxicity at the time of lumbar
`puncture had many of the highest concentrations of Hcy, HCA,
`and CSA.
`Conclusion: These data support our hypothesis that MTX-
`associated neurotoxicity may be mediated by Hcy and excito-
`toxic neurotransmiters.
`J Clin Oncl 15:2800-2806.
`ciety of Clinical Oncolgy.
`
`1997 by American So-
`
`/ETHOTREXATE
`
`(MTX) IS A MAINSTAY of
`
`MVI therapy for children with acute lymphoblastic leu-
`
`kemia (ALL)' and is also used widely in other neoplastic
`and nonneoplastic disorders. Clinical success notwith-
`standing, MTX therapy may be associated with significant
`neurotoxicity of uncertain etiology.2 5 MTX-associated
`neurotoxicity is classified in three characteristic forms:
`acute, subacute, and late. Acute neurotoxicity occurs
`within 1 day of administration and may be characterized
`by nausea, emesis, headaches, somnolence, confusion,
`and seizures. Subacute neurotoxicity typically occurs 7
`to 9 days following MTX exposure and manifests variably
`as seizures, affective disturbances, and focal neurologic
`deficits, usually transient, including paresis, anesthesia,
`pseudobulbar palsy, and visual disturbances. Late or
`chronic MTX-associated neurotoxicity occurs weeks to
`months following therapy and involves impairment of
`higher cognitive functions.
`
`From the Departments of Pediatrics, Neurology, and Pharmacol-
`ogy, The University of Texas Southwestern Medical Center; and
`Institute of Metabolic Disease, Baylor University Medical Center,
`Dallas, TX.
`Submitted October 23, 1996; accepted April 23, 1997
`Supported by grants from the American Cancer Society, Atlanta,
`GA to B.A.K. B.A.K. is an American Cancer Society Clinical Re-
`search Professor.
`Address reprint requests to Barton A. Kamen, MD, PhD, Depart-
`ment of Pediatrics, The University of Texas Southwestern Medical
`Center, 5323 Harry Hines Blvd, Dallas, TX 75235-9063; Email
`kamen@utsw.swmed.edu.
`© 1997 by American Society of Clinical Oncology.
`0732-183X/97/1508-0012$3.00/0
`
`Although MTX has been used for nearly 50 years, the
`pathogenesis of MTX-associated neurotoxicity remains
`unclear, although it is likely multifactorial.6 It has been
`proposed that this neurotoxicity may be in part mediated
`by the interference of MTX with the recycling of biopterin
`species, which thereby impairs synthesis of the neuro-
`transmitters dopamine and serotonin.' Studies have dem-
`onstrated perturbations in biopterin and neurotransmitter
`chemistry consequent to MTX exposure temporally corre-
`lated with neurologic symptoms8 ; however, these interest-
`ing associations require further investigation to establish
`causality. MTX is also known to promote adenosine elab-
`oration and accumulation. 9 Adenosine has neuromodula-
`tory properties and its accumulation in the CNS is associ-
`ated with headache, nausea, emesis, and somnolence.' 0
`We have recently shown that children who receive MTX
`for leukemia have significant elevations of adenosine in
`that methylxanthines,
`CSF and, when symptomatic,
`which are adenosine receptor antagonists, ameliorate
`acute toxicity." These models conceivably explain some
`features of MTX-associated neurotoxicity, but do not ade-
`quately account for the occurrence of others, including
`seizures, focal neurologic deficits, and chronic toxicity.
`MTX readily increases plasma homocysteine (Hcy) 12 ' 5
`and this increase may be a sensitive and responsive indi-
`cator of antifolate therapy. This observation is important,
`because Hcy is injurious to vascular endothelium 6"'7 and
`hyperhomocysteinemia is a strong, independent risk fac-
`tor for vascular disease,'" including stroke and carotid
`artery stenosis. Since reported consequences of MTX
`therapy include focal neurologic deficits,2 5 mineralizing
`microangiopathy,' 9 and radiographic ischemic white mat-
`
`2800
`
`Journal of Clinical Oncology, Vol 15, No 8 (August), 1997: pp 2800-2806
`
`Information downloaded from jco.ascopubs.org and provided by at Reprints Desk on September 12, 2016 from 216.185.156.28
`Copyright © 1997 American Society of Clinical Oncology. All rights reserved.
`
`Lilly Ex. 2033
`Sandoz v. Lilly IPR2016-00318
`
`

`
`METHOTREXATE
`
`INCREASES CSF Hcyt AND EAA NEUROTRANSMITTERS
`
`2801
`
`ter changes,20 ischemic vascular disease may be impli-
`cated in the pathogenesis of MTX-associated neurotoxic-
`ity. We propose that excess Hcy, in light of its role in
`vascular injury, may be in part the mediator of this
`toxicity. Moreover, excess Hcy can be metabolized to
`sulfur-containing excitatory amino acid (EAA) neuro-
`transmitters, including homocysteic acid (HCA) and cys-
`teine sulfinic acid (CSA), as well as the related com-
`pounds homocysteine sulfonic acid (HCSA) and cysteic
`acid (CA).2 ' 22 These endogenous agonists of N-methyl-
`D-aspartate (NMDA) receptors are likely important in the
`genesis of seizures 23 and, further, they may cause neu-
`ronal death and late MTX-associated sequelae via their
`excitotoxic properties. 4 25
`Determination of the CSF concentrations of Hcy and
`sulfur-containing EAAs in patients who receive MTX
`would be important to investigate further their roles as
`mediators of MTX-associated neurotoxicity. There are
`minimal published data on normal values for CSF Hcy,
`especially in the pediatric population. Normal ranges for
`CSF HCA, CSA, HCSA, and CA are not known. We
`therefore measured the Hcy and sulfur-containing EAA
`content of CSF from children who received MTX for
`cancer and from children not previously exposed to MTX.
`We hypothesized that the concentrations of Hcy and EAA
`neurotransmitters in the CSF of children who received
`MTX would be higher than in the control group.
`
`PATIENTS AND METHODS
`We conducted a retrospective analysis of stored CSF specimens
`from patients who had received lumbar punctures at the time of
`diagnosis of cancer or during the course of its treatment. Also ana-
`lyzed were stored CSF samples from patients with neurologic disease
`obtained for diagnostic purposes. For analysis of the data we defined
`two groups of patients: (1) a control population that included patients
`without any MTX exposure, and (2) a treatment group that included
`individuals who had received MTX within the 7 days preceding a
`lumbar puncture. A cut-off period of 7 days was chosen because
`many of the patients in this study received MTX according to weekly
`cycles; moreover, much of the acute and subacute neurologic toxicity
`associated with MTX occurs within 1 week of exposure.
`Informed consent was obtained from all patients (or parents),
`and MTX was administered according to several disease-specific,
`institutional review board-approved protocols. All lumbar punctures
`were obtained under the direction of these protocols; no extra lumbar
`punctures were obtained for the purposes of this study. Moreover,
`our frontline institutional protocol for acute lymphoblastic leukemia
`(since 1986) specifies that the scheduled lumbar puncture is per-
`formed at the end of a course of oral divided-dose MTX and before
`delayed leucovorin is given.2 6
`Not all patients had adequate CSF collection for all measurements
`and matched plasma samples were not available for the majority of
`the subjects. For patients from whom two separate CSF specimens
`were obtained, collections were separated by many weeks or months,
`and no patient had more than two samples of CSF assayed. None
`
`of the patients had leukemic CNS disease or exposure to cranial
`irradiation at the time the samples were obtained.
`Following collection, CSF samples were placed on ice at the
`bedside for transport and then stored at -20°C until they were ana-
`lyzed. Grossly bloody specimens were excluded from analysis due
`to the confounding presence of exogenous Hcy.
`The control group for Hcy determinations included samples of
`CSF from 34 children. Of these, 17 had lumbar punctures at the
`time of diagnosis of leukemia and the remaining 17 had lumbar
`punctures for the diagnosis of neurologic disease. The corresponding
`treatment group included 23 samples of CSF from 16 children. Of
`these 16 children, 15 had received MTX for the treatment of leuke-
`mia and one for the treatment of neuroblastoma.
`The control group for HCA, CSA, HCSA, and CA measurements
`included samples of CSF from 29 children; 12 had lumbar punctures
`at the time of diagnosis of leukemia and the remaining 17 had lumbar
`punctures for the diagnosis of neurologic disease. The corresponding
`treatment group included 16 samples of CSF from 12 patients. Of
`these 12 patients, 11 had received MTX for the treatment of ALL
`and one for the treatment of neuroblastoma.
`CSF Hcy was assayed by high-performance liquid chromatogra-
`phy (HPLC). The tributylphosphine/ammonium 7-fluorobenzo-2-
`oxa-1,3-diazide-4-sulfonate (TBP/SBD-F) method applied TBP for
`reduction and SBD-F as derivatization agent according to Araki et
`al,27 with the modifications reported by Ling et al28 for the microbore
`application, Hyland et al29 for CSF Hcy signal optimization scheme,
`and the internal standard quantitation according to Vestor et al.30
`The SBD-F derivatives were eluted isocratically from a Prodigy
`octadecylsilane (ODS)-3 100A 150- x 1.0-mm 5-pm microbore
`column (Phenomenex, Torrance, CA). Standards ranged from 5
`nmol/L to 2,000 nmol/L. Variance-stabilized regression analysis of
`calibration standards was used for statistical analysis of data.3 ' The
`standard curve was linear over the range tested (R2 =.992) with a
`mean residual of 8%. The within-day and between-day coefficients
`of variation were determined using a single pooled CSF sample
`aliquotted into 250-mL portions and left frozen at -20°C.
`CSF HCA, HCSA, CSA, CA, and other amino acids were deter-
`mined by reverse-phase HPLC coupled to electrochemical detection
`(OPA). 32
`after precolumn derivatization with o-phthaldialdehyde
`Modifications to the method were made for the separation and detec-
`tion of the excitotoxic sulfur amino acids from other amino acids
`present in CSF samples. Activated OPA reagent was prepared by
`the addition of 1 /tL of 3-mercaptopropionic acid to 200 L of OPA
`incomplete reagent solution (Sigma, St Louis, MO). CSF (4 L)
`was then reacted with activated OPA reagent, and after exactly 2
`minutes, the entire mixture was injected into a Hewlett Packard 1090
`gradient liquid chromatograph. The OPA-derivatized amino acids
`were separated on a reverse-phase C18 Hypersil column (250 mm
`x 4 mm, 5-gum particle size; Hewlett Packard, Wilmington, DE) and
`detected by a Coulochem II electrochemical detector equipped with
`a dual electrode analytical cell, model 5011 with El set at +250
`mV and E2 set at +700 mV (ESA Inc, Chelmsford, MA). The
`mobile phase consisted of two eluents: 0.15 mol/L sodium acetate
`adjusted to pH 5.3 with concentrated acetic acid, finally containing
`6% methanol (solvent A), and 0.15 mol/L sodium acetate adjusted
`to pH 5.7 with concentrated acetic acid, finally containing 70%
`methanol (solvent B). The flow rate was at 1.3 mL/min and separa-
`tions were performed at 400 C. Isocratic conditions with 100% solvent
`A were held at 15 minutes, then solvent B was increased to 100%
`in 35 minutes. Reequilibration to 100% solvent A lasted 10 minutes
`before the next sample was injected. Peak identity was confirmed
`
`Information downloaded from jco.ascopubs.org and provided by at Reprints Desk on September 12, 2016 from 216.185.156.28
`Copyright © 1997 American Society of Clinical Oncology. All rights reserved.
`
`Lilly Ex. 2033
`Sandoz v. Lilly IPR2016-00318
`
`

`
`QUINN ET AL
`
`0 0c
`
`o v
`
`1 A
`,IU
`
`100
`
`50
`
`_
`
`500-
`
`400-
`
`8 200.
`
`100.
`
`0.
`
`Treatment
`
`Control
`
`Treatment
`
`Fig 2. Effect of MTX on CSF HCA and CSA. HCA and CSA were not
`detected in CSF from controls (n = 291; however, MTX caused marked
`32.0, n = 16) and CSA
`accumulation CSF HCA (119.1 /umol/L
`(28.4 umol/L ± 7.7, n = 16) in the treatment group. Horizontal bars
`represent mean concentrations.
`
`range of 0 to 102 pimol/L. A highly significant correlation
`between HCA and CSA was detected by linear regression
`analysis (R2 = .99, P < .0001). Also measured were the
`CSF concentrations of the related compounds HCSA and
`.tmol/L + 0.38) and CA (0.18 moVL
`CA. HCSA (1.31
`± 0.06) were found in the treatment group (n = 16), but
`none was detected in controls (n = 29).
`Sulfur-containing EAA neurotransmitters are increased
`in the CSF when CSF Hcy is high. Figure 3 depicts the
`relationship between CSF Hcy and CSF HCA for all
`patients who had both compounds measured. No patient
`whose CSF Hcy was less than 0.355 pimol/L had detect-
`able quantities of HCA. Most patients whose CSF Hcy
`0.355 mol/L, and all patients with CSF Hcy
`was
`
`nf
`
`400.
`
`-! 300
`
`LL,go 200.
`
`100
`
`0 355iM
`
`. +
`
`Us
`
`1
`
`
`
`_, e
`
`
`r--- @
`0 10
`
`
`
`-- --- **H
`
`CSF Hcy (pM)
`
`_
`I
`10'00
`
`Fig 3. Relationship between CSF Hcy and HCA. No patient whose
`CSF Hcy was < 0.355 mol/L ( ---- ) had detectable CSF HCA. Most
`and all patients whose
`patients whose CSF Hcy was
`: 0.355 /mol/L.,
`CSF Hcy was > 0.930 mol/L, had elevated CSF HCA.
`
`2802
`
`4.
`
`3.
`
`2
`
`1.
`
`l.
`
`U-
`
`:*
`
`*
`
`Control
`
`Treatment
`
`Fig 1. Effect of MTX on CSF Hcy. The treatment group had a sig-
`nificantly (P = .0255) greater concentration of Hcy in CSF (0.814
`/mol/L ± 0.215 [mean + SEMI, n = 23) than the control group (0.210
`,mol/L ± 0.028, n = 34). Horizontal bars represent mean concentra-
`tions.
`
`by matching retention times based on the retention time of external
`standards, as well as by coelution after spiking human CSF with
`authentic standards. The lowest limit of detection for the various
`OPA-derivatized amino acids varied between 10 and 50 nmol/L
`using a signal-to-noise ratio of 3.
`CSF 5-methyltetrahydrofolate was determined by reverse-phase
`to electrochemical detection as previously de-
`HPLC coupled
`scribed.3 3
`One- and two-tailed Mann-Whitney tests and linear regression
`analysis were used for statistical analysis where appropriate.
`
`RESULTS
`Hcy is increased in the CSF of patients recently treated
`with MTX. The concentration of Hcy in CSF from the
`treatment and control groups is depicted in Fig 1. The
`control group (n = 34) had a mean Hcy concentration of
`0.210 /tmolL + 0.028 (SEM), a median of 0.155 jimol/
`L, and a range of 0.003 to 0.700 pimol/L. The treatment
`group (n = 23) had a mean Hcy concentration of 0.814
`/molL + 0.215, a median of 0.567 jimol/L, and a range
`of 0.008 to 3.862 /mol/L. A one-tailed Mann-Whitney
`test detected a significant difference between the groups
`(P = .0255).
`Sulfur-containing EAA neurotransmitters are increased
`in the CSF of patients recently treated with MTX. The
`concentrations of HCA and CSA in CSF from the treat-
`ment and control groups are depicted in Fig 2. HCA and
`CSA were not detected in CSF from the control popula-
`tion (n = 29). The treatment group (n = 16) had a mean
`HCA concentration of 119.1 /.mol/L + 32.0, a median
`of 96 ptmol/L, and a range of 0 to 416 mol/L. The
`treatment group (n = 16) also had a mean CSA concentra-
`tion of 28.4 tmol/L ± 7.7, a median of 24.9 Lmol/L, and a
`
`Information downloaded from jco.ascopubs.org and provided by at Reprints Desk on September 12, 2016 from 216.185.156.28
`Copyright © 1997 American Society of Clinical Oncology. All rights reserved.
`
`Lilly Ex. 2033
`Sandoz v. Lilly IPR2016-00318
`
`

`
`METHOTREXATE
`
`INCREASES CSF Hcyt AND EAA NEUROTRANSMITTERS
`
`2803
`
`Table 1. Effect of MTX on the Concentration of Other Amino Acids in the CSF
`
`Mean + SEM Value (mol/L}
`
`Group
`
`Asp
`
`Ser
`
`Glu
`
`Gin
`
`Gly
`
`Thr
`
`Arg
`
`Tau
`
`Control (n = 28)
`Treatment (n = 16)
`
`1.6 ± 0.2
`1.3 ± 0.2
`
`14.3 ± 0.8
`11.3
`1.2
`
`1.1 ± 0.1
`0.6± 0.1'
`
`944 ±+ 27
`836 ± 121
`
`4.0 ± 0.4
`8.0 ±+ 0.6t
`
`12.2 ± 0.8
`20.9 ± 26'
`
`10.9 ± 0.7
`10.1 ± 1.2
`
`4.5 ± 0.4
`4.0
`0.5
`
`Abbreviations' Asp, aspartate; Ser, serine; Glu, glutamate; Gin, glutamine; Gly, glycine; Thr, threonine; Arg, arginine; Tau, taurine.
`*P < .01.
`tP < .0001.
`
`levels greater than 0.930 mol/L, had detectable HCA in
`CSF. A similar relationship was found between Hcy and
`the other sulfur-containing EAAs CSA, HCSA, and CA.
`CSF was also analyzed for eight other amino acids
`(aspartate, serine, glutamate, glutamine, glycine, threo-
`nine, arginine, and taurine) in a subset of the previous
`patients (Table 1). No significant differences between the
`treatment (n = 28) and control groups (n = 16) for aspar-
`tate, serine, glutamine, arginine, and taurine were detected
`by a two-tailed Mann-Whitney test. A significantly lower
`concentration of glutamate and significantly higher con-
`centrations of glycine and threonine were detected by the
`two-tailed Mann-Whitney.
`The CSF content of 5-methyltetrahydrofolate was com-
`pared for the treatment (n = 28) and control groups (n
`= 17). Although both groups had CSF folate levels in
`the normal range (40 to 80 nmol/L), we found that the
`treatment group had a lower concentration of folate (54.2
`nmol/L + 11.8) than the control group (82.8 nmol/L ±
`9.0), which was demonstrated to be statistically signifi-
`cant (P = .0245) by a two-tailed Mann-Whitney test.
`The control population is biochemically homogeneous.
`The control population was analyzed for differences be-
`tween the patients with cancer and those with neurologic
`disease relative to the CSF concentrations of Hcy, HCA,
`CSA, HCSA, CA, folate, and the eight other amino acids.
`No significant differences were detected for Hcy (cancer
`0.185 pmol/L + 0.03 v neurologic disease 0.239 pmol/
`L + 0.05), folate (cancer 75.5 nmol/L ± 13.1 v neurologic
`11.8), or the eight other amino
`disease 83.3 nmol/L
`acids by a two-tailed Mann-Whitney test. With reference
`to HCA, CSA, HCSA, and CA, the two subgroups were
`identical.
`One patient was excluded from the control group be-
`cause she manifested significant baseline hyperhomocys-
`teinemia of uncertain etiology. Her plasma Hcy level be-
`fore antifolate therapy was 22.5 umol/L (normal range,
`1 to 10 pmol/L). This patient's CSF Hcy was also high
`(1.9 Mmol/L), and HCA and CSA were present at 154
`,umol/L and 36 mol/L, respectively. HCA and CSA were
`detected in no other patients without prior antifolate ther-
`apy. Also excluded were two patients with dystonia who
`
`had received treatment with L-dihydroxy-phenylalanine
`(L-DOPA), which may cause an increase in Hcy. These
`two patients had CSF Hcy concentrations of 0.760 zmol/
`L and 0.930 /lmol/L, but no detectable sulfur-containing
`EAAs.
`Three patients manifested neurologic toxicity at the
`time of a scheduled lumbar puncture. Patient no. 1 devel-
`oped seizures within 3 days of receiving oral MTX (25
`mg/m2 every 6 hours for four doses). Her CSF Hcy level,
`the highest recorded, was 3.862 umol/L. Her HCA level
`was 126 /mol/L and her CSA level was 31 mol/L.
`Patient no. 2 experienced an episode of ataxia, dysarthria,
`blurred vision, and confusion within 1 week of receiving
`intravenous MTX (2 g/m2 over 2 hours). Her CSF Hcy
`concentration was 0.690 mol/L; HCA and CSA levels
`were 269 gmol/L and 69.1 umol/L, respectively, both of
`which were the third highest measured values. Patient no.
`3 presented with a question of mild developmental delay
`at the time of diagnosis with acute lymphoblastic leuke-
`mia. During the course of his treatment, we noted marked
`neurodevelopmental regression, and computed tomogra-
`phy and magnetic resonance imaging of the cranium dem-
`onstrated evidence of mineralizing microangiopathy. Pa-
`tient no. 3 received his last dose of MTX (25 mg/m2
`orally every 6 hours for four doses) 1 week before lumbar
`puncture, and his CSF Hcy level was 1.680 mol/L. His
`CSF HCA and CSA concentrations were 416 gzmol/L and
`102 mol/L, respectively, the highest measured for both
`compounds. In light of these findings, the intensity of this
`patient's systemic MTX therapy has been decreased, and
`he no longer receives intrathecal MTX.
`Due to the small number of samples and several differ-
`ent regimens for systemic MTX (eg, 100 mg/m2 as 25
`mg/m 2 orally every 6 hours, and 2 g/m2 intravenously
`over 4 hours), no correlation between CSF levels of the
`previous compounds and MTX dose was readily apparent.
`
`DISCUSSION
`Hcy is at a metabolic crossroads. It may enter the acti-
`vated methyl cycle by remethylation to form methionine.
`is catalyzed by 5-methyltetrahydrofo-
`Methylation
`late:Hcy methyltransferase, wherein 5-methyltetrahy-
`
`Information downloaded from jco.ascopubs.org and provided by at Reprints Desk on September 12, 2016 from 216.185.156.28
`Copyright © 1997 American Society of Clinical Oncology. All rights reserved.
`
`Lilly Ex. 2033
`Sandoz v. Lilly IPR2016-00318
`
`

`
`QUINN ET AL
`
`occurrence of focal neurologic deficits,2 5 mineralizing
`microangiopathy,'9 and radiographic ischemic white mat-
`ter changes20 secondary to MTX therapy suggests that
`ischemic vascular disease may be involved. Hcy may be
`the mediator of this toxicity.
`A related aspect of Hcy metabolism is also likely to
`be important in the pathogenesis of MTX neurotoxicity.
`EAAs, such as HCA and CSA, may be oxidatively de-
`rived from Hcy and cysteine, respectively. These com-
`pounds are putative neurotransmitters and are endogenous
`agonists of the NMDA receptor, a subtype of glutamate
`receptor.' 2224 HCA is known to accumulate in excess in
`patients with classic homocystinuria.3 8 Similarly, rats
`made hyperhomocysteinemic by dietary Hcy supplemen-
`tation will accumulate HCA.39 CSA may also accumulate
`in states of Hcy excess that result from folate depletion
`due to shunting of Hcy through the transsulfuration path-
`way. Similar EAAs include HCSA and CA, closely re-
`lated to HCA and CSA, respectively. EAAs are believed
`to be important in the pathogenesis of seizures 23 -25 and it
`has been proposed that EAAs may be responsible for the
`seizures that occur in classic homocystinuria.40 Further-
`more, overexpression of EAAs may cause neuronal injury
`and death via excessive glutamate receptor activation.
`This mechanism of injury is called excitotoxicity and it
`has been implicated as a final common pathway for a
`wide range of chronic neurodegenerative disorders and
`acute neurologic processes, including stroke.25
`The presence of HCA and CSA in whole brain tissue
`has been demonstrated 2 1' 22 24 ; however, little is known
`about their concentrations in CSF. We report here the
`first published series of HCA and CSA determinations in
`CSF. We have found that HCA and CSA are not detect-
`able in the CSF of individuals without MTX exposure.
`However, we have shown that MTX markedly increased
`the concentration of these EAA neurotransmitters.
`Smaller elevations in the related compounds HCSA and
`CA were also observed, both of which were not detected
`in controls. Furthermore, all patients with neurologic tox-
`icity at the time of lumbar puncture had marked elevations
`of HCA and CSA, and the patient with the highest mea-
`sured values for both compounds presently suffers from
`severe chronic toxicity.
`There appears to be two subsets of patients within the
`treatment group: those with and those without detectable
`HCA and CSA (Fig 2). However, it should be noted that
`all individuals without detectable sulfur-containing EAAs
`had CSF Hcy concentrations less than 0.355 ,amol/L.
`Thus, there appears to be a threshold phenomenon for the
`appearance of HCA and CSA in the CSF. This observa-
`tion implies that accumulation of these EAAs occurs only
`
`2804
`
`drofolate serves as the methyl donor and cobalamin is a
`cofactor. This reaction provides the biochemical basis for
`the occurrence of hyperhomocysteinemia in folate and
`cobalamin deficiencies. Betaine:Hcy methyltransferase
`may also catalyze the methylation with betaine serving
`as the methyl donor. Alternatively, Hcy may enter the
`transsulfuration pathway by conjugation with serine to
`form cysteine. Hcy may also be oxidatively metabolized
`to HCA and HCSA.
`Despite the identification of hyperhomocysteinemia as
`a risk factor for vascular disease and increasingly frequent
`monitoring of plasma Hcy levels, little is known about
`the presence of Hcy in CSF. Hyland and Bottiglieri 29
`published a series of nine patients with motor neuron
`disease or peripheral neuropathy, but with normal serum
`B12 and RBC folate levels, in which the mean CSF Hcy
`concentration was 0.46 /mol/L (range, 0.28 to 0.66,u.tmol/
`L). Blom et a134 analyzed CSF from six patients with
`neurologic disease not related to Hcy metabolism and
`found a range of 0.007 to 0.020
`rtmol/L for Hcy. Also
`reported by the latter group was a CSF Hcy concentration
`of 3.5
`.tmol/L for one patient with severe cobalamin de-
`ficiency and related neurologic disease. Similarly, Stabler
`et a135 reported a range of 0.5 to 3.0 )umol/L for CSF
`Hcy in four patients with neurologic disease related to
`cobalamin deficiency. Finally, Bottiglieri36 found that
`CSF Hcy ranged from 1.377 to 4.732 .tmol/L in four B12-
`deficient patients with subacute combine degeneration,
`and it was 0.493 /mol/L in one folate-deficient patient,
`with an adult control range (n = 18) of 0.015 to 0.140
`,tmol/L.
`MTX can cause an initial biochemical, intracellular
`reduced folate deficiency, and ultimately an absolute de-
`ficiency.3 7 Since folate deficiency causes a secondary ele-
`vation in Hcy, we determined the effect of MTX on CSF
`Hcy. We report here the largest series of CSF Hcy deter-
`minations published to date. We have shown that CSF
`Hcy in individuals not exposed to MTX is low, in accor-
`dance with previous few reports.2 9
`3 4
`6 However, we
`'
`found that MTX significantly raised the concentration of
`Hcy in CSF, and this effect may be sustained at least 1
`week. Moreover, patients with neurologic toxicity had
`many of the highest measured values.
`Elevation of plasma and CSF Hcy consequent to MTX
`administration may be important in the pathogenesis of
`MTX-associated neurotoxicity. Hcy is a highly reactive
`amino acid that is directly toxic to vascular endothelium,
`and exposure to it promotes a prothrombotic vascular
`surface.' 16 17 Furthermore, hyperhomocys-
`endothelial
`teinemia is a strong independent risk factor for vascular
`disease,'8 including stroke and carotid artery stenosis. The
`
`-3
`
`Information downloaded from jco.ascopubs.org and provided by at Reprints Desk on September 12, 2016 from 216.185.156.28
`Copyright © 1997 American Society of Clinical Oncology. All rights reserved.
`
`Lilly Ex. 2033
`Sandoz v. Lilly IPR2016-00318
`
`

`
`METHOTREXATE
`
`INCREASES CSF Hcyt AND EAA NEUROTRANSMITTERS
`
`2805
`
`in states of Hcy excess. Moreover, it is unlikely that
`elevation of these EAAs resulted artifactually from spon-
`taneous, nonenzymatic oxidation of Hcy during specimen
`storage since their concentrations are 10- to 100-fold
`higher than that of Hcy. These data suggest a process
`of ongoing metabolism of excess Hcy and consequent
`accumulation of metabolites in vivo.
`As expected, no marked changes in the CSF concentra-
`tions of aspartate, serine, glutamine, arginine, and taurine
`were noted consequent to MTX exposure. The decreased
`concentration of glutamate may have been secondary to
`polyglutamation of MTX. The significance of the increased
`concentrations of glycine and threonine is not known.
`Certain individuals may be at risk for pronounced
`MTX-induced hyperhomocysteinemia,
`including those
`with polymorphisms of important enzymes related to Hcy
`metabolism. For example, the thermolabile variant of
`5,10-methylenetetrahydrofolate reductase (MTHFR) has
`only 35% to 40% of normal activity.41 Individuals with
`this allele manifest baseline hyperhomocysteinemia. Un-
`der the stress of iatrogenic folate deficiency, such as ad-
`ministration of an antifolate, the hyperhomocysteinemia
`and its adverse effects may be magnified. Similarly, poly-
`morphisms of cystathionine-P-synthase or other enzymes
`
`may be involved. 7 Coexistence of factor V Leiden with
`one of these mutations may pose an even greater risk.4 2
`The identification of a pharmacogenetic risk profile for
`MTX neurotoxicity, therefore, is an important endeavor.
`Currently, we are engaged in an extensive prospective
`risk profile analysis of patients with newly diagnosed
`cancer, to include baseline plasma and CSF Hcy, MTHFR
`phenotype, factor V Leiden analysis, and other determina-
`tions. Furthermore, we continue to analyze CSF prospec-
`tively for sulfur-containing EAAs in MTX-exposed indi-
`viduals.
`Fortunately, hyperhomocysteinemia is a treatable con-
`dition. Folate therapy will reliably reduce plasma Hcy
`levels; however, this would also rescue cells from the
`cytotoxic effects of MTX. There is another potential solu-
`tion: betaine, a readily available nutritional supplement
`used in the treatment of homocystinuria, 43 can serve as
`an alternate methyl donor for the remethylation of Hcy.
`Betaine therapy might, therefore, provide specific rescue
`and thereby improve the therapeutic index of MTX. It
`has been suggested that lowering Hcy levels may reduce
`cardiovascular risk. We postulate that reduction or pre-
`vention of hyperhomocysteinemia with betaine in patients
`who receive MTX may lower the risk of neurotoxicity.
`
`REFERENCES
`11. Bernini JC, Fort DW, Griener JC, et al: Aminophylline for
`1. Jonsson OG, Kamen BA: Methotrexate and leukemia. Cancer
`methotrexate-induced neurotoxicity. Lancet 345:544-547, 1995
`Invest 9:53-60, 1991
`12. Refsum H, Helland S, Ueland PM: Fasting plasma homocys-
`2. Allen JC, Rosen G, Mehta BM, et al: Leukoencephalopathy
`teine as a sensitive parameter of antifolate effect: A study of psoriasis
`following high-dose IV methotrexate chemotherapy with leucovorin
`patients receiving low-dose methotrexate treatment. Clin Pharmacol
`rescue. Cancer Treat Rep 64:1261-1273, 1980
`Ther 46:510-520, 1989
`3. Jaffe N, Takaue Y, Anzai T, et al: Transient neurologic distur-
`13. Refsum H, Ueland PM, Kvinnsland S: Acute and long-term
`bances
`induced by high-dose methotrexate
`treatment. Cancer
`effects of high-dose methotrexate treatment on homocysteine in
`56:1356-1360, 1985
`plasma and urine. Cancer Res 46:5385-5391, 1986
`4. Walker RW, Allen JC, Rosen G, et al: Transient cerebral dys-
`14. Broxson EH, Stork LC, Allen RH, et al: Changes in plasma
`function secondary to high-dose methotrexate. J Clin Oncol 4:1845-
`methionine and total homocysteine levels in patients receiving meth-
`1850, 1986
`otrexate infusions. Cancer Res 49:5879-5883, 1989
`5. Kay HE, Knapton PJ, O'Sullivan JP, et al: Encephalopathy in
`15. Refsum H, Wesenberg F, Ueland PM: Plasma homocysteine
`acute leukaemia associated with methotrexate therapy. Arch Dis
`in children with acute lymphoblastic leukemia: Changes during a
`Child 47:344-354, 1972
`chemotherapeutic regimen
`including methotrexate. Cancer Res
`6. Quinn CT, Kamen BA: A biochemical perspective of metho-
`51:828-835, 1991
`16. Rees MM, Rodgers GM: Homocysteinemia: Association of
`trexate neurotoxicity with insight on nonfolate rescue modalities. J
`a metabolic disorder with vascular disease and thrombosis. Thromb
`Invest Med 44:522-530, 1996
`Res 71:337-359, 1993
`7. Abelson HT: Methotrexate and central nervous system toxicity.
`17. Malinow MR: Homocyst(e)irie and arterial occlusive diseases.
`Cancer Treat Rep 62:1999-2001, 1978
`J Intern Med 236:603-617, 1994
`8. Millot F, Dhondt J-L, Mazingue F, et al: Changes of cerebral
`18. Clarke R, Daly L, Robinson K, et al: Hyperhomocysteinemia:
`biopterin and biogenic amine metabolism in leukemic children receiving
`An independent risk factor for vascular disease. N Engl J Med
`5 g/m2 intravenous methotrexate. Pediatr Res 37:151-154, 1995
`324:1149-1155, 1991
`9. Cronstein