~
nlli BRAIN RESEARCH
ELSEVIER
Brain Research 647 (1994) 172-176
Short Communication 19F
NMR calcium changes, edema and histology in neonatal rat brain slices during glutamate toxicity
Maryceline T. Espanol, Lawrence Litt *, Yan Xu, Lee-Hong Chang, Thomas L. James, Philip R. Weinstein, Pak H. Chan Departments of Pharmaceutical Chemistry, Anesthesia, Neurology, Neurosurgery and Radiology, and Cardiovascular Research Institute. University of California at San Francisco, San Francisco, CA, USA
(Accepted 10 March 1994)
Abstract
Respiring neonatal cerebrocortical slices (350/xm thick), loaded with the free calcium indicator 5F-BAPTA, were perfused in a 20-mm-diameter glass NMR tube with oxygenated artificial CSF, exposed to extracellular glutamate and studied at 4.7 Tesla with 19F NMR spectroscopy. 31p//1H NMR spectra, obtained concurrently, were used to assess slice integrity from determinations of intracellular pH, ATP, PCr, lactate and N-acetylaspartate. 60-min periods were induced of recoverable and nonrecoverable glutamate toxicity-defined from changes in NMR metabolites. In other NMR studies, where 5F-BAPTA was not used, metabolic toxicity was modulated by three glutamate receptor antagonists: dizocilpine, NBQX and kynurenic aci& Outcome measurements were made of edema, determined invasively in isolated slices from % swelling and water content and from histological changes in Nissl stains of slice sections. Edema was (1) detectable in all slices within minutes after onset of glutamate exposure, though never in untreated control slices, and (2) modulated differently by dizocilpine, NBQX and kynurenate. Correlations were observed between edema and NMR decreases in PCr and ATP. Nissl stains of sections from slices treated with the most protective agent, dizociipine, showed preservation of neuronal processes. As was expected in 7-day-old rats with immature NMDA receptors, 19F NMR spectroscopy revealed only small increases in free intracellular calcium ([Ca2+]~). These occurred late during glutamate exposure and reversed early during glutamate washout. The studies demonstrate that it is possible to study correlations between repeated noninvasive NMR spectra in ensembles of brain slices and invasive measures of early cellular responses. Key words: Glutamate; Brain slice; Brain edema; NMR; Calcium
Neurotoxicity and brain injury are known to result from activation of CNS glutamate receptors, according to mechanisms postulated for a wide pathological spectrum by the excitotoxic hypothesis [6,7]. Although N M R spectroscopy can be used noninvasively in vivo to measure toxic metabolic changes, special experimental advantages arise if N M R is instead used to study respiring brain slices ex vivo [4,9,12]. We recently completed a nondestructive 31p/1H N M R spectroscopy study of respiring neonatal cerebrocortical slices at 4.7 Tesla and found that energy failure began immediately after glutamate exposure [10,11]. Early energy failure agreed
* Corresponding author. Address: Department of Anesthesia, University of California at San Francisco, 521 Parnassus Avenue, San Francisco, CA 94143, USA. Fax: (1) (415) 476-9516. 0006-8993/94/$07.00 © 1994 Elsevier Science B.V. All rights reserved SSDI 0006-8993(94)00300-2
with previous N M R spectroscopy studies of N M D A and glutamate administration to live brain slices [2,12]. Our studies, however, were the first to concurrently monitor lactate, N-acetylaspartate (NAA) and administered glutamate. Additionally, we showed that energy failure was modulated differently by treatment with three glutamate receptor agonists: dizociline, kynurenate and NBQX. Using the same protocol, we now demonstrate it is possible to establish correlations between (1) N M R spectroscopy changes, (2) brain edema. an early marker of injury, and (3) histological measurements using light microscopy. Because increased free intracellular calcium ([Ca2+]i) is generally observed following activation of glutamate receptors, we used 19F N M R to monitor it (concurrently with 31p//IH N M R spectroscopy) in slices loaded with 5F-BAPTA [1,13].
M.T. Espanol et a l . / Brain Research 647 (1994) 172-176
Respiring cerebral cortical slices (350 ~ m thick) were prepared from 20 neonatal Sprague-Dawley rats (10-12 g, age 7 + 2 days; Simonson, Gilroy, CA) according to published procedures developed by us [9] and approved by the UCSF Committee on Animal Research. In brief, 80 tissue slices (total weight ~ 3.2 g) were transferred individually to a 20-mm-diameter glass N M R tube (Wilmad, Buena, N J) that served as a tissue chamber. In this chamber, slices were perfused with artificial cerebrospinal fluid (ACSF) at 37°C. The CSF flow rate was 15-20 ml/min. The bicarbonate pH buffer for oxy-ACSF was maintained by continuous bubbling with a 95% 0 2 / 5 % CO2 gas mixture that kept p C O 2 and extracellular pH constant (at 40 mm Hg and ~ 7.4, respectively). Extracellular fluid samples of perfusion medium were collected every 15 min from the entrance and exit of the N M R chamber and analysed with a Radiometer Model ABL 30 blood-gas analyser. The rationales, protocols and dose-response data that justify choosing 2 mM for the glutamate concentration for slices, which is substantially higher than concentrations used in cell culture investigations, are described in our previous study [10,11] which consisted only of 3 1 p / I H N M R spectroscopy measurements. Antagonist concentrations, also chosen as in our previous study, were 150/xM for dizocilpine, 1 mM for kynurenate and 6 tzM for NBQX. N M R spectra were obtained on a Nalorac Quest 4400 4.7 Tesla N M R instrument, operating at 81, 188 and 200 MHz, for 3lp, t9 F and ~H, respectively. The NMR tube containing perfused slices was positioned inside one of two custom-made, four-turn, solenoidal coils, 23 × 15 mm. For interleaved 31p/1H studies, a double-tuned coil was used as described elsewhere [10,11]. For trinuclear studies, a second coil of the same dimensions was used that was triple-tuned to each of the above frequencies. [Ca2+] i was detected in trinuclear N M R studies, where brain slices were loaded with 5F-BAPTA (obtained as an ester from Molecular Probes, OR; and prepared as a 50-mM stock solution in DMSO). The methods have been described by others [1,2]. Brain-slice spectra in Fig. 1 show two well-resolved 19F resonance peaks, corresponding to 5F-BAPTA that is either free
or bound to intracellular calcium. For each spectrum, the ratio of free to bound intracellular calcium was calculated from the ratio of 5F-BAPTA NMR signal intensities. Estimates of absolute values for [Ca2+] i can be obtained from the law of mass action if one knows
,1 i
@
i
A
"1
,j i
i
i''1 ©
J :
®
(~ Fig. 1. Representative 19F NMR spectra, obtained at 188 MHz, from one study of glutamate toxicity in perfused, respiring, cerebrocortical slices loaded with 5F-BAPTA. 31p and IH NMR spectra from these slices were indistinguishable from slices never exposed to 5F-BAPTA. Two 19F resonance peaks in each spectrum correspond to (1) 5FBAPTA that is bound to intracellular Ca 2+ and (2) 5F-BAPTA that is free. Chronological order of spectra is from bottom to top and corresponds to time epochs: (A) before glutamate exposure, (B) after 30 min of exposure to 2 mM glutamate, (C) after 60 min of exposure to 2 mM glutamate and (D) after 60 min of washout (recovery) with glutamate-free ACSF.
173
i 15.0
t
Ca+2-bound I 5FBAPTA /i
~0 0
Free 5FBAPTA
o'.o
5.0 PPM
-~'.o
-,o.o
174
M. 7: Espanol et al. / B r a i n Research 647 (1994) 172 ~176
K d for the reaction: C a 2 + + ( 5 F - B A P T A ) ~ Ca2+(5F BAPTA). [Ca2+]i is simply K d × the ratio of I~F NMR signal intensities [13]. However, complexity and controversy surround the determination of K d for different tissue types and compartments [13], whereas determinations of relative changes in [Ca 2+]i are independent of K d. Interleaved 3 1 p / I H / 1 9 F NMR spectra, acquired by using rapid frequency switching and versatile RF pulse programming, were obtained in a system that has but two frequency synthesizers [10,11]. For each frequency, the spectral width was 4000 Hz and 2K data points were obtained. Typical 3~P and ~gF pulse parameters during trinuclear N M R studies were: 30 izs for the 3tp (45 ° nutation) and 40/~s for the 19F ( 9 0 ° nutation); 1.8 s for the 3~p interpulse delay; and 300 ms for the J~F interpulse delay. All data were collected using quadrature phase detection. When doing trinuclear NMR, spin-echo 1H NMR experiments were initiated with a 1331-2662 binomial pulse duo whose excitation minimum was centered on the water resonance at 4.7 ppm, and whose excitation was maximum for metabolites having chemical shifts near the lactate peak (1.32 ppm). Typical tH pulse parameters during trinuclear NMR studies were: 40 ~s (90 ° nutation), 136 ms (spin-echt~ delay), 724/zs (interpulse delay) and 300 ms (repetition time). NMR signal intensities for each metabolite were determined by numerical integration of corresponding peak areas in optimal computer fits to the spectra (Nalorac Quest 4400 Curve Fitting Program). 3~p metabolite concentrations were measured relative to corresponding signal intensities in the control run. Relative ATP levels were determined from signal intensities for the/3-ATP peak at 16.3 ppm. Fully relaxed spectra (20-s interpulse delay) were obtained in special studies to obtain relaxation time corrections for different metabolites. Lactate signal intensities were measured relative to those for NAA. In summary, each trinuclear run lasted 12 min. The pulse sequence for nuclear excitations, {31p_1H_[ 19F_ 19F lqF_ 19F]}4' was repeated 50 × with null amplitudes for ~H excitation (first 10 min) and then two more times with null amplitudes for 31p and t9F excitations (last 2 min). This resulted in the collection of 800 free induction decays (FIDs) for t~F, 200 for 3tp and 8 for ~H. Edema studies in slices were conducted as separate, parallel experiments, without NMR measurements. Slices were obtained exactly as in N M R studies and exposed to glutamate and antagonists in identical protocols. 27 brain slices were obtained, segregated from others and studied individually, using separate test tubes, so that initial wet weight could be correlated with final parameters. Edema experiments thus began with the isolation of 27 slices. Three slices were removed from separate test tubes every 15 rain after the onset of the protocol, i.e., at t = 0 min, t = 15 min and
Table 1 Relative intracellular Ca 2~ concentrations las '; ~t control} allc~ exposure of slices to 2 mm glutamate t (min)
('ontrol
2 mM glutamalc
0
100+11
30 60 120 (washout)
101 + 7 99+ 8 11)0+ 7
99+, 168± I I:~ 286+t,1:* 99-k 7
Data shown are mean + S.D. (n = 3 experiments). All NMR [Ca 2 + ]i determinations were normalized relative to average for control slices at t = 0. Multiple comparisons were made using Bonferroni-corrected t tests. Time points having significant between-group differences are marked with asterisks (* P < 0.0007, ** P < 0.O001 ).
up to t = 120 min. Slices were blotted lightly with filter paper to quickly remove surface water and transferred to a clean, preweighed, dry culture tube. After measuring wet weights, slices were dried at 110°C for 16 h and dry-weight determinations were made. Calculations were ,made for each slice of the increase in wet weight and the final % water content. Data obtained for slices removed at t - 0 (no glutamate) were then used to calculate the % tissue swelling. Definitions of the preceding terms and formulae for calculations have been described previously by one of the authors [5].
Or.COO
A
~tO
~.
Control
Glut~lm
D~zocUplne Kwmnmate
NIOX
Fig. 2. A: relative increases m water contenl at end of a 60 rain exposure to 2 mM glutamate (solid vertical bar) and. then 60 min later, after glutamate washout washout with glutamate-free ACSF (gray vertical bar). Control data are for a 60-min perfusion with Mg2+-free ACSF (solid vertical bar; glutamate N O T used), followed by 60 additional min of washout with Mg2+-containing ACSF (gray vertical bar). B: relative concentrations of PCr ( solid vertical bar) and ATP (gray vertical bar), each after 60 min of washout with glutamate-free ACSF, Data shown are mean :I:S.D. (n = 3 experiments). Multiple comparisons were made using Bonferroni-corrected t tests. Time points having significant between-group differences are marked with asterisks (* P < 0.0083, ** P < 0.001 }.
M.T. Espanol et aL / Brain Research 647 (1994) 172-176
Histological studies were conducted in brain slices taken during N M R studies at three time points: t = 0 min (control), t = 60 min (end of 60-min exposure to 2 mM glutamate) and t = 120 min (end of glutamate washout). Three slices were taken at each time point and transferred to tubes containing 10% formalin in 0.9% saline solution. The tissues were washed, dehydrated and embedded in paraffin. Adjacent sections, 10/xm thick, were cut in the plane of the slices, stained
175
with Cresyl violet (Nissl stain) and examined and photographed using a phase-contrast light microscope (Nikon Diaphot T M D inverted microscope). Numbers of shrunken and dead neurons were counted in fields of 50-140 adjacent cells, corresponding to areas of 10 4/zm 2. Data are reported as mean +_ S.E. and n indicates the numbers of separate experiments. Multiple comparisons between groups were made using Bonferroni-
Fig. 3. Two phase-contrast micrographs of Nissl stained sections from slices removed at t = 120 min, after 60 min of washout with normal ACSF. A: glutamate (2 m M ) alone was used in 60-min exposure period. B: glutamate (2 mM) and dizocilpine (150 ~ M ) were used in 60-min exposure period. Neurons exposed to glutamate alone are swollen as indicated by arrow in A. Qualitatively, dizocilpine treatment resulted in less swelling and in intact neuronal processes which are indicated by arrow in B. Scale (bar, 20/~m) is same for both micrographs.
176
M. 7: E.spanol et al. / Brain Research 047 ( I q94) 17.? - ! 7~
corrected t tests [11]. Differences between groups were considered significant for P < 0.001. Slices loaded with 5F-BAPTA had no discernible changes in their 31p and tH control spectra. Table 1 summarizes glutamate-induced % increases in [Ca 2+]~ after 30- and 60-min exposures to 2 mM glutamate and after 60 min of glutamate washout. These changes are in qualitative agreement with [Ca2+]i changes found in fluorescence studies of cerebrocortical slices from 7day-old rats [3]. Fig. 2A shows increases in water content (edema) of cerebrocortical slices after 60 rain of glutamate exposure (with and without antagonists). Fig. 2B shows corresponding changes in NMR determinations of ATP and PCr. Fig. 3 shows that neuronal swelling and axonal loss is worse in untreated slices. The fact that glutamate caused only small NMR changes in [Ca2+] i (without the use of glutamate antagonists) is consistent with a recent cerebrocortical slice study that used a fluorescent intracellular probe and found: (a) significant age dependence of the [Ca2+]i response to extracellular glutamate; and (b) a similarly small glutamate-induced increase in [CaZ+]~ in 7-dayold rats, possibly due to immaturity of NMDA receptors [3]. In our study, [Ca2+] i increases occurred late during glutamate exposure and reversed early during glutamate washout. It has been suggested that transient increases in [Ca2+]i in adult slices might cause injury by activating Ca2+-dependent phospholipases and other enzymes [8]. Although changes in [Ca2+] i might not entirely explain energy failure and injury in our neonatal brain slices, our studies demonstrate the feasibility of using noninvasive NMR spectroscopy to monitor [Ca2+]i changes in adult slices and of concurrently using invasive histological and immunochemical methods to link NMR data with early intracellular events. This research was supported by the following NIH Research Grants: GM NS 34767 (to L. Litt), NS22022 (to P.R. Weinstein), NS14543 (to P.H. Chan), NS25372 (to P.H. Chan), RR03841 (to T.L. James) and the Whitaker Foundation (LHC). We wish to thank Novo
Nordisk (Denmark) for providing NBQX and the Merck Corporation (NJ) for providing dizociJpinc. Wc arc also grateful to P. Morris, H. Bachclard and R. BadarGoffer for helpful suggestions regarding the use ~t 5F-BAPTA for NMR determinations of l e a : ' ]i.
[1] Badar-Goffer, R., Ben-Yoseph, O., Dolin, 8,J. el al., Use of 1,2-bis(2-amino-5-fluorophenoxy)ethane-N,N,N',N'-tetra-acetic acid (5FBAPTA) in the measurement of frec intracellular calcium in the brain by ~F-nuclear magnetic resonance spectroscopy, J. Neurochem., 55 (1990) 878-884. [2] Ben-Yoseph, O., Bachelard, H.S., Badar-Goffer, P,. el at.. F,i"fects of N-methyl-D-aspartate on [Ca 2~ ]i and the energy state in the brain by 19F- and 31P-nuclear magnetic res~mance spectroscopy, J. Neurochem., 55 (1990) 1446-14411 [3] Bickler, P.E., Gallego, S.M. and Hansen, B.M, Developmental changes in intracellular calcium regulation in rat cerebral cortex during hypoxia, J. Cereb. Blood Flow Metab.. 13 (1993) 8l 1-819. [4] Brooks, K.J., Porteous, R. and Bachelard, lt.S., Effects of hypoglycaemia and hypoxia on the intracellular pH of cerebral tissue as measured by 31p nuclear magnetic resonance: J. Nenrochem., 52 (1989) 604-610. [5] Chan, P.H. and Fishman, R.A., Brain edema: reduction in cortical slices by polyunsaturated fatty acids, Science, 21)1 (1978) 358-360. [6] Choi, D.W., Methods for antagonizing glutamate neurotoxicity, Cerebrouase. Brain Metab. Ret'., 2 (1990) 105-147. [7] Choi, D.W., Excitotoxic cell death. J. Neurobiol., 23 (1992) 1261-1276. [8] Coyle, J.T. and Puttfarcken, P., Oxidative stress, glutamate, and neurodegenerative disorders, Science, 262 (1993) 689-695. [9] Espanol, M.T., Litt, L., Yang, G.Y. et al., Tolerance of low intracellular pH during hypercapnia by rat cortical brain slices: a 31p/IH NMR study, J. Neurochern,, 59 (1992) 1820-1828. [10] Espanol, M.T., Xu, Y., Litt, L. et al., Kynurenate, dizocilpine, and NBQX differently modulate glutamate-induced energy failure and Ca ++ in live brain slices, Z Cereb: Blood Flow Metab.. 11 (Suppl.) (1993) $748. [11] Espanol, M.T., Xu, Y., Lilt, L. et al., Modulation of glutamateinduced intracellular energy failure in neonatal cerebral cortical slices by kynurenic acid, dizocilpine, and NBQX. J. Cereb, Blood Flow Metab., 14 (1994) 269-278. [12] Jacquin, T., Gillet, B., Fortin, G. et al., Metabolic action of N-methyI-D-aspartate in newborn rat brain ex vivo: 3~p magnetic resonance spectroscopy, Brain Res., 497 (1989) 296-304. [13] Smith, G.A., Hesketh, R.T., Metcalfe, J.C. et al., lntracellular calcium measurements by 19F NMR of fluorine-labeled chelators, Proc. Natl. Acad. Sci. USA, 80 (1983) 7178-7182~