means of the an iterative approach (Hazlewood et aI.,. 1974). The T, was also ..... Bakay L, Kurland RJ, Parrish RG, Lee JC, Peng RJ, Barto kowsky HM (1975) ...
Journal o/ Cerebral Blood Flow and Metabolism
7:552-556 © 1987 Raven Press, Ltd., New York
Intracellular and Extracellular Spaces of Normal Adult Rat Brain Determined from the Proton Nuclear Magnetic Resonance Relaxation Times
Munetaka Haida, Masahiro Yamamoto, Hideshi Matsumura, Yukito Shinohara, and *Minoru Fukuzaki Department of Neurology, Tokai University School of Medicine, Isehara, Kanagawa, and *Hasumi Institute for Cancer Research, Tokyo, Japan
Summary: The nuclear magnetic resonance method was used to investigate the state of water molecules in normal rat brain tissue in vitro. The transverse magnetization decay curve (TMDC) of the fresh brain tissue of adult rats (8- or 10-weeks-old) was biexponential, which could be interpreted in terms of two distinct transverse relax ation times (T2). Several factors that may affect the
TMDC are discussed. It was concluded that the fast and slow components of T2 correspond to those of the water molecules of the intracellular and the extracellular spaces of normal rat brain tissue, respectively. Key Words: Brain-Nuclear magnetic resonance-Rats-Extracel lular space.
It is well known that the state of water in biolog ical systems is different from that of bulk water be cause of various intermolecular interactions. Changes in the state of water caused by alterations in the protein-ion-water interactions have been suggested to be one of the major causes of in creases in the colloid osmotic potential of brain tissue (Tomita et ai., 1979). The behavior and states of water molecules in brain tissue sampled after death in animals have been studied in vitro by sev eral investigators using the nuclear magnetic reso nance (NMR) method (Bakay et ai., 1975; Naruse et aI., 1982; Go and Hommo, 1975). They reported that the transverse relaxation time (T2) of normal rat brain tissue is a single component. However, from the evidence that the Tz of the muscle tissue is multicomponent (Hazlewood et ai., 1974), it would be expected that the T2 of the brain is also multi component. This discrepancy may have arisen from
their experimental procedures, because they sam pled the brain after the animal had been killed. The state of water in the brain might change very rap idly after death as a result of ischemia (Hossmann, 1971). Therefore, the state of water in normal brain tissue still remains uncertain. The purpose of this study was to obtain informa tion regarding the states of water in normal adult rat brain by means of brain biopsy. In the present study, we mainly made use of the T2 of NMR, since it is believed to be more sensitive to the environ mental changes around water molecules than the longitudinal relaxation time (T1) ( Farrar and Becker, 1971).
MATERIALS AND METHODS The T2 of water protons were obtained by Hahn's spin echo method. The free induction decay of the echo signal was Fourier transformed to obtain the water peak. The transverse magnetization decay curves (TMDC) of water signals for rat brain tissue and albumin solutions were measured in vitro at 2SOC with 100 mHz Fourier-trans form, NMR equipment (Nihon Denshi PS 100 and FT 100 system) with the accumulation of four successive echo signals. The repetition time was 8 s. The sampling delay time was 200 ms. The measurements were started within 1 min after tissue preparation and were completed within 15 min. Nine different echo times 2T (0, 80, 160,240,320, 400, 480, 560, 640 ms) were used to obtain TMDCs.
Received March 1987; accepted May 1987. Address correspondence and reprint requests to Dr. M. Haida at the Department of Neurology, Tokai University School of Medicine, Isehara, Kanagawa 259-11, Japan. A part of this work was presented at the 11th Symposium on Cerebral Blood Flow and Metabolism, Paris, 1983. Abbreviations used: NMR, nuclear magnetic resonance; T MDC, transverse magnetization decay curve.
552
RAT BRAIN INTRA- IEXTRACELLULAR SPACES BY NMR
553
It is well known that the TMDC follows an equation (Abragam, 1978): 2 , -y1DG tT1 A Mo exp - tl1 (I) 1 3
(
=
)
where A is the amplitude of the spin-echo signals, Mo is magnetization at t 0, t is sampling time (echo time), -y is the gyromagnetic ratio,D is the diffusion constant,Gis magnetic field gradient at the specimen, and T is the time interval between the 900 pulse and the 1800 pulse. Based on the relation t 2T of Hahn's spin-echo method, equa tion (1) was simplified by introducing a constant k l (k 2-y1DG /3). For monoexponential TMDC =
=
=
A
=
Mo exp( - 2TIT2 - kT3)
sample tube
(2) sample
and for biexponential TMDC A
=
MOf exp( - 2TITlf - k[T3) + Mos
exp( - 2TIT2s - ksT3)
seal
(2a)
where suffixes f and s denote the fast and slow compo nents, respectively. The parameters were obtained by fitting the TMDC to equation (2) or (2a) by the least-squares method, and MOf and MoS' T2f, and Tl5' and kf and ks were determined by means of the an iterative approach (Hazlewood et aI. , 1974). The T, was also determined by means of the inver sion recovery method with seven different inversion times (0, 100, 300, 600, 900, 1,200, 1,500 ms) and the rep etition time of 10 s for normal adult rats (n 10). This study used 20 normal Wistar rats (250-300 g), 8 weeks (n 6) and 10 weeks (n 14) old. Under light anesthesia by intraperitoneal injection of urethane (500 mg/kg body weight) and chloralose (50 mg/kg body weight), a hole was drilled in the right frontoparietal re gion of the rat skull. The dura and arachnoid were care fully incised. A 2-mm-thick brain tissue sample that was mainly gray matter was punched out with a sharp-edged Pyrex glass tube. The NMR measurements were made on each sample in a Pyrex glass tube sealed tightly with a waterproof material,Hematoseal (a compound for hemat ocrit measurements that was confirmed not to affect the NMR signals), and coaxially placed in an outer glass tube containing deuterium oxide (D 0) to stabilize the mag 1 netic field, as shown in Fig. !. Among the 20 rats, the 14 lO-week-old adult rats were used as a control group. Brain tissue samples were obtained by means of the following three different procedures from 8-week-old rats (n 6). Procedure a. Under light anesthesia, a hole was drilled in the skull, then the brain tissue was sampled as de scribed (brain biopsy) (n 2). Procedure h. Under light anesthesia, a hole was drilled in the skull as described. The rat was then killed by intra venous injection of KCl solution, and sampling was per formed immediately after injection (n 2). Procedure c. The rat was killed by KCl injection, then a hole was drilled in the skull, and the brain tissue was sampled usually -5 min after injection (brain necropsy) (n 2). To determine the effect of protein concentration on the T2 values of the slow and fast components, T2 measure ments were also performed on a series of human albumin solutions of various concentrations (0.5-500 giL). Sera of the lO-week-old rats were separated from hepa=
=
outer tube
FIG. 1. Sample tubes and specimen.
rinized blood samples that were obtained by venopunc ture of main vessel of rats.
RESULTS
=
Figure 2 is an example of the transverse magneti zation decay curve (TMDC) of 10-week-old rat brain tissue. The vertical axis shows the amplitude of the echo signal (arbitrary unit) on a logarithmic scale, and the horizontal axis shows 2T, where the T is the time interval between the 90° and 180° pulses. This curve is biexponential, as shown in Fig. 2, and this finding was consistently observed in 18 of the 20 rats (two rats subjected to procedure c gave monoexponential curves). The T2 values of the fast and slow components (means ± SD) obtained from 10-week-old rats were
=
=
=
=
+C ::J ...: :0 L 0 C '=>f-
U)
Z lJ.J f-
;;:;
0
200
400
600
27:(msec) FIG. 2. Transverse magnetization decay curve for 10-week old rat brain tissue.
J
Cereb Blood Flow Me/ab, Vol.
7,
No. 5,
1987
M. HAIDA ET AL.
554
57.0 ± 2.8 and 173.7 ± 70.2 ms, respectively (n 14). The fraction of the slow component with re spect to the total, MOf/(Mof ±Mos), was 4.05 ± 3.0% (n = 14). The T) of the normal rat brain tissue was a monocomponent, and its value was 1240. 2 ± 71.2 ms (n = 10). The results of the three different sampling proce dures for the 8-week-old rats are shown in Fig. 3a-c, respectively. The TMDCs for the samples obtained by procedures a and b were biexponential. On the other hand, the TMDC for the sample ob tained by procedure c was monoexponential. The relationship between T2 and the concentra tion of albumin solution is shown in Table 1. Higher concentrations of albumin correspond to shorter T2 values. Figure 4 shows, as an example, the TMDC of albumin solution of 500 giL. The curve is mono exponential with a slight diffusion effect, and its t2 value is 55.1 ms. Monoexponential dependency was seen throughout the whole range of albumin concentrations tested. The T) of rat serum was 2122.6 ± 234.4 and Tz was 396.8 ± 34.5 (n = 10).
DISCUSSION
=
:g:::l ...;
15
(a)� > f-
iI) Z W f-
� 0�----1� 07 0---- �20�0�-- �3�0�0�-- �04 '0 2 "t"(msE.'e)
-
C
:::l ...;
15 (b)
� > f-
in
z W fZ L-__��____��__��__
o
100
200
2
1:
300
� 400
(msE.'e )
C
:::l ...;
The present results show that the TMDCs for the brain tissue of 10-week-old rats are of multiexpo nential type and can be well interpreted as by biex ponential. Several factors affecting TMDCs have t? be considered. They can be summarized as expen mental artifacts, bound water and free water, tissue inhomogeneities (spatial inhomogeneities of the T2 relaxation times of the brain tissue, etc.), and water in the intracellular space and extracellular space of the brain tissue. Experimental artifacts. Since this experimental arrangement has coaxially arranged double lumens, a question may arise concerning the nature of these two components. One possibility is that the slow component of TMDC (Fig, 2) represents the signal of trace water contaminating the D20 in the outer tube. Such a signal from traces of water, if any, in the outer tube can be eliminated by adding a little CUS04 powder (dehydrated by baking) to the D20 in the outer sample tube. When done in this experi ment, the relaxation time of the slow component was unaffected, and the biexponential nature of the curve was retained. Thus, it can be concluded that the slow component of TMDC (Fig. 2) does not arise from traces of water in the D20 in the outer tube. Bound water andfree water. Water bound to high polymers has short T2 compared with free water. If exchange of the water molecules between these two states cannot occur within their relaxation times, biexponential TMDC would be expected. If the exchange between the two states occurs very rapidly, the sample would have a monoexponential TMDC with a relaxation time corresponding to the weighted average of the relaxation times of the two states, based on the fast exchange model of Zim mermann and Brittin (1957). Since the TMDCs of albumin solutions are monoexponential, their ex change must be very rapid. The bound water has an extremely short T2, �400 J..LS (Hazlewood et aI., 1974), and the instrumentation used in this experi ment could not detect such a fast decay. Moreover,
15
:u (e);;
TABLE 1. Transverse relaxation time
f-
in
�
- 0�--1�OO:-:---;;-20*'0�-�30�0:;----;4�0:r;0-' 2-r: (msE.'c)
FIG. 3. Transverse magnetization decay curves for a-week old rat brain tissue obtained by brain biopsy (procedure a), sampling (procedure b), immediately after the rat was killed . and brain necropsy (procedure c). For details, see Matenals and Methods section.
J
(T2) versus
albumin concentration
z w
Cereb Blood Flow Metab, Vol. 7, No.5, 1987
Albumin concentration 0.5 25 50 100 250 500
(giL)
T2 value
(ms)
2210 1490 986 473 202 55.1
RAT BRAIN INTRA- IEXTRACELLULAR SPACES BY NMR
C
:J
15
�
>f-
Vi
Z W fZ
•
100
a 21:"
(m5ec)
FIG. 4. The transverse magnetization decay curve of al bumin solution with a concentration of 500 giL.
if the T2f and T2s originated from bound water and free water, respectively, the monoexponential TMDC obtained by procedure c could not be ex plained. Tissue inhomogeneities. The T2 relaxation times of brain tissue may have some distributions be cause of tissue inhomogeneities. Widely distributed T2 values could make the TMDCs multiexponen tial, as suggested by Chang and Hazlewood (1975). Because exponentiality should be consistent in pro cedures a-c a possibility of the multiexponentiality arising from the histological inhomogeneities, in cluding different cell types of the brain tissue, can be ruled out on the basis of the monoexponential TMDC obtained by procedure c. Such inhomogen eities could arise, however, from several types of compartmentalization, as mentioned by Fullerton et al. (1982). Fullerton and colleagues proposed three types of compartmentalization in the tissue, i. e. , fundamental, compositional, and spatial com partmentalization. The fundamental compartmen talization was already discussed under bound water and free water, and the compositional compartmen talization can be ruled out by the procedures used in this experiment, which can separate the signal of water from that of the fat in the frequency spec trum. Spacial compartmentalization would be the most reasonable model to explain biexponential TMDCs in this experiment. In this model, there are two compartments in the brain tissue, giving two peaks in the T2 distribution. Macroscopic spatial compartmentalization, as mentioned by Fullerton et al. (1982), would be composed of the intracellular and extracellular spaces separated by the cell mem brane. This model will be discussed in the following in detail. Water of intracellular space and extracellular space. If there are two compartments separated by some barrier and the residence time of water mole cules in one compartment is long enough compared with the relaxation time of the other compartment, the TMDCs would be of biexponential type ex pressed by equation (2a). If exchange of the water
555
molecules between these two compartments is very rapid, TMDC of monoexponential type with a T2 value equivalent to the weighted average of T2 of each compartment would be expected (Zimmer mann and Brittin, 1957). The residence time of red blood cells (RBC) was sufficiently short to show monoexponential TMDC (Fullerton et aI. , 1982) and that of water of the in tracellular space of muscle tissue was relatively long, giving a biexponential TMDC (Belton et aI., 1972; Hazlewood et aI. , 1974). It seems reasonable for RBCs to have such a high membrane perme ability to water molecules because of their function of gas exchange. However, other tissues including muscle and brain appear to have longer residence time on the basis of the experimental finding of biexponential TMDCs of these tissues. Four compartments may be considered for the water molecules in the rat brain tissue (gray matter) by this model; intracellular (ic) space, extracellular (ec) space excluding intravascular space, intravas cular (iv) space, and cerebrospinal fluid space. Any significant contribution from cerebrospinal fluid space could be ruled out for the following two reasons. First, our sampling was limited mainly to the gray matter of rat brain tissue. Second, the T2 value of the slow component is equivalent to that of albumin solution with a concentration of �250 giL, which is too concentrated to correspond to cere brospinal fluid. The T2 value of the fast component of the TMDC of 10-week-old rat brain tissue was �60 ms, and this value coincides with the T2 value of albumin solution of 500 giL. The fast component may be assigned to the intracellular space of the brain tissue, because the intracellular space should have such a high protein concentration on the basis of a simple estimation according to previously pub lished data (Chao and Rumsby, 1977; Snyder et aI. , 1980). The slow component of TMDC (Fig. 2) can thus be assigned to the extracellular space or the intravascular space. However, the serum has a longer relaxation time as shown in the result, so the intravascular space can be ruled out. A monoexponential curve could be obtained by changing the experimental procedure from proce dure a or b to procedure c (see Materials and Methods), as shown in Fig. 3. The monoexponen tial curve of Fig. 3c can eliminate a possibility that the slow component also came from the intracel lular space of the cells other than glial or neural cells and be well explained by the shift of water molecules from the extracellular space to the intra cellular space that would occur after the animals were killed (Hossmann, 1971). This shift of water from the extracellular space to the intracellular
J
Cereb Bluud Fluw Melab, Vul. 7, No.5, 1987
556
M. HAIDA ET AL.
space, probably associated with cell swelling, may greatly reduce the volume of the extracellular space or reduce the Mo, below a detectable level, re sulting in the observation of monoexponential TMDC. The volume of the extracellular space has been estimated as 10-38% by various investigators with marked variation depending on the methods used, as discussed by Katzman and Pappius (1973). The results of this experiment show that the Mos fraction was �4%, which is rather small compared with the previously reported values. This can be explained by the shift of the water molecules from the extracellular space to the intracellular space after the sampling, which reduces the extracellular space (Haida et al. , 1986). The finding of a biexponential TMDC is contrary to the report by Naruse et al. (1982) who found that the TMDC for normal rat brain tissue is monoex ponential. This discrepancy may have arisen for the following two reasons. First, there is the difference in the sample preparation procedures. The results of Naruse et al. (1982) are compatible with the re sults of this study obtained by procedure c (i. e. , brain necropsy, with the tissue sample taken after the animals were killed). Second, the difference of experimental procedures should be stressed. In this experiment, the signals were averaged four times and the signal intensity of the water was evaluated based on the frequency spectrum, not based on the free induction decay. The procedures in this study should give a better signal/noise ratio, which would make it possible to detect a small amount of Mos or the biexponential TMDCs. Thus, it was concluded that the fast and slow components of the TMDC observed in normal adult rat brain tissue could be assigned to the signals from the water of the intracellular and the extracel lular spaces, respectively.
J
Cereb Blood Flow Metab, Vol. 7, No.5, 1987
REFERENCES Abragam A (1978) The principles of nuclear magnetism. Oxford: Clarendon Press, pp 61 Bakay L, Kurland RJ, Parrish RG, Lee JC, Peng RJ, Barto kowsky HM (1975) Nuclear magnetic resonance studies in normal and edematous brain. tissue. Exp Brain Res 23:241248 Belton PS, Jackson RR, Packer KJ (1972) Pulsed NMR studies of water in striated muscle. I Transverse nuclear spin relax ation times and freezing effects. Biochim Biophys Acta 286:16-25 Chang DC, Hazlewood CF (1975) Comment on PMR studies of tissue water. J Magn Resonance 18:550-554 Chao SW, Rumsby MG (1977) Preparation of astrocytes, neurones and oligodendrocytes from the same rat brain. Brain Res 124:347-351 Farrar T C, Becker ED (1971) Pulse and Fourier transform NMR. New York: Academic Press, pp 46-52 Fullerton GD, Potter JL, Dornbluth NC (1982) NMR relaxation of protons in tissues and other macromolecular water solu tions. Magn Reson Imag 4:209-226 Go KG, Hommo TE (1975) Water in brain edema. Arch Neural 32:462-465 Haida M, Yamamoto M, Taniguchi R, Ohsuga H, Fukuzaki M, Shinohara Y (1986) A rapid shift of water from extracellular space to intracellular space determined by NMR. Neurology 36 ( Suppl 1): 175 Hazlewood CF, Chang DC, Nichols BL, Woessner DE (1974) Nuclear magnetic resonance transverse relaxation times of water protons in skeletal muscle. Biophys J 14:583-606 Hossmann KA (1971) Cortical steady potential, impedance and excitability changes during and after total ischemia of cat brain. Exp Neurol 32:163-175 Katzman R, Pappius HM (1973) Brain Electrolytes and Fluid Metabolism. Williams & Wilkins, pp 33-47 Naruse S, Horikawa Y, Tanaka C, Hirakawa K, Nishikawa H, Yoshizaki K (1982) Proton nuclear magnetic resonance studies on brain edema. J Neurosurg 56:747-752 Snyder D, Stephen RC, Farooq M, Norton WT (1980) T he bulk isolation of oligodendroglia from whole rat forebrain: a new procedure using physiologic media. J Neurochem 34:16141621 Tomita M, Gotoh F, Sato T, Yamamoto M, Amano T, Tanahashi N, Tanaka K (1979) Determination of the osmotic potential for swelling of cat brain in vitro. Exp Neurol 65:66-77 Zimmermann JR, Brittin WE (1957) Nuclear magnetic resonance studies in multiple phase systems: life time of water mole cules in an absorbing phase on silica gel. J Phys Chem 6:1328-1333
