Journal of Neuroscience Methods 79 (1998) 115 – 121

Changes in blood – brain barrier permeability following neurotoxic

lesions of rat brain can be visualised with trypan blue

D.S. Reynolds, A.J. Morton *

Department of Pharmacology, Uni6ersity of Cambridge, Tennis Court Road, Cambridge CB2 1QJ, UK

Received 18 July 1997; received in revised form 6 October 1997; accepted 8 October 1997

Abstract

A simple method for measuring changes in blood–brain barrier (BBB) permeability following neurotoxic lesions is described.
In the brains of animals perfused transcardially with a trypan blue solution at the time of sacrifice, the presence of trypan blue
staining correlated with changes in BBB function seen with more traditional markers, such as albumin staining. Thus, trypan blue
appears to be useful as a marker for changes in BBB permeability. We have used this method to show increases in BBB
permeability in striatal lesions induced by three different neurotoxins: chronic systemic injection of 3-nitropropionic acid (3-NP)
and intrastriatal injection of either quinolinic or kainic acid. Trypan blue staining was seen in all three types of lesion, with both
the neuropil and some neurones being stained. In the kainic acid lesioned animals, trypan blue also stained hippocampal and
cortical neurones which are known to degenerate. Our findings suggest that trypan blue makes a more sensitive marker than
albumin for both BBB integrity changes and degenerating neurones. Furthermore, this method has the advantages over others of
being quick, economic and compatible with most subsequent histological and immunocytochemical staining. © 1998 Elsevier
Science B.V.

Keywords: Blood–brain barrier; Excitotoxicity; Neurodegeneration; 3-Nitropropionic acid; Rat; Striatum; Trypan blue

1. Introduction partate (NMDA; Dietrich et al., 1992), intrastriatal

injection of NMDA or endothelin-1 (Miller et al., 1996)
The mechanisms of neuronal death following trau- and 3-nitropropionic acid (3-NP; Nishino et al., 1995,
matic brain injury or neurotoxic insults are not fully 1997). Such changes in BBB permeability may con-
understood but changes in blood – brain barrier (BBB) tribute to the process of neurodegeneration. Therefore,
permeability may play an important role. Increases in a simple measure of BBB function would facilitate
BBB permeability have been demonstrated in a number experimental studies.
of different types of clinical brain damage, including The integrity of the BBB was first measured using in
cerebral ischemia (Olssen et al., 1971; O’Brien et al., vivo injection of dyestuffs, such as Evan’s blue and
1974; Petito et al., 1982; Nordborg et al., 1991, 1994) trypan blue (Broman, 1944; Broman et al., 1966). How-
and subarachnoid haemorrhage (Lindsberg et al., ever, these dyes were found to bind plasma proteins
1996). Increased permeability of the BBB has also been (Gregersen and Rawson, 1942; Tschirgi, 1950), and lose
shown in experimental models of neurodegeneration, their colour intensity during chemical reactions in the
including kainic acid-induced seizures (Zucker et al., tissue. As a result controversy arose as to their reliabil-
1983), intraventricular infusion of N-methyl-D-as- ity for measuring BBB integrity (for review see Dob-
bing, 1961). These difficulties led to the development of
* Corresponding author. Tel.: +44 1223 334057; fax: + 44 1223 alternative methods, including the in vivo injection of
334040. horseradish peroxidase (HRP, Reese and Karnovsky,

0165-0270/98/$19.00 © 1998 Elsevier Science B.V. All rights reserved.

PII S 0 1 6 5 - 0 2 7 0 ( 9 7 ) 0 0 1 6 8 - 4
116 D.S. Reynolds, A.J. Morton / Journal of Neuroscience Methods 79 (1998) 115–121

1967; Petito et al., 1982; Broadwell and Sofroniew, (3 nmol) was injected into the left striatum, in a 1 ml
1993), radiolabelled tracers (Zucker et al., 1983; Pont et volume, at a constant rate for a period of 2 min. The
al., 1995) or fluorescent dyes (Miller et al., 1996). coordinates used were anterior 1 mm and lateral 2.5
Additionally, immunocytochemistry has been widely mm from the bregma and ventral 5 mm from the dura
used to reveal albumin extravasation (Hamilton and mater (Paxinos and Watson, 1982). After each injection
Gould, 1987b; Nordborg et al., 1991; Broadwell and the needle was then left in situ for 2 min before being
Sofroniew, 1993; Nordborg et al., 1994; Nishino et al., slowly withdrawn before the incision was closed with
1995). However, although these methods give satisfac- interrupted silk sutures. Rats were sacrificed 1 week
tory results, difficulties arise when double labelling after lesioning.
studies are attempted. For example, perfusion with
HRP is not readily compatible with subsequent process- 2.3. Transcardial perfusion of trypan blue
ing of the brain for immunocytochemistry because the
brain cannot be perfuse-fixed with paraformaldehyde. All rats were perfused with a trypan blue solution
Furthermore, these methods require in vivo treatments and then fixed with paraformaldehyde. Trypan blue
of the animals prior to sacrifice and (particularly for solution (0.5%) was made by dissolving 1 g of trypan
HRP) are expensive. blue in 200 ml PBS, with gentle heat. The solution was
In this study we show that transcardial perfusion of allowed to cool to room temperature and then filtered
trypan blue at the time of sacrifice makes a reliable through cotton wool. Heparin (500 units) was added to
marker for BBB permeability changes in rat models of the filtrate, then the solution was placed on ice and
striatal neurodegeneration. There are several advan- used immediately. The temperature of the trypan blue
tages of using this system: no in vivo pre-treatment of solution was 10–12°C at the time of perfusion. Rats
the animal is required because no peroxidase is in- were anaesthetised with Avertin and perfused transcar-
volved, fixation with paraformaldehyde is possible and dially with 200 ml of trypan blue solution, followed by
trypan blue does not interfere with subsequent histolog- 300 ml of ice-cold paraformaldehyde (2% in PBS). The
ical and immunocytochemical staining. flow rate of the perfusate was maintained at 25 ml
min − 1. The brains were dissected and post-fixed
overnight in 2% paraformaldehyde and then cryopro-
2. Methods tected in 30% sucrose for 2 days. Subsequently, brains
were frozen in powdered dry ice and stored at −80°C
2.1. Systemic 3 -NP treatment until processed for histochemical and immunocyto-
chemical studies. Trypan blue staining was examined in
3-NP for injection was dissolved in phosphate 50 mm thick cryosections cut directly onto gelatin-
buffered saline (PBS, 100 mM) and adjusted to pH 7.4 coated slides. Trypan blue staining readily washed out
with NaOH, fresh stocks were made up each day. 3-NP of sections during processing and mounting. To min-
was administered by daily subcutaneous injection to imise this problem slides were rapidly dehydrated (three
rats at an initial dose of 12 mg kg − 1. After every 4 days washes, 30 s each) in alcohol, defatted in Histolene
of treatment, the dose was increased by 3 mg kg − 1 (Cellpath) and coverslipped with DPX mountant (Agar
until the rats developed a movement disorder character- Scientific). Adjacent sections (30 mm) were cut for
ised by a wobbly, uncoordinated gait which progressed immunocytochemistry and histochemistry. Trypan blue
to hind limb recumbency. At this point 3-NP treatment staining washed out of the sections during subsequent
ended and the rats were allowed to recover (this move- staining and hence did not interfere with these pro-
ment disorder is well correlated with striatal lesion cesses.
formation (Hamilton and Gould, 1987b; Beal et al.,
1993). For treatment 22 female Sprague-Dawley rats 2.4. Albumin immunocytochemistry
(200 –250 g) were used until they developed the charac-
teristic movement disorder. After the cessation of 3-NP Free-floating cryostat sections were stained for albu-
injections, rats were allowed to recover for up 4 weeks min immunoreactivity. Non-specific binding was
before being sacrificed. Control animals were treated blocked by incubating the sections at 4°C overnight in
with PBS alone, instead of 3-NP (n =18). blocking solution (3% normal deer serum in PBS con-
taining 0.2% Triton X-100) with 0.02% sodium azide
2.2. Intrastriatal injection of quinolinic or kainic acid added. Sections were then incubated in sheep poly-
clonal albumin antisera (Serotec) diluted to a concen-
Rats were deeply anaesthetised with Avertin (10 ml tration of 1:1000 in blocking solution for 1 week at
kg − 1, i.p.) and placed in a stereotaxic frame (Kopf) 4°C. Then sections were washed five times; for 5 min
with the incisor bar 3.3 mm below the interaural line. each; in wash solution (PBS containing 0.02% Triton
Using a 10 ml Hamilton syringe, QA (100 nmol) or KA X-100) and then incubated at 4°C overnight in HRP-la-
 D.S. Reynolds, A.J. Morton / Journal of Neuroscience Methods 79 (1998) 115–121 117

Fig. 1. Parallel coronal brain sections at the level of the striatum from a rat treated with either saline (a, c and e) or 3-NP (b, d and f). The cortex
(ctx), corpus callosum (cc) and striatum (st) have been indicated. Transcardial perfusion with trypan blue 1 week after the end of treatment
revealed very little staining in the saline-treated rat brain (a). In contrast, the 3-NP-treated brain displayed bilateral striatal lesions (arrows, b).
The increase in trypan blue staining correlated with an increase in albumin immunoreactivity in a parallel section (arrows, d). The area marked
by trypan blue and albumin staining corresponded to the lesion area, as shown by a decrease in NADPH diaphorase activity (arrows, f). There
was no specific albumin staining in control brain sections (c), nor did NADPH diaphorase staining reveal lesions. Bar =2 mm.

belled rabbit anti-sheep secondary antibody (Sigma) Sections for cresyl violet staining were mounted onto
diluted to 1:1000 in blocking solution without azide. gelatin-coated slides, dehydrated through graded alco-
The sections were then washed five times, for 5 min hol and defatted in Histolene for 20 min. They were
with wash solution and developed with 3,3%-di- then incubated in a 1% solution of cresyl violet acetate
aminobenzidine (0.5 mg ml − 1) in 50 mM Tris buffer for 30 min, differentiated in 1% acetic acid in 70%
(pH 7.6), containing 0.009% hydrogen peroxide. aqueous alcohol, before being dehydrated and mounted
in DPX.
2.5. NADPH diaphorase and cresyl 6iolet
histochemistry
3. Results and discussion
The presence and position of lesions was determined
using nicotinamide adenine dinucleotide phosphate
(NADPH) diaphorase and cresyl violet histochemical In saline-treated control rat brain sections (Fig. 1(a))
staining. NADPH diaphorase staining was performed there was no trypan blue staining visible in any region,
on free-floating cryostat sections by incubating them except in the circumventricular organs which lie outside
for 60 min in the dark, at 37°C in a reaction mixture the BBB (Fig. 2). The brain regions displayed intense
containing 50 mM PBS, 5 mM MgCl2, 2 mg ml − 1 cellular staining, as well as paler neuropil staining
NADPH (reduced form) and 1 mg ml − 1 nitro blue which had partially diffused into adjacent areas. Paral-
tetrazolium. lel sections stained for albumin revealed similar staining
118 D.S. Reynolds, A.J. Morton / Journal of Neuroscience Methods 79 (1998) 115–121

Fig. 2. Both trypan blue (a and c) and albumin immunocytochemistry (b and d) stained the circumventricular organs, which lie outside the BBB.
The subfornical organ (SFO) displayed strong trypan blue cellular staining (a), with some of the dye diffusing into the neighbouring ventral
hippocampal commissure (vhc). The choroid plexus (ChP) also stained strongly with trypan blue. Similarly, albumin immunoreactivity was seen
in the SFO and ChP (c). The arcuate nucleus (Arc) and median eminence (ME) stained intensely for both trypan blue (b), and albumin (d). D3V,
dorsal third ventricle; 3V, third ventricle. Bar = 500 mm.

only in the circumventricular organs (Fig. 1(c) Fig. 2). bodies within the striatal lesions (data not shown). This
Animals treated with 3-NP displayed bilateral striatal increased permeability to trypan blue persisting for at
lesions, whereas those treated with either quinolinic or least 4 weeks, although the amount of staining de-
kainic acid only had lesions in the striatum on the side creased between 1 and 4 weeks after lesion formation.
in which the toxin was injected (Fig. 1(d) Fig. 3). Similar results have been found previously, using albu-
Throughout the lesions cresyl violet staining showed a min immunocytochemistry (Hamilton and Gould,
marked loss of neuronal cell bodies and an increase in 1987a) These workers found that at 1 h after the onset
glial cells (Coyle and Schwarcz, 1976; Beal et al., 1986, of clinical symptoms of lesion development there was
1993). NADPH diaphorase activity was markedly de- no gross albumin immunoreactivity, although there was
creased within the lesion area and the boundary of the specific staining surrounding the capillaries within the
lesion was well delineated by this stain (Fig. 1(f), Fig. lesion area. After 2.5 h there was strong immunoreac-
3(c d)). There was intense bilateral trypan blue staining tivity throughout the lesion area at a gross level.
within the striatal lesions of 3-NP-treated rats (Fig. Nishino and colleagues have also shown that BBB
1(b)) and unilateral staining in those treated with both permeability is maintained for several weeks following
quinolinic or kainic acid (Fig. 3(a b)). Parallel sections 3-NP treatment (Nishino et al., 1995, 1997). Hence, the
from these animals also showed intense albumin stain- increase in trypan blue staining seen during striatal
ing throughout the lesions (Fig. 1(d)), suggesting that lesion formation in our study is likely to be due to an
trypan blue was a good marker of changes in perme- increase in BBB permeability.
ability of the BBB. In addition to the clear delineation of the lesion area
Damage to the integrity of the BBB occurred rapidly at a macroscopic level, trypan blue staining revealed
after 3-NP-induced lesion formation. The earliest time interesting details at a microscopic level. Within the
point we examined was 2 h after the onset of clinical striatal lesions induced by both chronic and acute treat-
symptoms. In these rats there was a marked increase in ments (Fig. 4(b c)), trypan blue strongly stained the
trypan blue staining of both the neuropil and cell neuropil and cell bodies, but not the fibre bundles. The
 D.S. Reynolds, A.J. Morton / Journal of Neuroscience Methods 79 (1998) 115–121 119

Fig. 3. Coronal rats brain sections from rats that received intrastriatal injections of either quinolinic acid (a, c and e) or kainic acid (b, d and f).
The trypan blue staining (a and b) was intense staining throughout the injected striatum (left side) after both treatments. Following quinolinic acid
injection, the neurodegeneration is restricted to the striatum, as shown by a loss of NADPH diaphorase staining (c). However, kainic acid injection
causes neuronal loss from extrastriatal areas, including the deep layers of the overlying cortex (arrowheads, d), and the CA3 cells of the ipsilateral
hippocampus (arrowheads, f). Trypan blue marks these dying neurones. Extrastriatal damage is not a feature of quinolinic acid injection and no
hippocampal neurones stain for trypan blue (e), although a few cortical neurones are marked. L, lesion; Hf, hippocampus; ME, median eminence.
Bar =2 mm.

neuropil staining was heterogeneous, with darker areas lesions. Following kainic acid injection, the ipsilateral
of staining distributed predominantly around the edges CA3 cells of the hippocampus stained positively for
of the lesions. Intensely stained cells were visible within trypan blue, whereas the contralateral hippocampal
the dye-stained neuropil, with the number, size and neurones were unstained (Fig. 3(f) Fig. 4(d)). Further-
morphology of cells suggesting that they are neurones. more, the deep layer cortical neurones overlying the
Since neurones marked by trypan blue are those damaged striatum were also dye-positive, but the con-
within the area that degenerates in these models, it tralateral cortical neurones were unstained (Fig.
seems likely that they are damaged. One can therefore 3(b)).This is consistent with the site of distal damage
speculate that the dye should mark other neurones that reported by other groups in the hippocampus and
are susceptible to degeneration. One such group of cortex (Wuerthele et al., 1978; Schwob et al., 1980;
neurones is the CA3 cells of the hippocampus, which Zaczek et al., 1980). To test that this staining was not
degenerate following kainic acid injection (Wuerthele et just due to damage to the neurones during the perfu-
al., 1978; Schwob et al., 1980; Zaczek et al., 1980). To sion procedure, the hippocampus of quinolinic acid
test this hypothesis we examined the CA3 hippocampal lesioned rats was also examined. Extrastriatal damage
neurones and the deep-layer cortical neurones in the is not a common feature of intrastriatal quinolinic acid
brains of rats that had received intrastriatal kainic acid injection and accordingly there was no trypan blue
120 D.S. Reynolds, A.J. Morton / Journal of Neuroscience Methods 79 (1998) 115–121

Fig. 4. Perfusion of trypan blue produced precipitate artefacts in blood vessels (a). These were clearly non-cellular and could be easily
distinguished from specific labelling of degenerating neurones. Within both the 3-NP (b) and kainic acid-induced (c) striatal lesions there was
heterogeneous neuropil staining with the fibre bundles remaining unstained (asterisks). Intensely stained cells could be seen within the dye-positive
neuropil; their number, size and morphology suggested that they are neurones. Following intrastriatal injection of kainic acid, the ipsilateral CA3
hippocampal cells degenerate and these neurones stain for trypan blue (d). Bar =100 mm.

labelling of hippocampal neurones in any of the quino- occurred. Mild damage to striatal blood vessels may
linic acid lesioned brains (although a few cortical neu- explain the increased number of trypan blue precipi-
rones were stained, particularly in the piriform cortex tates in 3-NP intoxicated rat brains. Similar artefacts
(Fig. 3(e))). Trypan blue appears to be a sensitive were found in the study of Broadwell and Sofroniew
marker of neurodegeneration. In fact it may be a better (1993) using albumin extravasation, but these were also
marker for changes in BBB integrity in distal regions clearly distinguishable from staining of BBB break-
than albumin staining, since it is very difficult to visu- down. With our method precipitate inclusions could be
alise hippocampal damage using albumin immunocyto- minimised by ensuring that the trypan blue solution
chemistry. cooled slowly and was used immediately after filtration.
Although occasional artefacts were seen following Our data suggest that transcardial perfusion of rats
transcardial perfusion with trypan blue, these were very with a trypan blue solution at the time of sacrifice
different and easily distinguished from the trypan blue provides a reliable marker of BBB permeability. Trypan
staining associated with lesions. The most common blue stains only regions outside the BBB in control rats,
artefactual staining appears to be largely due to precip- but stains neuropil as well as cells within the area of
itates of trypan blue becoming lodged in small blood striatal damage. In addition, in the kainic acid model,
vessels (Fig. 4(a)). An increased number of such inclu- distal regions that undergo degeneration (such as the
sions were found in 3-NP-treated rat brains, compared hippocampus) are also stained. The trypan blue method
to saline-treated brains. Nishino et al. (1995) noted that is an improvement over existing methods for examining
after prolonged 3-NP treatment, rats that did not de- BBB permeability, because it is sensitive and can be
velop striatal lesions displayed immunoreactivity for performed using standard perfusion techniques. More-
vascular elements around striatal blood vessels, suggest- over, trypan blue staining does not interfere with most
ing that a mild dysfunction of the BBB had already subsequent histochemical or immunocytochemical pro-
 D.S. Reynolds, A.J. Morton / Journal of Neuroscience Methods 79 (1998) 115–121 121

cedures, thus making it the ideal marker for BBB Lindsberg PJ, Ohman J, Lehto T, et al. Complement activation in the
central nervous system following blood – brain barrier damage in
permeability studies.
man. Ann Neurol 1996;40:587 – 96.
Miller D, Monsul NT, Vender JR, Lehmann JC. NMDA- and
endothelin-1-induced increases in blood – brain barrier permeabil-
Acknowledgements ity quantitated with Lucifer yellow. J Neurol Sci 1996;136:37–40.
Nishino H, Shimano Y, Kumazaki M, Sakurai T. Chronically admin-
istered 3-nitropropionic acid induces striatal lesions attributed to
This work was supported by grants from the Well- dysfunction of the blood – brain barrier. Neurosci Lett
come Trust and the Hereditary Disease Foundation. 1995;186:161 – 4.
We thank Mr Roger Hart for excellent photography Nishino H, Kumazaki M, Fukuda A, et al. Acute 3-nitropropionic
and Mrs Wendy Leavens for expert technical assis- acid intoxication induces striatal astrocytic cell death and dys-
tance. function of the blood – brain barrier: involvement of dopamine
toxicity. Neurosci Res 1997;27:343 – 55.
Nordborg C, Sokrab TEO, Johansson BB. The relationship between
plasma protein extravasation and remote tissue changes after
References experimental brain infarction. Acta Neuropathol 1991;82:118–26.
Nordborg C, Sokrab TEO, Johansson BB. Oedema-related tissue
Beal MF, Kowall NW, Ellison DW, Mazurek MF, Swartz KJ, damage after temporary and permanent occlusion of the middle
Martin JB. Replication of the neurochemical characteristics of cerebral artery. Neuropath Appl Neurobiol 1994;20:56 –65.
Huntington’s disease by quinolinic acid. Nature 1986;321:168 – 71. O’Brien MD, Jordan MM, Waltz AG. Ischemic cerebral edema and
Beal MF, Brouillet E, Jenkins BG, et al. Neurochemical and histo- the blood – brain barrier: distributions of pertechnetate, albumin,
logic characterization of striatal excitotoxic lesions produced by sodium and antipyrine in brains of cats after occlusion of the
the mitochondrial toxin 3-nitropropionic acid. J Neurosci midde cerebral artery. Arch Neurol 1974;18:461 – 5.
1993;13:4181 – 92. Olssen Y, Crowell RM, Klatzo I. Permeability of the blood–brain
Broadwell RD, Sofroniew MV. Serum proteins bypass the blood – barrier to protein tracers in focal ischemia and infarction caused
by occlusion of the middle cerebral artery. Acta Neuropathol
brain fluid barriers for extracellular entry to the central nervous
1971;18:89 – 102.
system. Exp Neurol 1993;120:245–63.
Paxinos G, Watson C. The Rat Brain in Stereotaxic Coordinates.
Broman T. Supravital analysis of disorders in the cerebral vascular
New York: Academic Press, 1982.
permeability in man. Acta Medica Scand 1944;118:79–83.
Petito CK, Pulsinelli WA, Jacobsen G, Plum F. Edema and vascular
Broman T, Branemark P-I, Johansson B, Steinwall O. Intravital and
permeability in cerebral ischemia: comparison between ischemic
postmortem studies on air embolism damage of the blood – brain
neuronal damage and infarction. J Neuropathol Exp Neurol
barrier tested with trypan blue. Acta Neurol Scand 1966;42:146 –
1982;41:423 – 36.
52.
Pont F, Collet A, Lallement G. Early and transient increase of rat
Coyle JT, Schwarcz R. Lesion of striatal neurones with kainic acid hippocampal blood – brain barrier permeability to amino acids
provides a model of Huntington’s chorea. Nature 1976;263:244 – during kainic acid-induced seizures. Neurosci Lett 1995;184:52–4.
6. Reese TS, Karnovsky MJ. Fine structural localization of a blood–
Dietrich WD, Alonso O, Halley M, Busto R, Globus MY-T. Intra- brain barrier to exogenous peroxidase. J Cell Biol 1967;34:207–
ventricular infusion of N-methyl-D-aspartate. 1. Acute blood – 17.
brain barrier consequences. Acta Neuropathol 1992;84:621 – 9. Schwob JE, Fuller T, Price JL, Olney JW. Widespread patterns of
Dobbing J. The blood – brain barrier. Physiological Rev 1961;41:135 – neuronal damage following systemic or intracerebral injections of
89. kainic acid: a histologial study. Neuroscience 1980;5:991–1014.
Gregersen MI, Rawson RA. The disappearance of T-1824 and struc- Tschirgi RD. Protein complexes and the impermeability of the
turally related dyes from the blood stream. Am J Physiol blood – brain barrier to dyes. Am J Physiol 1950;163:756.
1942;138:698 – 707. Wuerthele SM, Lovell KJ, Jones MZ, Moore KE. A histological
Hamilton BF, Gould DH. Nature and distribution of brain lesions in study of kainic acid-induced lesions in the rat brain. Brain Res
rats intoxicated with 3-nitropropionic acid: a type of hypoxic 1978;149:489 – 97.
(energy deficient) brain damage. Acta Neuropathol 1987a;72:286 – Zaczek R, Simonton S, Coyle JT. Local and distant neuronal degen-
97. eration following intrastriatal injection of kainic acid. J Neuro-
Hamilton BF, Gould DH. Correlation of morphologic brain lesions pathol Exp Neurol 1980;39:245 – 64.
with physiologic alterations and blood–brain barrier impairment Zucker DK, Wooten GF, Lothman EW. Blood – brain barrier
in 3-nitropropionic acid toxicity in rats. Acta Neuropathol changes with kainic acid-induced limbic seizures. Exp Neurol
1987b;74:67 – 74. 1983;79:422 – 33.