Specific Brain Protein Changes Correlated with Behaviourally Effective Brain Transplants

Journal of Neurochemistry Raven Press, Ltd., New York 0 1991 International Society for Neurochemistry

Specific Brain Protein Changes Correlated with Behaviourally Effective Brain Transplants

*Kathleen M. Wets, tJohn Sinden, ?Helen Hodges, $Yvonne Allen, and *R. M. Marchbanks

Departments of *Biochemistry now entitled Neuroscience, ?Psychology, and SNeuropathology, Institute of Psychiatry, De Crespigny Park, London, England

Abstract: The objective of this study was to identify cellular proteins that are associated with foetal brain transplants effective in reinstating memory function in adult rats with brain lesions. Quantitative memory deficits can be created in rats by lesioning the cholinergic projection system, using ibotenic acid. Previous work suggested that injection of cell suspen- sions prepared from presumptive cholinergic cells of foetal basal forebrain into adult brain, after such lesions, are most effective in restoring cognitive function. It was not clear, however, whether it was the cholinergic nature of the transplants that was critical for their success or whether other fac- tors were involved. In this study, the proteins present in transplanted tissues and control brains were analysed by two- dimensional polyacrylamide gel electrophoresis to identify markers for the cells that were specifically correlated with restoration of cognitive function. On each gel, the relative optical densities of the same 33 selected proteins were measured on an interactive computerised image analyser. The amount of each protein was compared between treatment groups and correlated with four behavioural measurements. Seven of the proteins analysed had levels of expression that were either related to transplantation or correlated with behavioural performance. The proteins of interest were divided into the following three groups: (1) transplant-related proteins,

(2) cholinergic transplant-specific proteins, and (3) behaviour- related proteins. Notable among the proteins of interest was one of the cholinergic transplant-specific proteins that was positively correlated with three of the four behavioural measurements and was also the only protein among those analysed that was significantly correlated with choline acetyltransferase (ChAT) levels. This has been identified, by immunoblotting, as dial fibrillary acidic protein, an astrocytic cell marker. These results suggest, therefore, that at least two cell types, astrocytes and ChAT+-staining cells, play an important role in the successful recovery of cognitive function. This study also identified possible protein markers for cognitive performance. The level of expression of two of the proteins analysed was not affected by lesioning or transplantation, but was significantly correlated with behaviour. One of these proteins, whose amounts correlated negatively with behavioural measurements, has been identified as neurone- specific enolase, a brain-specific neuronal cell marker. Key Words: Two-dimensional gel electrophoresis-Cholinergic- rich transplants-Memory-Brain proteins-Glial fibrillary acid protein. Wets K. M. et al. Specific brain protein changes correlated with behaviourally effective brain transplants. J. Neurochem. 57, 1661-1670 (1991).

The development of brain transplantation tech- niques offers a promising new approach to the allevia- tion of psychiatric and neurological illnesses that result from neuronal degeneration. It is now well established that transplants of neural tissue that survive well in host brain can ameliorate impaired motor function (Freed, 1983). There is also evidence that cognitive deficits produced by brain lesions in rats (Dunnett et al.,

Received December 19, 1990 revised manuscript received March 20, 1991; accepted April 15, 1991.

Address correspondence and reprint requests to Dr. R. M. March- banks at Department of Neuroscience, Institute of Psychiatry, De Crespigny Park, L,ondon SE5 8AF, U.K.

The present address of Dr. Y. Allen is Department of Anatomy and Cell Biology, University of Liverpool, L69 3BX, U.K.

1982, 1985; Arendt et al., 1988) or as a consequence of normal aging (Gage et al., 1984) can similarly be ameliorated by foetal brain tissue transplants (Gray et al., 1990). In each ofthese studies on cognitive function, successful transplants have included cholinergic cells, whereas control transplants lacking these cells were in- effective (Dunnett et al., 1982, 1985; Arendt et al., 1988, 1989). Our earlier studies have shown that as

Abbreviations used: ACh, acetylcholine; BF, basal forebrain; CUT, choline acetyltransferase; GFAP, glial fibrillary acidic protein; IEF, isoelectric focusing; IGV, integrated grey value; NP-40, Nonidet P- 40; NSE, neurone specific enolase; PBS, phosphate-buffered saline; PI, isoelectric point; SDS, sodium dodecyl sulphate; TCA, trichloroacetic acid; TGS, Tris, glycine, and SDS; 2D-PAGE, two-dimensional polyacrylamide gel electrophoresis.

1661

1662 K. M. WETS ET AL.

the cholinergic-rich transplants grew, in rats with lesions of the ascending forebrain cholinergic projection system, there was a concomitant recovery of cholinergic parameters [e.g., acetylcholine (ACh) and choline acetyltransferase (ChAT) levels] and of various cognitive functions (Arendt et al., 1988, 1989; Hodges et al., 1990).

Although these results show that transplants dissected from appropriate regions of the foetal brain are associated with the restoration of behavioural function and cholinergic parameters, the population of cells transplanted is, in fact, heterogenous. We have at- tempted to characterise transplanted tissue using a method that makes no assumptions about transmitter or cell type. In this study, the proteins in the transplanted tissue were analysed by two-dimensional polyacrylamide gel electrophoresis (2D-PAGE), which sep- arates cellular proteins according to their charges (PI) and molecular weights. 2D-PAGE coupled with sen- sitive silver stains and computerised scanning densitometry allows separation and quantification of a large number of proteins from small tissue samples.

2D-PAGE has been used successfully to analyse protein differences in discrete areas of rat brain affected by disruption of major neuronal pathways (Heydorn et al., 1 9 8 4 , 1 9 8 5 ~ 1986a,b; Jacobowitz and Heydorn, 1984). None of these studies, however, made parallel behavioural measurements. Although some work has been done on the biochemistry of transplanted tissues (e.g., Arendt et al., 1988, 1989), there have not been any reports of attempts to characterise the proteins from transplanted neural tissue using 2D-PAGE.

The point of applying this methodology to the study of transplanted tissue is that it allows for quantitative estimation of individual proteins, even though their identity may not be known. It thus allows for the study of the biochemical mechanisms of restoration of cognitive function without preconceptions as to the cell type(s) or neurotransmitter(s) involved. Protein changes that correlate with behavioural recovery can then be used as cellular markers.

MATERIALS AND METHODS

Chemicals Molecular mass standards (14-66 kDa and 29-205 kDa)

for sodium dodecyl sulphate (SDS)-PAGE, ibotenic acid, trypsin, and second antibodies (anti-mouse and anti-rabbit) immunoglobulin G (whole molecule) biotin conjugates from goat were from Sigma. Ampholines (pH 3.5-10 and 5-7) were purchased from Pharmacia-LKB. Coomassie Brilliant Blue R-250 and Nonidet P-40 (NP-40) were from BDH. The silver-staining kit was purchased from Bio-Rad. The nitrocellulose (0.45 pm) was from Anderman and Co. for Schleicher-Schuell. Mouse monoclonal anti-p-tubulin and streptavidin biotinylated horseradish peroxidase complex were from Amersham. Rabbit anti-cow glial fibrillary acidic protein (GFAP) and rabbit anti-human neurone-specific enolase (NSE) were from Dakopatts.

General design Tissue for 2D-PAGE was obtained from 22 male Sprague-

Dawley rats (OLAC, U.K.) selected from a larger group in which the effects of transplants have been assessed behaviourally and morphologically (Hodges et al., 1990, submitted). Beginning at age 6-7 weeks, the rats were initially trained to asymptotic performance in place and cue tasks in an eight- arm radial maze (Jarrard, 1986). They then received infusions of 0.1 M phosphate-buffered (pH 7.4) saline solution (PBS vehicle) or ibotenic acid in 0.1 M PBS in both the nucleus basalis and medial septal/diagonal band brain regions (i.e., at the source of the cortical and hippocampal cholinergic projections), and then were retested at intervals for 16 weeks, to assess the effects of the ibotenate lesions (Hodges et al., 1990).

Tissue for 2-D PAGE was taken from lesioned rats that were randomly assigned to four treatment groups, the first of which received cholinergic-rich foetal basal forebrain (BF) cell suspension transplants. These tissues were dissected on embryonic day 15 (E- 15) and injected into two sites in frontal cortex (n = 7/samples selected from nine rats). The second group, the transplant control group, received cholinergic-poor foetal hippocampal E- 17 suspension grafts in the cortex (n = 517). The lesioned (n = 6/8) and nonlesioned (n = 4/10) control groups were sham operated. After recovery from transplantation (2 weeks), rats were tested once a week, for 26 weeks, on both the radial maze tasks to investigate the time course of functional recovery from lesion-induced deficits. Rats were killed by cervical dislocation -7 months after transplantation and 12 months after lesioning, and their brains were removed and bisected. One set of half brains was used for histological examination (see Hodges et al., 1990, submitted), the other was reserved for measurement of ChAT activity and 2D-PAGE analysis. Samples of 20-30 mg (wet weight) were dissected freehand from frontal cortex to contain transplant material. Injection sites were normally visible and care was taken to avoid inclusion of surrounding nontrans- planted tissue. Equivalent areas were taken from rats without transplants, and all tissues were frozen at -7OOC until assayed. To assess the protein composition of normal BF tissue at an age equivalent to that of the transplants, further two comparison samples of BF tissue were dissected from adult (6 months) rats that were not behaviourally tested.

Behavioural testing Two eight-arm radial mazes were used, one above the other.

The upper maze had good visibility for cues around the room (posters, cupboards, etc.) and was used for the place task; the lower maze was used for the cue task, in which each arm contained a distinctively textured insert (sandpaper, wire mesh, carpet, etc.). Rats were trained to find sucrose (30%) at the end of four of the eight arms in both the place and cue tasks. For the place task, sucrose rewards were always in the same four arms; whereas for the cue task, the same four inserts were always rewarded, but were moved to different arms on each trial, so that the rat had to learn the association between cue and reward, regardless of the location of the cue. The following two types of error were scored reference, where a rat entered a never-rewarded arm for the first time on each trial, and working, where a rat revisited an arm already entered within a given trial. The former represented a failure of long-term memory for reward locations/associations that were constant across trials, whereas the latter involved a loss of trial-specific short-term memory. Thus, the maze tested

J. Neurochem., Vol. 57, No. 5, 1991

PROTEIN CHANGES CORRELATED WITH BEHA VIOUR 1663

four aspects of memory, short- and long-term in both spatial and nonspatial modes. For each trial, the rat was placed in the centre of the maze and allowed to explore freely for 5 min or until all four rewards had been obtained. Rats were given two consecutive trials per day on either the place or the cue task, with a 5-min interval between trials.

Lesioning Rats were anaesthetised with equithesin (3.0 ml/kg i.p.)

and injected bilaterally (using a stereotaxic apparatus) in both the nucleus basalis and the medial septal nucleus/diagonal band with ibotenic acid 10.0 mg/ml in 0.1 mol/L of PBS (pH 7.4). The lesion procedures and coordinates have been described elsewhere (Hodges et al., 1990, submitted). Controls were injected with PBS at the same sites.

Transplantation Cell suspension transplants were made using the methods

of Arendt et al. (1988) and Sinden et al. (1 990). Using a low- power microscope, E- 15 BF or E- 17 hippocampal tissue was dissected from foetuses removed by caesarean section from a dam overdosed with pentobarbitone. The tissue was dissociated by incubation for 20 min at 37°C with 0.25% trypsin in 0.6% glucose-saline, and then mechanically dissociated by gently pipetting through a fire-polished Pasteur pipette. Rats were anaesthetised as for lesioning, and holes drilled in the skull to allow penetration of a 10-pl Hamilton syringe needle into the neocortical sites bilaterally. The injection coordinates are described elsewhere (Arendt et al., 1989; Hodges et al., 1989). Control rats were sham operated, i.e., the skull was exposed and a hole drilled, but no injections were made.

ChAT measurement Weighed samples were homogenised in 4% NP-40 using

8.3 pl/mg of tissue (wet weight). Two aliquots (10 pl) of the homogenate were used for the ChAT assay and the remaining solubilised tissue was analysed by 2D-PAGE. ChAT activity was measured by a method derived from Fonnum et al. (1975) and described by Hodges et al. (1989).

2D-PAGE 2D-PAGE was performed according to O’Farrell (1975)

with minor modifications. Microdissected transplant and brain cortex samples were placed in 100 pl of buffer [5% 2- mercaptoethanol, 9.5 Murea, 10 mmol/L of L-lysine mono- hydrochloride, 2% (wt/vol) NP-40, 1% ampholines (pH range, 5-7), 1% ampholines (pH range, 3.5-lo), and 0.5% SDS]. The samples were homogenised by being drawn through 2 1 - and 25-gauge needles, and the supernatant, containing the solubilised proteins, was removed after centrifugation. Buffer containing 600 p g of protein from each sample was first separated according to the isoelectric points (PI) in 3.5% acrylamide gels (1 50 mm X 2.7 mm diameter) in glass rods. Iso- electric focusing (IEF) gel mix consisted of 5.5 g of urea, 0.38 g of acrylamide, 0.022 g of bisacrylamide, 2 ml of 10% NP- 40, 0.25 ml of ampholines (pH range, 5-7), 0.25 ml of ampholines (pH range, 3.5-lo), 6 p1 of tetramethylethylenedi- amine, and 12.5 pl of 10% (freshly prepared) ammonium persulphate made up to 10 mi. The solubilised samples were loaded on top of the IEF rods and run at 400 V for 16 h and 1,000 V for 4 h. Rod gels were then removed and equilibrated in 5 ml of SDS sample buffer [40% (vol/vol) glycerol, 1 % (vol/vol) 2-mercaptoethanol, 2.5% (wt/vol) SDS, 0. I25 M Tris-HC1 (pH 6.8), and 0.002% bromophenol blue], for 30 min on a slow shaker.

In the second dimension, proteins were separated by molecular weight, on slab gels 1.7 mm thick according to the procedure first described by Laemmli (1970), using a top stacking gel (50 mm high) and a bottom resolving gel (150 mm high). The stacking gel was 4.5% acrylamide, 0.125 M Tris-HC1 (pH 6 4 , and 0.1% SDS, and the resolving gel was 10% acrylamide, 0.38 M Tris-HC1 (pH 8.8), with 0.1% SDS. After equilibration, the IEF rod gels were placed on top of the stacking gel, along with molecular mass markers (14-66 kDa and 29-205 kDa) and sealed in place with 1% agarose. The slab gels were run overnight at 70 V in running buffer containing (per liter) 1 g of SDS, 14.4 g of glycine, and 3 g of Trizma base, until the dye front reached the bottom of the gel. The gels were then either stained or immunoblotted (described below). For image analyser processing, after re- moval of the gels from the glass plates, the proteins were fixed in 12.5% trichloroacetic acid (TCA) for 1 h and then stained with 0.25% Coomassie Brilliant Blue R-250, 40% methanol, and 7% acetic acid for 5 h. The gels were destained in 25% methanol and 7% acetic acid, then in 10% methanol and 7% acetic acid, until they were sufficiently destained for silver staining. The gels were silver stained using a kit based on the method of Meml et al. (1983). After staining, the wet gels were photographed and then placed in sealed plastic bags until they were processed on the image analyser. Proteins are given a two-digit number representing the apparent molecular mass in kilodaltons (kDa). Immunoblotting

The blotting procedure was performed as previously described by Jackson and Thompson (1 984). After running, gels selected for immunoblotting were fixed for 18 h in 500 ml of 100 g/L of TCA, 100 ml/L of acetic acid, and 200 ml/ L of propan-2-01, then in 500 ml of 100 ml/L of acetic acid and 200 ml/L of propan-2-01 for 6 h, and subsequently stained in 150 ml/gel of 1 .O g/L of Coomassie Brilliant Blue R-250 in 100 ml/L of acetic acid and 200 ml/L of propan-2-01 solution for 18 h at 4°C. Gels were destained by shaking at 4°C in 500 ml/gel of 70 ml/L of acetic acid and 300 ml/L of propan-2-01, which was changed four times (over 2 days) and then washed in four changes of 500 ml of water. To prepare for blotting, the gels were incubated for 0.5 h with shaking at 4°C in 500 ml of 0.025 mol/L of Tris base, 0.192 mol/L of glycine, and 10 g/L of SDS (TGS buffer). The gels were then placed in 4°C transfer buffer of 25 mmol/L of Tris, 192 mmol/L of glycine, and 20% MeOH, at pH 8.3 for 0.5 h. The gels were blotted onto nitrocellulose using the method for SDS-containing gels of Towbin et al. (1979), in a Bio- Rad Transblot apparatus at 250 mA for 4 h at room temperature without cooling. Nitrocellulose paper was removed and washed in PBS, with gentle shaking, for 1 h and then hung to dry. The dried nitrocellulose paper was placed in 20 mdml of BSA in PBS overnight, then incubated with primary antibody in PBS with 0.2% Tween 20 overnight. Blots were next incubated with biotin-conjugated second layer antibody in PBS and 0.2% Tween 20 for 1 h at room temperature. The blots were then incubated with streptavidin-biotinylated horseradish peroxidase complex in PBS with 10 mg/ml of BSA and 0.1 % Tween 20 for 1 h. The blots were developed in 0.625 mg/ml of diaminobenzidine in PBS with 1 pl/ml of 30% HzOz and 0.3 mg/ml of CoC12. Treatment of results

From all the spots visible on the gel, a smaller group was selected for this initial study due to the amount of processing

J. Neurochem., Vol. 57, No. 5, 1991


involved. Thirty-three spots were selected for analysis based on their reproducibility in different gels from the same sample and on their consistent resolution on the gels, thus reducing intergel variability.

The relative optical densitites of proteins present on the gels were determined by computerised scanning densitometry. On each gel, the area and intensity of the 33 protein spots were measured on an interactive computerised imaging facility (IBAS 2000, Kontron); their product gave an integrated grey value (IGV), which is proportional to the amount of protein present. Because the various protein spots differed from each other in both the intensity of staining and size on the gel, the IGVs have been normalised for each gel with respect to the average IGV of all proteins on that particular gel. Figure 1 shows the linear relationship between individual IGV values and the gel average IGV, thus justifying the nor- malisation procedure. Results in this study were defined as being significant if p I 0.05.

RESULTS

General findings The gels examined were of similar overall appearance

and were well resolved over the pH range of 5.0-7.0 and molecular mass range 20-100 kDa. In every gel, the same 33 proteins were analysed and expressed as IGV ratios (see Materials and Methods). A representative gel alone is shown (Fig. 2) to give a general idea of the position of the proteins of interest because statistical analysis of the IGV ratios obtained from each gel allows for a more objective quantification of the protein amounts (see Materials and Methods: Treat- ment of results).

Unnorrnallsed IGV

loo r

I /

-- 0 10 20 30 40 50 80 70

Gel Average IGV

FIG. 1. Plot of nonnormalised IGV values representing proteins of high (0). medium (m), and low (A) abundance against the gel average IGV indicating the linear relationship between IGV values and average overall gel-staining intensity.

FIG. 2. Representative 2D-PAGE gel of normal adult cortex from nonlesioned control rat brain. For orientation purposes, the gel has the ordinate and abscissa marked for molecular mass (MW) and isoelectric range (pl), respectively. Approximately 600 pg of protein is loaded into each gel with actin, representing 10% of total protein (Strocchi et al., 1981; Lim et al., 1983; Pearce et al., 1983; Kiessling et al., 1986). The above photo shows the 2D protein pattern over approximate pl 5-7 and molecular mass 24-70 kDa.

Behavioural results The effects of combined lesions of the nucleus basalis

and medial septal/diagonal band regions were to pro- duce substantial and significant increases in both place and cue reference and working memory errors in the radial maze. Cholinergic-rich BF transplants, but not cholinergic-poor hippocampal ones, into frontal cortex produced a highly significant reduction in all four types of memory error (Table 1) although the magnitude of transplant effects was greater in working memory performance (for details, see Hodges et al., 1990, submitted).

Group differences To analyse protein differences related to treatment

groups, the IGV ratios were averaged for the rat groups having similar treatments. Table 2 and Figs. 3, 4, and 5 summarise the statistical details of the proteins of interest for the different rat groups, which have also been pointed out with arrows on the gel in Fig. 2. Table 2 also gives, for each of the listed proteins, the values for normal adult BF, the brain region into which cholinergic-rich foetal transplant tissues would have normally developed. A comparison of frontal cortex proteins in lesioned and nonlesioned control rats shows that none of the proteins listed in Table 2 are significantly altered by the lesion procedure.

Transplant-related proteins Three proteins demonstrate substantial differences

when comparing transplant tissue and the lesioned control groups. Protein 24 (Fig. 2) is significantly increased in both types of transplanted groups compared with lesioned cortex (Table 2, Fig. 3). Protein 72 (Fig. 2) is also increased in both transplant types relative to control tissue (Table 2, Fig. 3); however, for both these

J. Neurochem.. Vol. 57, No. 5. 1991


TABLE 1. Statisticai summary of radiai arm maze performance and C U T levels of rats for the four experimental groupsa

Treatment group

Cholinergic-poor Cholinergic-rich Nonlesioned Lesioned transplant transplant

(n = 4) (n = 6) (n = 5) (n = 7)

Behaviour Place reference 1.1 f 0.5' 5.3 +- 0.3 5.5 f 0.9 3.3 t- 0.5 Place work 0.1 f 0.1 6.1 +- 2.0 5.6 f 3.0 1.4 f 0.5 Cue reference 0.9 t- 0.4 4.9 k 0.5 5.0 t- 0.5 2.7 f 0.4 Cue work 0.3 f 0.1 5.8 It 1.7 5.1 f 1.4 0.9 ? 0.3

ChAT 10.9 f 0.9' 6.2 ? 0.5 6.5 f 0.7 12.1 f 0.9

a Rats in each experimental group were selected as being behaviourally representative of that group. Average k SEM. Units are averages of errors made during last 3 weeks of testing. Average k SEM. Units are picomoles per minute per milligram of wet weight.

proteins, this increase only attains the concentration found in the normal adult BF. This implies that these proteins possibly originate from their precursor tissue, in the cholinergic-rich transplant, and that their level of expression is not affected by the displacement from their normal environment. As one would expect, the process of transplantation itself causes changes in the level of expression of some proteins. One such protein, 53 (Fig. 2), is negatively affected by transplantation. Lesioning slightly suppresses the expression of protein 53, and this suppression is even more pronounced after transplantation (Fig. 3).

Cholinergic transplant-specific proteins The amounts of two of the proteins examined dif-

fered significantly between cholinergic-rich and cholinergic-poor transplants and are therefore considered to be specific for one type of transplant. Protein 47 (Table 2, Fig. 2) is negatively affected by transplantation; however, this effect is more specific to cholinergic-

rich transplants (Fig. 4). The most interesting protein in this group, however, is protein 52 (Fig. 2), which increases significantly in cholinergic-rich transplants in comparison with all treatment conditions (Table 2, Fig. 4). ChAT content, which is associated with behavioural recovery (Table 1 ), is significantly positively correlated only with protein 52 ( p < 0.05), of the 33 proteins analysed.

Proteins correlated with behaviour For every protein, the correlation coefficient for each

of the four measures of memory performance was calculated using scores from individual rats. The histograms on the right-hand side of Figs. 3,4, and 5 show the magnitude of the regression coefficients for selected proteins of interest.

A positive correlation indicates that an increase in protein matches an improvement on the behavioural task. Protein 52, as mentioned above, is associated with cholinergic transplants and is positively correlated with

TABLE 2. Average IGV ratios of specific proteins of interest in cortex of lesioned and nonlesioned experimental rats and in normal adult basal .forebrain

Average IGV ratios

Approximate Cortex Basal

Protein Protein NL L forebrain ( k W PI (@)n (n = 4) (n = 6) (n = 2)

24 5.7 37 f 2 1.23 f 0.07' 1.06 f 0.1 1' 1.70 f 0.06' 47 5.8 73 f 8 2.52 f 0.26 2.27 f 0.23 2.03 t- 0.08 48 5.5 33 f 1 1.12 k 0.05 1.20 f 0.09 1.03 f 0.06 52 5.8 11 f 2 0.34 f 0.07 0.44 t- 0.08 0.32 f 0.01

3.10 t- 0.29 53 5.3 110 f 21 3.76 * 0.71 66 5.7 1 8 f 2 0.59 f 0.06 0.62 t- 0.09 0.85 f 0.13 72 Basic 1 3 f 1 0.41 f 0.04 0.48 f 0.07 0.63 f 0.02

3.40 t- 0.33

a Micrograms of protein per 10 mg of wet weight of transplant tissue; calculated by comparison

' Average IGV f SEM for groups of rats nonlesioned (NL) and lesioned (L). 'Average IGV f range for normal adult nucleus basalis.

with actin, assuming it constitutes 10% of total brain protein (Blitz and Fine, 1974).

J. Neurochem., Vol. 57, No. 5. 1991


Transplant Differences Behavioural Correlations

0.1 .

0.6 * * * 1 T

8 T

0.5 "O I T

-0.5 O n

-1.0 I I

0 6 1 T

-0'3 1 - o ' 6 zLl-2z

CRT vs CPT

Protein 24 0.2 .

T 0.1 I * -l-

-0.2 I

-O" I * -0.2 I

Protein 72

0 ' 4 1 * *

o.2 0 L;; - 0 . 4 +

P R P W C R C W

ChAT levels. This protein also correlates positively with three of the four measures of memory (Fig. 4). Protein 72 is a transplant-related protein, as opposed to a cholinergic-rich transplant-specific one, but has histogram patterns similar to protein 52, as seen in Figs. 3 and 4, on the left- and right-hand side. Protein 72, however, correlates positively only with place reference and working memory. Protein 24, which is also a transplant-related protein, also correlates significantly with the two place tasks (Fig. 3).

A negative correlation with behaviour implies that an increase in the level of expression of that protein has a deleterious effect on the rat's performance. Both proteins 47 and 53 correlate negatively with behavioural performance and are suppressed in cholinergic-

FIG. 3. Transplant-related proteins: proteins 24, 53, and 72. In Figs. 3-5, the histograms on the left-hand side, for each protein, show the difference between each type of transplant and lesioned rats. From left to right, the histograms represent cholinergic-rich transplants versus lesioned (CRT vs LES), cholinergic-rich versus cholinergic-poor transplants (CRT vs CPT), and cholinergic-poor transplants versus lesioned (CPT vs LES). The right-hand side shows the regression coefficients of each protein, normalised with respect to the average content of that protein, for the four behavioural tasks (PR, place reference; PW, place working; CR, cue reference; CW, cue working). *p = 0.05; **p < 0.05; "'p < 0.01.

rich transplants. Protein 53 is present in lower amounts in both types of transplanted tissue and correlates negatively with both place reference and place working memory (Fig. 3). Protein 47, however, is a cholinergic transplant-specific protein and, although its correlations with behaviour are weak, it generally correlates negatively, notably with place reference memory (Fig. 4).

In the course of examining these correlations, an- other type of protein emerges, one that is not affected in any significant manner by the lesion or transplant procedures, but that nevertheless correlates strongly with the behavioural measurements. Protein 48 (Fig. 2) appears treatment indifferent but shows a strong negative correlation with both kinds of place memory,

J. Nerrrochem , Vol 57, No 5, 1991

PROTEIN CHANGES CORRELATED WITH BEHA VIOUR

I

Transplant Differences

Protein 47

o '6 I

-0.3 ow FIG. 4. Chdinergic transplant-specific: proteins 47 and

between cholinergic-rich and cholinergic-poor transplants is always significant, as shown in the centre of the histograms on the left-hand side. See legend to Fig. 3 for details.

52. Notice how in this group of proteins, the difference -0.6

O . 7 * *

-0.3 1

and a weaker but still negative correlation with the two kinds of cue memory (Fig. 5). Protein 66 (Fig. 2), on the other hand, is the mirror image of protein 48 dis- cussed above, and its level of expression correlates positively with both kinds of place memory (Fig. 5).

Immunoblotting Tentative identification of some of the proteins of

interest can be done based on their PI values and molecular weights which are similar to those previously reported in other articles. To confirm identification of these proteins, some gels were Coomassie Blue-stained and blotted onto nitrocellulose paper, and the proteins were then immunoidentified with the appropriate antibody. The reason for using this method is that the Coomassie Blue-stained pattern on the gel is trans- ferred to the nitrocellulose thereby conveniently allow- ing for simultaneous viewing of the Coomassie Blue- stained proteins from the gel and the antibody-specific black-stained protein of interest. Using the Coomassie Blue-stained pattern as a benchmark ensures correct identification of the protein of interest. This method was used to confirm tentative identification of three proteins, i.e., protein 52 as being GFAP, protein 48 as NSE, and protein 53 as P-tubulin.

1667

Behavioural Correlations

0 .2

0 1

-0.2 '

Protein 52

0 4

0 2

0

-0 2

-0 4

* * * * * T T

P R P W C R C W

DISCUSSION

Recent studies in our laboratory have shown that cholinergic-rich E- 15 foetal BF cell suspension transplants into frontal cortex were able to improve all four measures of memory function in the radial maze task of Jarrard (I 986) after damage to the forebrain cholinergic system produced by ibotenate lesions (Hodges et al., 1990, submitted) or chronic alcohol consumption (Arendt et al., 1989). Cholinergic-poor E- 17 foetal hippocampal transplants were without effect and, in all experiments, behavioural improvement was dependent on recovery in cholinergic markers measured post- mortem.

This study shows that there are furthermore significant protein changes in transplanted tissue, some of which are associated with behavioural recovery. Given the range of PI and molecular weight examined, it is likely that there are protein changes that were not de- tected because they were either outside the range used or not abundant enough for the sensitivity of the silver- staining method. The objective of the study, however, was not to characterise all protein changes associated with transplantation, of which there are certainly many, but rather to find markers for behaviourally effective

J . Neurochem., Vol. 57, No. 5. 1991


Transplant Differences Behavioural Correlations

Protein 48

-0.3 O r - =

-0.6

0 3 o’6 1

CRT vs CPT

I -0 2 ’

FIG. 5. Behaviour-related proteins: proteins 48 and 66. See legend to Fig. 3 for details.

Protein 66

O’l 0 P -0.2 J-

P R P W C R C W

transplants. The proteins we have identified, whose changes correlate with function, are sufficiently abundant to be readily purified.

Almost all the proteins of interest in this study were correlated significantly with the place tasks. This may be due to a greater sensitivity of spatial memory to the brain protein changes that we have identified in this study. Alternatively, it may suggest a more important role for the frontal cortex in the processing of spatial, as opposed to cue, information in the radial maze. This complex phenomenon is the subject of further inves- tigation looking at other proteins and other cortical regions.

In our study, there were no significant differences between the lesioned and nonlesioned groups of rats in the 33 proteins analysed. Because the rats in this experiment were killed - 12 months after lesion sur- gery and 7 months after transplantation, our analysis would only identify long-term effects. Therefore, any short-term effects on the level of protein expression, caused by the lesion or transplant procedures, would not be discernible in these gels. The effects of lesions on rat brain proteins have been previously analysed using 2D-PAGE (Heydorn et al., 1984, 1985c, 1986u,b; Jacobowitz and Heydorn, 1984). Heydorn et al. ( 1 9854

examined protein changes in the hippocampus and oc- cipital cortex 9 and 35 days after radiofrequency lesioning of the nucleus of the diagonal band. They ob- served changes in proteins that were either of lower molecular weight or more basic than the range of proteins we examined.

In this study, seven proteins appear to change their level of expression in correlation with transplantation or behavioural performance. Using immunoblotting, we are able to identify three of these proteins. Increases in the amounts of protein 52 are specific to cholinergic- rich transplants, and it is positively correlated with three of the four measures of memory, as well as to ChAT content. On the gels, it appears in the region of GFAP, a glial cell marker (Strocchi et al., 1981; Pearce et al., 1983; Kiessling et al., 1986), and was stained with GFAP antibody on an immunoblot. Pate1 and Hunt ( 1 989) found that astrocytes promote a trophic factor important for the development of cholinergic cells. This suggests that glial cells may play an enabling role in cholinergic-rich transplants.

Protein 53, given its molecular weight, PI, and im- munoidentification, may be P-tubulin (Strocchi et al., 1981; Lim et al., 1983; Pearce et al., 1983; Kiessling et al., 1986). P-Tubulin is one of the most abundant

J. Neurochem.. Vol. 57, No. 5, 1991


proteins in brain [approximately 10% of total brain protein (Blitz and Fine, 1974)J and we estimate that per 10 mg of brain tissue, there is approximately 1 10 pg of this protein on each gel. Although P-tubulin does not differ significantly between the four treatment groups, the manner in which it changes is interesting. There is less of this protein in transplanted tissue compared with control cortical tissue, and its levels correlate negatively with some of the behavioural measures. As this is an abundant cytoskeletal protein, these results are difficult to interpret because they suggest that increased amounts of @-tubulin have a deleterious effect on cognitive performance, and yet it is decreased in both cholinergic-rich and cholinergic-poor transplants. One explanation would be that a large volume of the transplanted tissue is composed of glial cells, known to contain less P-tubulin in comparison with neuronal tissue. A similar explanation can be made for protein 47, which also appears to be suppressed in cholinergic transplants and correlates negatively with behaviour. Protein 47 has not yet been identified, but appears in the region of two other abundant brain proteins, creatine kinase (Lim et al., 1983; Pearce et al., 1983; Kies- sling et al., 1986) (73 pg per 10 mg ofbrain tissue) and soluble glutamic oxaloacetic transaminase (Heydorn et al., 198%) (0.4% of total soluble brain protein).

The amount of protein 48, on the other hand, is not affected by lesioning or transplantation, but is negatively correlated with all four measures of memory. It appears on the gels in the region of NSE (14-3-2) (Jack- son and Thompson, 1981; Lim et al., 1983; Heydorn et al., 1984,1985~; Kiessling et al., 1986), and is stained with NSE antibody on immunoblots. NSE is considered a neuronal marker (Pickel et al., 1976; Schmechel et al., 1978). These results, at first glance, appear con- tradictory because this neuronal marker correlates negatively with behaviour when ChAT levels correlate positively. CMT' cells only represent a subpopulation of the neuronal cells present in the transplant and surrounding tissue, however, and there is no correlation between amounts of NSE and ChAT. The other "behaviour-related" protein, 66, which correlates positively with behavioural performance, is to our knowledge not yet characterised.

Of the two remaining proteins, 24 and 72, both correlate positively with cognitive recovery, particularly on the place task, but have not yet been identified. They are both highly expressed in the transplants relative to the control cortical tissue, apparently as a result of their ectopic origins. We do not suppose, however, that they are also astroglial related proteins because preliminary observations show that their distribution does not parallel that of GFAP.

The use of 2D-PAGE in studying the protein components of behaviourally characterised brain transplanted tissue has shown that there are protein changes that correlate with the previously described behavioural improvements seen in cognitively effective brain trans-

plants. We have identified possible markers for cognitive performance (e.g., proteins 66 and 48), which are unrelated to lesioning or transplantation. Because the rats are not from an inbred strain, these proteins may represent the varying expression of a gene that confers cognitive advantage. The objective of this study, however, was to find cellular markers for cognitively successful transplants, and we believe that some of the proteins (e.g., proteins 52 and 24) we have identified may assist further studies on transplantation. The fact that GFAP (protein 52) correlates with ChAT and with cholinergic-rich transplants suggests that at least two types of cells, glial and ChAT'-staining cells, are involved in effective transplants. We are currently in- vestigating whether cholinergic BF transplants enriched with cultured glial cells may have additional positive functional effects.

Acknowledgment: We are grateful to the Departments of Pharmacology, Neurology, and Neuropathology for assisting our use of the IBAS image-analysing facility. We thank Drs. Christopher Perrett and Stephen Whatley for their help and advice with the initial two-dimensional gel work. K. M. Wets was supported by a grant from the Bethlem Maudsley research fund. The Wellcome Trust also provided support.

REFERENCES Arendt T., Allen Y., Sinden J., Schugens M. M., Marchbanks R. M.,

Lantos P. L., and Gray J. A. ( 1 988) Cholinergic-rich brain transplants reverse alcohol-induced memory deficit. Nafure 332,448- 450.

Arendt T., Allen Y., Marchbanks R. M., Schugens M. M., Sinden J., Lantos P. L., and Gray J. A. (1989) Cholinergic system and memory in the rat: effects of chronic ethanol, embryonic basal forebrain transplants and excitotoxic lesions of cholinergic basal forebrain projection systems. Neuroscience 33, 435-462.

Blitz A. L. and Fine R. E. (1974) Muscle-like contractile proteins and tubulin in synaptosomes. Proc. Nufl. Acad. Sci. USA 71, 4472-4476.

Dunnett S. B., Low W. C., Iversen S. D., Stenevi U., and Bjorklund A. (1982) Septa1 transplants restore maze learning in rats with fornix-fimbria lesions. Bruin Res. 251, 335-348.

Dunnett S. B., Toniolo G., Fine A., Ryan C. N., Bjorklund A., and Iversen S. D. (1985) Transplantation of embryonic ventral forebrain neurons to the neocortex of rats with lesions of nucleus basalis magnocellularis. I f . Sensorimotor and learning impairments. Neuroscience 16, 787-797.

Fonnum F. (1975) A rapid radiochemical method for the determi- nation of choline acetyltransferase. J. Neurochem. 24,407-409.

Freed W. J. (1983) Functional brain tissue transplantation: reversal of lesion-induced rotation by intraventricular substantia nigra and adrenal medulla grafts, with a note on intracranial retinal grafts. Bid. Psychiatry 18, 1205- 126 I .

Gage F. H., Bjorklund A,, Stenevi U., Dunnett S. B., and Kelly P. A. T. ( 1984) Intrahippocampal septa1 grafts ameliorate learning impairments in aged rats. Science 222, 533-536.

Gray J. A., Sinden J. D., and Hodges H. (1990) Cognitive function: neural degeneration and transplantation. Semin. Neurosci. 2, 133-142.

Heydorn W. E., Creed G. J., and Jacobowitz D. M. (1984) Effect of desmethylimipramine and reserpine on the concentration of specific proteins in the parietal cortex and the hippocampus of rats as analyzed by two-dimensional gel electrophoresis. J. Phur- macol. Exp. Ther. 229,622-628.

Heydorn W. E., Creed G. J., Marangos P. J., and Jacobowitz D. M.

J. Neurochem., Vol, 57, No. S. 1991

1670 K. M, WETS ET AL.

( 19854 Identification of neuron-specific enolase and nonneu- ronal enolase in human and rat brain on two-dimensional polyacrylamide gels. J. Neurochem. 44,201-209.

Heydorn W. E., Creed G. J., Wada H., and Jacobowitz D. M. (19856) Immunological evidence for existence of two subforms of soluble glutamic oxaloacetic transaminase (SCOT) in human and rat brain. Neurochem. Int. 7, 833-841.

Heydorn W. E., Nguyen K. Q., Creed G. J., and Jacobowitz D. M. (1 985~) Effect of reduction of cholinergic inout on the concentration'of specific proteins in different cortical regions of the rat brain. Brain Res. 339, 209-2 18. dorn W. E., Nguyen K. Q., Creed G. J., Kostrzewa R. M., and Jacobowitz D. M. (1986~) Effects of bilateral lesion of the locus coeruleus and of neonatal administration of 6-hydroxydopamine on the concentration of individual proteins in rat brain. Bruin Res. 367, 31-38. dorn W. E., Creed G. J., Nguyen K. Q., and Jacobowitz D. M. ( I 986b) Effect of 5,7-dihydroxytryptamine on the concentration of individual proteins in different areas of the rat brain. Bruin Res. 368. 193-196.

Hodges H., Thrasher S., and Gray J. A. (1989) Improved radial maze performance induced by the benzodiazepine antagonist ZK 93 426 in lesioned and alcohol-treated rats. Behuv. Phurmacol. 1, 45-55.

Hodges H., Allen Y., Sinden J., Lantos P. L., and Gray J. A. (1990) Cholinergic-rich transplants alleviate cognitive deficits in lesioned rats, but exacerbate responses to cholinergic drugs. Prog. Bruin Res. 82, 347-358.

Jackson P. and Thompson R. J. (1981) The demonstration of new human brain specific proteins by high resolution two-dimensional polyacrylamide gel electrophoresis. J. Neurol. Sci. 19,429- 438.

Jackson P. and Thompson R. J. (1984) The immunodetection of brain proteins blotted onto nitrocellulose from fixed and stained two-dimensional polyacrylamide gels. Electrophoresis 5,35-42.

Jacobowitz D. M. and Heydorn W. E. (1984) Two-dimensional gel electrophoresis used in neurobiological studies of proteins in discrete areas of the rat brain. Clin. Chem. 30, 1996-2002.

Jarrard L. E. ( 1 986) Selective hippocampal lesions and behavior: implications for current research and theorizing, in The Hip- pocampus, Vol. 4 (Isaacson R. and Pribram K., eds), pp. 93- 126. Plenum Press, New York.

Kiessling M., Dienel G. A., Jacewicz M., and Pulsinelli W. A. (1986) Protein synthesis in postischaemic rat brain: a two-dimensional electrophoretic analysis. J. Cereb. Blood Flow Metub. 6, 642- 649.

Laemmli U. K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680- 685.

Lim L., Hall C., Leung T., Mahadevan L., and Whatley S. (1983) Neurone-specific enolase and creatine phosphokinase are protein components of rat brain synaptic plasma membranes. J. Neu- rochem. 41, 1177-1 182.

Meml C. R., Goldman D., Sedman S . A,, and Ebert M. H. (1983) Ultrasensitive stain for proteins in polyacrylamide gels shows regional variation in cerebrospinal fluid proteins. Science 21 1,

OFarrell P. H. ( 1975) High resolution two-dimensional electrophoresis of proteins. J. Biol. Chem. 250,4007-402 1.

Patel A. J. and Hunt A. (1989) Regulation of production by primary cultures of rat forebrain astrocytes of a trophic factor important for the development of cholinergic neurons. Neurosci. Leit. 99,

Pearce B. R., Dutton G. R., and White F. P. (1983) Induction of a stress protein in developing cell cultures of the rat cerebellum. J. Neurochem. 41,291-294.

Pickel V. M., Reis D. J., Marangos P. J., and Zomzely-Neurath C. ( 1976) Immunocytochemical localization of nervous system specific protein (NSP-R) in rat brain. Bruin Res. 105, 184- 187.

Schmechel D., Marangos P. J., Brightman M., and Goodwin F. K. (1978) Brain enolases as specific markers of neuronal and glial cells. Science 199, 3 13-3 15.

Sinden J. D., Allen Y. S., Rawlins J. N. P., and Gray J. A. (1990) The effects of ibotenic acid lesions of the nucleus basalis and cholinergic-rich neural transplants on win-stay/lose-shift and win- shift/lose-stay performance in the rat. Behav. Bruin Res. 36,

Strocchi P., Brown B. A,, Young J. D., Bonventre J . A., and Gilbert J. M. (1981) The characterization oftubulin in CNS membrane fractions. J. Neurochem. 37, 1295-1307.

Towbin H., Staehelin T., and Gordon J. (1979) Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc. Natl. Acad. Sci.

1437- 1438.

223-228.

229-249.

USA 76,4350-4354.

J . Neurochem., Vol. 57, No. 5, 1991

Specific Brain Protein Changes Correlated with Behaviourally Effective Brain Transplants

Documents