Rat hippocampal GABAergic molecular markers are differentially
affected by ageing
Jose´ Vela,* Antonia Gutierrez,
Javier Vitorica* and Diego Ruano*
*Departamento de Bioquı´mica, Bromatologı´a, Toxicologı´a y Medicina Legal Facultad de Farmacia, Universidad de Sevilla,
Sevilla, Spain
Departamento de Biologı´a Celular, Facultad de Ciencias, Universidad de Ma´laga, Ma´laga, Spain
Abstract
We previously reported that the pharmacological properties of the hippocampal GABAA receptor and the expression of several subunits are modified during normal ageing. However, correlation between these post-synaptic modifications and pre-synaptic deficits were not determined. To address this issue, we have analysed the mRNA levels of several GABA-ergic molecular markers in young and old rat hippocampus, including glutamic acid decarboxylase enzymes, parvalbumin, calretinin, somatostatin, neuropeptide Y and vasoactive intestinal peptide (VIP). There was a differential age-related decrease in these interneuronal mRNAs that was inversely correlated with up-regulation of thea1 GABA receptor subunit. Somatostatin and neuropeptide Y mRNAs were most fre-quently affected (75% of the animals), then calretinin and VIP mRNAs (50% of the animals), and parvalbumin mRNA (25%
of the animals) in the aged hippocampus. This selective vulnerability was well correlated at the protein/cellular level as analysed by immunocytochemistry. Somatostatin interneu-rones, which mostly innervate principal cell distal dendrites, were more vulnerable than calretinin interneurones, which target other interneurones. Parvalbumin interneurones, which mostly innervate perisomatic domains of principal cells, were preserved. This age-dependent differential reduction of spe-cific hippocampal inteneuronal subpopulations might produce functional alterations in the GABAergic tone which might be compensated, at the post-synaptic level, by up-regulation of the expression of thea1 GABAAreceptor subunit.
Keywords: calcium-binding proteins, GABAAreceptor, inter-neurones, neuropeptides, reverse transcriptase competitive polymerase chain reaction.
J. Neurochem.(2003)85,368–377.
Inhibitory interneurones constitute a heterogeneous cell population that have a powerful inhibitory effect on excita-tory cells and other interneurones, contributing to the generation of oscillatory activity which seems to be important for the integration of synaptic inputs and memory formation (Buzsaki and Chrobak 1995; Paulsen and Moser 1998). In rat hippocampus, interneurones have been classified according to different criteria (morphology, location, innervation, specific protein expression) (Buzsaki and Chrobak 1995). The existence of such cell diversity raises questions about their function. Indeed, there is a relationship between the selective synaptic targeting of different domains of post-synaptic cells and the expression of specific molecular markers (see Freund and Buzsaki 1996 for review). Most perisomatic inhibitory cells express either parvalbumin (PV) or cholecystokinin and vasoactive intestinal polypeptide (VIP); mid-proximal dend-ritic inhibitory cells express calbindin D28k; distal denddend-ritic inhibitory cells express somatostatin (SOM) and 60–70% of this population also express neuropeptide Y (NPY); and
interneurones that specifically innervate other interneurones express CR and, frequently, VIP (Nunziet al. 1985; Feucht
et al. 1999; Maccaferriet al. 2000).
At the post-synaptic level, GABAAreceptors are formed
by particular heteropentameric combination ofa,b,c,d,e,p
and hsubunits, which confers specific kinetic and pharma-cological properties (Sieghart 1995; Sigel et al. 1990). Moreover, single-cell RT–multiplex PCR has revealed that a single neurone can express several isoforms of the different subunit families (Ruano et al. 1997), supporting
Received September 24, 2002; revised manuscript received December 18, 2002; accepted December 27, 2002.
Address correspondence and reprint requests to Dr Diego Ruano, Departamento de Bioquı´mica, Calle/Profesor Garcı´a Gonza´lez, 41012-Sevilla, Spain. E-mail: [email protected]
Abbreviations used: CaBP, calcium-binding protein; CR, calretinin; GAD, glutamic acid decarboxylase; HIPP, hilar perforant path-associ-ated; NPY, neuropeptide Y; O-LM, oriens/lacunosum moleculare; PV, parvalbumin; SOM, somatostatin; VIP, vasoactive intestinal polypeptide.
the hypothesis that GABAAreceptors of different molecular
composition might mediate GABAergic transmission by different interneurone subtypes (Nusser et al. 1995; Tietz
et al. 1999; Maccaferri et al. 2000). However, the precise location of the GABAA receptors in relation to synaptic
inputs is poorly understood (Nyiriet al. 2001).
We have recently demonstrated the existence of an age-dependent up-regulation in the expression of several GABAA
receptor subunits (Ruanoet al. 2000). These modifications were interpreted as a post-synaptic adaptive response, occur-ring duoccur-ring normal ageing, as a consequence of deficits in the GABAergic inputs. In the present study we have addressed this question directly by quantifying, by RT-competitive PCR and immunocytochemistry, the expression of several pre-synaptic markers of GABAergic neurones. These markers included glutamic acid decarboxylase (GAD) 65 and GAD67, as markers of the whole GABAergic population, and the calcium-binding proteins (CaBPs) and neuropeptides PV, CR, SOM, NPY and VIP, as markers of specific subpopulations of GABAergic neurones. Moreover, we have extended our previous work in order to clarify relationship between the pre-and post-synaptic modifications. Use of RT–competitive PCR and immunocytochemistry has allowed us to obtain a global view of the cellular and molecular modifications occurring within the GABAergic system during normal ageing.
Materials and methods
Isolation of hippocampus
Adult (3 months) and aged (24 months) Wistar rats were killed by decapitation and both hippocampi (150–160 mg wet-weight) were dissected, frozen in liquid N2and stored at )80C until use. All
animal experiments were carried out in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals, and were approved by the committee of animal use for research at Sevilla and Malaga Universities. All efforts were made to minimize the number of animals used and their suffering.
RNA extraction
Total RNA was extracted using the TripureTMIsolation Reagent kit (Roche Molecular Biochemicals, Germany), according to the instructions of the manufacturer. The contaminating DNA was removed by incubation with Dnase (Sigma, St Louis, MO, USA) and confirmed by PCR analysis of total RNA samples before RT. Integrity of the RNA samples was assessed by agarose gel electrophoresis. The yield of total RNA was determined by measuring the absorbance (260/280 nm) of ethanol-precipitated aliquots. The recovery of RNA was similar for both young and old hippocampi (Ruanoet al. 2000).
RT–competitive PCR
RT was performed in a final volume of 20lL using 1lg total RNA as template. After RT, samples were treated with Rnase and cDNA was purified using Microcon PCR cartridges (Millipore, Corporation, Bedford, MA, USA) and quantified by measuring the absorbance at
260/280 nm. As a control, samples from adult and aged hippocampi were reverse transcribed as above but in the presence of 2lM
digoxigenin–dUTP, dotted on nylon membranes and developed. The mean ± SDaged/adult ratio of 1.08 ± 0.25 (n¼5) demonstrated a similar RT efficiency for both ages (Ruano et al. 2000). As an additional control,b-actin was amplified in adult and aged rats, for which the aged/adult ratio was 0.99 ± 0.11 (n¼15).
Competitive PCR was performed essentially as described previously (Ruanoet al. 2000). Briefly, 100 ng cDNA was mixed with an increasing amount of each internal standard. Each internal standard was synthesized as described by Velaet al. (2001) and was different in size with respect to the specific PCR product. The range of internal standards used was established previously in control experiments using adult hippocampus.
The PCR reaction was performed using 1lMeach of 5¢and 3¢
primers from the following set (from 5¢to 3¢): GAD65 5¢: TCTT-TTCTCCTGGTGGTGCC, GAD65/67 3¢: CCCCAAGCAGCATC-CACAT (one mismatch with the sequence of GAD67, indicated by the underlined base; Bochetet al. 1994) (position 713 and 1085); GAD67 5¢: TACGGGGTTCGCACAGGTC (position 1159); SOM 5¢ATCGTCCTGGCTTTGGGC, SOM 3¢ GCCTCATCTCGTCCT-GCTCA (position 43 and 231); NPY 5¢: GCCCAGAGCAGAG-CACCC, NPY 3¢: CAAGTTTCATTTCCCATCACCA (position
)45 and 292); VIP 5¢: GCCTTAGCGGAGAATGACA, VIP 3¢: CCTCACTGCTCCTCTTCCCA (position 167 and 434); PV 5¢: AAGAGTGCGGATGATGTGAAGA, PV 3¢: ATTGTTTCTCCAG-CATTTTCCAG (position 115 and 480); CR 5¢: CTGGAGAAGG-CAAGGAAAGGT, CR 3¢: AGGTTCATCATAGGGACGGTTG (position 142 and 429); b-actin 5¢: CGGAACCGCTCATTGCC,
b-actin 3¢: ACCCACACTGTGCCCATCTA. After an initial dena-turation at 94C for 4 min, 32 cycles of PCR were performed (linear phase) consisting of 94C for 45 s, 60C for 45 s and 72C for 50 s followed by a final elongation period of 5 min at 72C. The PCR products were separated on a 1.7% agarose gel, photographed under UV illumination and film images were captured by a rotating scanner (Mustek Scan express 12.000 SP, Germany). The optical density of the competitor and target bands in individual lanes was measured using the program PCBAS 2.08 GmbH, Germany. The log of the ratio of amplified competitor to target fragments was plotted as a function of the log of the known amount of internal standard added to individual PCR reactions. Regression analysis and calculation of x-intercepts were performed with the program Sigmaplot 6.01 SPSS, Chicago, IL, USA. All cDNAs were determined at least in triplicate and, for each experiment, a minimum of one adult and one aged sample was processed and analysed in parallel. The linearity of film exposure was also tested.
Immunocytochemistry
Animals were perfused through the heart under deep anaesthesia (chloral hydrate) with 300 mL heparinized 0.9% saline, followed by 300–400 mL of a mixture of 4% paraformaldehyde, 75 mMlysine
and 10 mMsodium periodate in 0.1Mphosphate buffer, pH 7.4, as
described previously (Gutierrezet al. 1994a). Brains were removed, post-fixed overnight, cryoprotected in 30% sucrose and frozen in liquid N2. Free-floating coronal sections (30lm thick) were
pretreated with 3% H2O2/3% methanol in 0.1Mphosphate-buffered
saline, pH 7.4, for 15 min and with an avidin/biotin blocking kit (Vector Laboratories, Burlingame, CA, USA). Sections from adult
and aged animals were incubated overnight at room temperature (22C) with one of the following primary antibodies: CR goat polyclonal antiserum (Swant Bellinzona, Switzerland) diluted 1 : 4000, PV monoclonal antibody (Sigma) diluted 1 : 10 000, or SOM rabbit antiserum (Rodriguez et al. 1987) diluted 1 : 1000. Sections were rinsed and incubated for 1 h at room temperature in a 1 : 500 dilution of the corresponding biotinylated secondary antibody: anti-goat IgG (Vector Laboratories), anti-mouse IgG (Dako, Carpinteria, CA, USA); anti-rabbit IgG (Vector Laborator-ies), for CR, PV and SOM primary antibodies respectively. Finally sections were incubated with extra-avidin–peroxidase complex (1 : 500) (Sigma) for 2 h at room temperature. Immunoreactive products were visualized by staining with 0.05% 3,3¢ -diaminobenzi-dine, 0.01% hydrogen peroxide and 0.03% nickel ammonium sulphate in phosphate-buffered saline. Sections were mounted on gelatin-coated slides, air dried, dehydrated in alcohol, cleared in xylene, coverslipped with DPX resin and studied in a Nikon Microphot FXA microscope (Nikon, Japan) with bright-field illu-mination. Omission of primary antibody or use of pre-immune serum served as negative controls for all immunocytochemical studies.
Image analysis
The cellular density in the different hippocampal regions (CA1, CA2 and CA3) and dentate gyrus was determined with a computerized image analysis system (Visilog 5.2; Noesis S.A., France). Images of immunostained sections were taken with a Sony CCDvideo camera (Sony, Japan) connected to a Nikon Labophot microscope. The regional areas (CA1, CA2, CA3 and dentate gyrus) were delimited (by morphological criteria) and only immunopositive cells showing a stained unit area sufficient to ensure that the majority of their volume was contained within the selected area (discriminated by eye) were scored. Cell counting was done in a minimum of six slices per antibody and animal, separated by a minimum of four consecutive slices. Cell density values were obtained by dividing the number of immunoreactive neurones of each type by the total area of the corresponding hippocampal region studied. The density of each type of immunolabelled neurones from different hippocampal regions was averaged and expressed as the mean number of positive cells per mm2. It was not our intention to determine the total immunostained cell number in the whole rat hippocampus. We are aware that, if this was the aim, Cavalieri point-counting in combination with three-dimensional optical disector methods would be more appropriate. Our principal interest was to corroborate the molecular data at the morphological and protein levels. However, in order to avoid the potential artifacts related to age-related morphological differences, such as an increase in hippocampal volume, some control measurements were per-formed. We determined the hippocampal volume, the average area each hippocampal region and the average area of the scored cells. We observed only a small increase, mean ± SD8.9 ± 2.2%, in hippo-campal volume in aged rats, but this did not account for the decrease in average cell density observed with age (41 ± 19%; see Fig. 4).
Statistical analysis
One-way or multifactor ANOVA followed by Bonferroni post-hoc multiple comparison tests were used. Comparison of the expression levels of all molecular markers in both age groups was done using a discriminant analysis (Mardiaet al. 1989).
Results
Age-dependent modifications in the expression of the GABAergic markers
GAD65 and GAD67
The expression of both isoenzymes was significantly reduced in the aged group (ANOVA F1,18¼15.66, p< 0.01 and
F1,17¼18.10, p< 0.01 for GAD65 and GAD67
respec-tively). The inter-individual variability in the expression of both isoforms was notably higher in old than in adult rats (Fig. 1a), suggesting a heterogenoeus effect of ageing. For all animals tested (Fig. 1b), the age-dependent modification in both GAD65 and GAD67 mRNAs (expressed as percentage variation) was similar. Six aged rats displayed a decrease in the expression of both GAD65 and GAD67 isoenzymes, whereas the other two aged rats showed no differences (numbers 10 and 26). To rule out the possible differences in the amount of starting cDNA, we determined the expression of b-actin in parallel experiments. No differences were observed (Fig. 1b). These results demonstrated the existence of a severe age-dependent reduction in the expression of the specific enzymes for the synthesis of GABA in rat hippo-campus. However, this effect was heterogeneous in our aged rat population.
The GABAergic interneurones in rat hippocampus consti-tute a highly heterogeneous cell population characterized by the expression of different CaBPs and/or neuropeptides (Freund and Buzsaki 1996). We therefore analysed whether the diminution in GADexpression observed in aged hippocampus was reflected in the expression of several interneuronal markers.
Neuropeptides
SOM, NPY and VIP mRNAs were quantified in the same adult and aged rat populations. The interindividual variability in the expression of all three neuropeptides was higher in old than in adult rats (Fig. 1c). The expression of SOM and NPY decreased significantly during ageing in most of the animals examined (ANOVA F1,12¼15.19, p< 0.01 and F1,12¼
21.26, p< 0.01 for SOM and NPY respectively). For the whole aged population, the decrease in the expression of VIP was not statistically significant. The expression profile of SOM and NPY mRNAs in the aged population was different from that observed for VIP mRNA (Fig. 1c). Expression of SOM and NPY mRNAs declined over a broad range of aged hippocampi, whereas aged rats could be clearly segregated into two groups according to the levels of VIP mRNA: aged animals numbers 1, 12, 18 and 24 showed a remarkable decrease (32.2 ± 7.0 fg per 0.1lg cDNA (n¼5) vs. 6.1 ± 4.8 fg per 0.1lg cDNA (n¼4) for adult and aged rats respectively), whereas the other four aged rats, numbers 7, 10, 16 and 26, displayed no change compared with
adult values (32.2 ± 7.0 fg per 0.1lg cDNA (n¼5) vs. 35.1 ± 2.6 fg per 0.1lg cDNA (n¼4) for adult and aged rats respectively). The expression of VIP in the first group of aged rats was significantly different both from that in adults and that in the other aged group (ANOVA F2,13¼35.53,
p< 0.01; Bonferronip< 0.05; Fig. 1c).
CaBPs
The expression of PV and CR was quantified in the same adult and aged samples. The coefficients of variation for these markers were also higher in aged than in adult rats. The expression of PV mRNA was not significantly modified in the aged population (Fig. 1d); in only two aged animals were the levels of PV clearly reduced. However, CR mRNA was significantly reduced with respect to that in adult animals (ANOVAF1,13¼13.40, p< 0.01). As with VIP, the aged rat
population could be separated clearly into two groups according to the expression of CR mRNA (compare Figs 1c and 1d). Again, aged rats numbers 1, 12, 18 and 24 had a significantly lower level of CR expression than rats 10, 26, 16 and 7, and than the adult population (1.5 ± 1.2 fg per 0.1lg cDNA (n¼4) vs. 9.8 ± 1.0 fg per 0.1lg cDNA
(n¼4) and 13.1 ± 4.8 fg per 0.1lg cDNA (n¼5) respectively; ANOVA F2,13¼114.38, p¼0.001; Bonferroni
p< 0.01). These results suggest the existence of a clear relationship between the expression of VIP and CR in the aged population.
Relationships between the expression of GAD, neuropeptide and CaBP mRNAs in aged rats
Different molecular markers are frequently co-expressed in the same GABAergic cells (Freund and Buzsaki 1996). We therefore analysed whether the expression of molecular markers that are usually co-localized in the same cell population (CR/VIP and SOM/NPY) were similarly affected in aged rats. We first compared the aged/adult ratio of CR and VIP mRNA expression in each aged rat. As shown in Fig. 2a, animals with little or no diminution in VIP mRNA expression also showed little or no diminution CR mRNA expression (numbers 7, 10, 16 and 26). Similarly, animals that had a substantially lower level of VIP mRNA expression showed a similar low level CR mRNA expression (numbers 1, 12, 18 and 24). We also compared the aged/adult ratios for SOM and NPY mRNA expression (Fig. 2b). Apart from rats 1 (a)
(b)
(c)
(d)
Fig. 1Age-dependent change in the expression of molecular markers
of GABAergic neurones. (a) Quantification of mRNAs coding for the GAD65 and GAD67 isoforms in adult and aged rat hippocampus. (b) Percentage variation in GAD65 and GAD67 calculated individually in aged animals. The SD of the adult values (with respect to the mean adult value) is indicated in the figure as a dotted line. The percentage
variation for theb-actin gene in individual aged rats is also shown. (c) and (d) Expression levels of mRNAs coding for the neuropeptides SOM, NPY and VIP, and the CaBPs PV and CR in adult and aged rat hippocampus. Data are shown individually (symbols) and as mean ± SD for the two rat populations (bars). Note the difference in scale for both axes.
and 26, all other aged animals showed a diminution compared with adult values, in the level of expression of both SOM and NPY. However, as also shown in Fig. 2(b), a different degree of expression of these two markers could also be observed within this aged population (i.e. rats 7 and 24). Finally, the level of mRNA for PV, which is frequently expressed alone, decreased exclusively in rats 12 and 24 (Fig. 2c).
The age-dependent variations in the expression of the different markers showed ‘animal-dependent’ specificity. Taking advantage of the fact that the expression of all GABAergic markers in each individual adult and aged rat can be analysed by means of RT–competitive PCR, it is possible to compare individually the age-dependent variation in the expression of the different markers with expression in the total adult population. For this purpose, we applied a discriminant statistical technique to the values for the different molecular markers in individual adult and aged rats. The aged animals were classified according to the existence of significant variations (p< 0.05) compared with the adult values. The degree of alteration was heterogeneous (Fig. 2d). First, in aged rats 10 and 26, the level of GAD65 mRNA was no different from adult values (2082 ± 517 fg (n¼6) and 2107 ± 209 fg (n¼2), for adult and aged rats respectively). Only the expression of SOM showed a decrease in rat 26. The aged rat 10 displayed no modification in the expression of any molecular marker.
Second, aged rats 1, 7 and 16 showed an intermediate decrease in the expression of the enzyme GAD65 (853 ± 243 fg (n¼3); 59% decrease with respect to that in adults). In addition, rats 7 and 16 also a showed decrease in the expression of SOM and NPY, and rat 1 in the expression of NPY, CR and VIP.
Third, aged rats 12, 18 and 24 showed a substantial decrease in the expression of GAD65 (367 ± 214 fg (n¼3); 82% decrease compared with that in adults) and were highly affected, particularly rats 12 and 24, attending to the expression of all molecular markers (SOM, NPY, CR, VIP and PV). The amounts of GAD65 mRNA expressed were significantly different between these three groups of aged rats (ANOVAF2,13¼110.40,p< 0.01; Bonferronip< 0.01).
Importantly, global analysis of mRNA expression in the aged rats revealed that the expression of SOM and NPY mRNAs was more frequently affected by ageing (six of eight aged rats) than the expression of CR and VIP (four of eight) or that of PV (two of eight). Overall, these results demonstrate the existence of a differential vulnerability within the GABAergic system from aged hippocampus.
Pre- and post-synaptic relationship in the GABAergic system during normal ageing
During normal ageing expression of thea1 andc2 GABAA
receptor subunits is up-regulated (Ruano et al. 2000). Moreover, in aged rat hippocampus, the a1-containing GABAA receptors are more sensitive to GABA and/or
benzodiazepines (Ruano et al. 1995); (Griffith and Murch-ison 1995). These post-synaptic modifications might be a consequence of a pre-synaptic deficit in GABAergic tone. As the same aged animals were used in the previous work (Ruanoet al. 2000) and in the present study, it was possible to determine directly the relationship between the age-dependent decrease in pre-synaptic molecular markers and the age-dependent up-regulation of the GABAA receptor
subunit mRNAs. We first quantified the expression of thea1 GABAAreceptor subunit mRNAs in aged rats numbers 16,
(a) (b)
(c) (d)
Fig. 2 Correlation between the age-related
change in expression of neuropeptides, CaBPs and GAD in individual aged animals. The percentage variation, with respect to adult values, in the expression is shown for the CaBPs CR and VIP (a), the neuropep-tides SOM and NPY (b) and PV (c). Values for adult rats were calculated using the individual data respect to the mean value for each molecular marker. (d) Data for GAD, CaBPs and neuropeptides in individ-ual aged rats were combined. The mo-lecular markers that showed a significant diminution with respect to the adult values (p< 0.05) according to the discriminant analysis are shown in grey. The aged rat population was segregated into three groups, according to the expression of the GAD65 enzyme: (i) rats 10 and 26, (ii) rats 1, 7 and 16 and (iii) rats 12, 18 and 24.
18 and 24, which were not included initially in our previous analysis (Fig. 3). As shown in Fig. 3, there was a significant correlation between the age-dependent decrease in the expression of GAD65 mRNA and the age-dependent increase in expression of the a1 subunit mRNA. Animals showing a moderate or no change in the expression of the GAD65 enzyme, did not show modified expression of thea1 GABAAreceptor subunit. Conversely, aged rats displaying
large decreases in the expression of the GAD65 enzyme showed large increases in the expression of thea1 GABAA
receptor subunit. Of all the molecular markers analysed, only SOM and NPY displayed a similar inverse correlation (Fig. 3). These data strongly suggest that, in the aged rat hippocampus, the GABAergic targets have the ability to adapt post-synaptically as a consequence of pre-synaptic deficits in GABAergic inputs.
Immunocytochemistry of PV-, CR- and SOM-positive interneurones in adult and aged rat hippocampus In order to analyse the cell types affected and, importantly, if the decrease observed in the mRNA was also reflected in protein level, we evaluated the average density of immuno-reactive neurones for SOM, CR and PV in rat hippocampus from adult and aged rats. The distribution pattern of the three interneuronal subpopulations in the different regions of the hippocampal formation and dentate gyrus was similar in both adult and aged rats (not shown). The SOM-positive cells were restricted to the stratum oriens in the CA1 and CA2 regions, and showed more widespread distribution in the
CA3 region of the hippocampus and in the hilar region in the dentate gyrus (Figs 4a and c). As shown in Figs 4(b) and (d), there was a clear decrease in the number of SOM-positive cells in the aged rats. The extent of the change was variable in the different areas analysed. Figure 4(i) shows the quantification of the density of SOM-positive cells in adult and aged rats. The more affected areas were the hilar region of the dentate gyrus (37% on average) and the rostral portion of the CA2 subfield (41% on average) of the hippocampal formation. Importantly, the average density of SOM-positive cells was significantly diminished in all the regions analysed from aged rat A (Fig. 4i).
The CR-immunoreactive cells were most widely distri-buted along the hippocampal formation. The average density of CR-immunopositive cells was significantly reduced mostly in the dentate gyrus (50% and 28% for aged rats A and C respectively), CA1 (60% and 52% for aged rats A and C respectively) and the caudal portion of CA3 (75% and 33% for aged rats A and C respectively) (Figs 4e, f and j). No differences were observed in aged rat B.
Finally, PV-positive cells were located in stratum pyram-idale and stratum granulare and, to a lesser extent, in the stratum oriens and stratum radiatum. Quantification of PV-immunopositive cells demonstrated that the average density of interneurones expressing this CaBP was well preserved in aged animals (Figs 4g and h). Two of the three aged rats showed no differences in any of the hippocampal regions (Fig. 4k). However, aged rat A showed a significant depletion in the average density of PV-immunopositive cells in all regions analysed (58, 61, 60 and 71% for CA1, CA2, CA3 and dentate gyrus respectively).
In summary, these results demonstrate that the age-related decreases in the expression of molecular markers observed at the mRNA level are reflected at the protein level and, importantly, a similar differential age-related susceptibility was observed.
Discussion
We have extended and completed our previous studies of the age-related changes of the GABAergic system in rat hippocampus (Ruano et al. 1995, 2000; Gutierrez et al. 1996). The most relevant findings may be summarized as follows: (i) the expression of several molecular markers of different subpopulations of hippocampal interneurones decreased during normal ageing; (ii) the expression of SOM and NPY was more frequently affected than that of CR and VIP, whereas the expression of PV was rarely affected; and (iii) the GABAergic system might be adapted post-synaptically in response to pre-synaptic deficits during normal ageing in rat hippocampus.
Our results demonstrated a progressive decrease in the expression of different molecular markers of the GABAergic system. Indeed, the expression of SOM and NPY was more Fig. 3Correlation between pre- and post-synaptic markers of the
GABAergic system. The relationship between the age-dependent decrease in the mRNA expression of GAD65 and the neuropeptides SOM and NPY, and the age-dependent increase in the mRNA expression of thea1 GABAAreceptor subunit is shown. Expression of
a1 GABAAreceptor subunit mRNA in aged rats numbers 7, 10, 12, 16
and 26 was 131, 115, 150, 128 and 105% over the mean adult value respectively (taken from Ruanoet al. 2000). Data for aged rats num-bers 1, 18 and 24 are from the present work and were obtained by RT–competitive PCR. These values were 127, 145 and 153% over the mean adult values respectively.ANOVArevealed a statistically
signifi-cant relationship between botha1 versus SOM + NPY and GAD65 (p< 0.01) at the 99% confidence level (dotted lines).
frequently reduced than that of CR and VIP, whereas PV was well preserved in aged rats. However, we also found two aged rats in which the expression of PV was decreased. Importantly, in these two animals, the decrease in PV expression occurred in combination with a decrease in all other GABAergic markers.
We are aware that the expression of a particular molecular marker is not restricted to a single interneuronal population. However, it is also true that each marker is mostly expressed by a particular subset of interneuronal populations in hippocampus. Taken that we have analysed expression by RT–competitive PCR, using total RNA extracted from whole hippocampi, the observed age-related modifications in expression of the different molecular markers should reflect modifications in the cell types that more frequently
express these markers. Bearing in mind this limitation, our results suggest that the interneurones innervating distal dendrites of the principal cells (mostly SOM- and NPY-positive cells) and interneurones targeting other interneu-rones (mostly CR- and VIP-positive cells) are firstly affected, whereas the interneurones innervating the periso-matic region of principal cells (PV-positive cells) are less affected, or affected later during the normal ageing process. Therefore, the PV interneuronal subpopulation displays a higher degree of resistance to age-dependent neurodegener-ative processes than the SOM and CR interneuronal subpopulations. A higher level of susceptibility of SOM-positive neurones has also been observed recently in experimental temporal lobe epilepsy (Cossart et al. 2001) and in rats subjected to transient cerebral ischaemia (Bering
(a) (b)
(c) (d)
(e) (f)
(g) (h)
(i)
(j)
(k)
Fig. 4 Age-related modifications in the number of cells immunoposi-tive for SOM, CR and PV in rat hippocampus. Representaimmunoposi-tive sections of adult and aged hippocampus immunostained for SOM (a–d), CR (e and f) and PV (g–h) are shown. The density of both SOM-positive cell subtypes, the O-LM cells (a and b) and HIPP cells (c and d), was decreased in the aged hippocampus (b and d). It is noteworthy that the sections shown in (b) and (d) were from the same aged animal (rat B) as that in (h) (stained by PV antibody) in which no variation was observed. (e, f) In adult rat, CR-positive cells were more abundant in the hilus (e) and stratum lucidum of the CA3 subfield (not shown). The number of CR-positive neurones was significantly decreased in the hilus of aged rats (f). (g, h) PV-containing neurones were mostly located in the stratum pyramidale of hippocampal formation CA1, CA2
and CA3 subfields (not shown) and in the hilar region of the dentate gyrus of both adult (g) and old (h) rats. The number of PV-labelled cells and the pattern of immunoreactive fibre plexuses were similar at both ages. (i–k) Quantitative analysis of the cell density. Data are shown as mean ± SD of the cell density determined in at least six sections per animal. Data from adult rats were pooled because no significant dif-ferences were observed between animals. Results from aged animals are present individually. One-wayANOVAfollowed by Bonferroni post-hoc multiple comparisons test indicated significant differences (*p< 0.05 vs. adult) in the density of positive cells for each marker. G, granule cell layer of dentate gyrus; H, hilus; SO, stratum oriens; SP, stratum pyramidale; SR, stratum radiatum. Scale bars: 75lm (a–d) and 150lm (e–h).
et al. 1997). However, the reason for this increased susceptibility remains unknown.
An important question is whether this age-dependent decrease in the expression of different molecular markers reflects a cell degeneration process or, simply, a loss of their phenotypic properties (hypofunctionality). There is a large body of evidence to indicate that the number of principal cells remains unaltered (or slightly decreases) in normal aged hippocampus, even in aged rats that displayed spatial learning deficits (Gallagher et al. 1996; Rapp et al. 1996; Merrill et al. 2001). However, as far as we know, no such information (i.e. stereological counting) is available for non-principal cells. From the present data, we cannot determine whether the decreased expression of molecular markers, and diminution in immunoreactive cell density, reflects selective age-related GABAergic neuronal death or loss of the GABAergic phenotype (i.e. molecular markers). Several lines of evidence strongly support the existence of GABAergic cell loss during ageing: (i) two different molecular markers that frequently co-localize in the same cell type (such as SOM and NPY or CR and VIP) were similarly affected in most individual aged rats; (ii) expression of both GAD65 and the molecular marker mRNAs decreased in parallel in individual aged rats; and (iii) the age-related decrease in the molecular markers was detected at both mRNA and protein levels. However, in some aged rats (numbers 1 and 26) a parallel decrease in the expression of SOM and NPY was not observed, which may argue against a cellular death process. Therefore, we cannot completely exclude the existence of both processes, hypofunctionality and cellular death, of the hippocampal interneurones during ageing (see Feuchtet al. 1999). Indeed, the two processes might be sequential and, in both situations, the GABAergic tone should be compromised in the aged hippocampus.
The molecular mechanisms that underlie either GABAergic cell loss and/or hypofunctionality are currently unknown. However, for some neurodegenerative disorders, such as Alzheimer’s disease (in which ageing is the principal risk factor), it has been proposed that unfolded proteins accumu-late which, ultimately, might cause endoplasmic reticulum stress and cellular death (Mattson et al. 2000; Imaizumi
et al. 2001). In our aged hippocampal samples, we have recently demonstrated the up-regulation and accumulation of non-functional (unassembled) c2 subunit of the GABAA
receptor (Ruanoet al. 2000). This suggests that accumula-tion of non-funcaccumula-tional protein during normal hippocampal ageing may be responsible for the neuronal degeneration.
Expression of either GAD65 or SOM/NPY decreased concomitantly with an increase in the expression of thea1 GABAAreceptor subunit mRNA. This response seems to be
specific for the hippocampal formation as a decrease in both GAD65 and a1 subunit expression has been described in other brain areas, such as inferior colliculus (Gutierrezet al. 1994b; Casparyet al. 1999). An increase in the efficacy of
the GABAAreceptors, independent of age-dependent
mole-cular modifications, has also been reported in other brain areas (Griffith and Murchison 1995; Casparyet al. 1999).
According to immunocytochemical data, the SOM-posit-ive cells correspond to either the oriens/lacunosum molec-ulare neurones (O-LM cells) in the CA regions and the hilar perforant path-associated cells in the dentate gyrus (HIPP cells) or to the oriens-bistratified cells (Maccaferri et al. 2000). Both the O-LM and HIPP cell types are involved in feedback circuits (Katona et al. 1999). The age-related mRNA up-regulation in the a1 GABAA receptor subunit
was restricted to the stratum pyramidale and stratum granulare of the hippocampal formation (Gutierrez et al. 1996). The a1 increase (mRNA and protein) was directly reflected in the pharmacological properties of the GABAA
receptor (i.e. increase in density and sensitivity of receptors with high affinity for zolpidem) (Ruanoet al. 1995, 2000). Interestingly, the increase in the density of high-affinity zolpidem-binding sites was mostly restricted to the stratum lacunosum moleculare of CA1 and CA2 fields and to the outer molecular layer of the dentate gyrus (Ruano et al. 1995). It is noteworthy that the O-LM cells in the CA1 and CA2 regions, and the HIPP cells in the hilus, project their axons from the stratum oriens to the stratum lacunosum moleculare and from the hilus to the outer molecular layer of the dentate gyrus respectively (Freund and Buzsaki 1996). Their post-synaptic elements are mainly dendritic shafts and spines of GABA-negative principal cells (Katonaet al. 1999; Maccaferriet al. 2000). Thus, we propose that the pharma-cological modifications and up-regulation of the expression of the a1 GABAA receptor subunit constitute an adaptive
process of the principal cells, as a consequence of the possible loss of GABAergic input, or decreased GABA release, from both the O-LM and HIPP cells, to maintain their network properties (i.e. feedback inhibition).
Considering that hippocampal neuronal networks are composed of inhibitory neurones that control the excitability and synchronization of the principal cells (Cobbet al. 1995); Paulsen and Moser 1998; Feuchtet al. 1999), a specific and differential loss of inhibitory inputs might be reflected in some of the properties of the network (integrative properties, firing rate, general excitability, etc.). The O-LM neurones seem to be activated by repetitive action potentials, dis-charged by different pyramidal cells during theta frequency (Siket al. 1995; Feuchtet al. 1999). Moreover, it has been proposed that O-LM cells (SOM positive), along with axo-axonic and basket cells (both PV positive), may cooperate with theta frequency inputs to CA1 pyramidal cells in dendrites and perisomatic terminals respectively (Katona
et al. 1999; Maccaferriet al. 2000; Van Hooftet al. 2000). The differential vulnerability of these interneuronal popula-tions to the ageing process (SOM-positive cells more vulnerable than than PV-positive cells) might produce an imbalance between dendritic and perisomatic inhibition in
the principal cells of the aged rat hippocampus. This differential vulnerability of GABAergic cells to the normal ageing process might be implicated in the lower mean discharge rate of hippocampal theta cells located in the stratum oriens (probably corresponding to O-LM cells) with respect to those located on stratum pyramidale, observed in aged hippocampus (Mizumoriet al. 1992).
The reasons for this differential susceptibility of hippo-campal GABAergic neurones to the ageing process are unknown. Differences in afferent input, targeting the GAB-Aergic neurones, might represent a potential source of the differential vulnerability. Indeed, the density of inputs as well as the total number of excitatory plus inhibitory synapses was several times higher on PV cells than on calbindin or CR cells (Gulyas et al. 1999). On the other hand, a specific decrease in the number of SOM-immunopositive cells in rat hippocampus has been observed by cholinergic deafferenta-tion of subcortical pathways to the hippocampus, but not by noradrenergic or serotonergic deafferentation (Jolkkonen
et al. 1997). The preferential susceptibility of the SOM-and/or NPY-positive cells might reflect a lower degree of innervation as well as a decrease in the cholinergic inputs in the aged hippocampus.
In conclusion, our data provide the first evidence for a progressive loss of GABAergic input to the hippocampal GABAergic neurones during normal ageing that seems to be compensated at the post-synaptic level. Moreover, our results suggest a cellular and molecular basis for the higher basal neuronal excitability observed in the aged hippocampus.
Acknowledgements
This work was supported by grant 00/0314 from Fondo de Investigaciones Sanitarias (FIS) and a contract from the Ramon y Cajal programme from Ministerio de Ciencia y Tecnologı´a to DR and a fellow from FIS to JV.
References
Bering R., Draguhn A., Diemer N. H. and Johansen F. F. (1997) Ischemia changes the coexpression of somatostatin and neuro-peptide Y in hippocampal interneurons. Exp. Brain Res. 115, 423–429.
Bochet P., Audinat E., Lambolez B., Crepel F., Rossier J., Lino M., Tsuzuki K. and Ozawa S. (1994) Subunit composition at the single-cell level explains functional properties of a glutamate-gated channel.Neuron12, 383–388.
Buzsaki G. and Chrobak J. J. (1995) Temporal structure in spatially organized neuronal ensembles: a role for interneuronal networks. Curr. Opin. Neurobiol.5, 504–510.
Caspary D. M., Holder T. M., Hughes L. F., Milbrandt J. C., McKernan R. M. and Naritoku D. K. (1999) Age-related changes in GABA(A) receptor subunit composition and function in rat auditory system.Neuroscience93, 307–312.
Cobb S. R., Buhl E. H., Halasy K., Paulsen O. and Somogyi P. (1995) Synchronization of neuronal activity in hippocampus by individual GABAergic interneurons.Nature378, 75–78.
Cossart R., Dinocourt C., Hirsch J. C., Merchan P. A., De Felipe J., Ben Ari Y., Esclapez M. and Bernard C. (2001) Dendritic but not somatic GABAergic inhibition is decreased in experimental epi-lepsy.Nat. Neurosci.4, 52–62.
Feucht M., Fuchs K., Pichlbauer E., Hornik K., Scharfetter J., Goessler R., Fureder T., Cvetkovic N., Sieghart W., Kasper S. and Aschauer H. (1999) Possible association between childhood absence epilepsy and the gene encoding GABRB3.Biol. Psychi-atry46, 997–1002.
Freund T. F. and Buzsaki G. (1996) Interneurons of the hippocampus. Hippocampus6, 347–470.
Gallagher M., Landfield P. W., Mcewen B., Meaney M. J., Rapp P. R., Sapolsky R. and West M. J. (1996) Hippocampal neurodegenera-tion in aging.Science274, 484–485.
Griffith W. H. and Murchison D. A. (1995) Enhancement of GABA-activated membrane currents in aged Fischer 344 rat basal fore-brain neurons.J. Neurosci.15, 2407–2416.
Gulyas A. I., Megias M., Emri Z. and Freund T. F. (1999) Total number and ratio of excitatory and inhibitory synapses con-verging onto single interneurons of different types in the CA1 area of the rat hippocampus. J. Neurosci. 19, 10082– 10097.
Gutierrez A., Khan Z. U. and Deblas A. L. (1994a) Immunocyto-chemical localization of c2 short and c2 long subunits of the GABAA receptor in the rat brain. J. Neurosci. 14, 7168– 7179.
Gutierrez A., Khan Z. U., Morris S. J. and Deblas A. L. (1994b) Age-related decrease of GABAAreceptor subunits and glutamic acid decarboxylase in the rat inferior colliculus.J. Neurosci.14, 7469– 7477.
Gutierrez A., Khan Z. U., Ruano D., Miralles C. P., Vitorica J. and Deblas A. L. (1996) Aging-related subunit expression changes of the GABAA receptor in the rat hippocampus. Neuroscience74, 341–348.
Imaizumi K., Miyoshi K., Katayama T., Yoneda T., Taniguchi M., Kudo T. and Tohyama M. (2001) The unfolded protein response and Alzheimer’s disease.Biochim. Biophys. Acta – Mol. Basis Dis.1536, 85–96.
Jolkkonen J., Kahkonen K. and Pitkanen A. (1997) Cholinergic deaf-ferentation exacerbates seizure-induced loss of somatostatin-immunoreactive neurons in the rat hippocampus.Neuroscience80, 401–411.
Katona I., Acsady L. and Freund T. F. (1999) Postsynaptic targets of somatostatin-immunoreactive interneurons in the rat hippocampus. Neuroscience88, 37–55.
Maccaferri G., Roberts J. D. B., Szucs P., Cottingham C. A. and Som-ogyi P. (2000) Cell surface domain specific postsynaptic currents evoked by identified GABAergic neurones in rat hippocampus in vitro.J. Physiol.524, 91–116.
Mardia K. V., Kent J. K. and Bibby J. M. (1989) Discriminant ana-lysis, in Multivariate Analysis, pp. 300–332. Academic Press, London.
Mattson M. P., LaFerla F. M., Chan S. L., Leissring M. A., Shepel P. N. and Geiger J. D. (2000) Calcium signaling in the ER: its role in neuronal plasticity and neurodegenerative disorders.Trends Neu-rosci.23, 222–229.
Merrill D. A., Chiba A. A. and Tuszynski M. H. (2001) Conservation of neuronal number and size in the entorhinal cortex of behaviorally characterized aged rats.J. Comp. Neurol.438, 445–456. Mizumori S. J., Barnes C. A. and McNaughton B. L. (1992) Differential
effects of age on subpopulations of hippocampal theta cells. Neu-robiol. Aging13, 673–679.
Nunzi M. G., Gorio A., Milan F., Freund T. F., Somogyi P. and Smith A. D. (1985) Cholecystokinin-immunoreactive cells form
symmetrical synaptic contacts with pyramidal and nonpyramidal neurons in the hippocampus.J. Comp. Neurol.237, 485–505. Nusser Z., Roberts J. D. B., Baude A., Richards J. G., Sieghart W. and
Somogyi P. (1995) Immunocytochemical localization of thea1 and b2/3 subunits of the GABAAreceptor in relation to specific GABA-ergic synapses in the dentate gyrus.Eur. J. Neurosci.7, 630–646. Nyiri G., Freund T. F. and Somogyi P. (2001) Input-dependent synaptic
targeting ofa2-subunit-containing GABAAreceptors in synapses of hippocampal pyramidal cells of the rat.Eur. J. Neurosci.13, 428–442.
Paulsen O. and Moser E. I. (1998) A model of hippocampal memory encoding and retrieval: GABAergic control of synaptic plasticity. Trends Neurosci.21, 273–278.
Rapp P. R., Burwell R. D. and West M. J. (1996) Individual differences in aging: implications for stereological studies of neuron loss – comment.Neurobiol. Aging17, 495–496.
Rodriguez E. M., Hein S., Rodriguez S., Herrera H., Peruzzo B., Nualartand F. and Oksche A. (1987) Analysis of the secretory products of the subcommisural organ, inFunctional Morphology of Neuroendocrine Systems: Evolutionary and Environmental Aspects(Sharrer B., Korf H. W. and Hartwig H. G., eds), pp. 345– 360. Springer, Berlin.
Ruano D., Benavides J., Machado A. and Vitorica J. (1995) Aging-associated changes in the pharmacological properties of the benzodiazepine (w) receptor isotypes in the rat hippocampus. J. Neurochem.64, 867–873.
Ruano D., Perrais D., Rossier J. and Ropert N. (1997) Expression of GABAAreceptor subunit mRNAs by layer V pyramidal cells of the rat primary visual cortex.Eur. J. Neurosci.9, 857–862. Ruano D., Araujo F., Revilla E., Vela J., Bergis O. and Vitorica J. (2000)
GABAA and alpha-amino-3-hydroxy-5-methylsoxazole-4-prop-ionate receptors are differentially affected by aging in the rat hip-pocampus.J. Biol. Chem.275, 19585–19593.
Sieghart W. (1995) Structure and pharmacology of gamma-aminobutyric acidAreceptor subtypes.Pharmacol. Rev.47, 181–234. Sigel E., Baur R., Trube G., Mohler H. and Malherbe P. (1990) The
effect of subunit composition of rat brain GABAA receptors on channel function.Neuron5, 703–711.
Sik A., Penttonen M., Ylinen A. and Buzsaki G. (1995) Hippocampal CA1 interneurons: anin vivointracellular labeling study.J. Neu-rosci.15, 6651–6665.
Tietz E. I., Kapur J. and MacDonald R. L. (1999) Functional GABAA receptor heterogeneity of acutely dissociated hippocampal CA1 pyramidal cells.J. Neurophysiol.81, 1575–1586.
Van Hooft J. A., Giuffrida R., Blatow M. and Monyer H. (2000) Dif-ferential expression of group I metabotropic glutamate receptors in functionally distinct hippocampal interneurons. J. Neurosci. 20, 3544–3551.
Vela J., Vitorica J. and Ruano D. (2001) Rapid PCR-mediated syn-thesis of competitor molecules for accurate quantification ofb2 GABAAreceptor subunit mRNA.Brain Res. Brain Res. Protoc.8, 184–190.
