|Home | About | Journals | Submit | Contact Us | Français|
Caloric restriction enhances N-methyl-D-aspartate (NMDA) receptor binding and upregulates messenger RNA expression of the GluN1 subunit during aging. Old growth hormone receptor knockout mice resemble old calorically restricted rodents in enhanced life span and brain function, as compared with aged controls. This study examined whether aged growth hormone receptor knockout mice also show enhanced expression of NMDA receptors. Six or 23- to 24-month-old male normal-sized control or dwarf growth hormone receptor knockout mice were assayed for NMDA-displaceable [3H]glutamate binding (autoradiography) and GluN1 subunit messenger RNA (in situ hybridization). There was slight sparing of NMDA receptor binding densities within aged medial prefrontal and motor cortices, similar to caloric restriction, but there were greater age-related declines in GluN1 messenger RNA in growth hormone receptor knockout versus control mice. These results suggest that some of the functional improvements in aged mice with altered growth hormone signaling may be due to enhancement of NMDA receptors, but not through the upregulation of messenger RNA for the GluN1 subunit.
DECLINES in brain functions during aging, including memory, strength, sensation, balance, and motor coordination, affect almost half of the human population over 65 years of age (1). Memory is one of the earliest of the cognitive functions to show declines during aging (2), and deficits are seen in aged humans and nonhuman primates (see reviews 3,4), dogs (5), and rodents (6–9). Humans and rodents both exhibit deficits in spatial memory during aging (7–17). Spatial memory normally involves brain regions such as the hippocampal formation and temporal and prefrontal cortices (18–23). Normal aging is also associated with a decline in memory retention for inhibitory avoidance (24–26). This task engages the prefrontal cortex, specifically the anterior cingulate and insular cortices, and the hippocampal formation (27–29). Age-related alterations in posture and movement are also a major problem for the elderly, resulting in injury and death (1,30,31). Skilled and purposeful movements are controlled by the motor cortices (1). Analyses of age-related changes in the cerebral cortex and hippocampal formation may provide clues for some of the important physiological changes seen during aging.
The N-methyl-D-aspartate (NMDA) receptors, a subtype of glutamate receptor, are found in the highest densities in the cerebral cortex and hippocampal formation (32). This receptor is particularly important in neuronal plasticity (33). NMDA antagonists impair formation of long-term memory, including spatial memory (34–37), and block the initiation of long-term potentiation in the hippocampal formation (37–39) and cerebral cortex (40). NMDA receptors also play a role in inhibitory avoidance memory (41–43). In addition, pharmacological and genetic manipulations of the NMDA receptor indicate that it is important for motor coordination (44–46). Thus, the NMDA receptor may be important in age-related declines in these functions.
Aging animals exhibit declines in NMDA receptor binding densities and functions. NMDA-stimulated release of transmitters is decreased with increasing age (47,48). Long-term potentiation is also altered in aged rodents (49,50). Age-related declines in binding of glutamate and/or [(±)-2-carboxypiperazin-4-yl] propyl-1-phosphonic acid to NMDA binding sites have been noted in multiple regions of the cerebral cortex and hippocampus in mice, rats, dogs, and monkeys (51–56). Humans also exhibit declines with age in [3H]MK801 binding to the NMDA receptor complex in the frontal cortex (57). This age-related decrease in NMDA binding site density within the frontal cortex (58) and hippocampal formation (54,58) of rodents has been correlated with poor performance in long-term spatial memory tasks, such as the Morris water maze. In addition, age-related reductions in binding of [3H]MK801 in the channel of the NMDA receptor are positively correlated with deficits in memory retention for inhibitory (passive) avoidance (26).
Some of the age-related changes in the NMDA receptor appear to be due to alterations in the expression of receptor subunits (for review, see 59). The GluN1 (previous nomenclature: mouse, ζ1; other species, NR1, 60) subunit has the same distribution as NMDA-displaceable [3H]glutamate binding (61–63). The effects of aging on the GluN1 subunit are variable between studies and species (59). Significant reductions during aging in the GluN1 subunit messenger RNA (mRNA) and/or protein have been seen in the cerebral cortex and/or hippocampal formation of C57BL/6 mice (55,64–66) and the hippocampal formation of Fischer 344 and Fischer 344 × Brown Norway F1 rats (67–71) and aged macaque monkeys (72). However, there are other studies using C57BL/6 mice (73–75) or Fischer 344 × Brown Norway F1, Wistar or Long–Evans rats that show no age-related change in the expression of the GluN1 subunit in frontal cortex and/or hippocampal formation (76–78). Correlation studies, regardless of whether or not there are age-related changes, suggest that GluN1 subunit expression in aged individuals contributes to spatial memory performance (66,73,76,79). The variability in the effects of aging may indicate that the changes that are observed in the expression of the GluN1 subunit are not due to a programmed decline that occurs across species, but rather are the result of environmental factors and experiences during aging.
In support of this, exposure to a behavioral testing experience can not only lead to an increase in the expression of mRNA for the GluN1-a splice variant in the prefrontal cortex of aged mice only but also exacerbate the age-related decline in the GluN1-3 splice variant (73). Calorie restriction, to 60% of ad-libitum intake beginning at 3 months of age, resulted in the maintenance of slightly higher levels of [3H]glutamate binding to the NMDA site in regions of the cerebral cortex and hippocampus in older C57Bl/6 mice (80,81). This effect in frontal cortical and hippocampal regions is associated with improved spatial memory performance in the water maze (58,81). Caloric restriction can also lead to significant increases in mRNA expression of the GluN1 subunit in medial and lateral frontal cortex in middle-aged and old mice, as compared with ad-libitum fed animals (81). Fischer 344 rats also show a positive effect of caloric restriction on GluN1 subunit protein expression in the aged hippocampal formation (68). Because caloric restriction is not a very practical intervention for human aging, it is important to understand the mechanisms underlying its effects.
In addition to positive effects on memory performance (82), including spatial (81,83,84) and passive avoidance (85,86) tasks, caloric restriction leads to extensions of the median and maximum life spans, reduction of adult body size, a delay in puberty, a reduction in fertility, and reduced plasma insulin-like growth factor-1 and insulin levels (87). Several lines of dwarf mice exhibit these same characteristics (88–90). Laron dwarf (growth hormone receptor knockout [GHRKO]) mice have a targeted disruption of the mouse growth hormone receptor or binding protein gene (90). In addition to the similarities to calorie restriction mentioned above, old GHRKO mice also show improved spatial memory, retention of inhibitory avoidance information, and locomotion relative to age-matched control siblings (24,91).
The present study was designed to determine whether alterations in the growth hormone signaling pathway could produce positive effects on NMDA receptor binding and GluN1 mRNA expression during aging, similar to those previously seen with caloric restriction in mice (81,92). In horizontal sections, the region identified as frontal cortex in the calorie restriction studies could have included insular and primary and secondary motor and somatosensory cortices (81,92). These rostral cerebral cortical regions were analyzed separately in the present study. The region labeled cingulate cortex in the calorie restriction study was located medially and could have included infralimbic, prelimbic, and cingulate cortices (81). This combined region will be referred to as medial prefrontal in the current study. Dorsal and intermediate hippocampal formations were analyzed separately because differences occur in the effects of aging between these regions in C57BL/6 mice (74).
Male GHRKO mice and their control littermates were supplied by the breeding colony at Southern Illinois University. The colony was established by crossing 129Ola and BALB/c N (GHR+/–) animals with mice derived from crosses of C57BL/6 and C3H/J strains and has been maintained on this heterogeneous genetic background with minimal inbreeding by avoiding Brother × Sister matings. A heterogeneous strain background was chosen in order to reduce the likelihood of obtaining results that may be unique to only one inbred strain and to more closely resemble a natural population. The knockout dwarf (GHR–/–; GHRKO) and heterozygous, normal-sized (GHR+/–; control) mice used in the present study were obtained by crossing knockout (GHR–/–) males with heterozygous (GHR+/–) females. As littermates, the GHRKO and control mice were matched for both genetic background and environmental effects (e.g., intrauterine environment and maternal care). The knockout mice were differentiated from the controls based on phenotypic differences, that is, reduced body size, weight, and length, which have been previously described for GHRKO mice (89,90). Heterozygous mice are phenotypically similar to wild-type (GHR+/+) mice in body weight, body length, serum GH and insulin-like growth factor-1, and life span (89,90).
There were two age groups (6 and 23–24 months of age) and two genotypes (GHRKO and control) used in this study. The N for each group was 4 for 6-month-old controls, 7 for 6-month-old GHRKOs, and 5 for both the 23- to 24-month-old GHRKO and control mice. Animals were housed under 12:12 hour light:dark cycles, with controlled temperature (20–23°C) and access to pelletized food and water ad libitum at Southern Illinois University. Mice were anesthetized with pentobarbital (100 mg/kg intraperitoneally) and decapitated. The brains were quickly removed, frozen on dry ice, and shipped to the Magnusson laboratory. Coronal 20-μm sections were obtained through prefrontal, motor, and somatosensory cortices (between Bregma +2.22 and –0.94 mm, 93) and the dorsal hippocampal formation (from septal half of hippocampal formation, between Bregma –1.22 and –2.54 mm, 93) and horizontal 20-μm sections were obtained through the intermediate hippocampal formation (located between septal half and temporal quarter of hippocampal formation; approximately between 2.5 and 4 mm deep to the skull, 93) with the use of a Microm HM500 cryostat (Zeiss, Thornwood, NY). Sections were cold-mounted onto gelatin-coated glass slides. Due to technical difficulties, the CA1 and CA3 regions were missing from many of the intermediate hippocampal sections, so those regions were not analyzed. Each slide contained a section from each age and genotype group. The placement order of the groups on the slides was varied between assays to decrease variability due to slide handling.
Binding was performed as previously described (55) with a modification in the incubation time. Slides were preincubated in cold (4°C) 50-mM Tris acetate buffer (pH 7.0) for 30 min, followed by 2 × 10 min incubations in cold Tris acetate buffer (4°C, pH 7.0). Sections were then incubated in a solution of 100-nM [3H]L-glutamate (New England Nuclear, Boston, MA), 1-μM kainate, 5-μM α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid, and 100-μM 4-acetamido-4’-isothiocyanotostilbene-2,2’-disulfonic acid for 45 min at 4°C. Slides were rinsed four times in Tris acetate buffer (4°C) for a total of 30 seconds and dried by a stream of compressed air at room temperature. Unlabeled 200-μM NMDA was added to the incubation solution for determining nonspecific binding.
Slides were allowed to dry overnight and were apposed to tritium sensitive film, Hyperfilm-3H (Amersham, Piscataway, NJ), along with tritium standards (Amersham) at –20°C for 2 months. Exposed films were developed in D-19 developer (Kodak, Rochester, NY). Brain and standard images were captured using a Macintosh G4 computer with a Powerlook 2100XL scanner (UMAX, Taiwan) and NIH Image software. Quantitative densitometry was performed on the images from four sections for total binding and two sections for nonspecific binding from each animal with the use of NIH Image software. Tritium standards were used to convert optical density measurements to femtomole per milligram protein for each brain region analyzed. Specific binding was determined by subtracting nonspecific binding from total binding.
In situ hybridization was performed as previously described (55). Oligonucleotides were commercially prepared (Macromolecular Resources, Colorado State University, Fort Collins, CO) for the GluN1 subunit (Probe sequence: 5’ GCACAGCGGGCCTGGT-TCTGGGTTGCGCGAGCGCGACCACCTCGC; complimentary to nucleotide residues –54 to –10). Probes were tailed with 33P-dATP (ICN, Costa Mesa, CA) using terminal deoxyribonucleotidyl transferase (New England Nuclear, Boston, MA) and separated from unbound label with Microspin G-25 columns (GE Healthcare Bio-Sciences Corp., Piscataway, NJ). In situ hybridization was performed according to the method of Watanabe and coworkers (94). Each solution step was performed with gentle rotation, except for the fixation and hybridization steps. Slides were thawed, air dried, fixed in 4% paraformaldehyde in phosphate-buffered saline (pH 7.2, 25°C) for 15 minutes, placed in 2 mg/ml glycine in phosphate-buffered saline (pH 7.2, 25°C) for 20 minutes, and acylated in freshly prepared 0.25% acetic anhydride in 0.1-M triethanolamine (pH 8.0, 25°C) for 10 minutes. Slides were blocked for 2 hours in a prehybridization solution, which consisted of 50% formamide, 0.1-M Tris–HCl (pH 7.5), 4× saline and sodium citrate solution (SSC; 1× SSC = 150-mM NaCl and 15-mM sodium citrate), 0.02% Ficoll, 0.02% polyvinylpyrrolidone, 0.02% bovine serum albumin, 2% sarkosyl, and 250 μg/ml salmon testes DNA. Chemicals were purchased from Sigma Chemical Co. (St. Louis, MO). Slides were then successively washed for 5 minutes each in 2× SSC and 70% and 100% ethanol and air dried for 15 minutes. Hybridization was performed by placing 150 μl of prehybridization solution plus 10% dextran sulfate and 2 picomoles 33P-labeled oligonucleotide probe per milliliter of prehybridization solution onto the slides, covering the sections with Parafilm (American National Can, Greenwich, CT) and incubating them for 18 hours in a 42°C oven humidified with 5× SSC. Parafilm was removed, and slides were rinsed for 40 minutes in 2× SSC and 0.1% sarkosyl (25°C) and twice for 40 minutes each in 0.1× SSC and 0.1% sarkosyl (55°C) and then air dried. Nonspecific hybridization was determined by addition of 20-fold excess unlabeled oligonucleotide to the hybridization solution on some slides. Sections were exposed to Kodak Biomax film (Eastman Kodak Co., Rochester, NY) for 3 days, along with 14C standards (American Radiolabeled Chemicals, Inc., St. Louis, MO). Brain images were captured and analyzed as described for receptor binding. The standards were used to convert optical density to picomoles 33P per milligram tissue. Specific signal was determined by subtracting nonspecific from total hybridization.
The binding and mRNA data were normalized between groups, based on the 6-month-old control mice, as previously described (80), in order to decrease variability due to film apposition. Age-related differences in the densities of mRNA and receptor binding were analyzed separately for rostral cortex, dorsal hippocampal formation, and intermediate dentate gyrus by repeated-measures analysis of variane (Age × Genotype × Brain Region) followed by Fisher's protected least significant difference with the use of Statview software (SAS Institute, Inc., Cary, NC). The analyses of individual brain regions and within genotype age differences were part of the original experimental plan.
There was a near significant main effect of Age, F(1,17) = 3.4, p = .08, but no significant main effect of Genotype, F(1,17) = .003, p = .96, on NMDA-displaceable [3H]glutamate binding overall within the rostral cerebral cortical regions analyzed. There was, however, a significant interaction between Age, Genotype, and Brain Region, F(9,153) = 2.3, p = .02, on NMDA-displaceable [3H]glutamate binding within the cortical regions (Figures (Figures1A–D,1A–D, N and and2A).2A). There was a significant main effect of Age, when data were collapsed across Genotype, on NMDA-displaceable [3H]glutamate binding in layers I–III of the primary motor (p = .02) and combined primary and secondary somatosensory (p = .04) cortices and layers IV–VI of the primary motor cortex (p = .03); 6-month-old mice exhibited higher binding densities than 23- to 24-month olds (Figure 2A). There were significant decreases in binding densities of NMDA-displaceable [3H]glutamate between 6 and 23–24 months of age in control mice in layers IV–VI of the medial prefrontal (p = .02), secondary motor (p = .006), and primary motor (p = .01) cortices and layers I–III of the primary motor cortex (p = .006; Figures 1A, B, N and and2A).2A). In contrast, no significant effects of Age were noted in GHRKO mice in the cortical regions examined (Figures (Figures1C,1C, D, N and and22A).
There was no significant main effect of Age, F(1,15) = .09, p = .77, or Genotype, F(1,15) = .19, p = .67, on NMDA-displaceable [3H]glutamate binding overall within the dorsal hippocampal formation (Figures (Figures1E–H,1E–H, O and and2B).2B). There was also no significant main effect of Age, F(1,6) = .01, p = .92, or Genotype, F(1,6) = .92, p = .37, on NMDA-displaceable [3H]glutamate binding overall within the intermediate dentate gyrus (Figures (Figures1I–L,1I–L, P and and2B2B [far right]). No individual regions in the hippocampal formation showed any significant effects of Age or Genotype (p range = .14 to .96; Figure 2B). The N for the intermediate CA1 and CA3 regions was insufficient to report.
There was a significant main effect of Age, F(1,17) = 11.1, p = .004, but no main effect of Genotype, F (1,17) = 1.1, p = .3, on mRNA density for the GluN1 subunit of the NMDA receptor within the rostral cerebral cortex (Figures 3A–D and and4A).4A). There was a significant main effect of Age on the density of mRNA for the GluN1 subunit in cortical layers II–III of the medial prefrontal (p = .01), insular (p = .02), and primary motor (p = .006) cortices and in layers IV–VI of the medial prefrontal (p = .02), secondary motor (p = .002), primary motor (p = .001), and combined primary and secondary somatosensory (p = .04) cortices; 23- to 24-month-old mice had lower mRNA densities than the 6-month-old mice when the data were collapsed across genotypes (Figure (Figure4A).4A). The 23- to 24-month-old control mice had significantly lower densities of mRNA for the GluN1 subunit in layers II–III of the medial prefrontal cortex, as compared with 6-month-old control mice (p = .03; Figures 3A,B and and4A).4A). The 23- to 24-month-old GHRKO mice had significantly less mRNA density for the GluN1 subunit than the 6-month-old GHRKO mice in layers II–III of the primary motor cortex (p = .002) and in layers IV–VI of the medial prefrontal (p = .02), secondary motor (p = .006), primary motor (p = .007), and combined primary and secondary somatosensory (p = .04) cortices (Figures 3C, D and and44A).
There was no significant main effect of Age, F(1,16) = 1.2, p = .29, or Genotype, F(1,16) = .15, p = .70, on the mRNA density of the GluN1 subunit overall within the dorsal hippocampal formation or within the regions of the dorsal hippocampal formation (p range = .09 to .96; Figures 3E–H and and4B).4B). There was a significant main effect of Age, F(1,11) = 9, p = .01, but no main effect of Genotype, F(1,11) = 1.3, p = .3, on GluN1 mRNA density in the upper and lower blades of the intermediate dentate gyrus (Figures 3I–L and and4B).4B). The 6-month-old mice had higher mRNA densities for GluN1 than 23- to 24-month-old mice overall in the intermediate dentate gyrus (Figure 3I–L) and separately within the lower and upper blades of the intermediate dentate gyrus (Figure 4B), both when data were collapsed across genotypes and when the GHRKO mice were analyzed separately. The N for the intermediate CA1 and CA3 regions was insufficient to report.
There were three main findings from this study: (i) there were age-related changes in NMDA receptor binding density in motor and somatosensory cortices and in GluN1 subunit expression in many rostral cortical regions and intermediate dentate gyrus, when both GHRKO and control mice were averaged together; (ii) there was a slight sparing of NMDA receptor binding density within medial prefrontal and motor cortices of aged GHRKO mice; (iii) there was a greater age-related decline in the GluN1 mRNA expression in GHRKO mice in all of the same regions that showed the sparing in NMDA receptor binding density, as compared with controls.
Three out of ten rostral cortical regions showed a significant decline in NMDA-displaceable [3H]glutamate binding between 6 and 24 months of age in these mice, regardless of knockout-related genotype. Primary motor and somatosensory cortices appeared to be more susceptible to aging than prefrontal regions across both genotypes. An age-related decrease in NMDA binding site densities has also been demonstrated within the cortex and/or hippocampal formation of C57Bl/6 and BALB/c mice (55,80,95,96), Fischer 344 (52,53,97–101, but see 102), Long–Evans (54), Wistar (103), and Sprague–Dawley (104) rats, dogs (56), and rhesus monkeys (51). Declines in the binding within the channel of the NMDA receptor have also been described in the frontal cortex of humans during aging (57). The lack of age-related changes in glutamate binding within the hippocampal formation regions is consistent with a lesser effect of aging on NMDA binding sites in these regions as compared with the cerebral cortex in C57BL/6 mice (81,92). The mice in the current study, from a heterogeneous strain background, thus showed similar age-related changes in NMDA receptor binding density to many other strains and species. The heterogeneity in the effects of aging on NMDA receptor binding between different brain regions seen in the present study has also been documented in C57BL/6 mice (59,74).
There was a slight sparing of NMDA-displaceable [3H]glutamate binding in the old GHRKO dwarf mice (9–17% better maintenance of binding density across aging), as compared with control, normal-sized mice, within the deep layers of the medial prefrontal and secondary motor cortices and in all layers of the primary motor cortex. In these regions, normal-sized aged mice showed a significant decline in binding density (21–30%), as compared with the normal-sized young mice. The decreases in binding between young and old GHRKO mice (8–14% of young GHRKO binding) did not reach significance in the same regions. It is possible that the differences in percent of age-related decline in NMDA receptor binding would have been greater in magnitude if the GHRKO mice had been compared with wild-type mice, rather than heterozygotes. Caloric restriction also produced a similar pattern with respect to significant changes in horizontal sections of the rostral cerebral cortex; ad libitum-fed mice showed significant aging changes (13–22% reduction from young ad libitum-fed) but calorie-restricted mice did not (5–14% reduction from young ad libitum-fed; 58,81). Caloric restriction in aged mice led to 3–15% better maintenance of NMDA receptor binding densities within rostral cortex across aging, as compared with the old ad libitum-fed mice (58,81). Thus, alteration of the growth hormone signaling pathways led to a similar pattern to caloric restriction within rostral cortical regions with respect to improvement in glutamate binding to NMDA receptors in aged animals, both in the degree of reduction of the age-related change, as compared with old controls, and the percent decrease from binding densities found in the young of the same genotype.
Seven out of 10 of the cortical regions and both blades of the intermediate dentate gyrus showed a significant decline in mRNA expression of the GluN1 subunit of the NMDA receptor during aging when both GHRKO and control mice were considered together. The only regions that overlapped with the NMDA receptor binding changes across genotype were all layers of the primary motor cortex. Antagonism of NMDA receptors with MK801 can produce motor ataxias (105), so it is possible that receptor declines in motor cortex may contribute to movement problems in aged individuals (106). Age-related declines in GluN1 mRNA have also been seen within frontal cortex of C57BL/6 mice (55,65), but other studies with this same strain showed no aging changes in the mRNA for this subunit (73–75). Fischer 344 and Fischer 344 × Brown Norway F1 rats show declines in protein expression of the GluN1 subunit in the hippocampal formation during aging (67–71). Wistar and Long–Evans rats, however, show no age-related change in the expression of the GluN1 subunit in hippocampal formation (76,77). A decrease in the protein expression of the GluN1 subunit protein is also observed in the distal dendrites of the dentate granule cells in aged macaque monkeys, as compared with young adults (72). The mice in the current study, with a heterogeneous strain background, showed a strong effect of the aging process on mRNA expression of the GluN1 subunit of the NMDA receptor.
More significant effects of aging were seen in the intermediate dentate gyrus, as compared to the dorsal, regardless of knockout genotype. This pattern is also seen in the hippocampal formation of C57BL/6 mice with NMDA receptor binding and GluN2B subunit mRNA (74). Rat studies also show no significant effect of aging on NMDA receptor binding densities in the dorsal hippocampal formation (53,107). It is not known why NMDA receptors in the intermediate hippocampal formation should be more susceptible to aging than the dorsal. It is also not clear exactly what the functional consequences of this change are. Declines in NMDA receptor binding in the intermediate hippocampus of C57BL/6 mice show a significant correlation to spatial memory deficits during aging (58). NMDA receptors are important for the acquisition of spatial memories (8,108,109), but the dorsal hippocampal formation is sufficient for acquiring spatial learning (110). The intermediate hippocampal formation, along with the dorsal, is involved in retrieval of spatial memories (111), but NMDA receptors are not involved in retrieval, at least not in young animals (8,108,109). The role of NMDA receptors within the aged intermediate hippocampus in memory remains to be elucidated.
The deep layers of medial prefrontal, secondary motor, and somatosensory cortices and the whole thickness of the primary motor cortex exhibited a significant decrease in GluN1 mRNA in the GHRKO mice between young and old, but the changes were not significant in the control mice. Four of these five regions are the same regions that showed more significant age-related changes in NMDA receptor binding density in control mice than the GHRKO mice; the opposite pattern. The strain of Wistar-Kyoto rats show markedly higher binding of [3H]MK801 to NMDA receptors than the Wistar strain of rats (112) and C57BL/6 mice show higher surface expression of another NMDA receptor subunit, GluN2B, than the 129S6/SvEv strain (113). Thus, it is possible that the heterogeneous strain background, combined with a small N, may have produced variability between individuals with respect to the NMDA receptor that interfered with detecting significant differences in both binding and mRNA within the same regions within the present study. The fact that there appear to be differences in the magnitude of the change between binding and mRNA densities within the knockout and control groups, however, suggests that variability was not the major issue. Although C57Bl/6 and FVB/N mice do not show differences in basal expression of glutamate receptor subunits, they do show strain differences in the change in receptor expression following manipulation with kainate injections (114). It is possible that the manipulation of the GH receptor on a heterogeneous strain background may have differentially affected some of the knockouts with respect to the NMDA receptor.
There still remains the fact that within each genotype group, the age-related changes in NMDA receptor binding differed in magnitude from the changes in GluN1 mRNA (i.e., in the knockouts, when GluN1 mRNA decreased significantly with increased age, the NMDA-displaceable glutamate binding was not significantly different across ages and vice versa in the controls). Several possibilities might explain this opposing relationship: (i) the decline in GluN1 subunit expression could preferentially affect certain splice variants that could lead to an increase in agonist affinity for the receptor (115) in GHRKO mice, (ii) mRNA expression could be inversely related to the protein expression of the GluN1 subunit, or (iii) there may be other associated changes in the GluN2 family of NMDA receptor subunits that could have influenced either the number or the affinity of glutamate binding sites.
Growth hormone treatments in normal Sprague–Dawley rats lead to declines in GluN1 subunit mRNA in young rats and increases in middle-aged rats (116). In Brown Norway rats, growth hormone treatment enhances microvascular density and spatial working and reference memory (117). Aged Fischer 344 × Brown Norway F1 rats, however, show no influence of insulin-like growth factor-1 on GluN1 subunit expression, despite improvements in GluN2A and GluN2B subunits (78). These results suggest that growth hormone and its downstream factors can be beneficial during aging in rodents, but do not consistently act on the GluN1 subunit. The use of GH in aged humans is not currently recommended by the Growth Hormone Research Society, but they do encourage further study in model systems, such as the GHRKO mouse (118).
The enhanced decline in the GluN1 subunit mRNA within some regions of the rostral cerebral cortex in GHRKO mice compared with normal-sized control mice is the opposite of the effect of caloric restriction on this subunit mRNA. Within both medial and lateral rostral cortex, caloric restriction in middle-aged and old C57BL/6 mice resulted in an increase in GluN1 subunit mRNA as compared with middle-aged, old, and/or young ad libitum-fed mice (81). However, it is possible that behavioral testing experience is necessary to induce the increase (81). The mRNA for the GluN1-a splice variants was upregulated following a behavioral testing experience in the water maze (73).
Rodents undergoing caloric restriction resemble the GHRKO mice in many different ways. Both have increased life spans, decreased insulin-like growth factor-1 plasma expression, and show similar pathologies (119). In addition, caloric restriction was unable to further extend life span in the GHRKO mice, as it did in the Ames dwarf mice (120–123), suggesting that there is a shared mechanism. Although caloric restriction and GHRKO mice show many phenotypic similarities, there are many differences being discovered in the gene expression profiles that are altered by caloric restriction in control mice versus GHRKO mice (120,124,125). There are also differences in stress tolerance in fibroblasts and Akt phosphorylation in cardiac muscle cells between the diet intervention and the knockout mice (126,127). One major shared cellular mechanism between GHRKO and normal calorie-restricted mice appears to be increased insulin sensitivity (120,121,123).
Insulin can enhance NMDA receptor activity (128–130), at least in part via phosphorylation of the GluN2A and GluN2B subunits (128). The similar degree of sparing of NMDA receptor binding density in GHRKO mice, as compared with calorie-restricted aged mice, suggest that altered insulin sensitivity may have had some effect on NMDA receptor expression. The opposite effects of the two interventions on the mRNA expression of the GluN1 subunit suggest that it is unlikely that the effects of increased insulin sensitivity alone account for the GluN1 subunit expression enhancements of caloric restriction. As mentioned above, it may require an interaction between increased insulin sensitivity and behavioral experience to enhance GluN1 subunit expression.
The life span of C57BL/6 mice is extended by caloric restriction to a greater extent than DBA/2 mice (131,132). These two strains also show differences in metabolic and oxidative stress markers following caloric restriction (133,134). Although the DBA/2 strain was not part of the background of the heterogeneous strain mice, the differences seen in GluN1 subunit expression between caloric restriction in C57BL/6 and ad libitum-fed GHRKO mice could also be due to strain differences.
Aged GHRKO mice show enhanced spatial and inhibitory avoidance memory and better locomotor function, as compared with normal-sized mice (24,91). The small effect of the GHRKO genotype on NMDA-displaceable glutamate binding might suggest that the memory and motor improvements are unlikely to be related to NMDA receptors. However, the higher NMDA agonist binding that was seen in old, calorically restricted C57BL/6 mice showed a significant association with better reference memory with a similar degree of sparing of receptor binding density (58,81). It remains to be seen if the changes seen here in the NMDA receptor are associated with the memory and/or locomotor improvements in aged GHRKO mice. It is also possible that the reduction in GluN1 mRNA in the GHRKO mice might enhance memory performance in old mice. There is increasing evidence of a negative relationship between this subunit and memory within groups of old mice (59,79). Additional examples, involving antioxidant therapy, high blood pressure, and obesity, also suggest that not all changes that occur during aging are detrimental to health and longevity (135). The best learners among old Fischer 344 rats appear to be able to upregulate some transmitter or signaling molecules in the face of age-related declines in others (136). Therefore, it is possible that the improved function in the aged GHRKO mice might be due to an enhanced ability to compensate for other declines.
In conclusion, there was a slight sparing of NMDA receptor density in GHRKO mice, similar to that seen with caloric restriction in C57BL/6 mice. This suggests that the enhanced memory and locomotion of the old GHRKO mice may be due, at least in part, to an enhancement of NMDA receptors. However, in contrast to caloric restriction, there was a significant decline in mRNA expression of the major subunit, GluN1. These results suggest that alterations in growth hormone signaling pathways may enhance the expression of NMDA receptors in old age, but not through the upregulation of mRNA expression for the GluN1 subunit.
National Institute on Aging at the National Institutes of Health (AG016332 to K.R.M., AG019899 to A.B., and AG021215 to P.R.P.).
We would like to acknowledge the technical expertise of Xue Zhao in helping to conduct these experiments.