Age-Related Declines in Prefrontal Cortical Expression of Metabotropic Glutamate Receptors that Support Working Memory

Abstract

Glutamate signaling is essential for the persistent neural activity in prefrontal cortex (PFC) that enables working memory. Metabotropic glutamate receptors (mGluRs) are a diverse class of proteins that modulate excitatory neurotransmission via both presynaptic regulation of extracellular glutamate levels and postsynaptic modulation of ion channels on dendritic spines. This receptor class is of significant therapeutic interest for treatment of cognitive disorders associated with glutamate dysregulation. Working memory impairment and cortical hypoexcitability are both associated with advanced aging. Whether aging modifies PFC mGluR expression, and the extent to which any such alterations are regionally or subtype specific, however, is unknown. Moreover, it is unclear whether specific mGluRs in PFC are critical for working memory, and thus, whether altered mGluR expression in aging or disease is sufficient to play a causative role in working memory decline. Experiments in the current study first evaluated the effects of age on medial PFC (mPFC) mGluR expression using biochemical and molecular approaches in rats. Of the eight mGluRs examined, only mGluR5, mGluR3, and mGluR4 were significantly reduced in the aged PFC. The reductions in mGluR3 and mGluR5 (but not mGluR4) were observed in both mRNA and protein and were selectively localized to the prelimbic (PrL), but not infralimbic (IL), subregion of mPFC. Finally, pharmacological blockade of mGluR5 or mGluR2/3 using selective antagonists directed to PrL significantly impaired working memory without influencing non-mnemonic aspects of task performance. Together, these data implicate attenuated expression of PFC mGluR5 and mGluR3 in the impaired working memory associated with advanced ages.

Keywords: aging; metabotropic glutamate receptor; prefrontal cortex; prelimbic cortex; rat; working memory.

Figure 1.

Schematic of delayed response working memory task. Each trial of the delayed response task includes three phases. During the sample phase, one lever (left or right, pseudo-randomly varied between pairs of trials) is extended into the chamber. The rat must press the extended lever to enter the variable duration “delay phase” (delays are pseudo-randomly varied from 0 to 24 s within each block of seven trials). During the delay, the rat must nose poke continuously into the centrally located food trough. The first nose poke emitted after the expiration of the predetermined delay timer initiates the choice phase wherein both levers (left and right) are extended into the chamber. The rat must remember and press the same lever that was extended during the sample phase to receive a food reward (a 45-mg food pellet), and this is scored as a correct choice. Pressing the other lever is scored as an incorrect choice and no food reward is delivered.

Figure 2.

mGluR protein levels in mPFC of young and aged rats. A, Representative images of immuno-reactive bands detected using mGluR subtype-selective antibodies in whole mPFC membrane homogenates prepared from young and aged rats. B, Group I mGluRs (in blue). The protein level of mGluR5, but not mGluR1, was significantly lower in the mPFC of aged rats compared to young adults (*p < 0.05 vs young). C, Group II mGluRs (in red). The protein level of mGluR2/3 was significantly lower in the mPFC of aged rats compared to young adults (*p < 0.05 vs young). D, Group III mGluRs (in green). There were no significant changes to the protein levels of Group III mGluRs of aged rats compared to young adults (p > 0.05 vs young). B–D, Mean protein level (transformed to “% of young” after normalizing integrated intensity to α-tubulin, y-axis; see Table 2 for normalized, untransformed data) is plotted as a function of mGluR subtype (x-axis) and age group (separate bars; n = 7-8 young and n = 15 aged). Open circles represent values for individual rats and bars represent group means.

Figure 3.

mGluR gene transcript expression in the PrL and IL of young and aged rats. A, Group I mGluRs (in blue). Expression of GRM5, but not GRM1, was significantly lower in the PrL subregion of aged rats compared to young adults (**p < 0.01 vs young). B, Group II mGluRs (in red). Expression of GRM3, but not GRM2, was significantly lower in the PrL of aged rats compared to young adults (**p < 0.01 vs young). C, Group III mGluRs (in green). Expression of GRM4, but not GRM7 or GRM8, was significantly lower in the PrL of aged rats compared to young adults (**p < 0.01 vs young). D–F, Gene expression was not significantly different between young adult and aged rats in the IL subregion. In all panels, mean gene expression (transformed to “% of young” after normalizing raw C_t values to RPLP1; y-axis) is plotted as a function of gene (x-axis) and age group (separate bars; n = 6 young and n = 12 aged). Open circles represent values for individual rats and bars represent group means.

Figure 4.

Effect of micro-infusing MTEP (mGluR5 antagonist) into PrL cortex on performance in the delayed response working memory task. A, Histologically verified placements of injector tips used to micro-infuse the mGluR5 antagonist MTEP into the PrL cortex of young adult rats before testing in the delayed response task (n = 7 young rats). B, Micro-infusion of 0.3-µg MTEP significantly reduced choice accuracy relative to vehicle (n = 7; ***p < 0.05 vs vehicle, main effect of dose). C, Post hoc analysis comparing 0.3-µg dose of MTEP to vehicle. The 0.3-µg dose of MTEP impaired performance across all delays in all rats compared to vehicle performance, p < 0.001. D, The 0.3-µg dose of MTEP impaired performance at long delays (12–24 s) in all rats compared to vehicle performance. E, The number of trials completed did not change as a function of MTEP dose. In A, placements are mapped to standardized coronal sections corresponding to +2.70 and +3.20 mm from bregma according to the atlas of Paxinos and Watson (2005). In B, mean choice accuracy (y-axis) is plotted as a function of delay (x-axis) and dose (symbols/lines; refer to legend for specific dose). In C, D, mean choice accuracy (collapsed across all delays in C and long delays (12–24 s) in D; y-axis) is plotted as a function of the 0.3-µg dose of MTEP (x-axis; symbols/lines). Error bars represent the SEM.

Figure 5.

Effect of micro-infusing LY341495 (mGluR2/3 antagonist) into PrL cortex on performance in the delayed response working memory task. A, Histologically verified placements of injector tips used to micro-infuse the mGluR2/3 antagonist LY341495 into the PrL cortex of young adult rats before testing in the delayed response task (n = 10 young rats). B, Microinfusion of LY341495 significantly reduced choice accuracy relative to vehicle at all doses tested (n = 10; **p < 0.05 vs vehicle, main effect of dose; #p < 0.05 vs vehicle, dose × delay interaction). C, Post hoc analysis comparing 500-ng dose of LY341495 to vehicle. The 500-ng dose impaired performance across all delays in all rats compared to vehicle performance, p < 0.01. D, Micro-infusion of 500-ng LY341495 impaired performance at long delays (12–24 s) compared to vehicle performance. E, The number of trials completed did not change as a function of dose. In A, placements are mapped to standardized coronal sections corresponding to +2.70 and +3.20 mm from bregma according to the atlas of Paxinos and Watson (2005). In B, mean choice accuracy (y-axis) is plotted as a function of delay (x-axis) and dose (symbols/lines; refer to legend for specific dose). In C, D, mean choice accuracy (collapsed across all delays in C and long delays (12–24 s) in D; y-axis) is plotted as a function of the 500-ng dose of LY341495 (x-axis; symbols/lines). Error bars represent the SEM.