Postadolescent Changes in Regional Cerebral Protein Synthesis: An In Vivo Study in the Fmr1 Null Mouse

Introduction

The requirement for protein synthesis in long-term adaptive change in the nervous system is well documented (West et al., 2001). Inhibition of protein synthesis blocks long-term memory storage in both vertebrates and invertebrates (Agranoff, 1981). Activity-driven changes in gene expression are essential for maintenance of the late phase of long-term potentiation (Nguyen et al., 1994) and long-term depression (Linden, 1996) and promote neuronal survival during development (Mao et al., 1999). Molecular mechanisms underlying these processes may involve translation of newly transcribed targeted genes in neuronal soma and local translation of selected mRNAs in dendrites (Steward and Schuman, 2001).

One of the mRNAs translated in dendrites in response to activation of metabotropic glutamate receptors (mGluRs) is fragile X mental retardation-1 (Fmr1) (Weiler et al., 1997). Silencing of the gene for Fmr1 by a repeat expansion of (CGG)n trinucleotide in the 5′-untranslated region results in fragile X syndrome (FrX), an X-linked disorder characterized by mental retardation. In patients with FrX, fragile X mental retardation protein (FMRP), the gene product of Fmr1, is absent. FrX phenotype includes cognitive impairments ranging from learning disabilities to severe mental retardation (Rousseau et al., 1994), behavioral dysfunction such as hyperactivity, social anxiety, attention problems, and autistic-like behavior (Miller et al., 1999), and subtle physical abnormalities, including macroorchidism in males (Hagerman, 2002). A striking neuropathological feature of FrX syndrome is the long, thin, and tortuous appearance of dendritic spines (Rudelli et al., 1985; Hinton et al., 1991), a morphology similar to that seen early in development (Marin-Padilla, 1967).

Normally, FMRP is widely expressed in fetal and adult tissues and is most abundant in brain and testes (Devys et al., 1993). It is highly expressed in neurons but not in glia (Feng et al., 1997) and is found in cytoplasm, dendrites, and dendritic spines. FMRP is an RNA-binding protein that associates with translating polyribosomes (Feng et al., 1997), and accumulating evidence suggests that FMRP may act as a negative regulator of translation. Mammalian FMRP inhibits mRNA translation in rabbit reticulocyte lysates (Li et al., 2001) and in microinjected Xenopus oocytes (Laggerbauer et al., 2001). In Drosophila, the homolog of FMRP (dFMRP) binds futsch [the Drosophila homolog of MAP-1B (microtubule-associated protein 1B)] mRNA and negatively regulates futsch expression (Zhang et al., 2001). Recent evidence suggests that dFMRP is involved in the micro RNA (miRNA) and RNA interference (RNAi) pathways (Caudy et al., 2002; Ishizuka et al., 2002), implicating dFMRP in the RNAi/miRNA-directed processes of mRNA degradation, translation inhibition, and/or chromatin silencing. To what extent these effects of FMRP on gene expression are operating in brain in vivo has not been shown.

We used the Fmr1 null mouse (Dutch-Belgian Fragile X Consortium, 1994) to study cerebral protein synthesis in an animal model of FrX. In this study, we report that in adult Fmr1 null mice regional rates of cerebral protein synthesis (rCPS) measured in vivo are increased in selective brain regions compared with wildtype (WT) littermates, supporting the idea that FMRP acts as a suppressor of translation in brain in vivo.

Parts of this work have been published previously in abstract form (Qin et al., 2003).

Materials and Methods

Chemicals and materials were obtained from the following sources: l-[1-¹⁴C]leucine [specific activity (SA), 60 mCi/mmol] and l-[3,4,5-³H]leucine (SA, 50 Ci/mmol) from PerkinElmer Life Sciences (Boston, MA); Escherichia coli tRNA from Sigma (St. Louis, MO); vanadyl ribonucleoside complex and redistilled nucleic acid-grade phenol from Bethesda Research Laboratories (Gaithersburg, MD); and l-norleucine and 5-sulfosalicylic acid from Fluka (Buchs, Switzerland).

Animals. All procedures were performed in accordance with the National Institutes of Health Guidelines for the Care and Use of Animals and an animal study protocol approved by the National Institute of Mental Health Animal Care and Use Committee. FVB/NJ-Fmr1^tm1Cgr breeding pairs (heterozygous females and hemizygous males) were obtained from The Jackson Laboratory (Bar Harbor, ME). Heterozygous female and wild-type male offspring were mated to provide offspring in two experimental groups: hemizygous males and wild-type males. All mice were housed in a central facility and maintained under controlled conditions of normal humidity and temperature with standard alternating 12 h periods of light and darkness. Food (NIH-31 rodent chow) and water were provided ad libitum. Mature mice between 3 and 6 months of age were studied.

Genotyping. Genomic DNA was extracted from a section of tail taken from each animal (Puregene; Gentra System, Minneapolis, MN). The M2 (5′-ATCTAGTCATGCTATGGATATCAGC-3′) and N2 (5′-GTGGG CTCTATGGCTTCTGAGG-3′) primers (12 pmol) were used to screen for the presence or absence of the mutant allele in a PCR buffer containing 10 mm Tris-HCl, pH 8.3, 50 mm KCl, 2.5 mm MgCl₂, 0.2 mm dNTPs, and 2 U TaqDNA polymerase (AmpliTaq Gold; Applied Biosystems, Branchburg, NJ) as described previously (Qin et al., 2002). The PCR product at ∼800 bp indicated the presence of the knock-out allele.

Surgical preparation of animals. Mice were prepared for studies by insertion under light halothane anesthesia of polyethylene catheters (PE-10) into one femoral artery and one or two femoral veins as described previously (Smith and Kang, 2000). Mice were allowed to recover from the surgery overnight in clear plastic cylinders 13 cm in diameter. This apparatus permitted mice to move freely throughout the recovery and experimental periods but not gain access to the tubing. Food and water were available ad libitum.

Measurement of physiological variables. Mean arterial blood pressure, hematocrit, and arterial plasma glucose concentrations were measured to evaluate each animal's physiological state (Qin et al., 2002). Arterial plasma corticosterone concentration was determined by radioimmunoassay (MP Biomedicals, Costamesa, CA).

Determination of rCPS. rCPS was determined with l-[1-¹⁴C]leucine as the radiolabeled tracer used in conjunction with quantitative autoradiography to achieve spatial localization of the labeled protein in the brain (Smith et al., 1988). The operational equation of the method is as follows:

in which R_i is the rate of leucine incorporation into protein in tissue i; P_i^* (T) is the concentration of ¹⁴C fixed in tissue i at any time, T, after introduction of the tracer into the circulation; λ_i is equal to the fraction of leucine in the precursor pool for protein synthesis in tissue i that is derived from plasma; C_p^*(t) and C_p are the concentrations of labeled and unlabeled leucine in the arterial plasma, respectively; and t is the variable time. The evaluation of λ_i is performed in a separate series of experiments by measuring the steady-state ratio of leucine SA in the tissue tRNA-bound pool to that of the arterial plasma.

Mice were surgically prepared and catheterized, and their physiological states were monitored as described above. The experimental period was initiated by an intravenous pulse of 100 μCi/kg l-[1-¹⁴C]leucine contained in ∼40 μl of physiological saline. Timed arterial samples were collected during the following 60 min for determination of the time courses of plasma concentrations of leucine and [¹⁴C]leucine as described previously (Smith and Kang, 2000). At the end of the experimental interval, mice were killed by an intravenous injection of sodium pentobarbital, and brains were rapidly removed and frozen in isopentane cooled to -40°C with dry ice. Serial sections, 20 μm thick, were cut in a Leica (Deerfield, IL) 1850 cryostat at -18°C, thaw mounted on gelatin-coated slides, dried in a stream of air, fixed and washed in 10% formalin, and exposed to EMC-1 film (Eastman Kodak, Rochester, NY) along with calibrated [¹⁴C]methylmethacrylate standards as described previously (Smith and Kang, 2000). Autoradiograms were digitized with a pixel size of 11 μm by means of a 10-bit DVC-1310 digital camera (DVC, Austin, TX). Images were analyzed with an MCID Elite image processing system (Imaging Research, St. Catharine's, Ontario, Canada). The concentration of ¹⁴C in each region of interest was determined from the optical density versus ¹⁴C concentration curve determined from the calibrated plastic standards, and rCPS was calculated by means of the above equation. Brain regions were identified by reference to a mouse brain atlas (Paxinos and Franklin, 2001). Global CPS and brain volume were determined by analysis of autoradiograms of all sections of the entire brain digitized (42 μm pixel size) by means of a Multirad 850 Howtek Film Digitzer (Howtek, Hudson, NH). Weighted average CPS was determined as described above, and brain volume was determined from the total number of pixels and the calibrated pixel size.

Evaluation of λ_i. λ_i for whole brain (λ_WB) was evaluated in a series of steady-state experiments as described previously (Smith et al., 1988). A constant arterial plasma SA for [³H]leucine was maintained for 120 min by means of a programmed intravenous infusion of [³H]leucine designed to achieve and maintain a constant arterial plasma [³H]leucine concentration. The total infusion contained 3-8 mCi of [³H]leucine in a volume of ∼0.4 ml. The SAs of [³H]leucine in the arterial plasma and tRNA-bound pools in brain were determined as described below. Timed arterial blood samples (∼20 μl) were collected every 15 or 30 min during the infusion and centrifuged immediately to separate plasma, which was then deproteinized by addition of one-third of a volume of 16% (W/V) sulfosalicylic acid containing l-norleucine (0.04 mm) as an internal standard for amino acid analyses. The deproteinized samples were stored at -70°C until assayed for leucine and [³H]leucine concentrations. At the end of the infusions, the mice were decapitated, and their brains were quickly removed and chilled to 0°C in ice-cold 0.25 m sucrose.

Each brain was weighed and homogenized in 5 ml of 0.25 m sucrose (0°C) containing 10 mm vanadyl ribonucleoside complex to inhibit ribonuclease, 6 mg of uncharged tRNA as carrier, and l-norleucine (0.02 mm) added as an internal standard and centrifuged at 100,000 × g for 1 h. The tRNA-bound amino acids were purified as described previously (Smith et al., 1988). The pure tRNA fraction was dissolved in 50 mm sodium carbonate, pH 10, and incubated at 37°C for 90 min to deacylate the aminoacyl-tRNA. Deacylated tRNA was precipitated overnight at -20°C in ethanol and removed by centrifugation (12,000 × g, 20 min) (Smith et al., 1988). The supernatant solution, which contained the previously tRNA-bound but now free amino acids, was dried in a stream of N₂ and redissolved in 40 μl of 0.2 m sodium citrate, pH 2.2.

Deproteinized plasma samples and tissue fractions of deacylated aminoacyl-tRNA pools were assayed for leucine concentration by postcolumn derivatization with o-phthaldehyde and fluorometric assay with a Beckman model 7300 Amino Acid Analyzer (Beckman Instruments, Fullerton, CA), and column eluates were collected every 0.5 min and assayed for ³H with a model 2250CA Packard Liquid Scintillation Counter (PerkinElmer Life Sciences). SA of each sample was calculated from total ³H in all fractions in the leucine peak and the total measured leucine content in the peak.

Values of λ_WB were calculated as the ratios of the measured steady-state SA of leucine in the tissue tRNA-bound pool to that of the free leucine in arterial plasma. The time course of the SA in arterial plasma and the SA of tRNA-bound leucine in whole brain at the end of the experimental interval were determined as described above. Only animals with relatively constant plasma leucine SAs were included in the series.

Immunohistochemistry. For FMRP-immunohistochemical staining, mice were deeply anesthetized with halothane and perfused via the ascending aorta with 0.9% normal saline, followed by 4% paraformaldehyde. Brains were postfixed in 4% paraformaldehyde for 6 h at 4°C, immersed in 0.1 m PBS, pH 7.4, containing 20% sucrose for 48 h and frozen and stored at -75°C. Coronal sections 20 μm in thickness were cut with a cryostat and thaw mounted on gelatin-coated glass slides. Alternate sections were taken for thionin staining and immunohistochemistry. We used the immunoperoxidase procedure (M.O.M. kit; Vector Laboratories, Burlingame, CA) (Hsu et al., 1981) for FMRP immunohistochemistry with a monoclonal mouse anti-FMRP IgG (Chemicon, Temecula, CA) at 1:50.

The specificity of the antibody (monoclonal antibody 1a) was originally demonstrated by Devys et al. (1993). In the present study, the specificity of the primary antibody was confirmed with the following controls: (1) omission of the specific antibody and (2) use of normal mouse IgG instead of the specific antibody. We also used Western blot analysis of tissue extracts prepared from WT and Fmr1 null mouse brains to confirm the specificity of the primary antibody (Fig. 1). Equal amounts of protein (50 μg) were separated by PAGE and blotted onto polyvinylidene difluoride membranes (Invitrogen, Carlsbad, CA). Membranes were incubated overnight at 4°C with FMRP antibody (1:1000), followed by biotinylated secondary antibody (1:500) (Amersham Biosciences, Piscataway, NJ). The blot was then reacted with streptavidin-HRP-conjugated antibody (1: 5000), and immunoreactive bands were visualized with ECL Western blotting detection reagents (Amersham Biosciences).

Statistical analyses. Measurements were analyzed by means of two-way ANOVA with post hoc t tests.

Results

Discussion

FMRP is thought to play a role in regulating experience-driven protein synthesis. Particularly provocative is the idea that it may function to repress translation primarily in postsynaptic terminals undergoing plasticity. Before the present study, the influence of FMRP on translation in vivo was not known. We addressed this question by measuring rCPS in conscious WT and Fmr1 null mice with the quantitative autoradiographic l-[1-¹⁴C]leucine method. Our results show that, in the absence of FMRP, rCPS is increased in selective brain regions. Effects of the mutation are noteworthy in hippocampus, thalamus, and hypothalamus.

We studied WT and Fmr1 null mice at 4 and 6 months of age and found large and widespread effects of age on rCPS. We chose this age range because it was well passed the developmental period and considerably before effects of aging are observed. Our results in the mouse are in keeping with our finding that mean CPS decreased by 12%, albeit not statistically significantly, between 35 and 60 d of age in rat (Sun et al., 1995). That rCPS continues to decrease after maturation may reflect the process of synaptic pruning that is a hallmark of the postadolescent period. In humans, decreases in gray matter volume between adolescence and young adulthood (Giedd et al., 1999) may also reflect synaptic pruning.

In Fmr1 null mice, effects of age on rCPS parallel those in WT but at a higher level, suggesting that pruning takes place even in the absence of FMRP. The effects of genotype show regional selectivity that may be a function of the regional distribution of FMRP. Immunohistochemistry for FMRP in WT showed note-worthy labeling in supraoptic nucleus and CA1 hippocampal pyramidal cells, areas in which the effect of genotype on rCPS at 6 months of age was >23%. There is also some correspondence between the regions most affected and the distribution of group 1 mGluRs (Shigemoto and Mizuno, 2000). This is of interest in light of the proposal that many of the symptoms of fragile X syndrome might be attributable to overactive signaling by group 1 mGluRs (Bear et al., 2004).

We measured rCPS with the l-[1-¹⁴C]leucine method, which is based on a compartmental model for the behavior of leucine in brain that includes recycling of unlabeled leucine from the break-down of tissue protein into the precursor pool (Smith et al., 1988). The constant λ_WB is a measure of recycling. Defined as the fraction of leucine in the precursor pool for protein synthesis derived from arterial plasma, λ_WB was determined in a separate series of experiments in both WT and Fmr1 null mice. Conditions were the same as those in which we measured rCPS, but the animals were somewhat younger. We have shown in rats that λ_WB is constant between 1 and 24 months (Smith et al., 1995; Sun et al., 1995), so this age difference should not confound our results. Because values of λ_WB measured in both genotypes were similar, we used the average value. We assumed that the fraction of recycled leucine in the precursor pool is constant from region to region and used λ_WB to calculate rCPS in all regions (Sun et al., 1992).

We surveyed 75 brain regions to determine the extent of effects on rCPS, and we considered the possibility that analysis of so many regions in the same series of animals may generate some false positives. If we apply the very conservative Bonferroni correction for multiple comparisons, a p ≤ 0.0007 would be required for statistical significance, and 35 regions would show a statistically significant effect of age and no region would show a statistically significant effect of genotype. Because this may result in false negatives, we reported the actual F and p values computed by the two-way ANOVA.

We considered a differential response of the two genotypes to stress as a possible explanation for our results. Effects of stress on rCPS in rats have been reported (Giaume et al., 1995; Baubet et al., 1996). Children with FrX may be more sensitive to stress than their siblings (Wisbeck et al., 2000; Hessl et al., 2002), and Fmr1 null mice may be more sensitive to stress than WT (Beckel-Mitchener et al., 2003). Our animals recovered from the surgical implantation of catheters and acclimated to the experiment environment for 24 h before rCPS measurement. They were freely moving, and, apart from the sampling of arterial blood, mice were not subjected to stressors during the study. Whereas measured plasma corticosterone concentrations were higher than basal levels (Muller et al., 2003), indicating that our mice were responding to stress during the study, both genotypes were similarly affected. It is, therefore, unlikely that a differential response to stress could account for differences in rCPS between the genotypes.

Some of the regions most affected in the Fmr1 null mice are consistent with the behavioral manifestations in the mouse and, to some extent, the clinical symptoms associated with FrX. In particular, changes in the hippocampus are consistent with learning and memory deficits in FrX and with deficits in performance on passive avoidance tests (Qin et al., 2002) and spatial learning tests (Kooy, 2003) in Fmr1 null mice. Effects in hypothalamus and brainstem may be related to increased sensory sensitivity and autistic-like behavior seen in FrX (Hagerman, 2002). The hypothalamus supplies a major input to the dorsal motor nucleus of the vagus, and, via the parasympathetic system, the dorsal motor nucleus of the vagus regulates visceral functions. Parasympathetic dysregulation, hyperarousal, and hyperactivity are all characteristic of FrX patients (Boccia and Roberts, 2000; Hagerman, 2002). Fmr1 null mice are hyperactive (Qin et al., 2002), are susceptible to audiogenic seizures, and have an increase in prepulse inhibition and auditory startle responses (Chen and Toth, 2001).

One interpretation of our results is that increased rCPS in Fmr1 null mice reflects an increase in the number of neurons and/or in the size of the dendritic arbor and number of spines in the mutant animals. Neuron density in neocortex in patients with FrX, however, does not differ from controls (Hinton et al., 1991). Furthermore, in sensory cortex of both FrX patients and adult Fmr1 null mice, the number of elongated immature spines is increased and the number of mature shaped spines is decreased (Hinton et al., 1991; Irwin et al., 2002), but changes in density of dendrites or spines are not evident. However, in our studies, rCPS was not significantly affected in areas of sensory cortex. It remains a possibility that changes do occur in spine and dendrite densities in brain regions particularly affected with respect to rCPS. A study of Timm staining in dentate gyrus mossy fiber terminals in Fmr1 null mice showed an increased staining in stratum oriens and the molecular layer (Ivanco and Greenough, 2002), both regions with statistically significant effects on rCPS.

The results of our study considered in the light of known RNA-binding properties of FMRP (Feng et al., 1997), the effects of FMRP on mRNA translation in vitro (Laggerbauer et al., 2001; Li et al., 2001), and the regulation by FMRP of translation of MAP-1B during synaptogenesis in cultured neurons (Lu et al., 2004) suggest that FMRP does indeed act to suppress protein synthesis in brain in vivo. Whereas certain messages may be preferentially affected by FMRP (Brown et al., 2001), our study points to an effect on protein synthesis that may be more general. It is highly unlikely that we could detect an effect on a small number of mRNAs with the autoradiographic method unless the small number of mRNAs could account for a major fraction of overall protein synthesis at the time of the study.

Whether the effects that we report are in soma or dendrites is difficult to discern at the level of resolution (∼50 μm) of the [¹⁴C]leucine autoradiographic method (Smith, 1983). There is increasing evidence of effects of FMRP in dendrites and synapses. Synaptic activation via mGluRs increases the synthesis of FMRP (Weiler et al., 1997) and affects the trafficking of FMRP and Fmr1 mRNA in dendrites (Antar et al., 2004). Studies of mRNA granules suggest that in vivo mGluR-activated translation from granules is ongoing and occurs at a higher rate in the absence of FMRP (Aschrafi et al., 2005). Enhanced mGluR-dependent long-term depression (Huber et al., 2002) and increased microtubule stability (Lu et al., 2004) in preparations from Fmr1 null mice suggest a role for FMRP in stabilization of synapses. Studies in synaptoneurosomes show increased assembly of polyribosomes in WT, but not Fmr1 null, mice in response to mGluR activation (Weiler et al., 2004), indicating that the regulatory effects of FMRP on translation in dendrites may be complex. Although we cannot draw any conclusions about FMRP and protein synthesis in dendrites from our studies, our results provide the first in vivo evidence of a role for FMRP in regulating translation and a map of where its effects are greatest.

Dysregulation of protein synthesis is very likely close to the underlying etiology of symptoms of FrX. Measurement of rCPS in patients with FrX syndrome may provide a means of measuring the progress of the disease and efficacy of treatments. With development of the l-[1-¹¹C]leucine positron emission tomography method for measurement of rCPS in human subjects (Schmidt et al., 2005; Smith et al., 2005), we can now test this possibility in FrX patients.