INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Huntington's disease (HD) is a progressive, autosomal dominant neurodegenerative disorder characterized by motor and cognitive symptomatology (Haddad and Cummings 1997). The disorder is one of a series of neurological diseases caused by a mutation containing expanded polyglutamine (CAG) repeats (Gusella et al. 1997; Paulson and Fishbeck 1996; Reddy and Housman 1997). Little is known, however, about how the expanded polyglutamine repeat region in the encoded protein huntingtin results in disturbances of neuronal function. The discovery and cloning of the HD gene (The Huntington's Disease Collaborative Research Group 1993) has permitted the development of several different transgenic and knock in mouse models of the disease (Hodgson et al. 1999; Levine et al. 1999; Mangiarini et al. 1996; Reddy et al. 1998); this allows direct tests of altered neuronal function. Although neuronal death has been suggested as the ultimate pathology (Butterworth et al. 1998; Portera-Cailliau et al. 1995; Vonsattel et al. 1985), earlier pathological events causing malfunction of and miscommunication between neurons may be responsible for the majority of symptoms (Aronin et al. 1999; Beal 2000; Cha 2000; Chesselet and Levine 2000).

In the present experiments, we used the R6/2 transgenic to examine cellular electrophysiological and morphological alterations in the striatum. This transgenic mouse contains exon 1 and promoter sequences of the human HD gene inserted into the mouse genome, and carries 141-157 CAG repeats (Mangiarini et al. 1996). Affected mice demonstrate a progressive neurological syndrome that includes alterations in transmitter and receptor expression and signaling mechanisms (Bibb et al. 2000; Cha et al. 1998; Luthi-Carter et al. 2000; Menalled et al. 2000), motor deficits (Carter et al. 1999), and learning disabilities (Lione et al. 1999; Murphy et al. 2000). Striatal neurons in these mice develop inclusions, but the striatum does not exhibit marked cell loss (Mangiarini et al. 1996), although late onset neuronal degeneration has been observed in the striatum as well as a number of other areas (Turmaine et al. 2000). Striatal neurons in these mutant animals are hypersensitive to N-methyl-D-aspartate (NMDA) (Cepeda et al. 2001; Levine et al. 1999), and this effect occurs in parallel with development of the overt behavioral phenotype.

As a first step in examining physiological changes, we assessed the electrophysiological properties of medium-sized spiny striatal neurons in the R6/2 transgenic, a neuronal population affected in human HD (Vonsattel et al. 1985). The major purpose of these studies was to determine which physiological properties were altered and if such properties were changed before the appearance of overt behavioral symptoms in transgenic mice. Thus the present experiments examined electrophysiological and concurrent morphological alterations in the R6/2 transgenic at two age points, ~40 days of age and ~90 days of age. Mice at 40 days do not display an overt behavioral phenotype, although subtle behavioral changes have been observed (Lione et al. 1999). Mice at 90 days display a marked behavioral phenotype (Carter et al. 1999). A preliminary description of some of the electrophysiological outcomes has been published (Levine et al. 1999).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals

Initially, breeding triads [2 wild-type (WT) females and 1 transgenic male] of R6/2 transgenic mice were obtained from the Jackson Laboratories (Bar Harbor, ME) to establish a breeding colony. Subsequently, all mice used in these experiments were obtained from the R6/2 breeding colony maintained at UCLA. As indicated in the preceding text, experiments were performed on two age groups of R6/2 transgenics and their age-matched WT controls. We use the terms presymptomatic (prior to development of overt motor signs) and symptomatic (after motor abnormalities were visible) R6/2 mice to differentiate the two groups of transgenics. Although we use the term presymptomatic to define the younger group, studies have shown that R6/2 animals before and at this age display some behavioral alterations (Lione et al. 1999). The younger age group consisted of seven presymptomatic R6/2 transgenic mice (39.7 ± 0.7 days old, range: 37-42 days) and three age-matched WT mice (39.0 ± 0.6 days old, range: 38-40 days). The older group consisted of 15 symptomatic R6/2 transgenic mice (89.4 ± 1.3 days old, range: 81-101 days) and 10 age-matched WT mice (95.0 ± 2.4 days old, range: 82-104 days). All experimental procedures were carried out in accordance with the National Institutes of Health Guide for Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee at UCLA.

Slice preparation

Mice were decapitated under deep halothane anesthesia. Brains were removed into ice-cold low-Ca²⁺ oxygenated artificial cerebrospinal fluid (ACSF; composition in mM: 130 NaCl, 5 MgCl₂, 1 CaCl₂, 3 KCl, 1.25 NaH₂PO₄, 26 NaHCO₃, and 10 glucose), and sliced coronally at 350 µm. Slices containing striatum and overlying cortex were maintained in an incubation chamber filled with ACSF (composition in mM: 124 NaCl, 2 MgSO₄, 2 CaCl₂, 5 KCl, 1.25 NaH₂PO₄, 26 NaHCO₃, and 10 glucose) bubbled continuously with 95% O₂-5% CO₂ at room temperature for 1 h before being transferred to a laminar flow, thin layer submersion recording chamber.

Stimulation and recording

In the recording chamber, slices were perfused constantly with oxygenated ACSF (31-32°C; composition as in the preceding text but with 50 µM picrotoxin to block synaptic responses mediated by activation of GABA_A receptors) in an atmosphere of warm, moist 95% O₂-5% CO₂. Responses of individual cells were recorded using an Axoclamp 2A amplifier (Axon Instruments, Foster City, CA) and sharp microelectrodes (60-110 M) filled with 3 M K⁺-acetate, 5 mM KCl, and 2% wt/vol biocytin.

Resting membrane potential (RMP) was noted only after the cell had recovered from penetration and had stabilized (10 min after impalement). The current-voltage relationship was determined from the responses to increasing intensities of square-wave hyperpolarizing and depolarizing current pulses delivered through the recording electrode (1.0- to +1.0-nA steps in 0.1-nA increments, 300- to 350-ms pulse durations). Current-voltage curves were generated and fit with a third-order polynomial to determine input resistance. The first action potential (AP) produced by the lowest incremental depolarizing current step was analyzed. AP amplitude was measured from the steady-state membrane potential immediately preceding AP onset to the peak of the AP. Afterhyperpolarization (AHP) amplitudes were measured from the same steady-state membrane potential to the peak of the negativity immediately following the AP. The maximum rise and decay slopes and the width at half-amplitude were calculated using Clampfit 8.0 software (Axon Instruments). The decay time constants () of the electrode and membrane were measured by a double exponential fit of the decay to steady-state from an average of 20 voltage responses to a 100- or 200-pA hyperpolarizing square wave current step.

To examine synaptic responses, a bipolar stimulating electrode was placed in the corpus callosum to activate excitatory striatal afferents. Pairs of stimuli of increasing intensity (100-µs duration, 0.1- to 5-mA, 50-ms inter-stimulus interval) were delivered every 5 s. For most recordings, five traces were collected and digitally averaged at each stimulus intensity to improve the signal-to-noise ratio. Peak amplitudes of digitally averaged excitatory postsynaptic potentials (EPSPs) were measured, and input-output relationships were plotted and fit by a sigmoidal function. From each cell, the averaged EPSP whose peak amplitude lay between 20 and 50% of maximum on the input-output curve was further analyzed for between-group comparisons.

Morphology

After recording, neurons were filled with biocytin by current injection. Brain slices were fixed in 10% paraformaldehyde, subsectioned at 50 µm, and processed using standard techniques (Kita and Armstrong 1991). For Golgi-Cox, we used five additional symptomatic R6/2 mice and five age-matched WT littermate controls. Animals were anesthetized deeply with Avertin, and brains were dissected from the skull and immediately placed in Golgi-Cox fixative for 6-8 wk in the dark (Caddy and Biscoe 1979). The fixative consisted of 1% potassium dichromate, 1% mercuric chloride, and 0.45% potassium chromate. After dehydration, brains were embedded and sections were cut at 100 µm using a sliding microtome.

Data analysis

Electrophysiological data were digitized at 10 kHz, captured, and analyzed off-line using pClamp software (version 8.0, Axon Instruments). Electrophysiological and morphological data are presented as means ± SE. Curve fitting for electrophysiology was performed using Origin 6.0 software (Microcal Software, Northampton, MA). Sigmastat 2.03 software (SPSS, San Rafael, CA) was used to perform statistical analyses. t-tests were used when means from only two groups were compared. ANOVAs were used for multiple comparisons. Unless otherwise stated in the text, groups were compared using a two-way ANOVA for independent samples with genotype and age as the two main effects. For post hoc evaluations using ANOVAs, the Bonferroni t-test was used because this test is one of the more conservative approaches using multiple comparisons. Differences in distributions were assessed with a ² analysis. Differences for all statistical tests were considered statistically significant when P < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Presymptomatic R6/2 mice did not display overt behavioral abnormalities like limb clasping or gross changes in motor coordination and appeared similar to their age-matched WT controls. All symptomatic R6/2 mice had obvious abnormal motor behaviors including limb clasping, poor motor coordination, tremors, and some displayed seizures.

Electrophysiology

After impalement with the microelectrode, the RMP in all neurons stabilized over a period of a few minutes and remained constant for the remainder of the recording period. Only cells that maintained stable recordings for 15 min after stabilization were subsequently analyzed. Recordings from most cells lasted for 1 h. Data were obtained from 21 neurons from presymptomatic R6/2 transgenics and 10 neurons from age-matched WTs and from 20 neurons from symptomatic R6/2 transgenics and 18 neurons from age-matched WTs.

There were no differences in RMP in neurons obtained from presymptomatic R6/2 and WT mice (Table 1). In contrast, symptomatic R6/2 transgenic mice exhibited a markedly depolarized RMP compared with neurons recorded from the age-matched WT mice (Table 1). There was a statistically significant interaction between age and genotype (F = 12.6, df = 1/64, P < 0.001) that was due to the statistically significant difference in mean RMP between symptomatic R6/2 mice and their age-matched WTs (t = 5.75, P < 0.001; Table 1). Neurons from both pre- and symptomatic R6/2 mice also exhibited significantly higher input resistances than those of their age-matched WT controls (F = 15.6, df = 1/58, P < 0.001 for the main effect of genotype; Table 1; Fig. 1).

View larger version (32K): [in this window] [in a new window]   Fig. 1. Passive membrane properties. Left: negative and positive square-wave current pulses delivered through the recording electrode that produced hyperpolarizing and depolarizing voltage responses in striatal medium spiny neurons. Resting membrane potentials (RMPs) were 83 and 79 mV for age-matched wild-type (WT) and presymptomatic R6/2 neurons, respectively, and 79 and 72 mV for age-matched WT and symptomatic R6/2 neurons, respectively. Current-voltage relationships for these cells were plotted and fit using a third-order polynomial function to obtain the input resistance at rest as determined by the slope at the origin on the graphs (right). The presymptomatic and symptomatic transgenic R6/2 neurons exhibited increased input resistance compared with their age-matched WT controls. Input resistance values were 20.9 and 38.3 M for WT and presymptomatic neurons, respectively, and 22.0 and 40.7 M for WT and symptomatic R6/2 neurons, respectively.

The membrane and electrode time constants () were measured by fitting a double exponential curve to the voltage response to a 100- to 200-nA hyperpolarizing square wave current step. There were no differences in time constants in cells obtained from presymptomatic R6/2 and age-matched WTs. Cells from symptomatic R6/2 transgenic mice had significantly faster time constants compared with WTs, indicating reduced capacitance. There was a statistically significant interaction between age and genotype (F = 5.12, df = 1/65, P < 0.05) that was due to the faster time constant in the symptomatic R6/2 mice (t = 2.22, P < 0.05; Table 1). The portions of the fits attributed to the electrode time constants were not statistically different between the groups (data not shown).

All neurons examined were capable of firing APs. In most respects, APs were similar in R6/2 transgenics and their age-matched WT controls (Fig. 2A). AP threshold voltage, amplitude, half-amplitude duration, and maximum rise slope were not different in pre- or symptomatic R6/2 transgenics compared with their respective age-matched WTs. One parameter that was different in both groups of transgenics compared with their respective age-matched WTs was the intensity of current required to evoke an AP (rheobase). There was a significant main effect of genotype (F = 9.92, df = 1/39, P < 0.003; Table 1), indicating that neurons from R6/2 transgenics were more excitable than those from age-matched WTs. In addition, there was a significant main effect of age indicating that in both older groups the rheobase was lower than in the younger groups (F = 5.43, df = 1/39, P < 0.025; Table 1). The maximum decay slope of the AP also was reduced in symptomatic R6/2 transgenics compared with their age-matched WT controls. The statistically significant interaction between genotype and age (F = 4.23, df = 1/46, P < 0.05) was due to a lower mean decay slope of the AP in neurons from symptomatic R6/2 mice compared with their age-matched controls (t = 2.42, P < 0.05; Table 1, Fig. 2A).

View larger version (34K): [in this window] [in a new window]   Fig. 2. Active membrane properties. A: typical traces of transgenic (· · ·) and WT () action potentials (APs) are superimposed. APs in presymptomatic R6/2 transgenics were not different from WT (left). APs in symptomatic R6/2's were similar to WT in most respects, except that they decayed more slowly (see Table 1). B: each set of 3 traces depicts the voltage responses to 3 successively increasing amplitude depolarizing current pulses (100-pA increments), starting with the 1st trace exhibiting APs in each neuron (bottom trace, each set). Numbers next to each trace are the currents necessary to evoke the voltage response. As depolarization was increased, the latency to the first AP was reduced and the AP frequency increased, with the shortest interspike intervals near the beginning of the train. Symptomatic R6/2 transgenics (bottom right) generated shorter interspike intervals and exhibited greater spike frequency adaptation than their age-matched WT counterparts (see Table 1). In addition there were occasional apparent AP failures (inset, bottom right).

To further analyze APs, sequentially increasing depolarizing current steps (350-ms duration, 100-pA increments) were delivered through the recording electrode. Properties of the APs produced by the first three depolarizing traces exhibiting APs in each cell were analyzed in detail. Firing frequency generally increased with increasing depolarization for all groups (Fig. 2B). There were no differences in these parameters in neurons from presymptomatic R6/2 transgenics compared with those from age-matched WTs. The statistically significant interaction between genotype and age (F = 4.77, df = 1/44, P < 0.05) was due to the decreased number of APs during depolarizing current steps obtained from neurons in R6/2 transgenics compared with those from age-matched WTs (t = 2.37, P < 0.05; Table 1; Fig. 2B). Furthermore, in symptomatic R6/2 transgenics, intervals between APs were significantly shorter at the start of the train (Table 1; Fig. 2B) and showed more spike frequency adaptation than those from WTs (Fig. 2B). Again, the statistically significant interaction between genotype and age (F =10.8, df =1/44, P < 0.002) was due to the decreased mean interval at the start of the train in R6/2 transgenics compared with age-matched WTs (t = 4.35, P < 0.001). Occasionally, in symptomatic R6/2 transgenics small brief depolarizations, which did not generate action potentials, occurred especially during larger depolarizing current steps (Fig. 2B, expanded trace).

Synaptic responses were evoked by stimulation of afferents. In all groups, increasing the stimulus intensity produced synaptic responses in striatal neurons with increasing amplitudes and durations (Fig. 3). Three to five traces were averaged at each stimulus intensity, and input-output relationships were plotted and fit using a sigmoidal function. Due to AP contamination of EPSPs at higher stimulus intensities, three cells each from the presymptomatic R6/2 transgenic and WT groups produced input-output data that were not fit well by a sigmoidal curve and therefore were not included in the statistical comparisons of the input-output curves. These data were subsequently fit using empirically derived estimates so that representative traces with amplitudes between 20 and 50% of maximum for all cells could be chosen for measurement of EPSP characteristics.

View larger version (20K): [in this window] [in a new window]   Fig. 3. Excitatory postsynaptic potentials (EPSPs) evoked by stimulation of excitatory striatal afferents. Increasing stimulus intensity produced increasing EPSP amplitudes. Left: superimposed representative responses at ~20, 50, 80, and 100% maximum amplitude. Note that EPSPs in symptomatic R6/2 transgenics (bottom right) do not decay back to baseline within 45 ms (double arrows). Right: sigmoidal input-output curves fit to normalized data. Points show means ± SE of the stimulus intensity required to evoke EPSPs of 20, 50, and 80% of maximum amplitude. Symptomatic R6/2 transgenics required significantly greater stimuli to evoke proportional responses than did their age-matched WT counterparts.

The input-output relationships were similar between the presymptomatic R6/2 transgenics and their age-matched WTs (Fig. 3). In contrast, symptomatic R6/2 transgenics required a significantly greater stimulus intensity to achieve a particular EPSP amplitude. These data were analyzed with a two-way ANOVA with one repeated measure. There was a statistically significant interaction between genotype and current to achieve 20, 50, and 80% maximum EPSP amplitude (F = 26.6, df = 2/36, P < 0.001; Fig. 3, right). Individual comparisons at each percent of maximum EPSP value were also statistically significant (t = 2.58, P < 0.025; t = 5.29, P < 0.001; t = 8.00, P < 0.001 for 20, 50, and 80%, respectively). Thus the input-output curves were shifted to the right and had a shallower slope in the symptomatic transgenics compared with WTs.

Neurons from WT preparations showed a significant change in their input-output relationship as a function of age, with older animals requiring significantly lower amplitude stimuli to evoke proportional postsynaptic responses. These data also were analyzed with a two-way ANOVA with one repeated measure. There was a statistically significant interaction between age and current to achieve 20, 50, and 80% maximum EPSP amplitude (F = 9.7, df = 2/42, P < 0.001; compare Fig. 3, WT traces from top and bottom graphs). Individual comparisons at 50 and 80% of maximum EPSP value were statistically significant (t = 3.18, P < 0.025; t = 4.98, P < 0.001 for 50 and 80%, respectively). In contrast, a similar analysis indicated the R6/2 transgenic neurons tended to require relatively greater stimulus intensity to evoke proportionately similar amplitude responses with age (Fig. 3, compare R6/2 traces in top and bottom graphs). The differences were not statistically significant, however.

There were no statistically significant differences between neurons from presymptomatic R6/2 transgenics and their age-matched WTs for all synaptic response characteristics measured except paired-pulse facilitation (PPF; Table 2).

Peak EPSP amplitudes and durations at half amplitude also were similar between symptomatic R6/2 transgenic and age-matched WT neurons (Table 2). Charge transfer during the EPSP was estimated by measuring the area under the digitized EPSP curve during a 45-ms window starting at the base of the rising phase of the EPSP. Although EPSPs from neurons from symptomatic R6/2 transgenics tended to have larger areas than their age-matched WTs (Table 2), this difference was not statistically significant. Neurons from symptomatic R6/2 transgenics had a significantly longer 10-90% rise time than those from age-matched WTs. The statistically significant interaction between genotype and age (F = 4.24, df = 1/51, P < 0.05) was due to the longer mean rise time in the symptomatic R6/2 transgenics compared with age-matched WTs (t = 2.21, P < 0.05; Table 2). A single exponential curve fit to the decay phase of the EPSP showed that decay rate of the EPSP was not significantly different between symptomatic R6/2 transgenics and their age-matched WTs. EPSPs obtained from symptomatic R6/2 neurons decayed to a significantly depolarized level compared with those from age-matched WTs. There was a statistically significant interaction between genotype and age (F = 4.15, df = 1/42, P < 0.05) that was due to the more depolarized membrane potential in the R6/2 transgenics (t = 2.63, P < 0.05; Table 2; Fig. 3, double arrows).

PPF was measured by delivering stimuli to the corticostriatal afferents in pairs with an interstimulus interval of 50 ms. PPF was expressed as the ratio of the amplitude of the second EPSP divided by the amplitude of the first EPSP. While neurons from WTs in both age groups showed robust PPF, neurons from both pre- and symptomatic R6/2 transgenics displayed significantly less PPF with symptomatic R6/2 transgenics displaying almost no PPF as a group. The effect of genotype was statistically significant (F = 7.87, df = 1/41, P < 0.01; Table 2).

Pearson Product-Moment correlations were computed to determine relationships between altered electrophysiological properties in the symptomatic R6/2 transgenics and their age-matched WTs. Measures intercorrelated were RMP, input resistance, membrane time constant, PPF, and current to induce synaptic responses. In the WT group, there were no statistically significant correlation coefficients. In the symptomatic R6/2 transgenics, two correlation coefficients were statistically significant, RMP and PPF (R = 0.821, P < 0.04, n = 7) and input resistance and PPF (R = 0.942, P < 0.005, n = 7). These results indicate that changes in PPF vary with changes in RMP and input resistance. In the transgenic neurons, less PPF is associated with higher input resistances, and these cells had more depolarized RMPs. Thus changes in the passive membrane properties could affect how the neurons express synaptic plasticity.

Morphology

All electrodes contained biocytin, which was injected into each recorded cell in R6/2 transgenics and WTs at both ages. Morphological data were obtained from 21 neurons from presymptomatic R6/2 transgenics and 10 neurons from age-matched WTs and from 19 neurons from symptomatic R6/2 transgenics and 18 neurons from age-matched WTs. All of the recovered neurons had medium-sized somata. There were no obvious differences in the morphology of the neurons from the presymptomatic R6/2 transgenics and their age-matched WTs (Fig. 4, left). However, there were a number of differences in gross morphology of neurons from the symptomatic R6/2's compared with those from their age-matched WTs (Fig. 4, middle). There was a statistically significant decrease in cross-sectional somatic area (126 ± 9 vs. 101 ± 9 µm² for WT and R6/2, respectively, t = 2.049, df = 35, P < 0.05) an effect we also demonstrated previously (Levine et al. 1999). Neurons from symptomatic R6/2 transgenic mice also had thinner dendrites and reduced dendritic spines compared with those from age-matched WTs (Fig. 4, right). To determine spine density in the biocytin-filled neurons, three observers, who were blinded to the genotype from which the neurons were obtained, rated spine density on a single distal dendrite for each neuron using the following scale (0-4 spines/10 µm; 5-9 spines/10 µm; >9 spines/10 µm). Each rater chose a 10-µm segment from the dendrite that had the greatest spine density and was furthest from the soma. There were high inter-rater reliabilities for the resulting classification (R = 0.88-0.93). In the symptomatic R6/2 transgenic neurons (n = 19), 53% of the cells had 0-4 spines/10 µm, 37% had 5-9 spines/10 µm, while 10% had >9 spines/10 µm. In the corresponding WT group (n = 18), 6% of the neurons had 0-4 spines/10 µm, 38% had 5-9 spines/10 µm, while 56% had >9 spines/10 µm. The difference between the distributions was statistically significant (² = 12.7, df = 3, P < 0.025). A similar analysis was performed on neurons from the presymptomatic R6/2 mice and their age-matched WTs. In the presymptomatic R6/2 mice (n = 21 neurons), 10% of the cells had 0-4 spines/10 µm, 33% had 5-9 spines/10 µm, while 57% had >9 spines/10 µm. In the corresponding WT group (n = 10 neurons) 20% of the neurons had 0-4 spines/10 µm, 20% had 5-9 spines/10 µm, while 60% had >9 spines/10 µm. The difference between the distributions was not statistically significant.

View larger version (55K): [in this window] [in a new window]   Fig. 4. Morphology of biocytin-filled medium-sized spiny neurons. Left: 4 micrographs show that neurons from presymptomatic R6/2 transgenics were similar to those of their WT counterparts. Calibration at bottom right refers to all micrographs. Middle: 4 micrographs show symptomatic R6/2 transgenic neurons have decreased dendritic diameters and decreased spine density compared with age-matched WTs. Calibration at bottom left refers to left 2 micrographs and at bottom right to right 2 micrographs. Right: 3 micrographs illustrate enlargements of dendritic segments showing paucity of dendritic spines and reduced dendritic diameters in neurons from symptomatic R6/2 transgenics. Right image is a higher magnification of the dendrite shown in the middle panel from the R6/2 transgenic. Calibration in right micrograph also refers to the middle micrograph.

Golgi impregnation was used to further assess dendritic appearance, spine density, and cell morphology in symptomatic R6/2 mice and their age-matched WT controls. This analysis also demonstrated a low incidence of dendritic spines on the medium-sized striatal neurons (Fig. 5), a reduction in the diameter of the dendritic field (Fig. 5B), and a decrease in the diameters of the dendritic shafts among the neurons in the R6/2 transgenics (Fig. 5, A and C). Additionally, the thin dendrites of the medium-sized transgenic striatal neurons exhibited scattered varicosities along the shafts. These expansions were ~2-3 µm in diameter and 5 µm in length.

View larger version (59K): [in this window] [in a new window]   Fig. 5. Morphology of Golgi-stained medium-sized spiny striatal neurons. A: photomicrographs of Golgi-stained neuronal field showing somata (top) and dendrites (bottom) of WT (left 3 micrographs) and R6/2 transgenics (right 3 micrographs). Calibration in top right panel refers to both top panels. Calibration in bottom left panel refers to all 4 bottom panels. B: camera lucida drawings of typical WT and R6/2 transgenic medium-sized spiny neurons. Note the marked decrease in the extent of the dendritic field in the R6/2 neuron. C and D: spine density was quantified by counting spines from the Golgi-impregnated material using high magnification (×100). C: camera lucida drawings of typical dendritic segments at each of the 5 branch orders. For each case, 10 neurons were chosen in which the best impregnated dendrite could be traced over 5 orders of dendritic branching (inset). D: bar graphs of spine density at each order. There was a statistically significant decrease in spine density for neurons obtained from R6/2 transgenics compared with WTs at branch orders 2-5 (*, see RESULTS).

In the Golgi-impregnated material, the spine density on medium-sized neurons was quantified by counting spines using high magnification (×100) by an observer blinded to the genotype of the specimen. For each of the five cases in each group, 10 neurons were chosen in which the best impregnated dendrite could be traced over five orders of dendritic branching. The number of spines for each branch order was counted, and an average for each branch order for all neurons from each animal was computed (Fig. 5D). These data were analyzed with a two-way ANOVA (genotype vs. branch order) with one repeated measure (branch order). There was a significant effect of genotype on spine density (F = 5.63, df = 1/8, P < 0.05). Subsequent post hoc tests indicated spine densities were significantly reduced at branch orders 2-5 (t = 3.89, P < 0.05; t = 4.11, P < 0.025; t = 3.02, P < 0.05; t = 2.99, P < 0.05, respectively; Fig. 5D).

In the Golgi-impregnated material, medium and large-sized aspiny striatal neurons were evident in both R6/2 and WT striata. Although we did not examine these neurons in detail, there were no obvious gross morphological differences between groups (Fig. 6, A and B). Morphological differences between the transgenic and WT groups were not confined to medium-sized spiny neurons in the striatum. For example, substantial reductions of dendritic spine densities, dendritic diameters and extent of the dendritic arborizations were found among cortical and hippocampal pyramidal neurons (Fig. 6C).

View larger version (27K): [in this window] [in a new window]   Fig. 6. Morphology of Golgi-stained medium- and large-sized aspiny striatal neurons and cortical pyramidal neurons. A and B: medium and large aspiny neurons from symptomatic R6/2 transgenics were not distinguishable from WT. C: cortical pyramidal neurons from R6/2 transgenics showed numerous alterations including loss of spines, decreases in somatic size, and decreases in extent of dendritic fields.

Together, the analyses of the biocytin-filled and Golgi-impregnated neurons indicate that marked morphological changes have occurred in symptomatic R6/2 transgenics. Furthermore, the significant decrease in spine density provides a morphological substrate for the increase in current necessary to induce synaptic responses and further suggests a loss of inputs to medium-sized spiny striatal neurons.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The major findings of these experiments encompass two areas. First, there are alterations in the passive and active electrophysiological properties of medium-sized striatal spiny neurons in R6/2 transgenic mice. Increases in input resistance and decreases in rheobase were observed earliest, occurring in both pre- and symptomatic R6/2 transgenics. In the symptomatic transgenic mice, a number of other alterations occurred. These were depolarized RMPs, decreased membrane time constants, decreased AP decay slopes, and alterations in the ability to generate trains of APs. The second major area is a breakdown in afferent connectivity of medium-sized spiny neurons demonstrated by decreased responsiveness to activation of synaptic inputs. There was a significant shift in the input-output function such that increased intensities of white matter stimulation were required to induce striatal EPSPs. There was also a decrease in PPF in R6/2 mice suggesting alterations in synaptic plasticity. Morphological alterations were also present in medium-sized spiny neurons in the symptomatic R6/2 transgenics. These included decreased spine densities, smaller diameter dendritic shafts, smaller somatic cross-sectional areas, and decreased diameter of the dendritic fields. These morphological changes provide a basis for the changes in synaptic responses.

Passive and active membrane properties of medium-sized spiny neurons

Significant increases in input resistance occurred in both pre- and symptomatic R6/2 transgenics. Such changes generally indicate reduced ionic conductances across the cell membrane at rest as predicted by Ohm's law. Input resistance has been attributed to K⁺ conductances, although reductions in other conductances, such as the nonspecific leak or Cl currents, could also affect input resistance and cannot be ruled out (Hille 1992). We have observed a reduction in the inward and delayed rectifying K⁺ currents in the striatum of R6/2 mice that could contribute to the increased input resistance (Ariano et al. 2000). However, the decrease in K⁺ conductances occurred primarily in neurons from symptomatic mice. It is possible that mechanisms responsible for input resistance increases may be different in pre- and symptomatic R6/2 transgenics. Because none of the RMP, AP decay, or firing rate was changed in the presymptomatic mice, attenuated K⁺ conductances may be more prevalent in the distal dendrites where their contributions to cellular physiology may not be detected by somatic recordings. Progressive K⁺ current reductions correlated with development of the behavioral phenotype might account for the increasing effects on other electrophysiological parameters in the symptomatic transgenics. We also have observed a decrease in voltage-gated Ca²⁺ conductances that could contribute to the physiological alterations detected in symptomatic R6/2 mice (Cepeda et al. 2001). The decreased membrane time constant in R6/2 transgenics suggests a reduction in cell surface area. Indeed, our morphological studies indicate reduced cross-sectional somatic areas, reduced dendritic fields, thinner dendrites, and decreased dendritic spine density in neurons from symptomatic R6/2 mice.

Depolarized RMPs occurred in the symptomatic mice. K⁺ conductances are responsible for establishing the membrane potential (Hille 1992). The inwardly rectifying K⁺ conductance has been specifically implicated in establishing the membrane potential of medium-sized spiny neurons (Nisenbaum and Wilson 1995). As pointed out in the preceding text, we have detected a significant loss in the inward rectifying K⁺ current in symptomatic R6/2 (Ariano et al. 2000), supporting this inference. Another potential mechanism to explain the depolarized RMP in the HD model is a change in cellular energy balance that ultimately compromises the integrity of the membrane potential (Beal 2000). The depolarized RMP may indicate a metabolic deficiency because many of the mechanisms that maintain this potential are ATP dependent (Calabresi et al. 1995). This mechanism has been proposed to account for changes in animal models of HD (Brouillet et al. 1999; Grunewald and Beal 1999; Tabrizi et al. 2000) and in HD patients (Polidori et al. 1999).

Depolarized RMPs may predispose neurons to degeneration via excitotoxic mechanisms. For example, the Mg²⁺ block of the NMDA receptor may be reduced, predisposing NMDA receptors to become activated more frequently in R6/2 transgenics compared with WT mice. In support of this, we have found that NMDA receptor function is enhanced in a subpopulation of medium-sized spiny neurons in the R6/2 transgenic as well as in other HD transgenic and knock in mice (Cepeda et al. 2001; Laforet et al. 2001; Levine et al. 1999).

Few of the active membrane properties examined were altered in symptomatic R6/2 transgenics. The slower decay of the AP and shorter AP intervals are consistent with a reduction in voltage-dependent K⁺ currents (Hille 1992). Although the AP threshold was not altered in pre- or symptomatic mice, current intensity (rheobase) required to reach AP threshold was significantly lower in transgenics. This observation is consistent with the increase in input resistance. Theoretically, a lower rheobase would predispose neurons to generate APs more readily and perhaps ectopically. This would tend to increase the firing of medium-sized neurons in transgenic animals, possibly increasing the level of synaptic "noise" in the circuitry, thus interfering with functional synaptic communication. R6/2 mice do display specific learning impairments (Carter et al. 1999; Lione et al. 1999). Recent studies in patients with HD also indicate that subtle changes in function are detectable prior to the onset of overt motor dysfunction (Harris et al. 1999; Smith et al. 2000). There is also a reduction in the number of APs in neurons from symptomatic R6/2 neurons during the depolarizing current steps, and occasional brief depolarizations that resembled AP failures. Similar changes also are observed in aged striatal neurons (Cepeda et al. 1992).

Synaptic responses

The glutamatergic corticostriatal input provides the major source of excitation to medium spiny striatal neurons (Fonnum et al. 1981; Smith and Bolam 1990). Symptomatic mutants exhibited marked changes in the properties of EPSPs mediated by activation of striatal inputs. The altered input-output relationship in symptomatic R6/2 mice suggests a loss in glutamatergic synaptic inputs from cortex and possibly the thalamus. Similar changes in input-output relationships of the striatal EPSP have been observed in another HD transgenic model with 100 CAG repeats (Laforet et al. 2001). Decreased dendritic spine density on medium-sized neurons provides morphological evidence for losses of postsynaptic sites in the R6/2 transgenic. Excitatory striatal inputs, mostly those from cortex, end primarily on spines (Smith and Bolam 1990). One issue is whether alterations in excitatory striatal inputs are primary rather than secondary to striatal alterations in HD. There is evidence that mutant huntingtin accumulates in the cortex (Sapp et al. 1999) and marked cortical changes occur in human HD (DiFiglia et al. 1997). Thus the issue becomes whether the decrease in synaptic connectivity is caused first by a loss of cortical inputs, which in turn cause the spines to degenerate or if the spines degenerate first, thus disconnecting cortical input. Clearly, more experiments will have to be performed to resolve this issue.

The loss of spines coupled with the decrease in excitatory synaptic input provides a mechanism to explain the observation that striatal neurons in the R6/2 transgenic are protected from excitotoxicity (Hansson et al. 1999; Morton and Leavens 2000). There is evidence that generation of excitotoxicity in vivo requires an intact glutamatergic input to the target structure that is subject to excitotoxic challenge. Kainate-induced degeneration of striatal neurons is dependent on the integrity of the corticostriatal pathway (McGeer et al. 1978). Thus if this connection is reduced, striatal neurons may display protection from excitotoxicity.

The maximum amplitude of synaptic responses was not different between R6/2 and WT mice. However, it is difficult to draw conclusions from this because the alterations in both the RMP and input resistance in the transgenics would have opposing effects on the amplitude of the EPSP. The kinetics of the postsynaptic potentials also are changed. Symptomatic R6/2 transgenics display slower rise and incomplete decay of the EPSP by 45 ms. This phenomenon could be explained by nonsynchronous synaptic discharges in the R6/2 transgenic. More likely these changes indicate a larger contribution of a slower kinetic ionic current, such as one mediated by activation of NMDA receptors. Enhanced NMDA currents and responses associated with increased NMDA-R1 subunit expression have been found in this HD model (Cepeda et al. 2001; Levine et al. 1999), providing complementary evidence for increased NMDA receptor activation.

There was a marked reduction in PPF in R6/2 mice. Thus at frequencies that are likely to occur in vivo (i.e., 20 Hz), repetitive synaptic activity is impaired. PPF has been used as an indicator of presynaptic transmitter release probability at central synapses (Dobrunz and Stevens 1997; Zucker 1989). While the postsynaptic passive membrane abnormalities in symptomatic mutants may confound assumptions about such presynaptic properties in the R6/2 mice, previous findings from a CAG knock in mouse model of HD demonstrated decreases in transmitter release probability during repetitive stimulation in the hippocampus (Usdin et al. 1999). The enhanced "late phase" of the EPSP also may contribute to the absence of PPF in symptomatic mutants because the postsynaptic conductances that remain active at the time of the second stimulus would not be available to contribute to the second EPSP. Additionally, the concomitant reduction of input resistance during the tail of the first EPSP may serve to "shunt" subsequent synaptic currents.

PPF has been considered an index of synaptic plasticity in the corticostriatal pathway (Tang et al. 2001). Thus the present findings also suggest that striatal synaptic plasticity might be impaired in R6/2 transgenics. Abnormal synaptic plasticity associated with impaired spatial cognition has been observed in the hippocampus of the R6/2 transgenic (Murphy et al. 2000). Many of the cellular anomalies hypothesized to occur in HD have been associated with alterations in synaptic plasticity in the striatum (Centonze et al. 2000).

In conclusion, these experiments provide important information about the functional alterations that occur in medium-sized spiny neurons of R6/2 HD transgenics. Our findings suggest that two fundamental changes have occurred during the development of the overt behavioral phenotype: changes in passive membrane properties and in several properties of the AP and a major decrease in excitatory striatal synaptic inputs. From a functional perspective, these alterations will markedly change information processing in the striatum and its ability to integrate and relay information through the basal ganglia. From the standpoint of HD, changes in these functional properties suggest areas to target for pharmacotherapies to alleviate and reduce the symptoms of HD.