Alterations in functional thalamocortical connectivity following neonatal whisker trimming with adult regrowth.

Neonatal whisker trimming followed by adult whisker regrowth leads to higher responsiveness and altered receptive field properties of cortical neurons in corresponding layer 4 barrels. Studies of functional thalamocortical (TC) connectivity in normally reared adult rats have provided insights into how experience-dependent TC synaptic plasticity could impact the establishment of feedforward excitatory and inhibitory receptive fields. The present study employed cross-correlation analyses to investigate lasting effects of neonatal whisker trimming on functional connections between simultaneously recorded thalamic neurons and regular-spike (RS), presumed excitatory, and fast-spike (FS), presumed inhibitory, barrel neurons. We find that, as reported previously, RS and FS cells in whisker-trimmed animals fire more during the earliest phase of their whisker-evoked responses, corresponding to the arrival of TC inputs, despite a lack of change or even a slight decrease in the firing of thalamic cells that contact them. Functional connections from thalamus to cortex are stronger. The probability of finding TC-RS connections was twofold greater in trimmed animals and similar to the frequency of TC-FS connections in control and trimmed animals, the latter being unaffected by whisker trimming. Unlike control cases, trimmed RS units are more likely to receive inputs from TC units (TCUs) and have mismatched angular tuning and even weakly responsive TCUs make strong functional connections on them. Results indicate that developmentally appropriate tactile experience early in life promotes the differential thalamic engagement of excitatory and inhibitory cortical neurons that underlies normal barrel function.

Weaker feedforward inhibition accounts for less pronounced thalamocortical response transformation in mouse vs. rat barrels.

Feedforward inhibition is a common motif of thalamocortical circuits. Strong engagement of inhibitory neurons by thalamic inputs enhances response differentials between preferred and nonpreferred stimuli. In rat whisker-barrel cortex, robustly driven inhibitory barrel neurons establish a brief epoch during which synchronous or near-synchronous thalamic firing produces larger responses to preferred stimuli, such as high-velocity deflections of the principal whisker in a preferred direction. Present experiments in mice show that barrel neuron responses to preferred vs. nonpreferred stimuli differ less than in rats. In addition, fast-spike units, thought to be inhibitory barrel neurons, fire less robustly to whisker stimuli in mice than in rats. Analyses of real and simulated data indicate that mouse barrel circuitry integrates thalamic inputs over a broad temporal window, and that, as a consequence, responses of barrel neurons are largely similar to those of thalamic neurons. Results are consistent with weaker feedforward inhibition in mouse barrels. Differences in thalamocortical circuitry between mice and rats may reflect mechanical properties of the whiskers themselves.

Functional independence of layer IV barrels in rodent somatosensory cortex.

Layer IV of rodent primary somatosensory cortex is characterized by an array of whisker-related groups of neurons, known as "barrels." Neurons within each barrel respond best to a particular whisker on the contralateral face, and, on deflection of adjacent whiskers, display relatively weak excitation followed by strong inhibition. A prominent hypothesis for the processing of vibrissal information within layer IV is that the multiwhisker receptive fields of barrel neurons reflect interconnections among neighboring barrels. An alternative view is that the receptive field properties of barrel neurons are derived from operations performed on multiwhisker, thalamic inputs by local circuitry within each barrel, independently of neighboring barrels. Here we report that adjacent whisker-evoked excitation and inhibition within a barrel are unaffected by ablation of the corresponding adjacent barrel. In supragranular neurons, on the other hand, excitatory responses to the ablated barrel's associated whisker are substantially reduced. We conclude that the layer IV barrels function as an array of independent parallel processors, each of which individually transforms thalamic afferent input for subsequent processing by horizontally interconnected circuits in other layers.

Laminar differences in bicuculline methiodide's effects on cortical neurons in the rat whisker/barrel system.

Extracellular unit recordings were made at various depths within SmI barrel cortex of immobilized, sedated rats, in the presence and absence of titrated amounts of the GABA(A) receptor antagonist bicuculline methiodide (BMI). Principal and adjacent whiskers were moved singly, or in paired combination in a condition-test paradigm, to assess excitatory and inhibitory receptive field (RF) characteristics. Neurons were classified as regular- or fast-spike units, and divided into three laminar groups: supragranular, granular (barrel), and infragranular. BMI increased response magnitude and duration, but did not affect response latencies. The excitatory RFs of barrel units, which are the most tightly focused on the principal whisker, were the most greatly defocused by BMI; infragranular units were least affected. All three layers had approximately equal amounts of adjacent whisker-evoked, surround inhibition, but BMI counteracted this inhibition substantially in barrel units and less so in infragranular units. The effects of BMI were most consistent in the barrel; more heterogeneity was found in the non-granular layers. These lamina-dependent effects of BMI are consistent with the idea that between-whisker inhibition is generated mostly within individual layer IV barrels as a result of the rapid engagement of strong, local inhibitory circuitry, and is subsequently embedded in layer IV's output to non-layer IV neurons. The latter's surround inhibition is thus relatively resistant to antagonism by locally applied BMI. The greater heterogeneity of non-granular units in terms of RF properties and the effects of BMI is consistent with other findings demonstrating that neighboring neurons in these layers may participate in different local circuits.

Effects of baclofen and phaclofen on receptive field properties of rat whisker barrel neurons.

Extracellular single-unit recordings were made in somatosensory cortical barrels of fentanyl-sedated rats. Whiskers were deflected singly or in paired combinations. Iontophoretically-applied (-)-baclofen disproportionately reduced weak responses, and phaclofen disproportionately increased them, resulting in more tightly focused or more broadly focused receptive fields, respectively. Both drugs had only minor effects on surround inhibition. In light of previous findings, we conclude that GABAA and GABAB mechanisms both act to enhance spatial contrast, but that the former plays a much greater role in enhancing temporal resolution.

Quantitative effects of GABA and bicuculline methiodide on receptive field properties of neurons in real and simulated whisker barrels.

1. Carbon fiber multibarrel glass microelectrodes were used to record extracellular single-unit activity during microiontophoretic application of gamma-aminobutyric acid (GABA) or bicuculline methiodide (BMI) onto layer IV barrel neurons in the somatosensory cortex of fentanyl-sedated rats. Excitatory and inhibitory aspects of the neurons' receptive fields were quantified with the use of controlled whisker stimuli. The principally activating whisker and one of its immediately adjacent neighbors were deflected alone or in paired combinations involving a condition-test paradigm. 2. Units were distinguished electrophysiologically on the basis of the time course of their action potential waveforms. Data were obtained from 26 regular-spike units (RSUs; presumed spiny stellate cells) and 7 fast-spike units (FSUs; presumed GABAergic neurons). An average of 15.0 nA of GABA produced a one-third to one-half reduction in RSU responses evoked by the maximally effective stimulus. An average of 8.7 nA of BMI was needed to counteract this reduction. This amount of BMI, in the absence of exogenous GABA, was found to increase average RSU and FSU responses by 98 and 53%, respectively, relative to predrug levels. 3. For RSUs, the BMI-induced twofold increase in responses evoked by moving the principal whisker at the neuron's best deflection angle was accompanied by an almost threefold increase in responses evoked by similarly moving an adjacent whisker. Disproportionately large percentage increases were also seen for responses to nonpreferred directions of principal and adjacent whisker movement. BMI thus effectively increased receptive field size and decreased angular tuning. Similarly, responses to stimulus offsets, which are normally smaller than ON responses, were increased proportionally more. 4. Predrug responses of FSUs were more vigorous than those of RSUs. However, FSUs showed a similar inverse relationship between percentage increase with BMI and initial response magnitude, although the proportional increases were less pronounced. 5. GABA, like BMI, had the greatest proportional effects on those responses that were initially smallest. It produced results opposite those of BMI, effectively decreasing receptive field size and sharpening angular tuning. 6. A previously described computational model of a barrel was tested for its ability to reproduce quantitatively the effects of BMI and GABA. The application of BMI was simulated by decreasing the strength of the inhibitory inputs onto the particular cell under study in the model network. GABA microiontophoresis was simulated by adding a constant hyperpolarizing voltage. The model RSUs and FSUs displayed proportional changes in response magnitude that were quantitatively similar to those of their biological counterparts. 7. Surround inhibition was greatly attenuated by BMI application, both for the real and simulated barrel neurons. Disinhibition was less pronounced for the former, perhaps because, unlike the simulated neurons, they also possess GABAB receptors, which are unaffected by BMI. 8. We conclude that the inhibitory receptive field properties of barrel neurons can be explained by intrabarrel inhibition and that the expansion of receptive field size and loss of angular tuning with BMI is due to an enhanced effectiveness of convergent, multi-whisker thalamocortical input. Examination of the model neurons' behavior suggests that the altered activity in response to GABA or BMI application, respectively, can be explained by the nonlinear effects of shifting somal membrane potential away from or toward the neuron's firing threshold.

OFF response transformations in the whisker/barrel system.

1. Previous studies have demonstrated marked differences in the relative sizes of ON and OFF responses of neurons in the whisker/barrel system. In particular, OFF responses are unexpectedly large in thalamic neurons. Extracellular unit recordings were used to examine whether varying the time between stimulus onset and offset differently affects OFF responses of neurons in the trigeminal ganglion, ventrobasal thalamus, and somatosensory cortical layer IV. Controlled whisker stimuli were used to deflect individual vibrissal hairs in different directions. We hypothesized that, in part because of the gradual waning of central inhibition evoked by stimulus onset, OFF responses of thalamic and cortical neurons but not trigeminal ganglion cells would increase in size with longer duration stimuli, with relative changes being greatest in the cortex. 2. OFF response magnitudes for thalamic and cortical neuronal populations increased as the stimulus duration was increased from 200 to 1,400 ms. Increases were greater at nonoptimal deflection angles. Similarly, individual cells having smaller OFF responses for the short-duration stimulus tended to display proportionately greater increases when the stimulus was lengthened. OFF responses of trigeminal ganglion cells were largely unaffected by stimulus duration. 3. Barrel neurons were subclassified as regular-spike units (RSUs) or fast-spike units (FSUs) on the basis of the time course of their action potentials. ON and OFF responses were smaller in the former and, when the stimulus was lengthened, percentage increases in their OFF responses were greater than those in FSUs. Results illustrate nonlinear transformations of the thalamic input signal by RSUs, which are presumed to be excitatory barrel neurons, and extend previous findings of response similarities between thalamocortical units (TCUs) and FSUs, the latter of which are thought to be inhibitory. 4. The time course of OFF response suppression in cortical neurons suggests that stimulus onset evokes central inhibition having two components, a potent one lasting several tens of milliseconds and a weaker one lasting many hundreds of milliseconds. Background activity levels in cortex and thalamus were diminished for > or = 1,800 ms after whisker movement. 5. For TCUs, 200-ms stimuli were less likely than 1,400-ms stimuli to elicit an OFF response, but when responses occurred they consisted of a greater number of spikes timed closer together. By contrast, the 200-ms stimulus OFF responses of the RSUs and FSUs displayed longer interspike intervals than did their 1400-ms responses, with no change in the number of spikes per response.(ABSTRACT TRUNCATED AT 400 WORDS)

Thalamocortical response transformations in simulated whisker barrels.

Layer IV of rodent somatosensory cortex contains identifiable networks of neurons, called "barrels," that are related one-to-one to individual whiskers on the face. A previous study (Simons and Carvell, 1989) described differences between the response properties of thalamic and cortical vibrissa neurons and proposed that these transformations can be explained by several features of barrel anatomy and physiology: nonlinear neuronal properties, strongly responsive inhibitory and less responsive excitatory neurons, convergent thalamic inputs to cells of both types, and interconnections among barrel neurons. In the present study these features were incorporated into a computational model in order to test their explanatory power quantitatively. The relative numbers of excitatory and inhibitory cells and the relative numbers of synapses of thalamic and intrabarrel origin were chosen to be consistent with available light and electron microscopic data. Known functional differences between excitatory and inhibitory barrel neurons were simulated through differences in spike activation functions, refractory periods, postsynaptic potential decay rates, and synaptic strengths. The model network was activated by spike trains recorded previously from thalamic neurons in response to three different whisker deflection protocols, and output, which consisted of spikes generated by the simulated neurons, was compared to data from our previous neurophysiological experiments. For each type of whisker stimulus, the same set of parameter values yielded accurate simulations of the cortical response. Realistic output was obtained under conditions where each barrel cell integrated excitatory and inhibitory synaptic inputs from a number of thalamic and other barrel neurons and where the ratios between network excitation, network inhibition, and thalamic excitation were approximately constant. Several quantities are defined that may be generally useful in characterizing neuronal networks. One important implication of the results is that thalamic relay neurons not only provide essential drive to the cortex but could, by changing their tonic activities, also directly regulate the tonic inhibition present in the cortex and thereby modulate cortical receptive field properties.

Normalization of Na(+)-K(+)-ATPase activity in isolated membrane fraction from sciatic nerves of streptozocin-induced diabetic rats by dietary myo-inositol supplementation in vivo or protein kinase C agonists in vitro.

A myo-inositol-related defect in nerve Na(+)-K(+)-ATPase in experimental diabetes has been invoked in the pathogenesis of diabetic neuropathy, but the mechanism linking altered myo-inositol metabolism and Na(+)-K(+)-ATPase regulation in diabetic nerve is uncertain. Decreased Na(+)-K(+)-ATPase in diabetic rat nerve is normalized by aldose reductase inhibitors or dietary myo-inositol, which preserve normal nerve myo-inositol content in vivo. Decreased Na(+)-K(+)-ATPase in diabetic rabbit nerve is acutely reversed by exposure to protein kinase C agonists in vitro. This study explored the relationship between the myo-inositol-sensitive and protein kinase C-agonist-sensitive Na(+)-K(+)-ATPase defects in diabetic rat nerve. Ouabain-sensitive ATPase activity was measured in an enriched membrane fraction isolated from nondiabetic, streptozocin-induced diabetic, and myo-inositol-supplemented streptozocin-induced diabetic rats before and after the membranes were exposed to protein kinase C agonists in vitro. The decreased ouabain-sensitive ATPase activity in plasma membranes from untreated diabetic rats was increased after exposure to two structurally unrelated protein kinase C agonists; the normal ouabain-sensitive ATPase in plasma membranes from myo-inositol-supplemented diabetic rats was unaffected by protein kinase C agonists. The nonadditivity and implied equivalence of the Na(+)-K(+)-ATPase defect corrected by myo-inositol in vivo and by protein kinase C agonists in vitro are consistent with the postulated existence of a deficient myo-inositol-dependent phospholipid-derived protein kinase C agonist (presumably diacylglycerol) in diabetic nerve that regulates nerve Na(+)-K(+)-ATPase either directly or via a protein kinase C mechanism.

Intractable unphysiologically low adenylate energy charge values in synaptosome fractions: an explanatory hypothesis based on the fraction's heterogeneity.

Synaptosomes prepared and incubated in a variety of ways from rat cerebra exhibited intractable, unphysiologically low adenylate energy charge values (approximately 0.37-0.60), low total adenine nucleotide contents (approximately 8-10 nmol/mg protein), and much higher adenylate kinase apparent Keq values (approximately 3-8) as compared to intact brain tissue (values of approximately 0.90, 25 nmol/mg, and 0.74, respectively). Synaptosomes prepared from mouse, dog, and chicken cerebra had values essentially identical to those from rat. When incubated under oxygen in a physiological salt solution containing glucose, synaptosomes metabolized more glucose to lactic acid than to CO2, and the addition of 100 microM veratridine caused a two- to threefold stimulation of O2 uptake, lactate accumulation, and CO2 output. It is known that synaptosome fractions contain a substantial number (at least 30-45% by volume) of cytoplasm-containing particles devoid of mitochondria (henceforth termed "cytosolic particles"), and that approximately 80% of brain hexokinase is bound to the outer mitochondrial membrane. For the cytosolic particles, lacking oxidative phosphorylation, to maintain their "in vivo" ATP turnover would require about a 19-fold increase in the glycolytic rate, which is not possible due to limiting amounts of hexokinase, and thus these particles are postulated to be responsible for the high level of aerobic lactate accumulation and the intractable low energy charge values found in synaptosome fractions. The mitochondria-containing particles are postulated to have a normal energy charge, a submaximal glycolytic rate, and minimal lactate production, on the basis of the capacity of veratridine to stimulate synaptosomal O2 uptake and CO2 and lactate output. Calculations based on this "two populations of particles" hypothesis indicate that for synaptosome fractions in general, (1) the cytosolic particles contain approximately 35-64% of the total adenine nucleotides and maintain an energy charge of approximately 0.12; (2) the cytosolic particles and mitochondria-containing particles have adenylate kinase apparent Keq values of approximately 0.21-1.66 and 0.74, respectively, revealing that the higher apparent Keq values of the synaptosome fractions probably are not real departures from equilibrium: and (3) approximately 31-45% of synaptosome fraction protein is contained in debris, which, when taken into account, yields total adenine nucleotide contents in the cytosolic particles and mitochondria-containing particles of approximately 15-24 and approximately 11-19 nmol/mg of particle protein, respectively.

An examination of the in vivo distribution of brain hexokinase between the cytosol and the outer mitochondrial membrane.

These studies addressed the question of the in vivo distribution of rat brain hexokinase (HK), and whether physiologically relevant changes in the glycolytic rate are accompanied by changes in the distribution of HK. Homogenates of fresh tissue showed only 11-15% of the overt (assayable without added detergent) HK to be soluble (found in high-speed centrifugation supernatant fractions) when homogenization was begun within 15-20 s of sacrifice. Freeze-blown rat brain tissue also was used, coupled with a new technique wherein it was homogenized as it thawed in a buffered sucrose solution containing 1 mM EDTA. In tissue sampled 15 min (anesthetized) or 60 min (waking) after ip Nembutal injection (40 mg/kg), 23% of the overt HK and 79% of the total lactate dehydrogenase were soluble. The average phosphocreatine content of these and similar homogenates had decreased only 23% from in vivo levels, while ATP had decreased by 65%, due to the combined effects of a high level of endogenous ATPase, chelation of Mg2+ by EDTA, and the greater stability of Mg-ATP2- relative to Mg-ADP1-. These data indicated that the tissue experienced, at most, the equivalent of 6 s of complete ischemia prior to the completion of homogenization. Synaptosomes derived from rat and chicken cerebra were incubated at 37 degrees C in a physiological salt solution containing 10 mM glucose. Addition of veratridine has been shown to stimulate glycolysis and oxidative phosphorylation two- to threefold (H. T. Kyriazi and R. E. Basford (1986) J. Neurochem., in press), but did not alter the HK distribution, as 21% was found in the supernatant fractions of both control and veratridine-stimulated synaptosomes treated with digitonin. These results indicate that in brain tissue, large net movements of HK on and off the outer mitochondrial membrane do not occur, and thus play no role in the regulation of glycolysis.

An enzymatic inosine 5'-monophosphate assay of increased specificity.

The enzymatic inosine 5'-monophosphate assay described by Grassl [in, Methods of Enzymatic Analysis (H. U. Bergman, ed.), pp. 2168-2171, Academic Press, New York (1974)] is highly nonspecific, as ITP, ATP, ADP, AMP, and adenosine react stoichiometrically. The reactivity with the adenine derivatives is due to the tri- and diphosphatase activity of alkaline phosphatase (AP), coupled with adenosine deaminase (and possibly AMP deaminase) contamination of commercially available preparations of AP, purine-nucleoside phosphorylase, and/or xanthine oxidase. The inclusion of coformycin (0.05 microgram/ml), a potent inhibitor of these deaminases, completely eliminated the cross-reactivity. ITP, however, still reacted stoichiometrically due to the tri- and diphosphatase activity of AP. Meyer and Terjung [Amer. J. Physiol. 237 C111-C118 (1979)] introduced a modification of Grassl's procedure, substituting 5'-nucleotidase for AP. It has been found that this disallows reactivity with ATP, ADP, and ITP but that AMP and adenosine still react completely. Coformycin prevents this cross-reactivity. It is therefore recommended that the assay be carried out with 5'-nucleotidase (instead of AP) and coformycin, in order to achieve a more specific assay, and one more suitable for use with whole tissue extracts.