|Home | About | Journals | Submit | Contact Us | Français|
Primary afferent axons within the solitary tract (ST) relay homeostatic information via glutamatergic synapses directly to 2nd order neurons within the nucleus of the solitary tract (NTS). These primary afferents arise from multiple organ systems and relay multiple sensory modalities. How this compact network organizes the flow of primary afferent information will shape central homeostatic control. To assess afferent convergence and divergence, we recorded ST-evoked synaptic responses in pairs of medial NTS neurons in horizontal brainstem slices. ST shocks activated EPSCs along monosynaptic or polysynaptic pathways. Gradations in shock intensity discriminated multiple inputs and stimulus recruitment profiles indicated that each EPSC was unitary. In 24 pairs, 75% were 2nd order neurons with 64% receiving one direct ST input with the remainder receiving additional convergent ST afferent inputs (22% two; 14% three monosynaptic ST-EPSCs). Some (34%) 2nd order neurons received polysynaptic EPSCs. Neurons receiving only higher order inputs were uncommon (13%). Most ST-EPSCs were completely independent, but four EPSCs of a total of 81 had equal thresholds, highly correlated latencies and synchronized synaptic failures consistent with divergence from a single source ST axon or from a common interneuron producing a pair of polysynaptic EPSCs. We conclude that ST afferent inputs are remarkably independent with little evidence of substantial shared information. Individual cells receive highly focused information from the viscera. Thus, afferent excitation of 2nd order NTS neurons is generally dominated by single visceral afferents and therefore focused on a single afferent modality and/or organ region.
The integrated control of visceral organs supports homeostasis and depends upon coordination within autonomic networks concentrated in the brainstem (Loewy, 1990;Saper, 2002;Travagli et al., 2006). Information from the viscera arrives directly in the CNS via the vagus and glossopharyngeal nerves bundled within the solitary tract (ST). These cranial primary afferents synapse onto neurons within the nucleus of the solitary tract (NTS) and neuroanatomical evidence suggests a loose but overlapping viscerotopic distribution of these cranial afferents within the NTS (Loewy, 1990;Altschuler et al., 1991;Kubin et al., 2006). Single cranial visceral afferent axons branch within NTS with varicosities suggesting terminals in the vicinity of multiple neurons (Donoghue et al., 1982a;Donoghue et al., 1982b;Davies and Kubin, 1986;Kubin et al., 2006). This proximity of multiple afferent branches and terminals also suggests possible convergence of afferents onto single NTS neurons. Activation of different peripheral nerve trunks (e.g. vagus and carotid sinus nerve) evoke responses in vivo consistent with individual NTS neurons receiving convergent inputs from multiple afferent sources (Bonham and Hasser, 1993;Mifflin, 1993;Mifflin, 1996). Thus, visceral information may converge onto single NTS neurons or may be shared across multiple neurons. Understanding the patterns of afferent organization and synaptic distribution to individual NTS neurons is a fundamental aspect that will shape reflex characteristics and integration within autonomic networks. Key questions include the number and routing of afferent inputs, their synaptic weighting and the degree of convergence/divergence of ST-driven excitatory neurotransmission.
Assessments of afferent driven inputs to NTS neurons have generally relied upon single cell recordings. This approach cannot detect common or shared inputs across neurons. To better assess ST synaptic organization, we recorded from pairs of NTS neurons in horizontal brainstem slices in which afferent ST axons could be activated up to 3 mm from the recorded cells. We finely graded the intensity of remote ST stimulus shocks to progressively recruit synaptic inputs and then characterized each by their timing, amplitude and synaptic failures. Stimulus recruitment profiles revealed the number, weight and path of synaptic connections from ST axons to NTS neurons. Across pairs of neurons, stimulus recruitment profiles were also used to assess possible afferent divergence. The results depict medial NTS as overwhelming comprised of 2nd order neurons, of which, nearly half receive only a single monosynaptic ST input and the remainder received limited numbers of additional monosynaptic and/or polysynaptic convergent inputs. Paired recordings detected only two instances of ST-driven input divergence. Thus, 2nd order NTS neurons are highly segregated from their nearest neighbors within this compact region and likely dedicated to processing and distributing a highly focused message of peripheral organ status.
All animal procedures were performed with the approval of the Institutional Animal Care and Use Committee at Oregon Health & Science University and conform to the guidelines of the National Institutes of Health publication “Guide for the Care and Use of Laboratory Animals”.
Brainstem slices were prepared from adult (>180 g, average weight 345 ± 24 g, n = 20) Sprague Dawley rats (Charles River Laboratories, Inc., Wilmington, MA) as described previously (Doyle and Andresen, 2001). Briefly, rats were deeply anesthetized with isoflurane. The medulla was rapidly cooled and trimmed rostral and caudal to yield a brainstem block centered on obex. A rostral-caudal wedge of ventral brainstem was removed in order to orient the remaining tissue to yield a single, 250 µm thick, horizontal slice which contained the greatest length of ST axons together with the medial NTS. Slices were cut with a sapphire knife (Delaware Diamond Knives, Wilmington, DE) mounted in a vibrating microtome (VT1000S; Leica Microsystems Inc., Bannockburn, IL). The external solution was an artificial cerebrospinal fluid (ACSF) containing (mM): 125 NaCl, 3 KCl, 1.2 KH2PO4, 1.2 MgSO4, 25 NaHCO3, 10 dextrose and 2 CaCl2. Slices were secured with a nylon mesh in a perfusion chamber and perfused with ACSF at 34°C, 300 mOsm, bubbled with 95% O2-5% CO2.
Recording pipettes (2.8 – 4.2 MΩ) were guided to neurons in the medial sub-nucleus of caudal NTS using anatomical landmarks. Neurons were visualized (Doyle et al., 2004) using infrared illumination with differential interference contrast optics (40× water immersion lens) on an Axioskop-2 FS plus fixed stage microscope (Zeiss, Thornwood, NJ) with digital camera (Hamamatsu Photonic Systems, Bridgewater, NJ). Pipettes were filled with a low Cl− intracellular solution containing (mM): 6 NaCl, 4 NaOH, 130 K-gluconate, 11 EGTA, 1 CaCl2, 1 MgCl2, 10 HEPES, 2 Na2ATP, and 0.2 Na2GTP (pH 7.3 and 296 mOsm). As a consequence ECl = −69mV, IPSCs had small amplitudes at VH = −60mV. All recordings were made in open, whole cell patch configuration under voltage clamp using a Multiclamp 700B (Molecular Devices, Sunnyvale, CA). Signals were filtered at 10 kHz and sampled at 30 kHz using p-Clamp software (version 9.2, Axon Instruments, MDS Analytical Technologies, Sunnyvale, CA). Voltages were not corrected for the liquid junction potential (−13.9 mV).
The concentric bipolar stimulating electrode (200 µm outer diameter; Frederick Haer Co., Bowdoinham, ME) was placed on the visible ST approximately 1–3 mm rostral from recorded neuron cell bodies (Figure 1). Passing current via the stimulating electrode activated ST primary afferents. The remote placement of the electrode minimized the likelihood that electrical shocks would activate non-ST axons or local neurons (Doyle and Andresen, 2001;Doyle et al., 2004;Bailey et al., 2008). Once stable recordings of each neuron were established a series of test shocks graded in intensity were delivered to the ST. Minimal stimulus-intensity protocols with incremental recruitment can define discrete inputs to central neurons (Rancz et al., 2007;Allen and Stevens, 1994). Likewise, afferent shocks distant to the recorded cells produce discrete, unitary recruitment-intensity profiles for each input (Blitz and Regehr, 2005;Acuna-Goycolea et al., 2008). A Master-8 isolated programmable stimulator (A.M.P.I., Jerusalem, Israel) generated bursts of five ST shocks at 50 Hz every 3 s (shock duration 0.1 ms) for a minimum of 10 consecutive sweeps for each shock intensity trial and respective synaptic responses were designated EPSC1, EPSC2, etc.. All neurons were tested with stimulus shock intensities from 0 to 800 µA. From 0 to 100 µA, step changes were 10 µA (minimum) and from 100 to 800 µA, step changes were 100 µA (minimum). The order of shock intensity presentation was varied. Once synaptic responses were detected, then a more finely graded series of shocks were delivered to better define the threshold stimulus intensity value for each synaptic input. Plots of the event amplitudes (discriminated by arrival time and waveform) against stimulus intensity created stimulus recruitment profiles to compare across events. The stimulus recruitment profile of each synaptic event is considered to reflect the all-or-none axon activation characteristics of the ST fiber initiating the recorded EPSC.
Synaptic latency jitter and failure rates were considered reliable indices that distinguish between direct, contacts from ST afferents (monosynaptic) and indirect (polysynaptic) afferent pathways to NTS neurons. Analysis of latency was based on consideration of the responses to the first shock in the train of five (EPSC1). Variability in latency (jitter) for EPSC1 was calculated over 30–40 trials as the standard deviation (SD) of latency and served as a critical index of synaptic order. EPSC1 jitter of <200 µs were considered monosynaptic and jitters >200 µs indicated a polysynaptic connection from the ST (Doyle and Andresen, 2001;Bailey et al., 2006a). This criterion (synaptic jitter) matches second order NTS neurons identified using anterogradely dye-labeled baroreceptive terminals (Andresen and Peters, 2008). Polysynaptic pathways activated by ST shocks are particularly prone to synaptic failures even with infrequent stimuli (Andresen and Peters, 2008;Bailey et al., 2006a). Suprathreshold ST shocks that failed to evoke an identifiable EPSC were counted as synaptic failures and failure rates were calculated for EPSC1 over 40 trials as a percent of total ST shocks delivered at a constant, supra-threshold intensity. Neurons that received only polysynaptic inputs to ST shocks were classified as higher order NTS neurons. Neurons that exhibited no ST-evoked synaptic responses to high intensity shocks (up to 800 µA) were considered “not connected” to ST afferents and not analyzed further.
Recorded neuron pairs were located in close proximity (0–150 µm) and one important aspect of our analysis was to test for the possibility that one ST afferent might contact two neurons. To detect common innervation, analysis began with the identification of synaptic inputs triggered at similar ST stimulus thresholds. We predicted that synaptic events arising from a common ST axon would be evoked in both neurons at identical shock intensities. Two forms of divergence might be predicted: a single ST afferent dividing to contact two neurons or a common excitatory interneuron whose axon divides to contact the two recorded NTS neurons. For all events triggered at similar thresholds, we compared the thresholds, timing and failures of ST-evoked synaptic events. Similarly, we reasoned that failure of a given shock to trigger an action potential in a common afferent fiber would fail to evoke respective synaptic events in both recorded neurons. Thus, evidence of divergence of a single ST afferent contacting two neurons included both common thresholds and synchronous use dependent synaptic failures. Where failure rates normally increase from EPSC1 through to EPSC5, within each burst of shocks (Bailey et al., 2006a;Andresen and Peters, 2008). Coincidence of synaptic failures in both recorded neurons was calculated as a co-failure rate over 40 trials of the five shock burst (200 total test shocks).
Synaptic inputs activated by the same ST afferent axons were expected to have highly correlated or “synchronous” synaptic successes and failures between pairs of neurons. Comparisons between suspected common inputs were tested against a random association using a chi squared analysis. Input variables including latency, jitter, amplitude and failure rates for EPSC1 were not normally distributed and thus were compared between groups by rank values using a Kruskal-Wallis one way ANOVA on ranks with a Dunn’s post hoc test (SigmaStat, San Jose, CA). All data are represented as mean ± SEM.
Pairs of neurons (n = 48 neurons; 24 pairs) located in the medial NTS (Figure 1) were targeted for recordings and characterized based on their ST-driven EPSCs. For individual pairs of recordings, neuronal cell bodies were separated by up to ~150 µm (average distance between recording electrodes was 51 ± 7 µm). Most neurons received at least one low jitter ST afferent contact and were designated as 2nd order (36 of 48 cells). Neurons receiving only high jitter, polysynaptic input from the ST were designated ”higher order” and represented a relatively small subgroup of NTS neurons (6 of 48 cells). Additionally, six neurons did not respond to even the highest intensity ST shocks, and were considered "not connected" (Figure 1D). These unconnected neurons appeared otherwise healthy and exhibited spontaneous synaptic events. In all cases, unconnected neurons were recorded with a paired neuron that responded to ST shocks. The number of unconnected neurons was similar in proportion to previous studies and may, in part, represent neurons whose ST connections were damaged during slice preparation (Appleyard et al., 2007). For the remainder of this report, summary expressions will consider only those 42 neurons with ST-synced responses.
The exposed tip of the concentric bipolar stimulating electrode has a 200 µm diameter that covered most of the visible width of the ST in horizontal slices (Figure 1A). All ST-evoked synaptic responses were triggered by shocks from a single ST location in each slice. The lowest shock intensities evoked no synaptic response (Figure 2A, B; 30–50 µA). As current intensity was increased, each shock reliably activated a fixed latency EPSC in only one of these two neurons (Figure 2A, B; 60 µA shocks, black traces, neuron “a”, input 1, neuron “b”, red traces, no inputs activated). Analysis of every event in each neuron generated a stimulus recruitment profile for the individual cell and represents a summary of their full range of synaptic responses. In this pair (Figure 2), neuron “a” received a single monosynaptic input (left, black) and neuron “b” received a combination of two separate but converging monosynaptic inputs (Figure 2; 90 µA shocks, neuron “b”, inputs 1 & 2). Note that the responses of all three inputs to the two neurons were evoked without failures by each single shock to ST at suprathreshold intensities. Each input was identified with a unique stimulus intensity threshold, amplitude, latency distribution and all-or-none response (Figure 2B). The activation of the second convergent input to neuron “b” produced a compound EPSC that obscured the presence of input 1 on account of its shorter latency (Figure 2A, B). Interestingly, the EPSC waveform in neuron “b” at high intensity (90 µA) contained no discernible inflection that might indicate two discrete ST inputs. Thus, the two inputs to this neuron were only revealed through the graded shock recruitment protocol. No other ST-evoked event was identified at shock intensities >90 µA. These results indicate that these two neurons together received three ST-driven inputs and all were monosynaptic. Thus, neuron “a” and “b” (Figure 2) were classified as 2nd order neurons.
Each of the three ST-EPSCs elicited in the two recorded neurons (neurons “a” & “b”) was activated at discrete shock intensity thresholds. The trial-to-trial variations characteristically were much greater in EPSC amplitudes than in event timing (see Supplemental Figure 1, dynamic GIF). Low-jitter ST-EPSCs were remarkably reliable and did not fail across trials at EPSC1. Together, these measures indicate activation of three independent ST axons whose responses are distinguished in their timing and amplitude characteristics (Figure 2C). Once recruited, the EPSC amplitudes were unaltered by increasing the shock intensity (Figure 2B) and the latency and jitter of the individual ST-evoked EPSCs were constant. This result conforms to the expected outcomes from the discrete recruitment of single axons directly connected to each neuron (Figure 2C).
Shocks to ST often evoked synaptic events with inconsistent latencies and frequent failures. These characteristics are consistent with ST activation of polysynaptic pathways that were prone to interruptions in transmission along intervening neurons before reaching the recorded cells (Bailey et al., 2006a). In 22 (20 2nd, 2 higher order) of 42 medial NTS neurons, stimulus recruitment profiles revealed combinations of low- and/or high-jitter ST-EPSCs and typically each event had a different threshold (Figure 3). Increasing shock intensity independently recruited a total of six different ST-synced EPSCs in the two neurons (Figure 3A). The 2nd order neuron “a” (Figure 3A; black traces) received two high-jitter inputs (polysynaptic) with relatively low and high thresholds each and a low-jitter EPSC recruited at middle intensity thresholds (Figure 3B). Thus, neuron “a” of this pair received a primary afferent input and two additional convergent, polysynaptic EPSCs. Neuron “b” received three unique ST-EPSCs, but all were low jitter (Figure 3A & B). Neuron “b” possessed the largest number of monosynaptic inputs to a single neuron found across 24 pairs among 81 unique, ST-synced synaptic events. In the case of neuron “b” (Figure 3A; red traces), the closely matched latencies of inputs 1 and 2 fused into a compound EPSC, but input 3 arrived later and produced a distinct inflection rising from the late initial plateau of the three-event compound waveform (50–100 µA, Figure 3A). This example demonstrates that threshold stimulus intensity was not predictive of pathway complexity or latency (2nd vs. higher order). As with monosynaptic EPSCs, ST-driven polysynaptic inputs had singular distinct threshold shock intensities that were all-or-none and triggered consistent waveforms (i.e. amplitude and decay kinetics) across trials. Polysynaptic events varied in both their timing and waveform shape indicative of the more convoluted processing in the individual pathways. Polysynaptic EPSCs also summed in compound waveforms and their influence characteristically varied in time and magnitude. At the extreme, failures in individual polysynaptic pathways resulted in component EPSCs dropping out of the compound waveform, an effect most evident in varied failures from trial to trial (see Supplemental Figure 2, dynamic GIF). These results within and between cells are consistent with a synchronous activation of multiple ST axons during a single shock to the ST and an organization model in which only a limited number of these afferent pathways directly and indirectly synapse on each neuron (Figure 3C).
A relatively small group of NTS neurons recorded received only indirect ST-driven information (6 of 48 neurons). As with the high-jitter, polysynaptic ST inputs to 2nd order NTS neurons, ST-activated EPSCs to higher order neurons had high jitter and were failure-prone (Figure 4; neuron “a”). It is important to note that under our recording conditions, ST-evoked GABAergic inputs were not apparent, although recording at depolarized potentials revealed that many of these NTS neurons also received ST-activated polysynaptic IPSCs (data not shown) similar to previous reports (Andresen and Yang, 1995;Bailey et al., 2008;Doyle and Andresen, 2001). Such ST-driven IPSCs represent an additional source of polysynaptic convergence to these NTS neurons which we did not quantify in the present study.
High intensities of ST shocks activated multiple ST afferent axons. Differences in stimulus intensity thresholds for each event suggested that individual inputs were triggered independently through afferent ST axons with distinct thresholds. Every ST-driven input exhibited an all-or-none stimulus-response profile regardless of pathway; mono- or polysynaptic. Anatomical evidence suggests that single ST afferent axons can branch to distribute terminals within the vicinity of different neurons in NTS, e.g. (Kubin et al., 2006). We reasoned that, for our pairs of recorded neurons, the proximity of the cell bodies and their presumed dendritic distributions might increase the likelihood of receiving branches from a common ST afferent axon. Inputs triggering synaptic events that arise from a common axon would be expected to have comparable threshold intensities. Therefore, to detect common inputs, we inspected and compared EPSC thresholds across all potentially shared input combinations (47 potentially shared inputs in 13 neuron pairs). Seven synaptic event pairs in 6 pairs of neurons had sufficiently similar thresholds (within 10% of full intensity range) to be suspected as triggered by a common ST axon. Of these 14 suspect events, four EPSC pairs were polysynaptic and three EPSC pairs were monosynaptic. After full analysis, only 2 neuron pairs exhibited sufficient evidence to support common sources of initiation – one pair with linked polysynaptic EPSCs and one pair with linked monosynaptic EPSCs.
In one case of suspected common sourced EPSCs, paired neurons “a” and “b” were adjacent to one another and ST-evoked EPSCs were triggered in both neurons at the same threshold intensity (Figure 4A, B; input 2 in each neuron, blue triangles). Both suspected common EPSCs had similar latency and jitter (>200 µs) that were consistent with activation along a polysynaptic pathway (Figure 4B). Each of these neurons received additional ST-synced EPSCs which appeared independent with unique thresholds and jitter (Figure 4A–C; inputs 1 & 3). Examination of the records showed that the suspected common inputs failed frequently (Figure 4A; yellow circles, 50 & 60 µA traces) and raised the possibility that a single neuron diverged to connect to both neurons (Figure 4C). More focused analysis showed that input 2 failed in unison in both neurons repeatedly over multiple trials at a 100% co-failure rate (Figure 5A, B; and Supplemental Figure 2, an animated GIF). Interestingly, the amplitudes of successful transmission of input 2 EPSCs were uncorrelated across neurons suggesting that the fluctuation in release probability was independent across neurons (Figure 5B). However, the timing of input 2 EPSCs closely tracked changes in latency between the two EPSCs and produced a positive, linear correlation (p < 0.05, linear regression) from one neuron to the other (Figure 5C). The cross-neuron synchrony of synaptic failures and high quantitative correlation of EPSC latencies is most likely accounted for by activation of a single ST axon (identical threshold) that excited a single pathway to a common intervening interneuron (Figure 5; blue model pathway). Examination of the remaining three pairs of suspected common polysynaptic EPSCs exhibited low co-failure rates (31, 23 and 0%) and uncorrelated EPSC latencies (p > 0.05, linear regression), suggesting these EPSCs were not initiated by common source interneurons.
One pair of neurons received ST-EPSCs that co-varied to a degree consistent with reliance on a single common ST axon, i.e. a collateralizing branch. The suspected pair of neurons had cell bodies that were ~40 µm apart and both received a single, monosynaptic ST input (Figure 6). Both monosynaptic ST-EPSCs were recruited at the same threshold intensity with no failures. To facilitate synaptic failures, we subjected the ST-NTS responses to repeated bursts of 5 shocks at 50 Hz, a paradigm which substantially raises failure rates in most NTS neurons (Andresen and Peters, 2008). Analysis of repeated shocks showed that use-dependent synaptic failures that occurred in near unison in both neurons – co-failure rate of 87% for EPSCs2–5 (Figure 6C). The highly correlated incidences of both successes and failures between neuron “a” and “b” were not random (Chi square = 89.175, 1 df, p < 0.001). Thus, the synaptic response of one neuron was an excellent predictor of the EPSC in the second neuron. In contrast in two additional suspected neurons, co-failure rates for EPSC2–5 were 31 (Chi-square = 6.314, 1 df, p = 0.012) and 35% (Figure 3C; input 2 to each neuron; Chi-square = 3.043, 1 df, p = 0.081) and thus consistent with unrelated sources. Failures in common (Figure 6C) suggest a common site in which action potential conduction is interrupted along a single ST axon and the simplest configuration would be failure distal to a single ST axon branching to contact each neuron (Figure 6D). Thus given the low incidence of neurons demonstrating common inputs for neurons in close proximity, we conclude sharing information across neurons is relatively infrequent, compared to convergence, in NTS.
Our paired recordings from medial NTS suggest several patterns ST afferent organization with medial NTS neurons. Approximately three-quarters of the neurons in medial NTS (Figure 7A; 36 of 48 neurons) were contacted directly by ST afferents via low jitter, non-failing EPSCs and were classified as 2nd order NTS neurons. Of the remaining neurons studied, 13% (6 of 48) received only high-jitter, polysynaptic EPSCs and were classified as higher order NTS neurons. The remaining neurons (6 of 48) exhibited no ST-synced synaptic events.
In general, most 2nd order NTS neurons received a single monosynaptic contact from the ST (Figure 7A; 64% of 2nd order neurons, 23 of 36). Most of these 2nd order neurons received no polysynaptic EPSCs. However, in the remaining 2nd order neurons, additional convergent EPSCs were equally likely to be direct or indirect. Many 2nd order neurons received only a single ST-EPSC with no other synaptic inputs linked to ST (Figure 7A; 44% of 2nd order neurons, 16 of 35). Whether direct or indirect, the total number of inputs converging on a single 2nd order neuron rarely exceeded three (Figure 7A). Even in neurons classified as higher order cells; ST-synced inputs did not exceed three (data not shown).
A total of 81 excitatory inputs were characterized in 42 neurons (47 mono-, 34 poly-synaptic). Recruitment thresholds ranged from 10 to 500 µA and 80% of inputs were first activated at shock intensities below 200 µA. Despite the relatively short potential path length, the measured range of ST-EPSC latencies was quite large and extended from 2 to 15.1 ms. The mean latency for monosynaptic EPSCs was shorter than polysynaptic EPSCs (5.74 ± 0.22 and 7.11 ± 0.46 ms respectively, p<0.05 one-way ANOVA). Amplitudes of monosynaptic ST-EPSCs averaged more than twice that of polysynaptic ST-EPSCs to both 2nd and higher order neurons (Figure 7B, Kruskal-Wallis one way ANOVA on ranks, n = 62, with 19 amplitudes that could not be assessed due to compound EPSCs). Within neurons, monosynaptic ST-EPSCs were almost always larger than polysynaptic ST-EPSCs projecting to the same 2nd order neuron. In most neurons receiving both poly- and mono-synaptic inputs (10 of 12), polysynaptic EPSC had an average amplitude only 55 ± 22% of that of the largest monosynaptic EPSC to that neuron. In two neurons, polysynaptic EPSCs had larger amplitudes that the monosynaptic EPSC. The amplitudes of polysynaptic ST-EPSCs clustered below 200 pA (Figure 7B; inset). Monosynaptic EPSC1s rarely failed whereas polysynaptic ST-EPSCs failed frequently, typically >5%. Although the general associations between high failures and high jitter were strong, exceptions were encountered occasionally. A few polysynaptic inputs that exhibited high jitter (>200 µs) did not fail at all on EPSC1 (Figure 7B). In 4 cases, failure rates exceeded 5% despite the presence of low jitter EPSCs that met the monosynaptic criterion (Figure 7B). Thus across all neurons, medial NTS is overwhelmingly composed of 2nd order neurons directly connected to limited numbers of ST afferents with substantial numbers of convergent EPSCs arriving via generally weaker polysynaptic EPSCs.
As cranial visceral afferents enter the brain, the precise patterns of how those ST afferent axons contact NTS neurons determines the first stage of information processing. Analysis of stimulus recruitment profiles and the timing and waveforms of ST-evoked EPSCs within the medial NTS suggest three main points about the organization of ST-driven neurotransmission: 1. More than half of 2nd order neurons received convergent ST-driven information, albeit limited in number and as equally likely to be polysynaptic as monosynaptic. 2. In considering direct ST afferent contacts only, two thirds of 2nd order neurons received a single monosynaptic input. 3. Divergence of ST-driven signaling between pairs of neurons was infrequent. These synaptic patterns describe an ST-NTS network that is often driven by a single strong monosynaptic inputs and consequently by a single afferent type. A second group of 2nd order neurons, however, received multiple synaptic contacts that may mix afferent types. Thus, processing of cranial visceral afferent signaling in medial NTS relies on a highly parallel organization of afferent processing and this relatively sparse afferent information flow defines dedicated neurons with limited cross-afferent convergence, divergence and mixing.
Our analysis of ST-activated EPSCs began with discrimination of monosynaptic from polysynaptic pathways. This discrimination relied on quantitative measures of latency variability of ST-EPSCs and the principle that only direct coupling through a single synapse met the <200 µs jitter criterion. 2nd order NTS neurons identified anatomically by fluorescently labeled primary afferent terminals never exceeded 200 µs jitter (Andresen and Peters, 2008). Similarly, dye identified 2nd order neurons contacted by carotid chemoreceptor or vagal afferents conformed to this jitter criterion (Kline et al., 2002;Zhang et al., 2009). The 200 µs jitter criterion successfully identified polysynaptic IPSCs initiated by ST shocks in NTS (Bailey et al., 2008;Doyle and Andresen, 2001;Smith et al., 1998). A second but less discriminating criterion often uses transmission failures to functionally separate direct from polysynaptic pathways. Neurons which are activated by pairs of closely timed afferent shocks (e.g. 5 ms) are considered monosynaptically innervated in extracellular and intracellular CNS recordings (Donoghue et al., 1985;Miles, 1986;Scheuer et al., 1996). In our recordings, failure rates of single-shock ST-EPSC classified as monosynaptic using jitter were nearly always 0%. In a few cases, however, higher failure rates were encountered despite <200 µs jitter and were classified as monosynaptic. Failure rates as an index of synaptic order, however, can be misleading. Some ST-evoked IPSCs in NTS neurons reliably follow 200 Hz ST trains (Doyle and Andresen, 2001). Conversely in dye identified 2nd order neurons, ST-EPSC failure rates of 30–40% were possible to bursts of ST shocks (Doyle and Andresen, 2001). Failures of ST-EPSCs in some neurons may also be modified by neuropeptides through a presynaptic mechanism rather than pathway failures between neurons (Bailey et al., 2006b). Together, this independent evidence suggests that, while elevated failure rates are correlated with higher pathway complexity, failures are not as reliable an indicator of synaptic order as latency jitter.
A major goal of this work was to determine whether there were patterns in the organizational relationship between ST afferents and NTS neurons. We randomly sampled medial NTS neurons two at a time and determined their connections using ST initiated synaptic responses. The results suggest that neurons in the medial sub-nucleus are overwhelmingly 2nd order to visceral sensory neurons. Together such findings were consistent with a basic organizational module of a single ST axon leading directly to a single 2nd order neuron. Variations on that basic ST module included neurons that received 1–2 additional monosynaptic, polysynaptic or mixed inputs. In each case, monosynaptic ST-EPSCs had unitary, all-or-none intensity-recruitment profiles. The intensity-recruitment profiles for polysynaptic EPSCs were also all-or-none; indicating a single ST afferent activated the polysynaptic path. Electrical shocks to ST likely simultaneously activated large proportions of the thousands of ST axons. These basic ST modules were surprisingly isolated from this potentially enormous pool of local afferents. Polysynaptic inputs tended to be weaker and failed more frequently compared to monosynaptic inputs. Thus, polysynaptic inputs should have less influence on neuron activity than primary afferent inputs. Physiologically this configuration facilitates more faithful transfer of highly focused information and may be the exclusive organization for particular afferents such as aortic baroreceptors (Andresen and Peters, 2008). Our electrophysiological approach did not define the modality or organ source of ST afferent terminals in this study. Convergent inputs may represent a single modality (e.g. baroreceptor) or represent mixed sources (e.g. cardiorespiratory) from different organs and/or sensory modalities (e.g. chemoreceptive or mechanosensitive). Likewise, some particular organizational constructs such as multiple convergent monosynaptic inputs may be favored by particular NTS phenotypes such as catecholaminergic neurons (Appleyard et al., 2007).
The ST-NTS pattern of excitatory convergence contrasted with cortical examples in which robust convergence is common. In cortex, single inputs fail often, 22% for white matter tract inputs to cerebellar cortex (Kanichay and Silver, 2008) or 50% mossy fiber failures to CA3 hippocampal neurons (Lawrence et al., 2004). Although unreliable cortical release can have near quantal amplitudes, recruitment of multiple convergent inputs achieves reliable transmission (Allen and Stevens, 1994). At ST-NTS synapses, the converse prevails: single or few, unfailing excitatory inputs with large unitary amplitudes. Our studies indicate that ST afferent convergence is limited, but these basic ST connections depend on a reliable process known to have intrinsically high release probability for glutamate resulting in a very high safety factor (Bailey et al., 2006b;Andresen and Peters, 2008;Peters et al., 2008).
Some classes of ST afferent axons branch to send collaterals across NTS (Kalia and Richter, 1985) and retrograde activation profiling suggests similar branching of single aortic baroreceptors in medial NTS (Donoghue et al., 1982a). In our slices, we found two pairs of neurons with ST-EPSCs with equivalent shock thresholds that likely shared common inputs. Evidence in our studies suggested one case that likely arose from a common interneuron source - tightly correlated variations in latency and synchronous failures. In the other case, likely a single collateralizing ST axon produced low-jitter EPSCs with synchronous use-dependent failures in two neurons. We found no evidence to support single afferents driving both mono- and polysynaptic inputs. Thus, afferent divergence seems rare between closely located medial NTS neurons. Patterns or organization may differ in other sub-regions and our methods did not allow us to assess the cross sub-region divergence suggested by anatomical afferent tracing (Kubin et al., 2006). Our assay measures transmission from successfully depolarized terminals and, under other conditions, successful transmission, the number of contacts and the potential for branch point failures might change. See (Bailey et al., 2006b;Peters et al., 2008).
Potential caveats of our approach include that it likely underestimates the number of ST-synced EPSCs across neurons so that clearly the precise proportions of mono- and polysynaptic connections may also be affected by our methods of approach. First, some ST axons or receptive dendrites may be severed and this will result in underestimation of the number and amplitude of ST pathway connections. Such damage might be indicated by the 10–20% of NTS neurons that fail to respond to ST activation across slice studies (Appleyard et al., 2007;Bailey et al., 2008). Axon sectioning might also contribute to differences in event amplitudes if branches are cut. Second, detection of inputs may be limited in certain circumstances. For example, activation of compound or large events likely obscures detection of lower amplitude EPSCs recruited at higher shock intensities and lead to underestimation of ST-synced inputs (Hamann et al., 2003). Lastly, our study design deliberately minimized observation of IPSCs by recording near the chloride reversal potential. ST-evoked IPSCs represent an important uncounted population of convergent ST-activated polysynaptic pathways (Smith et al., 1998;Derbenev et al., 2006;Bailey et al., 2008).
Despite these caveats, the in vitro patterns of convergence resemble the proportions of convergent afferent inputs detected in intact animals using physiological or electrical activation of different afferent groups. Maximal shocks to multiple whole peripheral nerve trunks commonly trigger spikes within NTS identified as single afferent inputs in roughly 85% of cases (Donoghue et al., 1985;Bonham and Hasser, 1993;Ootani et al., 1995;Mifflin, 1996). Convergence among cardiovascular afferents - arterial baroreceptors with cardiac mechanoreceptors – is similarly rare (<13%) in NTS neurons using pharmacological or physiological activation (Paton, 1998;Seagard et al., 1999). In vivo studies in which synaptic order was evaluated suggest that input from one nerve trunk met monosynaptic criteria while all convergent inputs from additional nerves were polysynaptic (Mifflin, 1996). Such in vivo results suggest that most NTS neurons conform to the most common basic ST module described by our slice results. Obtaining intracellular recordings in vivo from these small NTS neurons is difficult and our present results establish several approaches in slices with sufficient resolution and control to discern timing, synaptic order and cross-neuron interactions that could not be obtained in an intact CNS.
ST synaptic contacts define the position most neurons within the NTS network occupy. The organization of information flow from primary cranial visceral afferents into the brain, at its most basic, consists of a single afferent exciting a single NTS neuron. Additions of mono- and/or polysynaptic inputs produce different configurations and subsets of organization that may be associated with particular projection pathways or efferent systems. This relatively compartmentalized organization may reflect the dedication of specific afferent modalities to particular pathways and may underwrite the network architecture for separation of discrete organ homeostatic control.
Supported by grants from the National Institutes of Health, HL-41119 (MCA) and Fellowship HL-88894 (JHP), National Health and Medical Research Council of Australia, Overseas Training C J Martin Fellowship #400405 (SJM).