Postnatal development of GABA-immunoreactive neurons and terminals in rat periaqueductal gray matter: A light and electron microscopic study

Deep learning-based synapse counting and synaptic ultrastructure analysis of electron microscopy images

BACKGROUND

Synapses are the connections between neurons in the central nervous system (CNS) or between neurons and other excitable cells in the peripheral nervous system (PNS), where electrical or chemical signals rapidly travel through one cell to another with high spatial precision. Synaptic analysis, based on synapse numbers and fine morphology, is the basis for understanding neurological functions and diseases. Manual analysis of synaptic structures in electron microscopy (EM) images is often limited by low efficiency and subjective bias.

NEW METHOD

We developed a multifunctional synaptic analysis system based on several advanced deep learning (DL) models. The system achieved synapse counting in low-magnification EM images and synaptic ultrastructure analysis in high-magnification EM images.

RESULTS

The synapse counting system based on ResNet18 and a Faster R-CNN model had a mean average precision (mAP) of 92.55%. For synaptic ultrastructure analysis, the Faster R-CNN model based on ResNet50 achieved a mAP of 91.60%, the DeepLab v3 + model based on ResNet50 enabled high performance in presynaptic and postsynaptic membrane segmentation with a global accuracy of 0.9811, and the Faster R-CNN model based on ResNet18 achieved a mAP of 91.41% for synaptic vesicle detection.

CONCLUSIONS

The proposed multifunctional synaptic analysis system may help to overcome the experimental bias inherent in manual analysis, thereby facilitating EM image-based synaptic function studies.

Electrophysiological Characteristics of Dorsal Raphe Nucleus in Tail Suspension Test

The dorsal raphe nucleus (DRN) is a major source of serotonin in the central nervous system, which is closely related to depression-like behaviors and is modulated by local GABAergic interneurons. Although serotonin neurons are known to be activated by struggling behavior in tail suspension test (TST), the exact electrophysiological characteristics are still unclear. Here, we combined in vivo electrode recording and behavioral test to explore the mice neuron electrophysiology in DRN during TST and observed that gamma oscillation was related to despair-like behaviors whereas burst fraction was crucial for survival-like behaviors. We reported the identification of a subpopulation of DRN neurons which change their firing rates when mice get into and during TST immobile states. Both increase (putative despair units, D units for short) and decrease (putative survival units, S units for short) in firing rate were observed. Furthermore, using optogenetics to identify parvalbumin-positive (PV+) and serotonin transporter-positive (SERT+) neurons, we found that SERT+ neurons were almost S units. Interestingly, those that have been identified PV+ neurons include ~20% of D units and ~50% of S units. These results suggest that electrophysiological characteristics incorporated in despair-like behavior studies can provide new insight into the study of anti-depression targets, and GABAergic interneuron is a complex key hub to the coding and regulation of local neural network.

Early adolescent adversity alters periaqueductal gray/dorsal raphe threat responding in adult female rats

Early adolescent adversity increases adult risk for anxiety disorders. The ventrolateral periaqueductal gray (vlPAG) and neighboring dorsal raphe (DR) are integral to threat prediction, and are responsive to acute stressors. Here, we tested the hypothesis that early adolescent adversity reshapes vlPAG/DR threat-related cue activity and threat probability signaling. Female, Long Evans rats experienced a battery of adverse adolescent experiences (n = 12), while controls did not (n = 8). Single-unit activity was recorded 50 + days following the final adverse experience, when the adult rats received fear discrimination consisting of danger, uncertainty and safety cues. Despite achieving fear discrimination that was equivalent to controls, vlPAG/DR threat responding was altered in adverse-experienced rats. Early adolescent adversity resulted in a greater proportion of cue-responsive neurons. Cue-excited neurons showed greater increases in firing and cue-inhibited neurons showed greater decreases. Even more, early adversity reduced flexible, threat probability signaling by cue-excited neurons and promoted more rigid, fear output signaling by cue-inhibited neurons. The results reveal long-lasting changes in vlPAG/DR threat responding resulting from early adolescent adversity.

GABAergic modulation of serotonergic neurons in the dorsal raphe nucleus

The dorsal raphe nucleus (DRN), located in the brainstem, is involved in several functions such as sleep, temperature regulation, stress responses, and anxiety behaviors. This nucleus contains the largest population of serotonin expressing neurons in the brain. Serotonergic DRN neurons receive tonic γ-aminobutyric acid (GABA)inhibitory inputs from several brain areas, as well as from interneurons within the same nucleus. Serotonergic and GABAergic neurons in the DRN can be distinguished by their size, location, pharmacological responses, and electrophysiological properties. GABAergic neurons regulate the excitability of DRN serotonergic neurons and the serotonin release in different brain areas. Also, it has been shown that GABAergic neurons can synchronize the activity of serotonergic neurons across functions such as sleep or alertness. Moreover, dysregulation of GABA signaling in the DRN has been linked to psychiatric disorders such as anxiety and depression. This review focuses on GABAergic transmission in the DRN. The interaction between GABAergic and serotonergic neurons is discussed considering some physiological implications. Also, the main electrophysiological and morphological characteristics of serotonergic and GABAergic neurons are described.

Central circuitry responsible for the divergent sympathetic responses to tonic muscle pain in humans

Experimentally induced tonic muscle pain evokes divergent muscle vasoconstrictor responses, with some individuals exhibiting a sustained increase in muscle sympathetic nerve activity (MSNA), and others a sustained decrease. These patterns cannot be predicted from an individual's baseline physiological or psychological measures. The aim of this study was to investigate whether the different muscle sympathetic responses to tonic muscle pain were associated with differential changes in regional brain activity. Functional magnetic resonance imaging (fMRI) of the brain was performed concurrently with microelectrode recording of MSNA from the peroneal nerve during a 40-min infusion of hypertonic saline into the ipsilateral tibialis anterior muscle. MSNA increased in 26 and decreased in 11 of 37 subjects during tonic muscle pain. Within the prefrontal and cingulate cortices, precuneus, nucleus accumbens, caudate nucleus, and dorsomedial hypothalamus, blood oxygen level dependent (BOLD) signal intensity increased in the increasing-MSNA group and remained at baseline or decreased in the decreasing-MSNA group. Similar responses occurred in the dorsolateral pons and in the region of the rostral ventrolateral medulla. By contrast, within the region of the dorsolateral periaqueductal gray (dlPAG) signal intensity initially increased in both groups but returned to baseline levels only in the increasing-MSNA group. These results suggest that the divergent sympathetic responses to muscle pain result from activation of a neural pathway that includes the dlPAG, an area thought to be responsible for the behavioral and cardiovascular responses to psychological rather than physical stressors. Hum Brain Mapp 38:869-881, 2017. © 2016 Wiley Periodicals, Inc.

Connections from the rat dorsal column nuclei (DCN) to the periaqueductal gray matter (PAG)

Electrical stimulation of the dorsal columns (DCs; spinal cord stimulation; SCS) has been proposed to treat chronic neuropathic pain. SCS may activate a dual mechanism that would affect both the spinal cord and supraspinal levels. Stimulation of DCs or DC nuclei (DCN) in animals where neuropathic pain has been induced causes activation of brainstem centers including the periaqueductal gray (PAG), which is involved in the endogenous pain suppression system. Biotinylated dextran-amine (BDA) was iontophoretically injected into the DCN to analyze the ascending projection directed to the PAG. Separate injections into the gracile nucleus (GrN) and the cuneate nucleus (CunN) showed BDA-positive fibers terminating in different regions of the contralateral PAG. GrN-PAG afferents terminated in the caudal and middle portions of PAG-l, whereas CunN-PAG fibers terminated in the middle and rostral portions of PAG-l. Based on the DCN somatotopic map, the GrN sends information to the PAG from the contralateral hindlimb and the tail and the CunN from the contralateral forelimb, shoulder, neck and ear. This somatotopic organization is consistent with earlier electrophysiological and PAG stimulation studies. These fibers could form part of the DCs-brainstem-spinal cord loop, which may be involved in the inhibitory effects of SCS on neuropathic pain.

Integrity of the dorsolateral periaqueductal grey is essential for the fight-or-flight response, but not the respiratory component of a defense reaction

Periaqueductal grey is believed to be one of the key structures of the central respiratory stress network. Previous studies established that stimulation of the periaqueductal grey, especially its dorsolateral division (dlPAG), evokes tachypnea as well as increases in other autonomic parameters and motor activity. We investigated the effects of blockade of the dlPAG with GABAA agonist muscimol on respiration during stress and presentation of brief alerting stimuli in conscious unrestrained rats. We found that integrity of the dlPAG is not essential for stress-induced increase in basal/resting respiratory rate or for generation of respiratory responses to brief alerting stimuli. However, blockade of the dlPAG reduced the amount of motor activity and concomitant high-frequency respiratory activity during restraint and the first 5min of novelty stress. We conclude that the integrity of the dlPAG is not essential for generation of respiratory component of the defense reaction, but it mediates expression of the fight-or-flight response including its respiratory component.

Differential distribution of parvalbumin- and calbindin-D28K-immunoreactive neurons in the rat periaqueductal gray matter and their colocalization with enzymes producing nitric oxide

The distribution, colocalization with enzymes producing nitric oxide (NO), and the synaptic organization of neurons containing two calcium-binding proteins (CaBPs) - parvalbumin (Parv) and calbindin-D28K (Calb) - were investigated in the rat periaqueductal gray matter (PAG). Parv-immunopositive (ParvIP) neurons were detected in the mesencephalic nucleus and rarely in the PAG. CalbIP neurons were found both in the dorsolateral (PAG-dl) and ventrolateral PAG (PAG-vl); their size ranged from 112.96 μm(2) (PAG-dl) to 125.13 μm(2) (PAG-vl). Ultrastructurally Parv and Calb immunoreactivity was mostly found in dendritic profiles. Axon terminals containing each of the two CaBPs formed symmetric synapses. Moreover both Parv and Calb were used to label a subpopulation of NO-producing neurons. Colocalization was investigated using two protocols: (i) a combination of Calb and Parv immunocytochemistry (Icc) with nicotinamide adenine dinucleotide phosphate diaphorase (NADPH-d) histochemistry (Hi) and (ii) neuronal NO synthase-Icc (nNOS) (immunofluorescence). Both techniques demonstrated a complete lack of colocalization of Parv and NADPH-d/nNOS in PAG neurons. Double-labeled (DL) neurons (Calb-NADPH-d; Calb-nNOS) were detected in PAG-dl. NADPH-d-Hi/Calb-Icc indicated that 41-47% of NADPH-d-positive neurons contained Calb, whereas 17-23% of CalbIP cells contained NADPH-d. Two-color immunofluorescence revealed that 53-66% of nNOSIP cells colocalized with Calb and 24-34% of CalbIP neurons contained nNOS. DL neuron size was 104.44 μm(2); neurons labeled only with NADPH-d or Calb measured 89.793 μm(2) and 113.48 μm(2), respectively. Together with previous findings (Barbaresi et al. [2012]) these data suggest that: Therefore the important aspect of the PAG intrinsic organization emerging from this and previous double-labeling studies is the chemical diversity of NO-synthesizing neurons, which is likely related to the different functions in which these neurons are involved.

Analgesia induced by localized injection of opiate peptides into the brain of infant rats

BACKGROUND

Stimulation of a variety of brain sites electrically or by opiates activates descending inhibitory pathways to attenuate noxious input to the spinal cord dorsal horn and produce analgesia. Analgesia induced by electrical stimulation of the periaqueductal grey (PAG) of the midbrain or medial rostral ventral medulla (RVM) matures late, towards the end or past the pre-weaning period. Descending facilitation takes precedence over inhibition. Yet opiates injected intracerebroventricularly or directly into the PAG induce analgesia relatively early in development. Our goal was to re-examine the role of opiates specific to individual receptor types in analgesia at several supraspinal sites.

METHODS

Antinociception was tested following microinjection of DAMGO (μ-opiate agonist), DPDPE (∂-opiate agonist) or U50,488 (κ-opiate agonist) into the PAG, RVM or dorsal lateral pons (DLP) in 3-, 10- and 14-day-old rats.

RESULTS

DAMGO produced analgesia at 3 days of age at each brain area; the RVM was the most effective and the dorsal PAG was the least effective site. DPDPE produced modest analgesia at 10 and 14 days of age at the ventral PAG, RVM or DLP, but not the dorsal PAG. U50,488H was ineffective at all sites and all ages.

CONCLUSIONS

Antinociception could be elicited at all three sites by DAMGO as early as 3 days of age and DPDPE at 10 and 14 days of age. The degree of analgesia increased gradually during the first 2 weeks of life, and likely reflects the maturation of connections within the brain and of descending inhibitory paths from these sites.

Periaqueductal gray neuroplasticity following chronic morphine varies with age: role of oxidative stress

The development of tolerance to the antinociceptive effects of morphine has been associated with networks within ventrolateral periaqueductal gray (vlPAG) and separately, nitric oxide signaling. Furthermore, it is known that the mechanisms that underlie tolerance differ with age. In this study, we used a rat model of antinociceptive tolerance to morphine at two ages, postnatal day (PD) 7 and adult, to determine if changes in the vlPAG related to nitric oxide signaling produced by chronic morphine exposure were age-dependent. Three pharmacological groups were analyzed: control, acute morphine, and chronic morphine group. Either morphine (10mg/kg) or equal volume of normal saline was given subcutaneously twice daily for 6½ days. Animals were analyzed for morphine dose-response using Hot Plate test. The expression of several genes associated with nitric oxide metabolism was evaluated using rtPCR. In addition, the effect of morphine exposure on immunohistochemistry for Fos, and nNOS as well as nicotinamide adenine dinucleotide phosphate diaphorase (NADPH-d) reaction at the vlPAG were measured. In both age groups acute morphine activated Fos in the vlPAG, and this effect was attenuated by chronic morphine, specifically in the vlPAG at the level of the laterodorsal tegmental nucleus (LDTg). In adults, but not PD7 rats, chronic morphine administration was associated with activation of nitric oxide function. In contrast, changes in the gene expression of PD7 rats suggested superoxide and peroxide metabolisms may be engaged. These data indicate that there is supraspinal neuroplasticity following morphine administration as early as PD7. Furthermore, oxidative stress pathways associated with chronic morphine exposure appear age-specific.

Immunocytochemical localization of calretinin-containing neurons in the rat periaqueductal gray and colocalization with enzymes producing nitric oxide: a double, double-labeling study

The pattern of distribution and colocalization of the calcium-binding protein calretinin (Cal) and of enzymes producing nitric oxide (NO) was examined in the rat periaqueductal gray matter (PAG) using two different experimental approaches, by combining Cal immunocytochemistry with NADPH-diaphorase (NADPH-d) histochemistry and with NOS immunocytochemistry, respectively. Cal-immunopositive neurons were found throughout the rostrocaudal extension of both dorsolateral (PAG-dl) and ventrolateral PAG (PAG-vl). Double-labeled neurons were found only in PAG-dl. The first experimental approach indicated that 33-41% of the NADPH-d-positive (Nadph+) cells were immunoreactive for Cal, whereas NADPH-d activity appeared in 19-26% of the Cal-immunopositive (Cal(IP) ) neurons. Two-color immunofluorescence revealed that ∼39-43% of NOS-immunoreactive (NOS(IR) ) neurons were double-labeled with Cal and ∼23% of Cal(IP) neurons expressed NOS immunoreactivity. Measurement in semithin sections of the size of the three neuronal populations found in PAG-dl, showed that Cal(IP) neurons had a cross-sectional area of 94.7 μm², whereas Nadph+ neurons and double-labeled neurons were slightly smaller, having a cross-sectional area of 90.5 and 91.4 μm², respectively. On electron microscopy, Cal(IP) axon terminals formed either symmetric or asymmetric synapses; although the latter synapses were more numerous, both types contacted preferentially Cal(IP) dendrites. These experiments suggest that PAG-dl is characterized by a high degree of heterogeneity.

The gray area between synapse structure and function-Gray's synapse types I and II revisited

On the basis of ultrastructural parameters, the concept was formulated that asymmetric Type I and symmetric Type II synapses are excitatory and inhibitory, respectively. This "functional Gray synapses concept" received strong support from the demonstration of the excitatory neurotransmitter glutamate in Type I synapses and of the inhibitory neurotransmitter γ-aminobutyric acid in Type II synapses, and is still frequently used in modern literature. However, morphological and functional evidence has accumulated that the concept is less tenable. Typical features of synapses like shape and size of presynaptic vesicles and synaptic cleft and presence of a postsynaptic density (PsD) do not always fit the postulated (excitatory/inhibitory) function of Gray's synapses. Furthermore, synapse function depends on postsynaptic receptors and associated signal transduction mechanisms rather than on presynaptic morphology and neurotransmitter type. Moreover, the notion that many synapses are difficult to classify as either asymmetric or symmetric has cast doubt on the assumption that the presence of a PsD is a sign of excitatory synaptic transmission. In view of the morphological similarities of the PsD in asymmetric synapses with membrane junctional structures such as the zonula adherens and the desmosome, asymmetric synapses may play a role as links between the postsynaptic and presynaptic membrane, thus ensuring long-term maintenance of interneuronal communication. Symmetric synapses, on the other hand, might be sites of transient communication as takes place during development, learning, memory formation, and pathogenesis of brain disorders. Confirmation of this idea might help to return the functional Gray synapse concept its central place in neuroscience.