Cheng et al., 2017 - Characterization of a thalamic nucleus mediating habenula responses to changes in ambient illumination. BMC Biology   15:104 Full text @ BMC Biol.

Fig. 1

Overview of the habenula response to light ON and OFF. a, b Dorsal view of the fore and midbrain of 5-day-old elavl3:GCaMP6f fish, imaged with wide-field fluorescence microscopy at 200 Hz. The time since start of illumination is shown at the top right. The wedge indicates the ratio of fluorescence relative to the first frame. There is an increase in GCaMP6f fluorescence in the left habenula (arrowhead in b). c Maximum ΔF/F0 value in the left and right habenula after onset of light in five different fish. Each circle is one fish and the line joins data points from the same fish. df Two-photon imaging of the habenula in a 7 dpf GAL4s1011t, UAS:GCaMP6s fish, at 13 Hz. d Average of the time-lapse sequence, showing anatomy. The dorsal left neuropil is indicated by the yellow arrowhead. e Spatial distribution of responses to pulses of light. Pixels are color-coded according to the temporal pattern of response, as indicated in panel f. f Centers of clusters obtained from running K-means on the time series of pixels in panel d. Cluster centers are colored in shades of blue for responses to light ON and magenta and orange for responses to light OFF. The presence of light is indicated by the blue bars. g, h Neuropil response summarized from imaging 10 fish exposed to 7 pulses of blue light. g Pixels in the left neuropil from all fish could be classified into three main classes. They are pseudo-colored and overlaid on an average image of a 6 dpf fish. The largest response was a transient response to light ON (blue). A sustained response to light ON (cyan) and OFF (magenta) were also seen. Responses were reproducible in all 10 fish. h Average traces obtained from neuropil pixels, shown here for two pulses of light. i Percentage of cells active to light ON and OFF in the habenula is correlated with the percentage of active pixels in neuropil. The transient and sustained neuropil responses were combined into ON. Percentage of active cells or pixels were calculated by dividing the number of cells/pixels active to the stimulus by the total number of segmented cells or neuropil pixels. Each circle per category is one fish. The bold lines show best fit (linear regression), r is the correlation coefficient. jk Cumulative probability of peak ∆F/F0 response in cells and neuropil pixels responding to light ON (j) and light OFF (k). The response in the neuropil precedes the response in cells. P value and test statistic (D) were obtained by a Kolmogorov–Smirnov test between categories in the first 5 seconds. In panel j, the diamond ♦  (left) is the result of comparison between neuropil transient On and cell On, while plus + (right) between neuropil sustained On and cell On. rHb right habenula, lHb left habenula, Pa pallium, OT optic tectum, Pi Pineal, a anterior, p posterior. Scale bar = 25 μm

Fig. 2

Response of habenula neurons to pulses of light. ad The dorsal habenula response to 7 pulses of blue light in 10 fish (GAL4s1011t, UAS:GCaMP6s, 5–7 dpf). a Heatmaps from 5 example fish showing responses in cells that were classified as ON, OFF, or Inhibitory (Inh). The colors indicate ∆F/F0, as shown in the color bar. Responses in each fish are sorted in ascending order of mean ∆F/F0. Black horizontal lines separate each fish. The bold vertical lines correspond to light onset while the dashed lines indicate offset. The presence of light is also indicated by the blue bars. The height of the heatmaps represents the number of cells as indicated by the vertical scale bar. b Overlay of cells segmented from all fish. A small circle was drawn around the centroid of the segmented cell. Three main classes of activity are shown. Cyan indicates cells responding to light ON (ON), green cells are inhibited by light (Inh), and magenta cells are activated in the absence of light (OFF). Hollow circles did not show an evoked response. The gap in the left habenula indicates the neuropil region. c Averaged traces from the cells in panel b, showing the response of different classes for the first two pulses of light. d Boxplots showing distribution of cells responding to different classes in the left and right habenula. Each circle is one fish and the line joins data points from the same fish. P values and test statistic (Z) were obtained using Wilcoxon signed rank test. e K-means clustering of pixels in the habenula of fish from 5 to 10 dpf as indicated. Pixels are colored blue if they respond to light ON and magenta if they respond to light OFF. Data from each fish was analyzed separately. Individual traces of the cluster centroids are not shown here but are similar to Fig. 1f. All fish have a response to light onset in the dorsal left neuropil. Anterior is to the top in all panels. rHb right habenula, lHb left habenula

Fig. 3

Anatomical characterization of thalamic neurons projecting to the habenula. a An example of DiD injection (cyan) into the dorsal neuropil of the left habenula (yellow arrowhead). The dorsal neuropil of the right habenula contains afferents from the entopeduncular nucleus (labeled by the SqKR11Et line; magenta) and has no DiD-labeled neurons, indicating specificity of the injection. b Dorsal view of the thalamic region of a 7-day-old fish following DiD injection into the dorsal left habenula neuropil. Arrowheads indicate retrogradely DiD-labeled neurons that express eGFP (shown in yellow) under the control of the vGlut2a GAL4 driver. c Lateral view of the fish in panel b, showing DiD-labeled neurons on the right side of the brain. d Dorsal view of another larva, in which the dorsal right habenula had been injected with DiD. Retrogradely labeled neurons are located in the entopeduncular nucleus. e A double transgenic fish, with glutamatergic neurons shown in green (e'), and GABAergic neurons shown in magenta. The neuropil in the anterior thalamus (arrow) contains magenta label (e’’), indicating the presence of GABAergic fibers. f Lateral view of a 7 dpf larva following injection of CM-DiI into the dorsal left habenula and labeling with anti-GAD 65/67. The region of the neuropil containing CM-DiI-labeled neurites (red; arrowheads; f’’) is labeled with the GAD65/67 antibody (cyan; f’). All panels except c are single optical sections. Pa pallium, rHb right habenula, lHb left habenula, fr fasciculus retroflexus, Th thalamus, EN entopeduncular nucleus, OT optic tectum, otr optic tract, ZLI zona limitans intrathalamica, AC anterior commissure, a anterior, p posterior, d dorsal, v ventral

Fig. 4

The response of thalamic neurons to irradiance change. ae Activity in five different focal planes of a 5-day-old fish expressing GCaMP6s in thalamic neurons (arrows). Numbers indicate depth. The colors represent K-means cluster centers shown in panel f, with blue indicating ON responses and magenta indicating OFF responses; cyan pixels have a response to both light ON and OFF. gj Quantitative analysis of response of anterior thalamic neurons of GAL4s1020t, UAS:GCaMP6s fish (5–6 dpf) to pulses of blue light. Note that this driver is not expressed in afferent retinal ganglion cells. g Heatmaps of individual cells in five example fish, showing major classes of responses seen in thalamic neurons: sustained or transient excitation to light ON, light OFF, or both ON and OFF (yellow). g and g’ have different scales. h Responses in cells from 10 fish at three different focal planes. Four pulses of blue light were given and imaging was done at 7 Hz. Segmented cells in all fish overlaid and colored by their response. i Traces showing mean responses of cells in panel h for two blue pulses. j Percentage of cells responding to light ON, OFF, or both ON and OFF. k Neuropil responses to pulses of light. Pixels with different response classes from all fish were pseudo-colored and overlaid on an average image from a 5-dpf, GAL4s1020t, UAS:GCaMP6s fish. l Average traces of responses in panel k. m Percentage of neuropil pixels responding to light ON, OFF, or both ON and OFF. nr Thalamus response to blue and red light. n Spatial distribution of responses, color coded according to the K-means cluster centers in panel o, with blue pixels showing a sustained response to light ON, while magenta pixels and orange pixels are a mixture of responses to both ON and OFF. Z is the Z-score. p Heatmaps of cells responding to three pulses of blue light followed by three pulses of red light in n = 6 GAL4s1020t, UAS:GCaMP6s fish. Cells were classified as responding to light ON or OFF. While the same cells responded to both blue and red light, the amplitude of responses were lower to red light. q Peak amplitude of response during light ON and OFF is higher to blue light than red light. Each circle represents one fish and lines join data points from the same fish. Crosses and diamonds represent median amplitude. r Histogram showing amplitude of responses during blue (blue traces) and red (red traces) light ON (left panel) and OFF (right panel). Each trace is response distribution from all cells in a single fish. P values and test statistic (D) were obtained using Kolmogorov–Smirnov test on cumulative response distribution from all fish shown in the inset in r. Th thalamus

Fig. 5

Anatomical and physiological characterization of the anterior thalamic neuropil. a. Lateral view of a 6-day-old fish following injection of DiD (cyan) into the dorsal neuropil of the left habenula and DiI (yellow) into the right retina. Arrows indicate terminals from retinal ganglion cells in the vicinity of fibers from habenula afferents. See Additional file 5: Movie 5. b Illustration of a fish larvae, showing the region imaged in panel a (red box) and in panels c and d (black box). ch Response in the anterior thalamic neuropil to pulses of light. c Average projection of a lateral view of an elavl3:GCaMP6f fish, showing the thalamic neuropil (arrowhead). d, e The response to four pulses of blue light. Colors show the K-means cluster centers represented in panel e. The regions responding to light ON and OFF are distinct in the thalamic neuropil. Responses can also be seen in the habenula. fh Quantitation of the anterior thalamus neuropil response to light pulses in eight fish. f Contours show a bivariate kernel density estimate of neuropil pixel location for responses to ON (shades of blue) and OFF (shades of red) of blue light in eight fish. The two variables here are x and y of neuropil pixels. The orientation is same as panel d. Crosses indicate median location of response to light ON, while diamonds indicate median location of response to light OFF in each fish. The dorso-ventral and anterior-posterior positions of the median centers are shown in panels g and h, respectively. Each circle is one fish and lines join data points from the same fish. These panels show that ON and OFF responses have a different location, with OFF responses in a more anterior-ventral location. P values and test statistic (D) were obtained using Kolmogorov–Smirnov test on cumulative distribution of pixel location to light ON and OFF from all fish. a anterior, p posterior, d dorsal, v ventral, Hb habenula, Th thalamus

Fig. 6

Effect of eye removal on thalamus response to light ON and OFF. ai Activity in the thalamus in response to pulses of blue light in control (n = 3) and enucleated (n = 4) fish. a, b Heatmaps showing activity in individual cells in control and enucleated fish classified as having response to light ON or OFF. c A comparison of the percentage of active cells in control and enucleated fish. The response to light ON is reduced in fish lacking eyes, while the response to light OFF is comparable to controls. de Histogram of mean response amplitude in cells in control and enucleated fish during (d) light ON and (e) OFF. Each trace is one fish. The amplitude of response to light ON is reduced in enucleated fish. Insets show cumulative histogram from all fish. fi Pixels in the anterior thalamic neuropil of control (f) and enucleated (h) fish, that are active to light ON (cyan) or OFF (pink), were combined and overlaid. Panels f and h show a dorsal view of the thalamus. The average traces from the colored pixels in f and h are shown in g and i, respectively. Control fish have a response to light ON and OFF, whereas enucleated fish only have a response to light OFF. j Percentage of active neuropil pixels in control and enucleated fish. kl Cumulative probability of mean response amplitude in pixels of control and enucleated animals to light ON (k) and OFF (l). Mean response during light OFF is not significantly different in enucleated and control fish. m Dorsal view of a 6-day-old fish, following injection of DiD into the dorsal left habenula neuropil and CM-DiI into the pineal organ. See Additional file 6: Movie 6 for the complete z-stack. n Lateral view of the right side of a 6-day-old fish, showing anterogradely labeled fibers from the pineal (red) and retrogradely labeled fibers from the habenula (cyan). The arrow indicates a pineal axon in the neuropil of the anterior thalamus. o, p Response in the pineal organ to pulses of blue light (n = 4 fish). Only OFF responses can be detected. (o) Pixels showing OFF responses are combined from all fish and overlaid. (p) Average trace from the colored pixels. The habenula is shown here for orientation only; habenula neuron responses have been masked. qs Responses in the habenula to light OFF in control (q) and enucleated (r) fish (GAL4s1011t, UAS:GCaMP6s, n = 4 fish). Each row in the heatmaps represents an individual cell. s Percentage of cells showing an OFF response. Each circle is one fish and lines join data points from same fish before and after enucleation. Although reduced in number, there are still cells that display an OFF response. D-statistic and P values in panels d, e, k, and l were obtained using the Kolmogorov–Smirnov test on response amplitude distribution. Panels f, h, m, and o are single optical sections; n is a projection spanning 19.25 μm. rHb right habenula, lHb left habenula, Th thalamus, a anterior, p posterior, d dorsal, v ventral. Scale bar = 25 μm

Fig. 7

The effect of lesioning the anterior thalamic neuropil on habenula response to light. a Dorsal view of an elavl3:GCaMP6f fish, showing lesion bubbles in the anterior thalamic neuropil created by a femtosecond laser (arrows). The bubble reflects the two-photon laser, and is thus captured in a separate channel from GCaMP6f fluorescence. a’, a’’ Close-up of the anterior thalamus neuropil before (a’) and during (a’’) the lesion. The cavity has not yet formed. b Heatmap showing habenula cell responses before (left) and after (right) lesioning in three fish. The scale of the bottom right panel is different from the others as indicated. c, d The cells segmented from all three fish are drawn as circles and overlaid. Responding cells before and after the lesion are colored as indicated. e, f Histogram showing distribution of mean intensity in habenula neurons during light ON (e) and OFF (f) before and after the lesion. Insets show cumulative distribution from all fish. P values and test statistic were obtained using the Kolmogorov–Smirnov test. g Comparison of percentage of cells responding to light ON and OFF before and after the lesion. Each circle is one fish and lines join data points from the same fish. Th Thalamus, Pa Pallium, lHb left habenula, rHb right habenula, a anterior, p posterior. Images are all single optical sections

Fig. 8

GtACR1 expression in the thalamus disrupts habenula response to light ON and OFF. a A 6-day-old fish expressing GtACR1-eYFP under the control of the GAL4s1020t driver. GtACR1-expressing cells in the anterior thalamus (colored orange-purple) are indicated by the yellow arrowheads. Puncta of GtACR1-eYFP are visible. There is a low level of GCaMP6f expression, shown in green. The anterior margin of the thalamus is indicated by the dashed line. b A more dorsal focal plane, with habenula afferents labeled by the sqKR11Et line (magenta). GtACR1-eYFP puncta are indicated by the arrowheads. The asterisks indicate autofluorescent pigment cells. The habenulae are outlined. ck Comparison of dorsal left habenula neuropil response to pulses of blue light in controls and fish expressing GtACR1 in the thalamus. c, d Response in the neuropil of control (c; n = 7) and GtACR1-expressing siblings (d; n = 11). Blue represents fast ON, cyan represents slow ON, whereas magenta represents OFF response. e, f The average of the colored pixels from c and d, respectively. Shaded areas show 95% confidence intervals. g, h Boxplots show the distribution of percentage of neuropil pixels showing a response to light ON or OFF in fish expressing GtACR1 (h) or control siblings (g). Each circle is one fish and lines join data points from the same fish. P values and test statistic (U) were obtained using Mann–Whitney U test between the distribution of pixels in control and GtACR1 fish of the same response class. i, j Histograms showing neuropil response to light ON and OFF in individual fish expressing GtACR1 (j) and control siblings (i). There is a leftward shift in the distribution for response to light ON in GtACR1 fish. k Interpolations of the histograms in panels i–j using a smoothing spline fit to show the overall distribution per category. Pa pallium, Th thalamus, hc habenula commissure, rHb right habenula, lHb left habenula, a anterior, p posterior

Fig. 9

Effect of optogenetic stimulation of the thalamus on habenula activity. a Expression of ChR2-eYFP in the thalamus (arrowheads) of a 5-day-old GAL4s1020t, UAS:ChR2-eYFP, elavl3:GCaMP6f fish. b, c Activity in the habenula of a ChR2-expressing fish, with (b) and without (c) blue LED stimulation of the thalamus. The images show the maximum projections of F/F0 images for a 25-second period after blue LED illumination, following subtraction of maximum projections of the period before illumination (i.e., difference in activity before and after stimulation). df Heatmaps showing temporal activity from habenula neurons segmented in fish with (e, f) and without (d) ChR2. In panels e (n = 3 fish) and f (n = 2 fish), a blue light pulse was given at the time indicated by the black dashed line and at the specified frequency. g Cumulative distribution of mean response amplitude, 10 seconds after stimulation in ChR2-expressing and control fish and a randomly chosen 10 second period in fish with no stimulation. All stimulation frequencies were combined. The fish with ChR2 showed increased ∆F/F0 after optogenetic stimulation. Test statistic and P values were obtained using the Kolmogorov–Smirnov test. The gray * (bottom) is the result of comparison between control siblings and Chr2-expressing fish, while the black * (top) is the result of comparison between no stimulation and Chr2-expressing fish. h Mean amplitude before and after optogenetic stimulation at different frequencies. Each square stands for a stimulus trial. Scale bar = 25 μm. Pa pallium, a anterior, p posterior, lHb left habenula, rHb right habenula

Fig. s7 The effect of the parapineal lesion on habenula response to blue light. a Visualization of the parapineal (yellow arrow), which is located adjacent to the left habenula and innervates the dorsal neuropil. b, c Two-photon lesioning of the parapineal. b Before lesioning. c After lesioning, which led to formation of a bubble (arrow). d, e Habenula cells segmented from five fish, overlaid on top of each other, showing responses before and after the lesion. Cells responding to light ON are shown in blue and to OFF in pink. f, g Heatmaps of the habenula cells, in the five fish, responding to light ON and OFF before (f) and after (g) lesioning the parapineal. Horizontal black lines divide data from different fish. h Percentage of cells showing ON and OFF responses before and after parapineal lesioning. i, j Histogram showing distribution of mean intensity in habenula neurons during light ON (e) and OFF (f) before and after lesion. Insets show cumulative distribution from all fish. P values and test statistic (D) were obtained using the Kolmogorov–Smirnov test. pp parapineal, lHb left habenula, rHb right habenula, cv circumventricular organ, a anterior, p posterior. Scale bar = 25 μm.

Fig. s8 Examples of signals that were excluded from visualization of K-means clusters. ae Pixels showing stimulus-independent activity in the thalamus, at five different focal planes. Pixels are colored according to the traces in panel f. For clarity, these signals were excluded from the visualization of clusters representing light-evoked activity shown in Fig. 4a–e. g Stimulus-independent activity in the habenula. Pixels are colored according to the traces in panel h. For clarity, these signals were excluded from the visualization of clusters representing light-evoked activity shown in Fig. 1e–f. f, h Cluster centers that did not represent light-evoked activity in the thalamus and habenula, obtained by running K-means on the time series of pixels in panel a–e and g. Th Thalamus, lHb left habenula, rHb right habenula, a anterior, p posterior.

Acknowledgments:
ZFIN wishes to thank the journal BMC Biology for permission to reproduce figures from this article. Please note that this material may be protected by copyright. Full text @ BMC Biol.