Figures and data
Glutamatergic neurons and SST+ GABAergic neurons project from MEC to LEC
(A) Horizontal brain section through MEC of a GAD67๐๐บ๐น๐ mouse showing GFP+ neurons (green) together with retrogradelly labelled neurons (light cyan) from LEC. Inset: Schematic of the retrograde tracer injection in LEC. Scale bar, 500 ยตm. (B) SST+ GABAergic neurons in MEC project to LEC. Arrowheads indicate examples of triple-labelled neurons (SST+, GAD67๐๐บ๐น๐, FG) from the experiment in A. Scale bar, 100 ยตm. (C) Proportions of GABAergic neurons (SST+ and non-SST+) and glutamatergic neurons in the MEC-LEC pathway (neuron counts from 3 mice). (D) Delivery of an equal mix of AAV1-CAG-FLEX-tdTomato and AAV1-CamKII-eGFP-WPRE-rBG into MEC of SST๐ถ๐๐ mice (upper left schematic) selectively labels SST+ neurons with tdTomato and glutamatergic neurons with GFP. Sub, subiculum. Scale bar, 500 ยตm. (E) Different laminar innervation patterns in LEC of glutamatergic axons and SST+ axons from the experiment in D. Wm, white matter. Scale bar, 100 ยตm.
Target preference of medial entorhinal axons in LEC
(A) Left: Virus injection in MEC of a wild-type mouse. Right: High magnification image showing YFP-labelled photosensitive (ChR2) axons in LEC. MML, DG middle molecular layer. Scale bars, 500 ยตm (left) and 100 ยตm (right). (B) Membrane current recordings of a representative PNiia to the indicated pharmacological treatments during optogenetic stimulation (blue bars). The average membrane current traces (red/black) are superimposed on the individual traces (grey). (C) Laser-evoked EPSCs (black) and IPSCs (red) in representative PNiia, PNiib and PNiii recorded without (top, control ACSF) or with glutamatergic synaptic blockers (bottom, DNQX/APV). (D) Normalized IPSC amplitudes plotted against the location of photostimulation. Boxes indicate 25th-75th percentiles, red line is median, whiskers extend to all data points not considered outliers. All data points are superimposed on the box plots. Each value represents data from one cell resulting from photostimulation of the given layer. (E) Patch-clamp recorded neurons together with photosensitive MEC SST+ axons in a semicoronal slice of LEC from an SST๐ถ๐๐ mouse. PER, perirhinal cortex. Schematic: Injection to virally label MEC SST+ neurons with ChrimsonR-tdTomato. Scale bars, 200 ยตm. (F) Membrane current recordings of the numbered neurons in E during photostimulation (orange bars). (G) Membrane current amplitudes recorded from all neurons in the experiment in E. (H) Charge transfers in response to optogenetic activation when clamping neurons at 0 mV. Non-responding neurons have charge transfer values of 0 pC. (I) Same as in H, but for data obtained when clamping neurons to -50 mV. The color code for PNiia, PNiib and PNiii in (H) and (I) follow the color code as defined in (G).
Target-specific control of postsynaptic neuron activity by medial entorhinal axons in layer I
(A) Biocytin-filled principal neurons recorded from layers IIa, IIb and III. Scale bar, 100 ยตm. (B) Different laser-evoked responses of PNiia versus PNiib and PNiii (PNiia, hyperpolarizing; PNiib, PNiii; depolarizing). Neurons were depolarized by somatic current injections (yellow) while MEC axons were photostimulated with light (blue bar). Schematic upper left: A 100 ms laser stimulation (blue) of MEC axons (ChR2, green) in LEC layer I during current-clamp recordings. (C) Action potential firing across multiple sweeps of the photostimulation (blue bar) protocol in representative neurons. Top: Voltage recordings. Middle: Raster plots. Bottom: Histograms. Note the transient suppression of spiking activity in PNiia and the increase in spiking activity in PNiib and PNiii during photostimulation. (D) Mean firing rate (mean number of action potentials per bin) of all tested PNiia, PNiib and PNiii. Plots are mean (black) ยฑ standard error of the mean (grey). (E) Box plots showing the deviation in action potential firing relative to the average firing activity without laser stimulation. Box plots as in Fig. 2. Outliers are shown as grey points. (F) Donut plots showing the proportion of laser-evoked responses. (G) Laser-evoked hyperpolarizing responses are resistant to application of glutamatergic synaptic blockers (DNQX/APV). The recording is from a representative PNiia. (H) Left: Laser-evoked depolarizing responses recorded from a representative PNiii. Right: The responses disappear when blocking glutamatergic synaptic transmission with APV and DNQX. RMP, resting membrane potential. In the rightmost panels the average voltage trace (black) is overlaid the individual voltage traces (grey).
Cortical glutamatergic inputs to principal neurons in layer IIa
(A) Labelled axons in superficial layers of LEC following anterograde tracer injections into PER (left), cLEC (middle) and PIR (right). BDA, biotinylated dextran amine. See Supplementary Fig. 7 for images of the injection sites. (B) Confocal images of neurons retrogradely labelled from LEC in a GAD67๐๐บ๐น๐ mouse (same case as the experiment in Fig. 1A). FG+ cell bodies were found in MEC, PER, cLEC and PIR. In this brain, co-localization of GFP and FG was evident in PER and MEC (white square insets). Bar graphs show the proportions of neurons retrogradely labelled with FG or AAV2-CAG-tdTomato that also expressed GFP (FG/tdTomato and GFP; mean ยฑ standard deviation, n = 3 mice). (C) Electrophysiological interrogation of afferent pathways in wild-type mice. Optogenetic activation (blue bars) of photosensitive axons from PIR (top) or cLEC (middle) in LEC elicited monosynaptic EPSPs in PNiia. Insets show average membrane potential traces in control ACSF (black), and ACSF added with TTX (blue) or TTX + 4-AP (orange). Activation of PER neurons (bottom) through UV-photostimulation (purple bars) of caged glutamate evoked EPSPs in PNiia in LEC. All membrane potential traces show the average trace (black) superimposed on individual trials (grey). Note that the recordings in the three panels are from different mice. Scale bars are 100 ยตm in (A) and (B).
Synaptic inputs from MEC diminish EPSPs of principal neurons in layer IIa
(A) Experimental design of dual photostimulation within layer I of LEC while recording from PNiia. Depicted is an example of a biocytin-filled PNiia showing its typical dendritic arborizations in layers I/IIa. (B) Example PNiia recording showing EPSPs evoked by glutamate uncaging alone (left, right) or together with MEC input (middle). The average membrane potential trace (black) is superimposed on the individual traces (grey). The dotted blue line indicates the peak of the average EPSP in response to photolysis of glutamate alone, showing a decrease in EPSP amplitude when pairing glutamate release with MEC activation. (C) Quantification of EPSP amplitude for all individual trials recorded from the neuron in B. (D) Average voltage traces from the recording in B. (E) Scatter plots of EPSP amplitude (left) and halfwidth (right) for individual sweeps of the experimental protocol from all recorded PNiia (n = 38). (F) Relative deviations in EPSP amplitude between responses elicited when glutamate was released alone (purple) and when glutamate release was paired with photostimulation of MEC axons (purple and red). Data show relative deviations for all recorded PNiia (n = 38). Box plots as in Fig. 1. (G) Same as in F, but for EPSP halfwidth.
Primary antibodies used in the study.
Secondary antibodies used in the study.
SST+ neurons project from MEC to LEC, but only very sparsely in the opposite direction, from LEC to MEC
(AโB) Conditional labelling of MEC SST+ neurons through injection of AAV9-CAG-FLEX-GFP into MEC of an SST๐ถ๐๐ mouse. The labelled neurons give rise to an axonal projection innervating the most superficial layers of LEC (as well as local MEC projections). (CโD) Conditional labelling of LEC SST+ neurons through injection of the Cre-dependent virus AAV1-CAG-FLEX-tdTomato into LEC of an SST๐ถ๐๐ mouse. Processes of labelled LEC SST+ neurons, as shown in C, were confined to LEC, with barely any transport to neighboring MEC. A single labelled LEC SST+ axon was observed in dorsal MEC in this experiment, as shown in D. Scale bars are 500 ยตm (A,C,D) and 1000 ยตm (B).
Medial entorhinal PV+ neurons and VIP+ neurons do not innervate LEC.
(A) An injection of AAV5-hSyn-FLEX-splitTVA-2A-mCherry-2A-B19G into MEC of a PV๐ถ๐๐ mouse labels specifically PV+ neurons. Labelled processes were confined to MEC, and no labelling was evident in neighboring LEC. (B) Same as A, but for the specific labelling of VIP+ neurons in a VIP๐ถ๐๐ mouse using the Cre-dependent virus AAV1-CAG-FLEX-tdTomato. (C) Same as A, but for the specific labelling of SST+ neurons in an SST๐ถ๐๐ mouse using the Cre-dependent virus AAV9-CAG-FLEX-GFP. All scale bars are 200 ฮผm.
Layer-dependent distribution of excitatory and inhibitory synaptic inputs
Heatmaps of median membrane current amplitudes (normalized values) of neurons recorded in layers IIa, IIb, and III as a function of the layer of photostimulation (X axis). These color-coded excitation-and inhibition profiles based on the median normalized EPSC/IPSC amplitude values were calculated from all groups of responding principal neurons recorded in the presence of control ACSF and ACSF added with TTX/4-AP. The anatomical layer associated with each photostimulation position was determined post-recording.
Pharmacological scrutiny of laser-evoked outward inhibitory currents
(A) Membrane current recordings of a representative PNiia during different pharmacological treatments showing that laser-evoked IPSCs are intact in the presence of the glutamatergic synaptic blockers DNQX (10 ยตM) and APV (50 ยตM; 1. DNQX/APV) but disappear after combined exposure to DNQX/APV and bicuculline (10 ยตM), an antagonist of the GABAa receptor (2. DNQX/APV + bicuculline). The synaptic responses return after wash-out of bicuculline (3. DNQX/APV). (B) Quantification of membrane charge transfers of all neurons exposed to DNQX/APV and bicuculline. Data points are average laser-evoked responses collected from 11 neurons in 2 mice.
Synaptic inputs from MEC to GABAergic neurons in LEC
(A) Biocytin-filled neurons (magenta) in LEC of a GAD67๐๐บ๐น๐ mouse that received an injection of AAV1-hSyn-ChrimsonR-tdTomato into MEC. Recorded neurons that belong to genetically labelled GAD67๐๐บ๐น๐ neurons (green) or principal neurons are indicated by blue and white numbers, respectively. Scale bar, 100 ยตm. (B) Membrane current recordings of the neurons in A kept at different holding potentials (black, Vhold โ -50 mV; red, Vhold โ 0 mV). Average traces (dark colors) are superimposed on the individual traces (light colors). Orange bars are laser stimuli used to active ChrimsonR+ MEC axons in LEC. (C) Percentage of recorded GAD67๐๐บ๐น๐ neurons with laser-evoked IPSCs and EPSCs. (D) Swarm charts showing normalized amplitudes of postsynaptic currents between pairs of simultaneously recorded GAD67๐๐บ๐น๐ neurons (green) and principal neurons (black) exposed to control ACSF. Blue dots are median values. (E) Absolute amplitudes of postsynaptic currents for GAD67๐๐บ๐น๐ neurons (left) and principal neurons (middle) exposed to control ACSF and ACSF added with glutamatergic blockers (DNQX/APV) in GAD67๐๐บ๐น๐ mice. For comparison, data for the same pharmacological interventions are shown for principal neurons (n = 17 PNiia, n = 27 PNiib, n = 21 PNiii) recorded in wild type mice (right). Median values are shown in red (left and middle diagrams) and black (right diagram).
Synaptic inputs from MEC give rise to different responses in principal neurons in layer IIa compared with principal neurons in layer III recorded in the same slice
(A) Six recorded and biocytin-filled principal neurons in layer IIa and layer III. The three PNiia were recorded simultaneously, after which the three PNiii were recorded simultaneously. Voltage recordings of the neurons were carried out while medial entorhinal AAV1-hSyn-ChR2-YFP tagged axons in layer I were activated by a single 100 ms blue light pulse (insets). Scale bar, 100 ยตm. (B) Different impact of laser stimulation on the spiking activity of the recorded neurons in A. Note the hyperpolarizing deflections of the membrane potential during laser stimulation in the three PNiia and the depolarizing deflections during laser stimulation in the three PNiii. Neurons were subjected to depolarizing somatic current injections to induce baseline action potential firing. (C) The number of action potentials fired by the neurons in A, B during laser stimulation (% relative to the number of action potentials without laser stimulation).
Anterograde tracer injections into PER, LEC and PIR
(A) A schematic mouse brain illustrating the injection site of the anterograde tracer (BDA, white circle) in PER. A horizontal section through the ipsilateral hemisphere shows the cytoarchitecture of the hippocampal-parahippocampal region (light blue) and BDA labelling (white) in PER. Inset shows the detailed lamination pattern of PER. This case corresponds to the example shown in Fig. 4A. (B) Same as in (A), but for a virus injection into LEC. This case corresponds to the example shown in Fig. 4A. This injection also partially leaks into neighboring MEC and PER, however, this is unlikely to influence results as we have not observed a clear contralateral innervation of LEC after injections into MEC or PER. (C) A schematic mouse brain illustrating an injection of the anterograde virus (black circle) in PIR. A fluorescent section containing the injection site (black) is overlaid with a neighboring Nissl stained section (blue). Inset show the cytoarchitectural laminar profile of PIR. This case corresponds to the example shown in Fig. 4A. All scale bars (red) are 1000 ยตm. M, medial. L, lateral. A, anterior. P, posterior.
Convergence of cortical inputs to principal neurons in layer IIa
(A) A schematic illustration of the experiment to combine extracellular electrical stimulation of PER with optogenetic activation of axons from either PIR, cLEC or MEC. (B) Bar graph showing the proportion of tested neurons that responded to stimulation of PER and either PIR, cLEC or MEC, i.e., neurons that received convergent synaptic inputs from two tested brain areas. Numbers indicate responding/tested neurons. (C) Electrophysiological data showing postsynaptic potentials in recorded PNiia in response to local stimulation of PER neurons or stimulation of axons from PIR. An example cluster of biocytin-filled and recorded neurons (left) display experimentally evoked postsynaptic potentials to optogenetic activation of PIR axons locally in LEC (middle) and following electrical stimulation of neurons in PER (right). The average resting membrane potential is indicated below each voltage trace. Average membrane potential traces (black) are superimposed on traces from the individual trials (grey). Scale bar is 100 ยตm. (D) Same as in (C), but for interrogation of synaptic inputs from PER and cLEC. (E) Same as in (C), but for interrogation of synaptic inputs from PER and MEC.
