Ultrasound Used for Diagnostic Imaging Facilitates Dendritic Branching of Developing Neurons in the Mouse Cortex

Abstract

Neuronal differentiation and synaptogenesis are regulated by precise orchestration of intrinsic and extrinsic chemical and mechanical factors throughout all developmental steps critical for the assembly of neurons into functional circuits. While ultrasound is known to alter neuronal migration and activity acutely, its chronic effect on neuronal behavior or morphology is not well characterized. Furthermore, higher-frequency (3-5 MHz) ultrasound (HFU) is extensively used in gynecological practice for imaging, and while it has not been shown harmful for the developing brain, it might be associated with mild alterations that may have functional consequences. To shed light on the neurobiological effects of HFU on the developing brain, we examined cortical pyramidal cell morphology in a transgenic mouse model, following a single and short dose of high-frequency ultrasound. Layer V neurons in the retrosplenial cortex of mouse embryos were labeled with green and red fluorescent proteins by in utero electroporation at the time of their appearance (E14.5). At the time of their presumptive arrival to layer V (E18.5), HFU stimulation was performed with parameters matched to those used in human prenatal examinations. On the third postnatal day (P3), basic morphometric analyses were performed on labeled neurons reconstructed with Neurolucida. Low-intensity HFU-treated cells showed significantly increased dendritic branching compared to control (non-stimulated) neurons and showed elevated c-fos immunoreactivity. Labeled neurons were immunopositive for the mechanosensitive receptor TRPC4 at E18.5, suggesting the role of this receptor and the associated signaling pathways in the effects of HFU stimulation.

Keywords: dendrite; fetus; neurogenesis; neuron; ultrasound.

FIGURE 1

In utero electroporation revealed GFP-labeled neurons in the limbic area of the telencephalon. Coronal sections of mouse brain sampled at (A) E18.5 and (B) P3, after the labeling by utero electroporation was carried out at E14.5. (A) The GFP-labeled neurons already populated their presumptive layer in the cortex and the hippocampus which is shown in the upper right panel that is illustrated with the enframed area in the coronal view of the GFP-labeled mouse brain from E18.5 shown in the upper left panel. The labeled neurons showed polarized morphology, with a visible apical dendrite (arrows), and in some cases basal dendritic processes can also be obtained (arrowheads). (B) Upper panel shows the GFP (green) labeling pattern includes the retrosplenial cortex and the hippocampus (enframed area), the contralateral-sided somatosensory cortex (arrow), and scattered labeling in the hypothalamus (arrowhead). The retrosplenial cortex and the CA1 region of the hippocampus are shown with higher magnification which is illustrated in the framed area in the upper panel. In the retrosplenial cortex, almost all the GFP-positive pyramidal neurons were found in layer 5 and the hippocampal CA1 pyramidal cells were also labeled. Dashed lines illustrate the borders of regions of the retrosplenial cortex. L1-L6b, Cortical layers; cing, cingulum; fp, posterior forceps of the corpus callosum; dhc, dorsal hippocampal commissure; DG, dentate gyrus; CA1, hippocampal region CA1. Scale bar: 20 μm in (A), 500 μm in upper left panel, 100 μm in upper right panel and (B), and 700 μm in upper in (B).

FIGURE 2

Computer-aided reconstruction of pyramidal cells in the retrosplenial cortex. Maximum intensity projection of 200-μm confocal optical sections forms the frontal cortex where the neurons labeled with Brainbow vector at E14.5 by in utero electroporation. Samples were taken on the third day of postnatal stages from a non-treated control (A) and a mouse stimulated by a single dose of ultrasound (US) at E18.5 age of fetuses (C). Arrows show those of pyramidal neurons chosen randomly from layer 5 which were analyzed by Neurolucida. The reconstructed pyramidal cells from the control (B) and US-stimulated samples (D) are illustrated (15–15 pyramidal cells). Scale bar: 200 μm.

FIGURE 3

Ultrasound stimulus increases the number of dendrites. Morphometric analyses after reconstruction of the neurons in P3 indicate that a single short-term intrauterine ultrasound (3 MHz) stimulus at E18.5 (US-treated) increased the numbers of dendrites by approx. 1–2 compared to the non-stimulated (Control) group (A). The dendritic morphology was not changed significantly between the US-treated and Control groups. Only a slight and not significant gain was measured in the numbers of nodes (B) and the length of the internodal dendritic segments (C), and their tortuosity (D) was not changed in the US stimulus. **p < 0.01, ns: not significant (both Mann–Whitney). Boxes are indicating the middle 50% of the data with the median. Bars are indicating the upper and lower 25–25% of data while the asterisk labels are the outlier data. The means are indicated as small center squares in the boxes.

FIGURE 4

Ultrasound stimulus does not cause any changes in dendrite morphology and soma size. US treatment at E18.5 (at that time the GFP-labeled neurons were 4 days old) was ineffective to alter the length of the dendrites (A,B) of the GFP-labeled pyramidal cells at P3. The thickness of the dendrites (C) and the size of the soma (D) were also the same between the US-treated and the non-treated control groups. ns: not significant (Mann–Whitney). Boxes indicate the middle 50% of the data with the median. Bars indicate the upper and lower 25–25% of data while the asterisk labels are the outlier data. The means are indicated as small center squares in the boxes.

FIGURE 5

A single stimulation with ultrasound activated neurons but did not increase BDNF expression. Confocal images show immunofluorescent staining of c-Fos (red) and BDNF (green) in the retrosplenial cortex (A,B). Immunoreactivity of c-Fos could not be observed in the non-treated control (A) which was markedly increased upon a single US stimulus (B). BDNF immunoreactivity was restricted to the blood vessels and the meninges (arrows in A), which remained unchanged according to data of real-time PCR using templates from whole-brain RNAs (C). Scale bar: 20 μm. ns: not significant (Mann–Whitney).

FIGURE 6

Repeated ultrasound stimuli elevate c-Fos and BDNF expression in the cortex. Confocal images showing the retrosplenial cortex from P30 mice sampled from the non-treated control (A–D), the US-stimulated at E18.5 group (E–H), and the US-stimulated five times group (I–L). Each section was multiplied stained, and the histochemical reactions were separated with different fluorescent color channels and illustrated in grayscale. The GFP labeling was carried out at E14.5, and the samples show the mature pyramidal cells of the retrosplenial cortex from the non-stimulated control (B), the US-stimulated at E18.5 group (F), and the case of repetitive (once per week) US-stimulated group (J). The c-Fos (C,G) and BDNF (D,H) immunoreactivities were similar between the control and the single US-stimulated group. A marked increase in c-Fos and BDNF immunopositivity could be observed in the 5 times US-stimulated group (K,L) compared to the non-stimulated control sample (C,D). Sections were also stained with DAPI (A,E,I) for visualizing better the cortical layers. Scale bar: 100 μm.

FIGURE 7

Weekly repeated ultrasound stimuli elevated c-Fos and BDNF expression in the cortex. Single-plane (0.51-μm-thick) confocal images showing an area with higher magnification of the retrosplenial cortex are shown in Figure 6 focusing on the layer 5 GFP-labeled pyramidal neurons (D–F), c-Fos (G–I), and BDNF (J–L); immunohistochemistry and DAPI staining (A–C) in separated fluorescent color channels and merged colors (M–O). Faint immunopositivity of BDNF and c-Fos was observed in the non-stimulated control and the single US-stimulated samples (J,K,G,H). Visible c-Fos immunoreactivity was observed in the group which was stimulated with US 5 times (I, arrows) in GFP-labeled neurons and in other cells (O), and the BDNF immunoreactivity was also increased (L). Scale bar: 20 μm.

FIGURE 8

TRPC4 is expressed in GPF-labeled pyramidal cells. Confocal stacks from the E18.5 cortex in separated and merged fluorescent color channels show immature neurons labeled with GPF (green) at E14.5. The nuclei of the cells were labeled with DAPI (blue). TRPC4 immunoreactivity (red) showed a dot-like pattern in the cytoplasm and some cases also in the (leading) processes of the GFP labeled neurons (arrows). Scale bar: 20 μm.