Amira 4

For 3D-reconstruction, serial paraffin transversal sections (10 μm) of fetal pancreatic specimens at W17 and W21 were immunostained for glucagon and podoplanin as described in histology and immunohistochemistry. The sections were digitalized using the Panoramic MIDI digital scanner (3DHISTECH Ltd., Budapest, Hungary) and the reconstruction was conducted with the Amira 4.1 software (Visage Imaging, Berlin, Germany).

Fertilized White Leghorn chicken (Gallus gallus) eggs were incubated in a humidified atmosphere at 37.0°C until HH11–13 [7] (link). The embryos were isolated (HH11 n = 6; HH12 n = 10; HH13 n = 13), fixed overnight in 4% paraformaldehyde, included in paraffin, sectioned and stained for hematoxylin and eosin (H&E) as previously described [18] (link). Sections were imaged on an Olympus AX70 microscope (Olympus) equipped with an Olympus XC50 camera (Olympus). For 3D reconstructions, serial paraffin transversal sections of freshly isolated embryos at stage HH12 and HH13 stained with H&E were digitalized using a Panoramic MIDI scanner (3D Histech) and reconstructed with Amira 4.1 software (Visage Imaging).

In order to visualize the stained axons and the surrounding neuropil regions, the confocal image stacks were reconstructed by means of the software AMIRA 4.1 (Visage Imaging, Fürth, Germany). The neuropil regions were reconstructed by using the segmentation editor and the neuron axons by using the skeleton module. Of the successfully stained preparations, 10 were reconstructed by means of the AMIRA software − five including stained afferents emanating from the medial sensillum and five from the lateral.
In order to compare the target areas of gustatory receptor neurons that originated from the medial and lateral sensilla styloconica, respectively, the reconstructed neurons and surrounding neuropils structures from individual preparations were brought into the same frame. This frame was made from the preparation being best labeled via the synapsin immunostaining technique. Briefly, the label images of the preparation with stained neurons were affine- and elastically registered to the corresponding label images of the frame. The resulting transformation parameters for the neuropil structures were subsequently applied to the reconstructed neurons. The results were then carefully evaluated by comparing the confocal images with the obtained model.