Appearance, metabolism, blood supply, and identification of the elephant trigeminal nuclei

A, ventral view of the brain of Burma, a 52-year-old female Asian elephant.

B, ventral view of the brainstem of Burma. Note the large pons (upper part of the photograph). Posterior to the pons a pair of prominent bumps are visible, the putative trigeminal nuclei (TN) of the elephant. The trigeminal nuclei bumps are observed, wherein in humans and other mammals the inferior olive is found and was referred to as inferior olive by previous authors (Shoshani et al., 2006; Maseko et al., 2013; Rasenberger, 2019). The inferior olive can be identified at an unusual lateral position in elephants, however (IO).

C, 60 µm coronal section through the trigeminal nuclei of Raj, a four-year-old elephant bull, stained for cytochrome-oxidase reactivity, a mitochondrial enzyme, the reactivity of which reflects constitutive metabolic activity. The trigeminal nuclei show some of the strongest cytochrome-oxidase reactivity (indicated by the brown color) in the elephant brain and strong cytochrome-oxidase reactivity is typical for the trigeminal nuclei of many mammals.

D, a drawing of putative trigeminal subnuclei stained in C. Note the compact shape of the trunk module, which is unlike the inferior olive of mammals.

E, drawing of a coronal section stained for cytochrome-oxidase reactivity through the brainstem of Indra a female African elephant, borders of nuclei are outlined.

F, upper, micrograph of cytochrome-oxidase reactivity in the putative trunk fingertip representation. Lower, drawing of cytochrome-oxidase reactive erythrocytes in blood vessels. Note the high density of vessels inside of the fingertip but not in the surrounding tissue.

G, drawing of the entire brainstem section. The putative trunk module stands out from the rest of the brainstem in terms of vessel density.

H, quantification of blood vessel length in various parts of the brainstem. Note that not all vessels contain erythrocytes and are stained, i.e., measures of blood vessel length are lower bound estimates.

I, Brightfield micrograph of a parasaggital section through the trigeminal trunk module. Note the compact cellular architecture.

J, Brightfield micrograph of a parasaggital section through the inferior olive. Note the banded cellular architecture that is characteristic of the mammalian inferior olive.

L = lateral; V = ventral; P = posterior.

Peripherin-antibodies reveal climbing fibers and the olivo-cerebellar tract connecting the cerebellum to the inferior olive but not to the trigeminal nucleus

A, ventral view of the elephant brainstem with key structures schematically superimposed.

B, schematic of the elephant brainstem. We indicate the positions of the peripherin-stained sections shown in C-F and our delineation of the olivo-cerebellar tract based on peripherin-reactive putative climbing fibers.

C, left, overview picture of a brainstem section stained with anti-peripherin-antibodies (white color). Anti-peripherin-antibodies stain climbing fibers in a wide variety of mammals. The section comes from the posterior brainstem of African elephant cow Bibi; in this posterior region, both putative inferior olive and trigeminal nucleus are visible. Note the bright staining of the dorsolateral nucleus, the putative inferior olive according to Reveyaz et al. (the trigeminal nucleus according to Maseko et al., 2013). C, upper right, high magnification view of the dorsolateral nucleus (corresponding to the upper red rectangle in A). Anti-peripherin-positive axon bundles (red stars, putative climbing fibers) are seen in support of the inferior olive hypothesis of Reveyaz et al. C, lower right, high magnification view of the ventromedial nucleus (corresponding to the lower red rectangle in A). The ventromedial nucleus is weakly positive for peripherin but contains no anti-peripherin-positive axon bundles (i.e. no putative climbing fibers). Note that myelin stripes – green stars, weakly visible as dark omissions – are clearly anti-peripherin-negative. The region around the ventromedial nucleus is devoid of peripherin-reactivity and this is true throughout the brainstem. This observation suggests that the putative trigeminal nucleus does not receive climbing fiber input.

D, anti-peripherin-antibody (bright color) stained section below the cerebellar peduncle. Putative climbing fibers (arrow) can be seen to be budding off the olivo-cerebellar (red star) tract into the cerebellar peduncle.

E, anti-peripherin-antibody (bright color) stained section of the elephant cerebellum. The cerebellar white matter is bright, the cerebellar granule cell layer has a light grey appearance, and the cerebellar molecular layer has a dark grey appearance. Accordingly, peripherin-reactivity mirrors the distribution of cerebellar climbing fibers (white matter > granule cell layers > molecular layer).

F, high magnification view of an anti-peripherin-antibody stained section of the elephant cerebellum. Climbing fibers are apparent by their elongated axonal pattern (arrow). Purkinje cells appear as bright dots. Even higher magnification revealed that the inside of Purkinje cells is dark (anti-peripherin-negative), i.e. the bright dot appearance of Purkinje cells reflects the ensheating of Purkinje cells by climbing fibers. wm = white matter; gcl = granule cell layer; pcl = Purkinje cell layer; ml = molecular layer.

Overview of trigeminal nuclei in coronal and horizontal sections of African elephant cow Indra

A, micrograph of Nissl-stained coronal 60-µm-section through the right-hemispheric principalis trunk module of elephant cow Indra. The principalis is the most anterior and by far the largest trigeminal representation in elephants.

B, drawing of color-coded trigeminal modules belonging to the principalis nucleus.

C, drawing of an African elephant head with different facial structures color-coded according to the trigeminal modules they putatively correspond to in A, B.

D, left micrograph of Nissl-stained horizontal section through the left-hemispheric trigeminal nuclei of elephant cow Indra. The section is positioned at a dorsal level of the trigeminal nuclei. Right, drawing of trigeminal modules shown in the left micrograph; for the principalis (Pr5) module the same color code as in B, and C has been used. Sp5o refers to the trunk module directly posterior to the principalis nucleus. The facial nucleus and the putative inferior olive were also identified.

E, conventions as in D. Horizontal section through the mid-level of the trigeminal nuclei. Sp5i TM and Sp5c TM, refer to putative trunk modules posterior to the Sp5o TM.

F, conventions as in D. Horizontal section through the ventral level of the trigeminal nuclei.

Pr5, principal trigeminal nucleus; Sp5o, spinal trigeminal nucleus pars oralis; Sp5i, spinal trigeminal nucleus pars interpolaris; Sp5c, spinal trigeminal nucleus pars caudalis; TM = trunk module; A = anterior; L = lateral; V = ventral.

Cellular organization of the putative trunk module in the elephant trigeminal nuclei.

A, Golgi stained 200 µm coronal section through the trunk module of Raj, a four-year-old Asian elephant bull.

B, bundles of very thick axons are revealed by the Golgi staining (red arrows). Bundles were oriented orthogonal to the main axis of the module except for the putative finger representation, where they curved around.

C, Golgi-stained neurons are also observed albeit at a low frequency.

D, well-stained and well-preserved neurons reconstructed with a Neurolucida system. We show 47 neurons reconstructed from three adjacent coronal Golgi sections superimposed. Note the small size of the neurons relative to the module.

E, left (green), ten putative astrocytes. These small cells are the most abundant cellular element in the trigeminal nucleus. Right, four neuronal reconstructions are shown at higher magnification. Putative principal cells (three shown in black, total n = 41) had large somata and branched dendrites. A few cells (one cell shown in red, total n = 6) had small somata and unbranched dendrites. Dendritic trees were weakly polarized.

F, upper, raw polar plot of the orientation of neuronal dendritic segments (from all putative principal cells and putative interneurons, n = 47) relative to the soma confirms the common elongation of dendrites. Lower, when cells were aligned to the local axon bundle orientation an even stronger polarization of dendrites is evident.

G, antibody staining of neurons (green fluorescence, NeuN-antibody) and nuclei of all cells (blue fluorescence, DAPI) of a coronal section through the putative trunk module of Indra, a 34-year-old female African elephant. Neuron density is low.

H, high magnification view of the section shown in G, non-neural cells outnumber neurons by about a hundredfold (data refer to neuron and non-neural cell counts from three elephant trigeminal nuclei).

I, upper, somata drawing from a Nissl stained 60 µm coronal section through the putative trunk module of Indra. Lower, cells from the medial, the putatively proximal trunk representation of the module, and the lateral, the putatively distal trunk representation. Note the soma size difference.

J, plot of soma area along the length of the module. Neurons were sequentially measured along the axis of the module. Each dot refers to one of 1159 neurons in the section; red running average (across 40 neurons) of soma area. Cells are significantly larger in lateral (putatively distal trunk representation), unpaired T-test.

FT, putative dorsal Finger Tip representation; V = ventral; L = lateral.

Trunk module myelin stripes form a precise map of trunk folds

A, a brightfield image of a freshly cut 60 µm coronal section through the center of the putative trunk module of adult female elephant Indra. Neurons are evident as small white dots and whitish myelin stripes are readily apparent even in this unstained tissue. B, a fluomyelin stain (green fluorescence) confirms that stripes contain myelin. High-resolution brightfield microscopy (not shown) and Golgi stains (Figure 2B) confirm that the stripes consist of axon bundles.

C, a line drawing (red) of myelin stripes superimposed to the micrograph shown in A.

D, upper, enlarged view of line drawing (red) of myelin stripes shown in C; we quantified dorsally ending, ventrally ending, and full transversal stripes. Such numbers match the number of trunk folds quantified in E.

Lower, based on the idea of a match of trunk folds with myelin stripes one can compute magnification factors across the trunk module. Neural data (green) refer to distances between dorsally ending myelin stripes (i.e., neural distances were measured along the dorsal border (green line) of the module). Trunk distances were measured between dorsal trunk folds.

E, drawing of the dorsal folds of the trunk of Indra.

F, the composite photograph of the dorsal trunk of Indra. We counted dorsal folds, ventral folds (not visible here), and folds that fully transversed the right trunk side (not visible here). Trunk folds match in number, orientation (typically transversal), and patterning with myelin stripes seen on the trunk module.

V = ventral; L = lateral.

Morphological properties of elephant trigeminal cells in the Asian elephant Raj.

Data (mean ± SD) comes from Golgi-stains soma diameter which was defined as the maximal Feret diameter. Data relate to n = 20 for putative astrocytes, n = 41 for putative principal neurons, and n = 6 for putative interneurons. In unpaired t-tests, all morphological parameters were significantly different between groups. See Figure 2.

Microscopic organization of myelin stripes and absence of a strong relation of stripes to trigeminal neurons

A, upper, schematic of the trunk module with myelin stripes (grey) of African elephant Bambi and targeting of the 8 mm tissue punch (dashed red circle). Lower, synchrotron radiation (red flash, DESY, Hamburg) was directed to the area of the punch (black box) and imaged. V = ventral; L = lateral.

B, sketch of the parallel beam setup of the GINIX endstation (P10 beamline, DESY, Hamburg). In this geometry, a dataset of the trunk module was acquired at an effective voxel size of 650nm3

C, dimensions of the imaged volume shown as a volume rendering (0.65 µm isotropic voxel size)

D, transparent image volume. Two myelin stripes were followed through the volume image (highlighted red axon bundle). Four large diameter (∼ 15 µm) axons were also reconstructed and could be followed through the entire volume image (blue).

E, axon bundles (red) and reconstructed axons (blue) in isolation.

F, image section in the coronal plane at the center of the myelin stripe. The myelin stripe (pink overlay) is readily visible, the reconstructed axon is highlighted in blue, and the bundle has a width of about 7 myelinated axons.

G, cross-section through the axon bundle (pink overlay). The virtual section is cut orthogonal to the coronal plane. The reconstructed axon is highlighted in blue.

H, image section orthogonal to the coronal plane, the virtual section is cut in an anterior-posterior direction parallel to the axon bundle (pink overlay). The myelin stripe is readily visible, the reconstructed axon is highlighted in blue, and the bundle has a height of about 7 myelinated axons.

I, left, myelin stripes. Right, measurement of the thickness (width orthogonal to the main stripes axis) along the dorsoventral axis of stripes; data are the average of ten measurements along ten myelin stripes that fully transversed the module. Stripe thickness is fairly constant, there is no evidence of stripe tapering as would be expected if axons bud off into the tissue.

J, upper, measurement of the thickness (width orthogonal to the main stripes axis) of myelin stripes across the putative trunk module; only full transversal stripes were measured. Stripe thickness is fairly constant. Lower, neuron number between full transversal myelin stripes. Neuron number varies more than 100-fold and is very low (zero) between medial stripes (in the putatively proximal trunk representation). The fact that stripe thickness changes little across the module, while neuron number between stripes changes massively argues against a relationship between stripe thickness and trigeminal neuron number.

Differences between the putative trigeminal trunk modules of Asian (Elephas maximus) and African (Loxodonta africana) elephants

A, left, ventral view of the brainstem of African elephant Indra. Right, ventral view of the brainstem of Asian elephant Dumba. The anterior-posterior length of the trigeminal nuclei from the pons is indicated as a black line. The average trigeminal nuclei length refers to 10 African and 10 Asian trigeminal nuclei, the p-value refers to a Mann-Whitney test. The trigeminal nuclei bump is more elongated in African than in Asian elephants.

B, neuron number (left) and volume (right) of the putative principalis trunk module in Asian and African elephants. Trigeminal nuclei come from three African and three Asian elephants; p-values refer to unpaired t-tests.

C, micrograph of a cytochrome-oxidase-stained section through the putative trunk module of African elephant Indra.

D, drawings of the outlines and myelin stripes from cytochrome oxidase or Nissl stained sections through the putative trunk module of African elephants; the top drawing was made from the micrograph shown in C. The black line refers to the point of greatest width along the direction of myelin stripes on the putative trunk shaft (the putative finger was not considered in the width analysis).

E, micrograph of a cytochrome-oxidase-stained section through the putative trunk module of Asian elephant Raj.

F, drawings, of putative trunk modules from Asian elephants; conventions as in D.

G, length, and width of putative trunk module in African and Asian elephants. p-values refer to t-tests.

H, upper, drawing of the trunk of an African elephant. Lower, drawing of the trunk of an Asian elephant; note that Asian elephants have more folds. Arrows mark the point of greatest width of the putative trunk module (black lines in D, F) projected back on trunk positions in African (upper) and Asian (lower) elephant trunks. We also highlighted in color the dorsal (red) and ventral (pink) trunk tip and wrapping zone of Asian elephants in green and the analogous trunk part of African elephants in light green. The extent of the trunk wrapping zone was determined from photographs of Asian elephants wrapping objects. Specifically, we defined the wrapping zone as the trunk parts in contact with large objects (mangos, melons, fodder beets) being wrapped.

I, object grasping/pinching behavior in African (upper) and object wrapping strategy in Asian (lower) elephants (adapted from Kaufmann et al., 2022).

Overview of elephants and treatment of the corresponding specimen

Optical fractionator counts of the principalis trunk module

Cell count estimates were derived using the optical fractionator; every twentieth section was sampled. See text for details.

Optical fractionator counts for other trigeminal modules

Cell-count estimates were derived using the optical fractionator; every twentieth section was sampled. See text for details.