Quantitative comparison of cerebral artery development in metatherians and monotremes with non‐human eutherians

Abstract

A quantitative comparison of the internal diameters of cerebral feeder arteries (internal carotid and vertebral) and the aorta in developing non‐human eutherians, metatherians and monotremes has been made, with the aim of determining if there are differences in cerebral arterial flow between the three infraclasses of mammals such as might reflect differences in metabolism of the developing brain. There were no significant differences between eutherians and metatherians in the internal radius of the aorta or the thickness of the aortic wall, but aortic internal radius was significantly smaller in developing monotremes than therians at the < 10 mm body length range. Aortic thickness in the developing monotremes also rose at a slower rate relative to body length than in metatherians or eutherians. The sums of the internal calibres of the internal carotid and vertebral arteries were significantly lower in metatherians as a group and monotremes compared with non‐human eutherians at body lengths up to 20 mm and in metatherians at > 20 mm body length. The internal calibre of the internal carotids relative to the sum of all cerebral feeder arteries was also significantly lower in monotremes at < 10 mm body length compared with eutherians. It was noted that dasyurids differed from other metatherians in several measures of cerebral arterial calibre and aortic internal calibre. The findings suggest that: (i) both aortic outflow and cerebral arterial inflow may be lower in developing monotremes than in therians, particularly at small body size (< 20 mm); (ii) cerebral inflow may be lower in some developing metatherians than non‐human eutherians; and (iii) dasyurids have unusual features of cerebral arteries possibly related to the extreme immaturity and small size at which they are born. The findings have implications for nutritional sourcing of the developing brain in the three infraclasses of mammals.

Introduction

The three infraclasses of mammals have strikingly distinct modes of reproduction that have implications for central nervous system development. The young of monotremes (platypus and echidnas) develop in a leathery‐shelled egg before emerging at a body length of about 14 mm, in a developmental state similar to a eutherian fetus. Monotreme young must then repeatedly locate the nipple‐less areola to feed, and their brains develop slowly over a period of many months (Griffiths, 1968, 1978). Metatherians are also born in an immature state and rely on a prolonged period of maternal lactational support in a maternal pouch or abdominal depression, but metatherian young exhibit diverse dependencies depending on the order they belong to (Tyndale‐Biscoe & Renfree, 1987; Tyndale‐Biscoe, 2005). Some are very small at birth (e.g. dasyurids, about 10–15 mg body weight with a body length of only 5 mm; Hill & Hill, 1955; Nelson et al. 2003) and have only a rudimentary embryonic forebrain, whereas others are much larger at birth (e.g. diprotodontids, 200–2000 mg body weight and body lengths up to 17 mm) and have a newborn forebrain similar to a late gestation rodent fetus (Tyndale‐Biscoe, 2005). Evidence indicates that metatherians in general have a slower pace of neural development compared with eutherians (Darlington et al. 1999; Workman et al. 2013). Eutherian young all have a period of maternal support by a placenta, although the state of neural maturation at birth and the subsequent period of dependency can vary markedly between eutherian families.

An important question in comparative brain development is how various developing mammalian brains differ in their metabolic demands. The very different modes of growth and support for mammalian young in the three infraclasses raise the question of whether the metabolic demands of the developing brain differ between the three groups. Do monotreme and metatherian young of a given body size have the same cerebral metabolic demands as eutherians? Are differences in body size at birth among metatherians associated with differences in cerebral metabolic costs? These questions are fundamental to understanding the metabolic demands of brain development in different mammals, and hence the link between brain development, maternal metabolism and evolution. It is not technically possible to directly measure the metabolic demands of the embryonic brain in a variety of species, but there may be structural clues that can be used to indirectly assess brain metabolism. For all mammals, nutrients and oxygen to the developing brain must pass through the internal carotid and vertebral arteries. Poiseuille's law states that the volume rate of flow through a vessel is proportional to the difference in pressure along the vessel length, proportional to the internal radius of the vessel raised to the fourth power, and inversely proportional to the length of the vessel and viscosity of the fluid. This means that the internal radii of cerebral arteries play critically important roles in determining blood flow, such that a doubling of internal radius leads to a 16‐fold increase in flow. Small changes in the internal radius of particular arterial pairs could allow the developing embryo to direct blood flow (and hence nutrients) to key areas. Quantitative analysis of changes in the internal radii of cerebral arteries may therefore provide structural clues to cerebral arterial flow and developmental changes in the metabolic demands of the brain and its components.

In a previous study it was shown that human embryos have a pattern of growth in arterial supply to the developing brain that is rather different from non‐human eutherians (Ashwell & Shulruf, 2015). Although the developing human brain appears to receive less arterial flow at embryonic sizes (< 22 mm body length) than do other eutherian embryos of a similar body size, internal carotid and vertebral arterial flow is higher in human fetuses (body length > 30 mm) than in developing non‐humans of the same body size. Increased flow to the developing human brain relative to non‐humans is achieved by simultaneous increases in both aortic and cerebral feeder artery internal calibre. This enhanced flow appears to be associated with an explosive expansion of the pallial and ganglionic proliferative cell population after 22 mm body length (Ashwell, 2015b). These findings naturally raise the question of how developing cerebral feeder artery calibre varies across mammalian infraclasses and how this relates to reproductive strategies.

The aim of the present study was to analyse the internal calibre of the major arteries supplying the embryonic and early postnatal brain in metatherians and monotremes, and compare this with non‐human eutherian embryos. In order to reflect the impact of vessel internal radius on flow (by Poiseuille's law), the internal radius raised to the fourth power (i.e. r ⁴) or the sum of that for paired arteries (Σ r ⁴) was used in the comparisons. Also, the r ⁴ of the aorta as an indicator of cardiac output was compared with the average thickness of the aortic wall as an indicator of developmental change in aortic pressure. The general hypothesis was that the internal calibre of cerebral arteries supplying the developing metatherian and monotreme brain would be consistently less than that for arteries supplying the brains of eutherians. It was also hypothesized that the proportion of aortic blood flowing to the developing brain would be greater in eutherians than in metatherians and monotremes.

Materials and methods

Specimens

Most of the data for this study were obtained using sections of the head, neck and upper trunk in more than 600 embryos and fetuses from the Museum für Naturkunde (MfN) in Berlin. Specimens of Macropus eugenii and Rattus norvegicus were from the author's collection (KWA). Details of the species analysed are set out in Tables 1 and 2. The supporting information table contains the full list of specimens with measurements. Unfortunately the eutherian sample does not include specimens from non‐hominoid eutherian species with large adult brain size (e.g. cetaceans, proboscidians). Human embryos were excluded from the analysis because they appear to have an unusual pattern of cerebral artery development compared with other eutherians (Ashwell & Shulruf, 2015). Embryos and fetuses with abnormalities or data indicating potential pathology were also excluded from the study.

Table 1.

Summary of non‐human eutherian specimens from the MfN and author's collection

Order	Family/superfamily	Species	No. of specimens
Hyracoidea	Procaviidae	Procavia capensis	1
Xenarthra	Bradypodidae	Bradypus tridactylus	1
	Myrmecophagidae	Tamandua tetradactyla	2
Cingulata	Dasypodidae	Dasypus novemcinctus	2
Macroscelidea	Macroscelididae	Macroscelides sp.	1
Afrosoricida	Chrysochloridae	Chrysochloris sp.	1
Soricomorpha	Soricidae	Crocidura sp.	8
		Sorex sp.	16
		Suncus caeruleus	3
	Talpidae	Talpa europaeus	33
Erinaceomorpha	Erinaceidae	Erinaceus europaeus	33
Chiroptera	Vespertilionidae	Pipistrellus kuhlii	1
	Pteropodidae	Rousettus (Xantharpya) amplexicaudatus	45
		Pteropus edulis	1
Scandentia	Ptilocercidae	Ptilocercus sp.	1
Primata	Lemuridae	Lemur melanocephalus	2
	Cheirogaleidae	Microcebus murinus	3
	Indriidae	Propithecus sp.	1
	Galagidae	Galagoides demidovii	5
	Lorisidae	Loris sp.	5
	Tarsiidae	Tarsius sp.	22
	Cercopithecidae	Cercocebus cynomolgus	5
		Macaca fascicularis (irus)	3
		Nasalis larvatus	1
		Papio sp.	1
		Presbytis cristata (pyrrhus)	1
		Semnopithecus sp.	6
		Trachypithecus barbei	1
	Atelidae	Alouata sp.	1
	Callitrichidae	Callithrix (Hapale) jacchus	4
	Pitheciidae	Pithecia sp.	2
Carnivora	Felidae	Felis catus	30
	Canidae	Canis lupus familiaris	13
	Mustelidae	Mustela erminea	2
	Viverridae	Nandinia binotata	1
Artiodactyla	Suidae	Sus scrofa	20
	Tragulidae	Tragulus javanicus	1
	Bovidae	Bos taurus	1
		Capra hircus	1
		Ovis aries	23
Pholidota	Manidae	Manis javanica	15
Rodentia	Muridae	Rattus norvegicus	31
Lagomorpha	Leporidae	Oryctolagus cuniculus	56

Table 2.

Summary of metatherian and monotreme specimens from the MfN and author's collection

Order	Family/superfamily	Species	No. of specimens
Diprotodontia	Acrobatidae	Acrobates pygmaeus	1
	Petauridae	Petaurus sp.	1
	Pseudocheiridae	Pseudocheirus convolutus	1
	Macropodidae	Macropus eugenii	32
		Macropus giganteus	1
		Macropus parma	1
		Macropus rufogriseus	5
		Petrogale penicillata	4
		Thylogale thetis	1
	Phascolarctidae	Phascolarctos cinereus	12
	Vombatidae	Vombatus ursinus	6
	Phalangeridae	Trichosurus vulpecula	49
Peramelemorphia	Peramelidae	Isoodon obesulus	11
		Perameles nasuta	12
Dasyuropmorphia	Dasyuridae	Dasyurus viverrinus	39
Didelphimorphia	Didelphidae	Didelphis of various species	17
	Caluromyidae	Caluromys philander	1
Monotremata	Ornithorhynchidae	Ornithorhynchus anatinus	12
	Tachyglossidae	Tachyglossus aculeatus	8

The embryonic and early fetal specimens from MfN had been fully dehydrated, embedded in paraffin or celloidin, and sectioned at 8 μm or greater thickness in the transverse plane, before being stained with carmine, hematoxylin, hematoxylin and eosin, reduced silver, or alcian blue and nuclear red. Data for the tammar wallaby and laboratory rat were obtained from sections prepared previously for published studies in those species (Ashwell, 1986; Ashwell et al. 1996; Song et al. 2000; Hassiotis et al. 2002; Ashwell & Paxinos, 2007; McCluskey et al. 2008).

Photography and measurements

The material at MfN was photographed with the aid of either a Zeiss Axioplan2 fitted with an AxioCam MRc5 camera, or with a Leica M420 macroscope fitted with an Apozoom 1 : 6 lens and Leica DFC490 camera. All images were calibrated by photographing a scale bar at the same magnification. Photographs of the vessels of R. norvegicus were photographed with a Dino‐Lite digital microscope linked to a MacBook Pro running dinoxcope v1.3.3. Arterial internal diameter was measured in sections through M. eugenii heads with the aid of magellan 3.1 (Halasz & Martin, 1995).

Most of the specimens used in the study carry unique identifying museum catalog numbers that give no clue as to body length. This means that photomicrographs could be taken and measurements made with the observer blind as to the size of the specimens. imagej was used to make measurements on images of cerebral arteries after calibration (Rasband, 1997). Vessels were followed through adjacent sections to determine the course of the vessel relative to the section plane. Where the artery in question passed at right angles to the section plane and had a circular profile, the internal diameter was measured as the distance between the internal endothelial surface of one side and that of the other, along the diameter of the vessel profile. Where the vessel passed obliquely through the section plane, the internal diameter was measured along the minor axis of the elliptical profile of the arterial lumen.

The internal diameter of the aorta was measured at a point between the ductus arteriosus and the first cephalic branch (brachiocephalic trunk, its precursor or variant). The external diameter of the aorta (outer adventitia of one side to adventitia of the other) was also measured to calculate average aortic thickness as a proxy of aortic pressure (see above).

The internal diameter of the internal carotid artery was measured at a point alongside the pituitary gland or its precursor. All the ungulate embryos used in this analysis were at a developmental stage before the formation of the rete mirabile, so all specimens had a pair of internal carotid arteries at this point. In older embryos, this region is defined by the development of the cavernous segment of the internal carotid artery. The vertebral arteries were usually analysed as they passed through the foramina transversaria of the third or fourth cervical vertebra, a point at which the vast majority of vertebral arteries have entered the foramina transversaria in the species studied. Occasionally the vertebral artery was measured higher up in the neck or within the cranial cavity when the mid‐cervical level was not available.

Analysis and statistics

By Poiseuille's law, volume rate of flow is proportional to the internal radius of the vessel raised to the fourth power. The fourth power of the internal radius in μm (r ⁴, or the sum of this for vessel pairs, e.g. Σ ic r ⁴) was therefore used to compare putative flow between vessels. Log transformation was applied where values of both vessel diameter and embryo size had large ranges. The mean aortic wall thickness in μm was calculated as the difference between internal and external diameters along the minor axis of the elliptical or circular profile, divided by two. Non‐parametric statistics (Mann–Whitney test) were applied to test for significant changes in values of r ⁴ between groups. When comparing the three mammalian groups, the developmental continuum was divided into three body size ranges: (i) < 10 mm body length (a period before cortical plate aggregation in almost all mammals; Ashwell, 2015a,b); (ii) 10–20 mm body length (a period during which cortical plate aggregation begins in most mammals; Ashwell, 2015a,b); and (iii) > 20 mm body length (a period when the cortical plate expands and pallial connections begin to form; Ashwell, 2015a,b).

Results

Identification of cerebral arteries

All the arteries used in the quantitative analysis could be readily identified in all the species, and the internal carotid and vertebral arteries of most metatherians and the two monotremes were qualitatively indistinguishable from those in eutherians of the same body size. Nevertheless, there were some qualitative differences noted between the dasyurids (metatherian carnivores such as the quolls) and other metatherians (bandicoots, wombats, koalas, possums, kangaroos), monotremes and eutherians. Figure 1 shows examples of the internal carotid and vertebral arteries in representative diprotodontid, peramelid and dasyurid pouch young at 14–15 mm body length. In all the developing eutherians, both monotremes and all the metatherians except the dasyurids, the internal carotid arteries were relatively delicately walled vessels, surrounded by loose connective tissue (Fig. 1A,B). By contrast, the internal carotid arteries of the dasyurids (Fig. 1D) were of consistently narrow lumen, with thick walls relative to lumen size, and surrounded by relatively dense connective tissue. A similar contrast was noticeable for the vertebral arteries (Fig. 1D–F) in that the vertebral arteries of dasyurids had a narrow lumen and were surrounded by much denser connective tissue than in the eutherians, monotremes or other metatherians.

Figure 1.

Photomicrographs contrasting the internal carotid arteries (A–C) and vertebral arteries (D–F) in the pouch young of a representative diprotodontid metatherian (A, B; brush‐tailed possum), peramelid metatherian (C, D; bandicoot) and dasyurid (E, F; eastern quoll) of 14–15 mm body length. All photomicrographs were taken at the same magnification, and the scale bar in (A) applies to all images. Note the contrast in arterial lumen size, relative wall thickness of the artery and density of surrounding connective tissue in the dasyurid (eastern quoll) pouch young compared with the possum and bandicoot.

Developmental changes in aorta r ⁴ and aortic wall thickness

Figure 2 shows the log₁₀ of the aortic r ⁴ plotted against log₁₀ body length for non‐human eutherians, metatherians and monotremes. In all three infraclasses, aortic r ⁴ rose with body length, but the correlation was negligible (R ²: eutherians – 0.028; metatherians – 0.075; monotremes – 0.064). There was a wide range in aortic r ⁴ values for each of the groups, and there was no significant difference in aortic r ⁴ between metatherians as a group and non‐human eutherians across the body length range studied here (Fig. 3). However, aortic r ⁴ was significantly lower in monotremes in the < 10 mm body length range (early pre‐hatching stage) than in non‐human eutherians of the same body length (Fig. 3). This difference (a factor of 10^0.66) potentially amounts to a fourfold difference in aortic flow if all other factors were equal, but occurs within a wide range for both mammalian groups.

Figure 2.

Graph showing the log₁₀ of the aortic r ⁴ plotted against log₁₀ body length for non‐human eutherians, metatherians and monotremes. Note the similar slope to non‐human eutherians, but lower y‐intercept, for the regression line for monotremes.

Figure 3.

Graph showing the mean values (with 95% confidence intervals) of aortic r ⁴ (internal radius raised to the fourth power) for non‐human eutherians, metatherians and monotremes at three body sizes (< 10 mm body length, 10–20 mm body length, > 20 mm body length).

The log₁₀ of aortic wall thickness is plotted against log₁₀ body length in Fig. 4. This is being used as a proxy for aortic pressure, as the internal structure of the aortic wall appears to be similar in the embryos of all the mammals studied. All three mammalian groups show a rise in aortic thickness with body length with significant correlation (R ²: non‐human eutherians – 0.398; metatherians – 0.477; monotremes – 0.208). The regression lines for metatherians and non‐human eutherians are very similar in slope and not significantly different. On the other hand, aortic thickness rose more slowly with body length among the monotremes compared with the non‐human eutherians and metatherians, suggesting that aortic pressure rises more slowly with increasing body size in developing monotremes compared with the therians.

Figure 4.

Plot of log₁₀ aortic wall thickness against log₁₀ body length for non‐human eutherians, metatherians and monotremes. The regression lines for non‐human eutherians and metatherians have very similar slopes and y‐intercepts, but that for monotremes has a lower slope. Nevertheless, values for all three mammalian groups are very similar at the 10 mm body length size.

Developmental changes in r ⁴ of the internal carotid and vertebral arteries

Figures 5 and 6 show plots of log₁₀ r ⁴ for internal carotid (ic r ⁴) and vertebral artery (vert r ⁴) pairs, respectively, against log₁₀ of body length for non‐human eutherians, metatherians and monotremes. There was a negligible correlation between ic r ⁴ and body length for all three mammalian groups (R ² for ic r ⁴: non‐human eutherians – 0.088; metatherians – 0.014; monotremes – 0.002). Correlation with body length was also negligible for vert r ⁴ for metatherians and monotremes, but poor for eutherians (R ² for vert r ⁴: non‐human eutherians – 0.123; metatherians – 0.012; monotremes – 0.068). However, values of r ⁴ for metatherians and monotremes are consistently lower than non‐human eutherians. When the sum of r ⁴ for the cerebral feeder arteries (i.e. paired internal carotid and vertebral arteries; Σ ic r ⁴ + vert r ⁴) was calculated for all three infraclasses at three different body sizes (Fig. 7), significant differences were found between the three mammalian groups. In the < 10 mm body length range, Σ r ⁴ for the four cerebral feeder arteries was significantly lower in both metatherians and monotremes compared with non‐human eutherians. This difference continued into the 10–20 mm body length and > 20 mm body length ranges for the metatherians, although the paucity of monotreme specimens in the latter two size groups made the difference from non‐human eutherians not statistically significant.

Figure 5.

Plot of log₁₀ r ⁴ for internal carotid artery (ic r ⁴) pairs against log₁₀ of body length for non‐human eutherians, metatherians and monotremes. Both metatherians and monotremes have lower values for ic r ⁴ than non‐human eutherians in the 10–100 mm body length range.

Figure 6.

Plot of log₁₀ r ⁴ for vertebral artery (vert r ⁴) pairs against log₁₀ of body length for non‐human eutherians, metatherians and monotremes. Both metatherians and monotremes have lower values for vert r ⁴ than non‐human eutherians in the 10–100 mm body length range.

Figure 7.

The mean values (with 95% confidence intervals) of the sum of r ⁴ for the cerebral feeder arteries (i.e. paired internal carotid and vertebral arteries; Σ ic r ⁴ + vert r ⁴) calculated for all three infraclasses at three different body sizes. Note the significantly lower values for metatherians and monotremes compared with non‐human eutherians at all three body size phases.

The proportion of ln r ⁴ for ic arterial pairs over the total r ⁴ for the internal carotid and vertebral artery pairs was calculated for the three body length periods as an indicator of proportional flow through the rostral cerebral circulation. The results are shown in Fig. 8. The proportion of feeder flow to the brain that passes through the internal carotid arterial pair is significantly lower in both monotremes and metatherians compared with non‐human eutherians at the < 10 mm body length range. The trend is still evident at the 10–20 mm body length range, but not statistically significant because of paucity of monotreme specimens. At > 20 mm body length, there was no significant difference in the proportion for non‐human eutherians and metatherians.

Figure 8.

The mean values (with 95% confidence intervals) of the proportion of ln Σ r ⁴ for ic arterial pairs over the total Σ r ⁴ for the internal carotid and vertebral artery pairs calculated for the three body length periods for the three infraclasses of mammals. Note the lower values of this ratio for metatherians and monotremes at the < 10 mm body size phase, suggesting a lower proportional flow through the rostral cerebral circulation in those mammals compared with non‐human eutherians.

When trends were analysed for the three infraclasses in separate regression analyses for the three periods, the monotremes were found to exhibit a distinctly different trend from the therians (Fig. 9). The therians had a relatively poor correlation of the ln proportion with body length over all three body length size ranges (R ² < 0.100). By contrast, monotremes have a high correlation of the ln proportion of ic r ⁴ relative to all cerebral feeder flow for < 10 mm body length (R ²= 0.516) and 10–20 mm body length (R ²= 0.980). What is more, there was a distinct decline in the proportional flow through the internal carotid arteries with increasing log₁₀ body length for monotremes in the < 10 mm body length range (Fig. 9A). This was followed by a significant rising trend in proportional flow through the internal carotid arteries for the 10–20 mm body length range (Fig. 9A). There were insufficient specimens to properly analyse for trend in the > 20 mm body length range (Fig. 9C).

Figure 9.

Linear regression trend lines for the ln of the proportion of Σ r ⁴ for ic arterial pairs over the total Σ r ⁴ for the internal carotid and vertebral artery pairs, calculated for the three body length periods for the three infraclasses of mammals. The trend lines have a similar slope for metatherians and non‐human eutherians in (A) the < 10 mm and (B) 10–20 mm body length phases, but the trend lines for monotremes are rather different, showing at first a steep fall (< 10 mm body length phase) followed by a consistently steep rise (10–20 mm body length phase).

The log₁₀ Σ r ⁴ for internal carotid and vertebral artery pairs is plotted against log₁₀ r ⁴ of the aorta in Fig. 10. This correlation is meaningful for all three mammalian groups (R ²: non‐human eutherians – 0.284; metatherians – 0.316; monotremes – 0.285) with similar slopes. However, the regression lines for metatherians and monotremes are lower than that for non‐human eutherians. This is consistent across all body ranges for metatherians, amounting to a threefold difference in the ratio of cerebral feeder Σ r ⁴ to aortic r ⁴ between metatherians and non‐human eutherians (albeit within a wide intragroup variation). The ratio is similar for monotremes relative to non‐human eutherians in the 10–20 mm and > 20 mm body length ranges.

Figure 10.

Log₁₀ Σ r ⁴ for internal carotid and vertebral artery pairs plotted against log₁₀ r ⁴ of the aorta. The slopes of the regression lines for non‐human eutherians and metatherians are similar but vertically displaced relative to each other, suggesting a lower proportion of aortic flow distributed to cerebral feeder arteries in metatherians as a group. The regression line for monotremes also has a lower slope than in non‐human eutherians and suggests lower cerebral feeder flow than in eutherians at larger values of aortic r ⁴.

Internal calibre of the aorta and cerebral feeder arteries in dasyurids

The qualitative observations of blood vessel diameter in metatherians (Fig. 1) led us to test for a significant difference in aortic r ⁴, aortic wall thickness and r ⁴ of the internal carotid and vertebral arteries in 39 dasyurids as compared with all other metatherians (154 specimens). This analysis was conducted for the three body length periods used previously (i.e. < 10 mm, 10–20 mm, and > 20 mm body length). Non‐parametric statistics were applied (Mann–Whitney test) and the results summarised in Table 3. It was found that the aortic r ⁴ of dasyurids was significantly smaller than non‐dasyurid metatherians at < 10 mm greatest length (GL) and > 20 mm body length, but not significantly different at 10–20 mm body length. Despite the smaller aortic r ⁴ in dasyurids, there were no significant differences in aortic wall thickness between dasyurids and other metatherians at any body length. Internal carotid r ⁴ was also significantly smaller in dasyurids compared with non‐dasyurid metatherians at all body lengths, and vertebral artery r ⁴ was significantly smaller than non‐dasyurid metatherians at < 10 and at 10–20 mm GL, but not in larger pouch young (> 20 mm GL).

Table 3.

Summary of the results of Mann–Whitney U‐tests for differences in vessel calibre and thickness between the dasyurid and all other metatherians

Measurement	Values of P; dasyurid vs. non‐dasyurid metatherians
	< 10 mm GL	10–20 mm GL	> 20 mm GL
Aortic r 4	< 0.001	0.514	0.004
Aortic wall thickness	0.600	0.408	0.793
Internal carotid r 4	< 0.001	< 0.001	0.026
Vertebral r 4	< 0.001	0.007	0.637

GL, greatest length.

The finding for dasyurids led to ask whether non‐dasyurids have similar aortic, ic r ⁴ and vert r ⁴ compared with non‐human eutherians. In other words, could the current findings of a significant difference between metatherians and non‐human eutherians be due simply to the unusual features of one group of metatherians, i.e. the dasyurids. Figures 11, 12 and 13 show aortic r ⁴, ic r ⁴ and vert r ⁴ plotted against body length for the three periods examined (a: < 10 mm; b: 10–20 mm; and c: > 20 mm body length). The values for R ² for each of the linear regressions are shown in Table 4. It was found that the values of each were similar at all three body length periods for non‐dasyurid metatherians and non‐human eutherians (Figs 11, 12 and 13), although trend‐lines did differ for some of these analyses. Nevertheless, the regression lines across the animals did not significantly differ (Figs 11, 12 and 13). It can be concluded from this analysis that, with respect to the available data, values for aortic r ⁴, ic r ⁴ and vert r ⁴ in non‐dasyurid metatherians are similar to non‐human eutherians.

Figure 11.

Log₁₀ aortic r ⁴ for non‐dasyurid metatherians and non‐human eutherians plotted against body length for three body length phases (A: < 10 mm; B: 10–20 mm; C: > 20 mm body length). Regression lines and upper and lower confidence intervals are shown for each analysis. Fine lines have been used for these for non‐dasyurid metatherians, bold lines for non‐human eutherians.

Figure 12.

Log₁₀ internal carotid artery (ic) r ⁴ for non‐dasyurid metatherians and non‐human eutherians plotted against body length for three body length phases (A: < 10 mm; B: 10–20 mm; C: > 20 mm body length). Regression lines and upper and lower confidence intervals are shown for each analysis. Fine lines have been used for these for non‐dasyurid metatherians, bold lines for non‐human eutherians.

Figure 13.

Log₁₀ vertebral artery (vert) r ⁴ for non‐dasyurid metatherians and non‐human eutherians plotted against body length for three body length phases (A: < 10 mm; B: 10–20 mm; C: > 20 mm body length). Regression lines and upper and lower confidence intervals are shown for each analysis. Fine lines have been used for these for non‐dasyurid metatherians, bold lines for non‐human eutherians.

Table 4.

Summary of the results (R ²) of linear regression analyses for vessel calibre (aortic r ⁴, ic r ⁴ and vert r ⁴) for non‐dasyurid metatherians and non‐human eutherians

Measurement	R 2 for linear regression
	< 10 mm GL		10–20 mm GL		> 20 mm GL
	Non‐dasyurid metatherian	Non‐human eutherian	Non‐dasyurid metatherian	Non‐human eutherian	Non‐dasyurid metatherian	Non‐human eutherian
Aortic r 4	0.158	0.070	0.004	< 0.001	0.019	0.053
Internal carotid r 4	0.032	0.147	0.107	< 0.0001	0.209	0.222
Vertebral r 4	0.046	0.222	< 0.001	0.006	0.004	0.040

GL, greatest length.

Discussion

Limitations of the analysis

The collections at the MfN Berlin include a range of rare mammalian embryonic specimens that would be very difficult, or impossible, to obtain given the restrictions placed on collecting native mammals in the modern world. Nevertheless, there are some technical limitations arising from the way in which the tissue was processed. All the MfN specimens were collected and processed in the early 20th centuries and were fully dehydrated, which will inevitably cause shrinkage. Sectioning was relatively thick by modern standards (10 μm thickness or more). Staining was with appropriate methods for the time (e.g. hematoxylin and eosin, carmine), and it is not possible to apply modern immunohistochemical techniques to the tissue. In the present study it has been assumed that all specimens have been equally affected by dehydration and the analysis has been confined to measurements that can be reliably made in thick histological sections. Although comparison within the dataset can be made dependably, it may not be appropriate to make comparisons with the dimensions of unfixed vessels. The rodent and macropod specimens included in the dataset were also processed using traditional histological techniques, and were all fully dehydrated and paraffin embedded.

There are also significant drawbacks with the use of archived museum sections to analyse arterial development. The measurements are being made of a fixed and sectioned artery, which can at best only represent a snapshot in the presumably dynamic nature of cerebral arteries. Nevertheless, it is unlikely that in vivo information will be available concerning flow in embryonic mammalian cerebral arteries in the near future, so analysis of fixed tissue is the best current option for a study of this kind.

It should also be borne in mind that this sort of analysis cannot make any comment on the other components of Poiseuille's law. The aortic pressure that drives flow through the cerebral feeder arteries cannot be directly measured, but no significant difference in aortic thickness (a proxy for aortic pressure) between eutherians and metatherians across a body length range of 4–60 mm was found. Although developing monotremes had a lower slope for the regression of log₁₀ aortic thickness against log₁₀ body length than the therians, their aortic thickness is not significantly different from therians in the 10–20 mm body length range. Viscosity, oxygenation or nutrient concentrations in the blood to the developing brains of the diverse species considered here cannot be commented on.

Developmental changes in aortic r ⁴ and aortic wall thickness

It was observed that aortic r ⁴ and aortic wall thickness in non‐human eutherians and metatherians were very similar across the range of body lengths available, but that monotremes had lower aortic r ⁴ across the body length range studied and a steeper rise of aortic thickness with body length than the therians. Although the monotreme regression line for aortic thickness relative to body length is steeper than for therians, all three mammalian groups have essentially identical aortic thickness in the central part of the body length range. These observations suggest that therians have very similar cardiac and aortic hemodynamics, but that aortic hemodynamics among the monotremes may be developmentally different. Without an opportunity to conduct in vivo studies of the cardiac function of monotreme embryos and hatchlings, it is impossible to be more specific. This is unlikely to be feasible in the near future.

Developmental changes in cerebral arterial blood flow based on r ⁴ of the vessels

The current observations of ic r ⁴ and vert r ⁴ in metatherians and monotremes lead to the conclusion that both groups may have lower flow to the developing brain than non‐human eutherians of a similar body length. Furthermore, the proportional flow to the rostral circulation (internal carotid arterial pair) appears to be lower in metatherians and monotremes than in non‐human eutherians, particularly at < 10 mm body length. The further re‐examination of the metatherian data suggested that it is the unusual features of the cerebral arteries of pouch young dasyurids that account for the differences between metatherians and non‐human eutherians. In other words, the non‐dasyurid metatherians appear to have similar values for aortic r ⁴, ic r ⁴ and vert r ⁴ to non‐human eutherians of a similar body size. This implies that the arterial flow to the developing brains of non‐dasyurid metatherians is not apparently different from that to the non‐human eutherians, a group that includes a diverse range of taxa from soricomorpha to primates to ungulates. This in turn raises the question of what is functionally different about developing dasyurids and monotremes that could account for the apparent dissimilarities in developmental cerebral arterial supply.

What is unusual about dasyurids?

It was found that the dasyurids (actually specimens of the eastern quoll, Dasyurus viverrinus) had low aortic r ⁴, ic r ⁴ and vert r ⁴ compared with non‐dasyurid metatherians (and by implication non‐human eutherians), but no difference in aortic wall thickness was found, suggesting that arterial pressure may be similar in dasyurids to non‐dasyurid metatherians (and may be similar in all therian young of a similar body size). This implies low arterial flow to the brains of developing quolls. Qualitative examination of the histology gives the impression that the cerebral vasculature of quoll pouch young (i.e. internal carotid and vertebral arteries) are not only of small lumen, but also protected along their course by much denser encircling connective tissue than in the other therians, or even the monotremes. It is almost as if the vessels are armoured against external forces. Why would the cerebral arteries of a developing dasyurid be so different?

In fact, developing dasyurids show a number of features that set them apart from other metatherians. The most obvious is the extremely small size at which they are born. The dasyurids are arguably the most immature mammals at birth (Hill & Hill, 1955; Nelson et al. 2003), with a body size no bigger than a grain of rice (< 5 mm body length). As might be expected given the small birth size, the forebrain of newborn dasyurids is not greatly advanced beyond the neural tube stage. Nevertheless, there are some parts of the nervous system of newborn dasyurids that are actually developmentally advanced relative to body size. These are a sensory apparatus that contributes towards location of the teat and attachment (the trigeminal sensory pathway; Ashwell, 2015c) and the spinal cord locomotor apparatus that is essential for movement towards the milk source (Ashwell & Shulruf, 2014).

It may be that the forebrain of the dasyurid pouch young has either lower metabolic needs, or is permitted lower nutrient supply, than that in non‐dasyurid metatherians of the same body size. There are some quantitative data currently available concerning the pace of the parietal cortex relative to body length in metatherians that may throw some light on this. Cortical plate aggregation occurs at a slightly greater pallial thickness in dasyurids compared with other metatherians, although the rate of pallial growth in general is not greatly different from other metatherians (Ashwell, 2015a). Involution of the proliferative compartments also appears to occur at a greater pallial thickness in dasyurids than in other metatherians, suggesting a slower pace of cortical maturation. Taken together, these observations might suggest a lower rate of nutrient usage in the developing dasyurid forebrain, but it is not possible at present to make definitive conclusions on this because there are no quantitative data on development of subpallial parts of the dasyurid forebrain compared with other mammals.

What is unusual about monotremes?

The current observations suggest that monotremes have some distinct differences in cerebrovascular supply compared with eutherians. In particular, aortic internal radius was significantly smaller in developing monotremes than therians at the < 10 mm body length range. Aortic thickness in the developing monotremes also rose at a slower rate relative to body length than in metatherians or eutherians. The sums of the internal calibres of the internal carotid and vertebral arteries were significantly lower in monotremes compared with non‐human eutherians at body lengths up to 20 mm. Finally, the internal calibre of the internal carotids relative to the sum of all cerebral feeder arteries was also significantly lower in monotremes at < 10 mm body length compared with eutherians.

Monotremes are distinguished from all other mammals by developing for some time in a leathery‐skinned egg, before emerging to rely on milk secreted from a nipple‐less areola (Griffiths, 1968, 1978). The lateral pallium of the monotreme forebrain also appears to develop at a slower pace relative to body length than in therians (Ashwell, 2015a). Pallial thickness growth is slower in developing monotremes than in therians, and the proliferative compartments of the cortex (ventricular and subventricular zones) remain in evidence to a much greater body length and pallial thickness (Ashwell, 2015a). These observations suggest that nutrient utilisation by the developing monotreme cortex may be slower than in therians and could explain the quantitative differences in cerebral vasculature that have been observed.

Concluding remarks

The current findings highlight the unusual aspects of cerebral arterial development in dasyurids and monotremes. Metatherians are often lumped together in discussion of brain evolution (Darlington et al. 1999), but there is at least as much developmental diversity among metatherians as in eutherians, and several aspects of brain development in non‐dasyurid metatherians are comparable to eutherians. The special developmental features of dasyurids and monotremes may account for the apparently small calibre of the arterial vessels supplying their developing brains.

Supporting information

Table S1. Summary of data on vascular diameter and thickness for mammals used in this study.