Individual and age-related variation of cellular brain composition in a squamate reptile

Abstract

Within-species variation in the number of neurons, other brain cells and their allocation to different brain parts is poorly studied. Here, we assess these numbers in a squamate reptile, the Madagascar ground gecko (Paroedura picta). We examined adults from two captive populations and three age groups within one population. Even though reptiles exhibit extensive adult neurogenesis, intrapopulation variation in the number of neurons is similar to that in mice. However, the two populations differed significantly in most measures, highlighting the fact that using only one population can underestimate within-species variation. There is a substantial increase in the number of neurons and decrease in neuronal density in adult geckos relative to hatchlings and an increase in the number of neurons in the telencephalon in fully grown adults relative to sexually mature young adults. This finding implies that adult neurogenesis does not only replace worn out but also adds new telencephalic neurons in reptiles during adulthood. This markedly contrasts with the situation in mammals, where the number of cortical neurons declines with age.

1. Introduction

Even though numerous studies have pointed out the importance of brain neuron number, density and distribution as a proxy for brain processing capacity [1–4], individual variation in brain composition is virtually unknown. Data are available only for the laboratory mouse Mus musculus [5] and the guppy, Poecilia reticulata [6]. Both studies uncovered substantial individual variation in neuronal density, but despite that, guppies with bigger brains have more neurons on average, which is not true in mice. Unlike birds and most mammals (and the guppy), reptiles generally continue to grow for longer periods of time, even after attaining sexual maturity. Limited information is available as to what happens to the brain at the cellular level. Studies in the Nile crocodile (Crocodylus niloticus) have shown that the brain continues to grow, albeit with significant negative allometry [7] and the number of neurons grows with an even shallower slope, approaching an asymptote long before reaching maturity [8]. A postnatal increase in the number of neurons in several cortical areas has also been reported in the lizard Podarcis hispanica [9].

However, the existence of substantial neurogenesis throughout life [10–14] (reviewed in [15]) might potentially result in a higher variation in neuronal numbers in adult animals, complicating quantitative comparative studies. To assess this, we tested individual variation in the brain (divided into cerebral hemispheres and rest of the brain) mass and number of neurons and non-neuronal cells in a model reptile. As a suitable species, we used the Madagascar ground gecko (Paroedura picta), a fast-growing reptile, attaining sexual maturity at around 3 months of age and fully grown between 1 and 2 years [16]. We hypothesized that reptiles may have higher individual variation in the number of neurons compared with mammals owing to their extensive adult neurogenesis.

When evaluating within-species variation, we should account for the fact that there are, in essence, two components of this variation, one being the inter-individual differences within a population, and the other being differences between populations potentially under different selective pressures and adapted to different environments [17,18]. Many studies assessing individual variation within a species are limited to one captive population. On the one hand, this can mitigate problems with mixing animals of unknown history, subjected to different environmental conditions. On the other hand, it might also lead to underestimating the actual variability of the trait. To overcome this issue, we included two captive populations of recent wild ancestry (F1, and in a few cases F2 generation) and assessed both interpopulation and intrapopulation variation. Age and sex are other important factors contributing to within-species differences. Sexual dimorphism in relative brain size is present in some squamate species but not others [19] and sex differences in the volume of specific brain parts (e.g. [20–23]) are well documented in reptiles. Owing to prolonged neurogenesis, the numbers of brain neurons in reptiles might potentially decrease with age at a slower pace than in mammals, or even increase. To be able to assess the effect of sex and age, we used animals of both sexes in close to equal proportion and included three different age groups from one population, encompassing hatchlings, young adults and fully grown adults.

2. Material and methods

(a). Animals

For this study, we used the Madagascar ground gecko (Paroedura picta), a well-studied reptile, e.g. [24–27]. We compared two sources of variation—intrapopulation and interpopulation—for that purpose, we obtained 10 adult animals (5 males, 5 females) from a private breeder (population A) and 10 animals (5 males, 5 females) from the breeding facilities of the Faculty of Science of Charles University, Prague (population B). All the individuals were fully grown adults between 1 and 2 years of age, close to maximum body size (mean snout–vent length (SVL): 84 mm, 93% of the maximum reported SVL [28]). We included a group of fourteen 14-day-old hatchlings (5 males, 9 females) and ten 6-month-old sexually mature but not fully grown animals (herein referred to as ‘young adults') (5 males, 5 females) from the breeding facilities of the Faculty of Science of Charles University (population B) to compare the variation in different age groups from the same population.

(b). Brain processing

The animals were euthanized by anaesthetic overdose and perfused with 4% paraformaldehyde. Brains were removed, post-fixed and divided into the telencephalon (cerebral hemispheres, excluding olfactory bulbs) and the rest of the brain. To explore differences between brain regions, in a subset of individuals (adults from population A) we analysed the distribution of neurons in six brain divisions, namely in the cerebral hemispheres, olfactory bulbs, diencephalon, optic tectum, cerebellum and brainstem (electronic supplementary material, figure S1). In these brain components the total numbers of cells, neurons and non-neuronal cells were estimated using the isotropic fractionator [29], a fast and accurate technique that provides results comparable to unbiased stereology [30–32]. The protein NeuN was used as a nuclear marker of neurons [33]. Further details are provided in the electronic supplementary material.

(c). Statistical analysis

All the statistical tests were performed in the base package of R 3.6.2. [34]. To ensure normality, the continuous variables were log₁₀-transformed. The relationships between cell numbers or density and brain region or body mass were analysed using linear models. The differences between group means were tested with one-way ANOVA and Tukey's post-hoc test and the variances were compared with Bartlett's test. Although the assumption of homogeneity of variances was violated in a few cases, similar results were obtained when using non-parametric tests.

3. Results

(a). The Madagascar ground gecko brain in cell numbers

The 20 fully grown animals varied 2.26-fold in body mass (10.54–23.8 g), 1.41-fold in brain mass (73.6–103.6 mg) and 1.94-fold in number of brain neurons (2.9–5.6 millions) and 2.28-fold in number of other cells (2.8–6.3 millions) (electronic supplementary material, table S1). Unlike in mice [5], brain mass and body mass were correlated (F_1,18 = 12.76, p = 0.002, R² = 0.38), the number of brain neurons and brain mass were correlated (F_1,18 = 14.21, p = 0.001, R² = 0.41; figure 1a), but there was not a significant correlation between the number of neurons and body mass (F_1,18 = 3, p = 0.1; figure 1b), whereas the number of other brain cells and body mass were weakly correlated (F_1,18 = 5.21, p = 0.035, R² = 0.18).

Figure 1.

Intraspecific neuronal scaling and comparison between two captive populations for the Madagascar ground gecko. (a,b) Number of brain neurons (black symbols) and number of neurons in the cerebral hemispheres (red symbols) plotted as function of brain mass (a) and body mass (b). Shapes of points denote the two populations of fully grown animals. (c–f) Brain mass (c), number of brain neurons (d), number of non-neuronal cells (e) and neuronal densities (f) are compared between fully grown adults of the two populations. The horizontal bars denote the means, and 95% confidence intervals are shown for the population mean. Values for males are in blue and values for females in red.

In addition, we assessed brain cell distribution in six brain parts in adults of population A (electronic supplementary material, table S2). The number of neurons was significantly correlated with structure mass only in the cerebellum (F_1,8 = 5.517, p = 0.006, R² = 0.57). In all structures, neuronal density decreased with increasing structure mass (p < 0.001 in all cases).

(b). Intrapopulation and interpopulation variation and sex differences

We also assessed brain variation within the two populations (electronic supplementary material, table S1; figure 1c–f). The intrapopulation variances in both groups are comparable, except for brain mass, which has a significantly higher variance in population A (Bartlett's K² = 4.76, p = 0.03). The population means were significantly different in all the parameters (p < 0.03 in all cases) except neuronal density (F_1,18 = 3.77, p = 0.068), with population B having larger brains with more cells. There were no statistically significant differences in variance or mean between the sexes in any of the traits examined (in all cases p > 0.3, table S3; figure 1c–f).

(c). Differences in age groups within population B

The 6-month-old geckos had significantly smaller brains than the fully grown geckos (Tukey HSD: p < 0.0001), but they did not significantly differ in the number of whole-brain neurons (Tukey HSD: p = 0.78) and non-neurons (Tukey HSD: p = 0.94). The 14-day-old geckos had significantly smaller brains and fewer neurons and other brain cells than either of the adult groups (Tukey HSD: p < 0.001 in all cases) (electronic supplementary material, table S4; figure 2a,b).

Figure 2.

Comparison between three age cohorts. Brain mass (a), number of brain neurons (b), telencephalon mass (c) and number of telencephalic neurons (d) are compared across the three age groups of population B. The horizontal bars denote the mean and 95% confidence interval for the population mean. Values for males are in blue and values for females in red.

The 6-month-old geckos had a significantly smaller telencephalon than the fully grown geckos (Tukey HSD: p < 0.001) and fewer telencephalic neurons (Tukey HSD: p < 0.001), but there was no difference in non-neurons (p = 0.52). The 14-day-old geckos had again significantly smaller telencephalon and fewer neurons and other cells than either of the adult groups (Tukey's HSD: p < 0.001 in all cases) (electronic supplementary material, table S4; figure 2c,d).

Variance in brain mass in the 14-day-old group was significantly smaller than that of both adult groups (across all groups, Bartlett's K² = 8.01, p = 0.018), while the variance in brain neurons and non-neurons was homogenous across age groups. Variance in the number of telencephalic neurons in the 14-day-old group is significantly smaller than that of both adult groups (across all groups, Bartlett's K² = 11.81, p = 0.003). On the other hand, variance in neuron numbers in the rest of the brain is the same across all groups (Bartlett's K² = 0.005, p = 0.998).

4. Discussion

We assessed the individual variation in several brain traits both within and across two populations of Madagascar ground geckos and also evaluated sex differences and compared three different age groups within one population.

Across-population variation is larger than within-population, but not substantially so. Surprisingly, variation across adult geckos from both populations is comparable to within-population variation in 19 laboratory mice of the same strain, sex and age [5] in brain mass (1.41-fold versus 1.33-fold), number of neurons (1.94-fold versus 1.63-fold) and number of other cells (2.28-fold versus 2.98-fold); within-population variation in the geckos is even smaller than in mice in some traits (see electronic supplementary material, table S1). Combined with the data on the guppy and limited information on individual variation in mammals and birds [1,6,35] and our own unpublished data, it does not seem that there is generally a higher variability in brain traits in species with prolonged neurogenesis, although more species need to be investigated.

Differences between the population means were significant in most traits, except for neuronal density. The population B had larger brains with more cells overall, but similar neuron to glia ratio and similar average cell size. Interpopulation variation is thus mostly driven by the number of cells and probably genetically determined. It has to be noted that both populations were captive and were not currently under any particular artificial selection and we do not have information about the selective pressures shaping their ancestral populations. It is also possible that the differences are due to plasticity [15,36], as the housing conditions of the two populations were similar but not identical (see electronic supplementary material, Methods for details). Contrary to previous findings [15,36], smaller brains were observed in population A, which was kept in larger enclosures in small groups, and larger brains in population B, kept in smaller enclosures individually. One can speculate that, rather than acting as environmental enrichment, group housing in this species may have created social stress. In the wild, the sources of variation are likely more complicated.

In contrast with mice [5], geckos show a significant positive relationship between body mass and brain mass and between brain mass and the number of neurons, but not directly between the number of neurons and body mass, weakening the notion that larger bodies need more neurons to control them [37]. The neuronal density also goes down with increasing brain mass, resulting in a shallow slope of the relationship between brain mass and neuron counts.

We found no sex differences in any of the brain measures, even though sexual dimorphism in brain size has been reported in reptiles, e.g. [19,21–23]. Surprisingly, there was also no difference in body mass, although males in this species tend to be moderately larger [38]. This might be a bias of our particular sample. It is possible that in species with more pronounced sexual dimorphism, sex differences in brain neuronal density might arise, owing to brains in the larger sex ‘keeping up' with the body in size but not necessarily adding more neurons, which are metabolically expensive (e.g. [39]) and likely not needed to control a larger body.

Within population B, we additionally compared three different age cohorts. Between hatchlings and fully grown adults, the change in brain size was 4.5-fold, in the number of neurons 2.3-fold, in the number of other cells 2.02-fold, in telencephalon mass 5.41-fold, in telencephalic neurons 2.94-fold and in other telencephalic cells 2.74-fold. These changes are consistent with those reported in the Nile crocodile [8] in that the brain size grows much more quickly than the number of neurons and glial cells. However, only subadult crocodiles were included in that study, whereas our sample included fully grown adults. Increase in the number of brain neurons during early postnatal life and adolescence has been reported also in rodents [40–42].

While there were large differences between the hatchlings and adults in every measure, young and fully grown adults did not differ in the total number of brain neurons and non-neurons. This is potentially important for comparative studies, since it implies that including adult but not fully grown animals may not significantly affect the results in terms of absolute numbers of brain cells. However, fully grown adults had significantly more neurons in the telencephalon than sexually mature young adults, implying that adult neurogenesis [10–12] does not only replace worn out but also adds new telencephalic neurons in reptiles during adulthood. This markedly contrasts with the situation in mammals, where the number of cortical neurons declines with age [40,41].

The variance in brain mass in the 14-day-old group was significantly smaller than in adults, which is not true for the variance in brain neurons and other cells, suggesting that the higher variation in brain mass in adults is largely due to differences in cell size, dendritic arbours and connections, resulting in more bulk in the ‘rest of brain’. However, in the telencephalon, the variance in the number of neurons in hatchlings is significantly smaller than that of adults, implying differences in the rate of neurogenesis and/or neuronal death in the telencephalon are responsible for a substantial portion of the individual variation seen in adults. In any case, neuronal plasticity might play an important role in intrapopulation differences.

5. Conclusion

Our study provides the first data on inter-individual and interpopulation variation in the number of neurons in reptiles and suggests that despite the reptile brain growing in adulthood, within-species variation in neuronal numbers and densities is not substantially higher than in mammals. Furthermore, young adults do not have significantly lower numbers of neurons than fully grown adults, except for the telencephalon. Including adult but not fully grown animals thus should not significantly bias comparative studies based on numbers of neurons, although using very young individuals might skew neuronal densities and the number of neurons relative to body size. However, potential differences between populations might be a source of concern when directly comparing species, as examining just one population may mask small interspecific differences or create the appearance of a difference where there is none.

Ethics

The experiments presented in this work were approved by the Institutional Animal Care and Use Committee at Charles University in Prague and by the Ministry of Education, Youth and Sports of the Czech Republic (Permission No. MSMT-11721/2018-2).

Data accessibility

All data are available in the electronic supplementary material.

Authors' contributions

K.K. and A.P. conducted experiments and analysed data, K.K. and P.N. conceived of the project, L.K. provided experimental animals, and all authors wrote the manuscript. All authors approved the final version of the manuscript and agree to be held accountable for the content therein.

Competing interests

We declare we have no competing interests.

Funding

This study was supported by the Grant Agency of Charles University (810216) and Czech Science Foundation (18-15020S).