Population density does not affect seasonal regulation of reproductive physiology in male water voles

Abstract

Most small rodent species display cyclic fluctuations in their population density. The mechanisms behind these cyclical variations are not yet clearly understood. Density-dependent effects on reproductive function could affect these population variations. The fossorial water vole ecotype, Arvicola terrestris, exhibits multi-year cyclical dynamics with outbreak peaks. Here, we monitored different water vole populations over 3 years, in spring and autumn, to evaluate whether population density is related to male reproductive physiology. Our results show an effect of season and inter-annual factors on testis mass, plasmatic testosterone level, and androgen-dependent seminal vesicle mass. By contrast, population density does not affect any of these parameters, suggesting a lack of modulation of population dynamics by population density.

1. Introduction

The water vole Arvicola amphibius is a small rodent widely spread across Eurasia. Two ecotypes can be distinguished: the most common semiaquatic form, and the fossorial form, Arvicola terrestris, which is more restricted to southwestern Europe [1,2]. This fossorial ecotype mainly inhabits mid-mountain grasslands and orchards, where it digs a large network of galleries [2–4]. Populations of this fossorial ecotype exhibit cyclical population dynamics. Vole density thus varies from a few individuals per hectare to more than 500 individuals per hectare during outbreak phases. Each cycle lasts between 5 and 7 years on average in this species [2,5–7].

Most small rodent populations display cyclic density fluctuations [8,9]. Many studies have explored the role of extrinsic and intrinsic factors in these events, such as predation, parasitism, food resources, environmental conditions, stress, mortality and genetics, to identify the causes of such cyclical density dynamics [10]. Despite a better understanding of the factors leading to these density fluctuations, the mechanisms responsible for the periodicity and the amplitude of the outbreaks, as well as those causing the transition between peak and decline phases, remain unclear.

Regarding the population dynamics of the water vole, several studies have notably examined predation, landscape influence, agricultural practices, parasitism and pathogen-induced stress [6,11–16]. However, the impact of vole population density on reproductive physiology is still poorly documented.

In common voles and meadow voles, it has been shown that density has a direct impact on reproduction [17,18]. Indeed, during phases of high density, reproduction decreases in these voles leading to stabilization of the population. A similar effect may also exist in water voles; therefore, we hypothesized that the regulation of reproductive activity is dependent on population density with a decrease in reproductive function under high water vole density.

To test this hypothesis, we performed 3-year monitoring of male reproductive physiology in different water vole populations located in the mountains of the French Jura. Samples were collected at the beginning and end of the typical breeding period of water voles, i.e. in spring and autumn, and different parameters of male reproductive physiology were measured including testis masses, plasma levels of testosterone produced by testes and masses of seminal vesicles and lateral scent glands whose development depends on testosterone [19]. We assessed the impacts of population density, and seasonal and inter-annual factors on these parameters.

2. Material and methods

(a) . Study area, vole population density and sampling

From 2003 to 2005, capture sessions were performed twice a year, in spring (April and May) and autumn (October and November), at four to 14 sites (figure 1a) in the study area that covered 250 km² in the mid-mountain region of east-central France (Franche-Comté, Jura, Nozeroy canton: 46°47′N, 6″03′E). Two hundred and seventy-four sexually mature male fossorial water voles, Arvicola terrestris (Linnaeus 1758), were directly trapped in their gallery digs in open grassland (electronic supplementary material, table S1).

Figure 1.

Seasonal and inter-annual variations in body condition and reproductive physiology of the male water vole. (a) Population density (index method, %, see Giraudoux et al. [20] for a description of the method) for each capture session throughout the 3 years of monitoring, and the correspondence with the estimated number of individuals per hectare was estimated following Giraudoux et al. [20]. (b) Inter-annual changes in the body condition index (BCI). (c) Body mass and length significantly differed between voles caught in spring and autumn over the 3 years. (d–g) Inter-annual or/and seasonal changes in relative testis mass, plasma testosterone level and relative seminal vesicle mass. No change was detected in the relative size of the lateral scent glands. The results are shown as a Tukey's boxplot with outliers represented by circles and means indicated by a cross. Scaled (log) data were log-transformed before zero-centred scaling. Season and year correspond to the main effects tested in linear models (table 1), **, p < 0.01; ***, p < 0.001; N/A, data not available.

To investigate the effect of population density, captures were made in populations with different levels of individual abundance (figure 1a; electronic supplementary material, table S1). The vole population density was estimated during each capture session for each population following the index method [20]. The presence of mounds indicating vole activity was assessed at 5 m intervals along two perpendicular transects of 250 m. The number of intervals with mounds thus represents the density of the vole population as a percentage.

According to previous studies, analyses were performed on males with a body mass of ≥ 75 g which is a proxy for sexual maturity [3,21–24]. To minimize the effects of seasonal body mass variations in analyses, an organ-to-body mass ratio was used to study the relative mass of testes and seminal vesicles. Similarly, the lateral scent gland (LSG) surface was divided by body size to provide a relative surface index as previously described [22]. The body condition index (BCI) is the ratio of body mass to body length.

(b) . Testosterone assay

Blood samples were collected from 145 males at sacrifice. Plasma was obtained by centrifugation at 3000g for 30 min and stored at −20°C before the hormonal assay. The testosterone concentration was determined using the enzyme immunoassay developed by our laboratory (Laboratoire Phénotypage et Endocrinologie, UMR PRC, INRAE, Nouzilly, France). The testosterone assay was performed for each sample on 25 µl of plasma without an extraction process. The antibody against testosterone has been previously characterized [25] and was the same as that used for the radioimmunoassay previously described [26,27]. A parallelism test was performed to validate the testosterone assay on vole plasma (electronic supplementary material). The sensitivity of the testosterone assay was 0.1 ng ml⁻¹, and the inter-assay coefficient of variation was less than 5.8%.

(c) . Statistical analyses

Linear models were used to investigate the influence of year, season, and population density on different physiological parameters of voles. The plasmatic testosterone level, the relative mass of testes and seminal vesicles, and LSG sizes were log-transformed before zero-centred scaling. The year factor was not analysed for seminal vesicle mass due to the lack of data collected after spring 2004. The normality of the residual distribution was verified with the Shapiro–Wilk test, and the collinearity between main effects was tested using variance inflation factors with a limit of 2. For all variables, the ΔAICc (Akaike information criterion) was less than 2 between a model with a population density factor and the most parsimonious model. Population density factors were kept in the models shown in this study. Outliers (one value for relative testis mass and two values for relative seminal vesicle mass) were identified by the ROUT method with a false discovery rate set at 1% and were removed from analyses. Seasonal differences in body mass and body length were analysed by Mann–Whitney tests. Spearman's rank correlation coefficient was used to measure the relationships between reproductive parameters and vole population density. GraphPad Prism 9.4 (GraphPad Software Inc., San Diego, California, USA) was used to perform statistical analyses and draw graphs.

3. Results

(a) . Seasonal and inter-annual variations in reproductive function

The linear model indicated that the seasonal factor did not significantly affect the BCI of voles (table 1, figure 1b). However, the median mass and length of the voles caught over the 3 years of the study were significantly higher in spring than in autumn (figure 1c). By contrast, the linear model shows that season significantly affects all of the reproductive parameters studied (table 1). The relative testicular mass, blood testosterone level, and relative mass of seminal vesicles were higher for voles caught in spring than for those caught in autumn (figure 1d–f). A significant effect of year, reflecting inter-annual variations, was detected for the BCI, relative mass of testes, and plasma testosterone level (table 1, figure 1b,d,e). Values for these three parameters were significantly higher in voles trapped in 2004 than in those trapped in 2005. No season or year factor effects were detected in relative LSG size between voles trapped in spring and autumn (table 1, figure 1g).

Table 1.

Estimation of the parameters of the linear models tested to explain the physiological variations observed in the male water vole.

	estimate	s.e.	|t|	p value
body condition index
intercept	0.5992	0.008209	73	<0.0001 ***
year (2004)	0.01216	0.006985	1.741	0.0828 ns
year (2005)	−0.02426	0.00924	2.626	0.0091 **
season	−0.01247	0.007687	1.622	0.1059 ns
pop density	−0.00001602	0.0001167	0.1373	0.8909 ns
relative testes mass
intercept	0.823	0.1341	6.136	<0.0001 ***
year (2004)	−0.2351	0.1143	2.056	0.0408 *
year (2005)	−0.6582	0.1516	4.341	<0.0001 ***
season	−1.204	0.126	9.555	<0.0001 ***
pop density	−0.002523	0.001907	1.323	0.1869 ns
plasmatic testosterone level
intercept	−0.6710	0.2458	2.730	0.0071 ***
year (2004)	0.4038	0.2042	1.977	0.0499 *
year (2005)	−0.4137	0.2359	1.754	0.0816 ns
season	0.9454	0.1959	4.825	<0.0001 ***
pop density	0.0007445	0.002974	0.2503	0.8027 ns
relative seminal vesicle mass
intercept	0.439	0.1144	3.837	0.0002 ***
season	−1.02	0.1531	6.66	<0.0001 ***
pop density	−0.001801	0.002599	0.6929	0.4892 ns
relative lateral scent glands size
intercept	0.1307	0.1804	0.7248	0.4692 ns
year (2004)	−0.1499	0.155	0.967	0.3344 ns
year (2005)	−0.3411	0.2045	1.668	0.0965 ns
season	−0.287	0.1715	1.673	0.0955 ns
pop density	0.00256	0.002583	0.991	0.3226 ns

(b) . Vole population density and reproductive function

Linear models showed no effect of population density on the physiological parameters studied (table 1). No correlation with population density was detected, in either spring or autumn, for relative testicular mass (figure 2a), plasma testosterone level (figure 2b), relative seminal vesicle mass (figure 2c) or relative LSG size (figure 2d).

Figure 2.

Effect of population density on male water voles' reproductive physiology. No correlation was detected between vole population density and (a) relative testis mass, (b) plasma testosterone level, (c) relative seminal vesicle mass or (d) lateral scent gland size. r_s, Spearman's correlation coefficient.

4. Discussion

This study explored the impact of different factors including season, year and population density on the reproductive physiology of male water voles. Our results show that the BCI does not change according to the season in these populations. The significant difference in body mass and length between the spring and autumn may suggest a modification in the age structure of the populations between these two seasons. The proportion of young adult voles in autumn would be higher than that in spring. This could be due to the accumulation of juveniles born during the breeding season (typically from March to October) and the loss of the oldest individuals, as described in other populations of water voles [21,28].

A seasonal rhythm in male reproductive function was observed over the 3 years of monitoring. The greater mass of testes in voles trapped in spring compared to autumn is consistent with the findings of previous studies [19,22,24]. This result suggests greater testicular activity in the spring than in the autumn. This is supported by the higher level of plasma testosterone produced by testes that was measured in these wild water voles caught in spring in comparison to those caught in autumn. In spring, the increase in the masses of seminal vesicle which produce seminal fluid and are highly androgen-dependent [29,30], also reflects variations in testicular activity between seasons as previously described [22,24]. Surprisingly, no significant seasonal effect was detected for the relative size of LSG, while these sebaceous glands are known to be androgen dependent, as shown in the water vole and other rodent species [19,31,32]. However, it should be noted that the maximal development of the testes was observed later, around June, while their regression was complete in January [19,22,24]. In April and May, plasma testosterone levels would therefore not have peaked, and consequently, LSGs would not have reached full development in water voles at the time of capture. This result is consistent with previous data that did not show significant differences between spring and autumn; only finer-scale annual monitoring would allow seasonal variations to be observed [19,22]. Mammals living in temperate latitudes use yearly variation in day length (photoperiod) as the most predictive environmental cue to adapt their physiology to seasonal changes in the environment, notably for synchronizing their reproductive activity [33,34]. We have shown in a previous study that photoperiod is strongly involved in seasonal regulation of reproductive physiology in male water voles, although other factors may modulate this effect [22].

In this study, linear models showed significant year-dependent changes in the BCI and reproductive physiological parameters, such as relative testis mass and plasmatic testosterone level. These results support the involvement of other environmental factors in the regulation of reproductive physiology that may modulate the impact of photoperiod on male water voles, as previously suggested by studies reporting winter reproduction in Swiss Jura vole populations during particularly mild winters or continuous breeding in populations located in northwestern Spain where temperature and food availability are favourable throughout the year [23,35,36]. Moreover, it has been shown in other voles and rodent species that temperature and food quantity and quality affect male reproductive physiology. Indeed, in the common vole, testicular growth was higher in males fed alfalfa harvested in spring rather than autumn [37]. In hamsters exposed to an intermediate photoperiod, such as in spring or autumn, food restriction induced testicular regression [38]. In Syrian hamsters housed under short-day conditions, corresponding to the non-breeding season, testicular regression was slower in males housed at 20/25°C than in those housed at 5/10°C [39–42].

In water voles, seasonal breeding cycles overlap with pluri-annual cycles of population fluctuation [6]. In this study, we explored whether population density affects the reproductive function of male water voles. Our analyses did not reveal any effect of population density on the BCI or reproductive parameters of male water voles. Testicular development, steroidogenesis activity, nor the development of androgen-dependent seminal vesicles or lateral glands showed changes correlated with vole population density. Nevertheless, while male reproductive physiology does not seem to be strongly influenced by population density, this may not be the case for females. The cyclical density dynamics could be due to intrinsic individual quality transmitted by mother to offspring voles. In high-density populations, offspring individual quality was lower than in low-density populations, resulting in a decrease in the population growth rate [43–45]. In water voles, chronic stress during outbreak and decline, reflected by a higher level of basal corticosterone, might lead to reproductive capacity impairment of the female and accelerated senescence of immune function [12].

In conclusion, our long-term population monitoring showed seasonal regulation of reproductive physiology in male water voles. This result is consistent with previous observations in other water vole populations spread across Europe [19,22,24]. Our data also indicate that inter-annual factors modulate the seasonal regulation of male reproductive physiology. However, our study provides no evidence that population density can affect reproductive physiology in male water voles. To reach a more definitive conclusion, the impact of population density on reproductive physiology in female water voles remains to be explored.

Ethics

This study was carried out in accordance with the animal care guideline of Institut National de la Recherche Agronomique, INRA [46] and the French and European regulations on care and protection of laboratory animals (French Law 2001-486 issued on 6 June 2001). The procedures were approved by the Departmental Veterinary Service (B34-169-1), an institution accredited by the French National Ministry of Agriculture and Fisheries.

Data accessibility

Data are available from the Dryad Digital Repository: https://doi.org/10.5061/dryad.hhmgqnkkn [47].

The data are provided in the electronic supplementary material [48].

Authors' contributions

K.P.: conceptualization, data curation, formal analysis, investigation, visualization, writing—original draft, writing—review and editing; C.M.: investigation, writing—review and editing; M.B.: data curation, formal analysis, writing—review and editing; D.C.: investigation, writing—review and editing; A.-L.L.: investigation, methodology, validation, visualization, writing—review and editing; P.C.: methodology, visualization, writing—review and editing; M.M.: methodology, writing—review and editing; N.C.: conceptualization, investigation, methodology, supervision, writing—review and editing; M.K.: conceptualization, methodology, supervision, writing—review and editing.

All authors gave final approval for publication and agreed to be held accountable for the work performed therein.

Conflict of interest declaration

We declare we have no competing interests.

Funding

This work was support by the Ministère de l'Agriculture et de l'Alimentation, the DRAAF Auvergne-Rhône-Alpes as well as the Région Auvergne-Rhône-Alpes.