Our experiments resulted in a clear demonstration of the adverse effects of ALAN exposure on fitness of both nocturnal and diurnal species, manifested by reduced survival rates and reproductive success. Mortality rates of diurnal A. russatus were significantly higher when exposed to blue ALAN, compared to the control groups (Fig. 1.C). While the controls did not reach the median of survival until the end of the experiment, A. russatus exposed to white and blue ALAN reached the median at similar time (after 160 days of exposure), and yellow light exposed A. russatus reached the median 2 months later (Fig. 1. C). These results suggest that the blue, short wavelength light (420–520 nm), which is also included in the white light but almost absent in the yellow light, had an adverse effect on the mice. Blue light is known to have the largest effect on the eye's intrinsically photosensitive, melanopsin containing, retinal ganglion cells (ipRGCs) which result in the direct suppression of melatonin and influences biological rhythms24. Light intensities as low as 0.028 lux (monochromatic blue light) and 0.3 lux (white light) are capable of suppressing melatonin in rodents and birds, respectively21. Therefore, we suggest that our results indicate an involvement of biological rhythms in the effect.
The results of nocturnal A. cahirinus are more complex and suggest it suffered from competition in addition to the ALAN effect. Mortality rates in the control groups were significantly higher in A. cahirinus compared to A. russatus (Fig. 1. B P = 0.002). However, when A. russatus survival rates decreased significantly (under blue ALAN exposure) and as a result competition pressure reduced, survival rates of A. cahirinus were significantly higher compared to white ALAN exposure, where competition was presumably more intense. We suggest that the combined effects of competition and ALAN exposure affected all the other results we obtained for A. cahirinus, including reproductive success and fecal cortisol metabolites levels (see below).
This effect of competition may seem surprising at first, as it seems to contradict the observation that A. cahirinus competitively excludes A. russatus to diurnal activity25. However, the mechanism of exclusion is resource (food) mediated 26, and in direct confrontations A. russatus is more aggressive than A. cahirinus 27. Under the current experimental conditions, where food was ad libitum available, there was no competition on food, and we suggest that the more aggressive A. russatus outcompeted, increased stress levels and reduced survival rates and reproductive success of A. cahirinus.
In the second year of the experiment, we had two high mortality incidents in two physically distant enclosures exposed to white ALAN, one enclosure in January 2019, and the other in March 2019. All individuals of both species from these two enclosures died within days. The dead animals were not wounded or underweight and the cause of death was undetermined in pathological examination. We hypothesize that disruption of biological rhythms (daily and/or seasonal) caused a mismatch between the internal clock and environmental challenges, possibly including a misalignment of the immune system, which resulted in higher susceptibility to pathogens 15, as reported in hamsters 28, birds 29–32 and aphids33. Energy invested in immune function is often balanced against energy invested in reproduction, to optimize fitness 20. Since exposure to ALAN resulted in reproduction year-round, especially in A. cahirinus, it is possible that it compromised immune function during winter. Moreover, chronic activation of the HPA axis, manifested as high cortisol levels (see below) could have caused dysregulation of the immune system 34 leading to poor defense against pathogens, and as suggested in birds 30–32, inhibition of melatonin secretion by exposure to ALAN may be the underlying mechanism. However, this hypothesis and the role of melatonin in our system remains to be studied.
The effect of ALAN on reproductive success was striking (Fig. 2, Supplementary table 2). In their natural habitat, both species breed during summer: young individuals of A. cahirinus are observed from February until September, and young A. russatus are observed from April to July 35. While the reproductive activity in both species was affected by ALAN, the nature of response was different. A. russatus reproductive success was strongly affected by ALAN: the total number of pups of A. russatus exposed to all tested wavelengths significantly decreased by at least half compared to the control group (Supplementary table 2). Yet, in all treatments, reproduction appeared almost only during summer, which is the reproductive season of both species 35.
In A. cahirinus we again see a combined effect of ALAN exposure and interspecific competition, and a different effect of ALAN: The total number of pups of A. cahirinus seems to be affected mostly by competition. It had high number of newborns in the enclosures were A. russatus had low survival rates and fewer pups (blue treatment), and low number of pups in the control group, where A. russatus thrived most. Moreover, ALAN resulted in loss of seasonality of reproduction in A. cahirinus, which was reproductively active year-round under all ALAN exposed enclosures, as opposed to the control group, whose reproductive seasonality corresponded to their natural reproductive timing. In long day breeders like Acomys, seasonal reproduction is timed by day length, so we conclude that for A. cahirinus and somewhat for A. russatus, exposure to light pollution at all tested wavelengths led to a false perception of summer day length and resulted the loss of seasonality and continuous reproduction. Yet, overall reproduction (total number of offspring) was negatively affected by exposure to ALAN.
An effect of light pollution on reproductive timing was previously described 36–38, but these studies mostly focused on physiological reproductive state and did not measure reproductive output. Our experimental enclosures allowed us to continuously measure reproductive output and demonstrate the negative effect of ALAN on reproduction. However, our result may underestimate the actual effect under natural conditions: unlike natural conditions, the mice in our enclosures had ad libitum access to food and water. It is possible that in the wild, pups born during the winter, when temperatures and arthropods (Acomys preferred food) availability are lowest 26,39, and when spiny mice use torpor to reduce energy expenditure 40–43, would have low survival rates, further decreasing fitness.
The effect of ALAN on all measured parameters may be mediated by two hormones – melatonin and glucocorticoids (GC). Daily and seasonal timing cues rely on melatonin secretion which plays a key role in the biological regulation of daily and seasonal rhythms. Melatonin secretion is confined to the night by the circadian clock and is inhibited by light in both diurnal and nocturnal mammals 44,45. Therefore, light at night can reduce or stop melatonin secretion, and hence disrupt the natural cycle of all downstream biochemical and physiological processes influenced by melatonin. In the current study we did not measure melatonin secretion, and its role in the observed effects of ALAN in this setup remain to be studied.
We did measure the effect of ALAN on Acomys main glucocorticoid - cortisol23. We found that exposure to blue and white light at night increased the baseline GC levels of diurnal A. russatus (Fig. 2. A, C, D). GC are involved in various processes, acting as neurotransmitters and neuromodulators, activating, and regulating numerous processes related to stress response and homeostasis, and synchronization of peripheral clocks 46,47. Consistent with our data (Fig. 2. B), cortisol normally exhibits a dimorphic difference with higher concentration of GC among females48. Cortisol follows a circadian rhythm with tendency to reach high concentration in the blood at or just before wakening time, and decrease during the active hours49. Acomys cortisol daily pattern in fecal material is 12 hours delayed compared to blood concentration, with high concentration of fecal cortisol at 21:00 for A. russatus and 16:00 for A. cahirinus 23. Setting the traps and feces collection was done in the beginning of activity time of each species, representing 12 hours earlier, meaning the trough of the daily pattern. Therefore, elevated cortisol level may indicate chronic stress of the individual or a shift in the circadian pattern of GC.
Cortisol displays a seasonal pattern in some mammals, which is associated with reproductive activity50,51 and with metabolic adjustment to changes in energy demand and thermoregulation 52,53. To distinguish between seasonality in cortisol and the reaction to light pollution, we compared cortisol level of each group (species and sex) at each season. We found that in summer, cortisol level of A. russatus females exposed to blue and white ALAN were significantly higher (P < 0.05) and marginally significant in females exposed to yellow light (P = 0.55) compared to control (Fig. 2. C). Male A. russatus exposed to white light in winter had significantly lower level of cortisol compared to control (Fig. 2. D). Under free-living and semi-natural conditions, spiny mice of both species have lunar cycle in fecal cortisol metabolites levels, with increased levels during moon lit nights 23 as well as in response to artificially increasing light levels at night to moon levels 54. Full moon light levels (natural or artificial) also resulted in reduction of activity levels, foraging and food consumption in both species 23,54,55, and increased inter-specific aggressive interactions in A. cahirinus 54. However, these high cortisol levels were temporary and showed lunar cycle, while in the current experiment the exposure to ALAN and its effect on cortisol levels are chronic. High or low levels of cortisol may indicate that exposure to ALAN leads to an unbalanced activation of HPA axis with regards to season. Chronic dysregulation of the HPA axis is known to have many adverse physiological consequences and may results with a weaken immunity response34 which leaves the individual more susceptible to infectious diseases and parasites 56, and may explain the higher mortality rates in the groups exposed to blue and white ALAN. In A. cahirinus we found no significant changes in fecal cortisol levels between males and females, treatments, or seasons. We suggest that cortisol levels of all A. cahirinus were constantly high in this experiment as a result of competition with A. russatus, which masked the effects of ALAN.
We hypothesize that in A. russatus, blue and white light which increased basal level of cortisol, activated the HPA axis and resulted in the inhibition of reproductive activity. The response to yellow ALAN was weaker in all parameters. The results for A. cahirinus are again more complex. Reproductive output of A. cahirinus is normally higher than that of A. russatus 57,58. Yet, in our experiment, the total number of A. russatus pups was almost double that of A. cahirinus. This result, together with the higher mortality rate of A. cahirinus supports our hypothesis that A. cahirinus suffered from a combined effect of competition and ALAN exposure. This hypothesis also suggests that cortisol levels were high in all A. cahirinus groups chronically, and therefore we did not find any effect of sex, season, or treatment as we did in A. russatus.
Exposure to blue and white, and to a lesser extent yellow ALAN, changed the inter-specific interactions between the species and resulted in population decline and reduced fitness in both species. Such conditions could promote the introduction or spread of alien species that are less sensitive to illumination, such as the house mouse54.