Multi-Level Processes and Retina–Brain Pathways of Photic Regulation of Mood

The influence of light on mood through the sleep homeostasis is well illustrated in SAD. Several studies comparing healthy patients and patients with SAD indicate that the homeostatic “process S” is directly modulated by light. First, patients with SAD compared with healthy controls show differences in sleep duration during Winter [83,84]. Secondly, studies have shown significant differences in sleep architecture between patients with SAD and healthy controls in winter compared to summer: patients with SAD, compared with healthy controls, have lower delta activity or slow-wave sleep (SWS) and lower sleep efficiency in winter compared to summer, suggesting that photic or photoperiodic information play a key role in sleep homeostasis in SAD [7,83]. Thirdly, in SAD patients, LT improves sleep efficiency and delta sleep [84,85]. This positive effect of light on delta activity is similarly observed after administration of either morning or evening light, suggesting that light impacts the delta power spectrum and that positive effect does not result from modulation of the biological clock [84,85]. Light’s impact on EEG delta activity may so rather result from SCN-independent effects of light. In addition, some studies suggest that this light effect on delta activity could be melanopsin-dependent and may implicate the VLPO sleep-regulatory structure. Indeed, in patients with SAD, it has been observed that increases in EEG delta activity depends on the spectral component of light with blue light being more effective than red light in decreasing depressive symptomatology [86,87]. In mice the absence of melanopsin affect light effects on sleep and the electrocorticogram activity, and light through melanopsin retinal ganglion cells activates the VLPO [86,87]. Taken together, these results emphasize the close relationship between seasonal changes of sunshine, daylength, sleep need and mood. These results are also suggesting the hypothesis of direct effects of light on the VLPO through activation of ipRGCs (Figure 1). Further studies using animal models, healthy volunteers and patients affected with seasonal depression are needed in order to understand the complex relationships between light, sleep homeostasis and mood (Table 1). Nevertheless, clinical studies have already shown that sleep deprivation, which increases the sleep pressure, has a rapid-acting antidepressant effect [88]. Interestingly, this clinical efficacy is improved by adding other chronotherapeutic approaches, such as LT [89].

Lack of alertness or excessive daytime sleepiness (EDS) is a complaint frequently met in all affective disorders such as seasonal affective disorder but also in non-seasonal major depressive disorder and non-seasonal bipolar disorder [90]. EDS impairs concentration, memory and mood, which further negatively impact global functioning and wellbeing, and may be associated with suicidal ideation in patients with MDD [91]. Light is known to exert an acute alerting action in humans. This alerting response to light in humans depends on factors such as the illuminance levels, exposure duration, timing and wavelength of light [7,29,92]. Neuroanatomical and neurophysiological findings have shown that, in addition to its powerful alertness influence through the SCN [7,93], light may enhance alertness through direct mechanisms activating alertness centers in the brain [29,94,95]. Badia et al. were among the first to explore how light modulates alertness by applying a 90 min bright light therapy (BLT) of 5000 lux or a 90 min dim light (DL) of 50 lux during the subjective night [29]. They showed that alertness measured by EEG beta activity (i.e., a rapid waked EEG activity) was immediately increased during BLT session. Since Badia et al., it has been discovered that light could exert its alerting effect by simultaneously inhibiting the sleep-inducing VLPO nuclei and activating the monoaminergic arousal system implicated in mood [87]. These effects are attenuated and delayed in the absence of melanopsin, demonstrating that light would directly exert this alerting effect via the ipRGCs [87]. In addition, more recently Milosavljevic et al. observed that in mice the activation of ipRGCs not only increased alertness and arousal, but also had immediate effects on diverse brain structures implicated in mood, attention and alertness [96]. These brain structures included: (i) the dorsomedial hypothalamus/dorsal hypothalamic area (DMH/DHA); DMH/DHA is known to provide inhibitory control of the VLPO [96] and is known to be excited during periods of wakefulness [48,97]; (ii) the hypothalamus paraventricular, dorsomedial, and lateral hypothalamus that are active during wakefulness; (iii) the thalamus, including the paraventricular thalamus nucleus (PVT) and the intralaminar thalamic nucleus (ILT) both independently associated with enhanced alertness and vigilance [98,99]; and (iv) the limbic system including the amygdala and the nucleus accumbens. The amygdala has been reported to be light activated in humans [100] and also to receive direct inputs from ipRGCs [33]. The nucleus accumbens is a region associated with enhanced motivation and reward-seeking behavior [101]. Finally, in humans, several studies using positron emission tomography (PET)119 or functional magnetic resonance imaging (fMRI), observed that light also induced changes in multiple brain areas involved on both alertness and mood such as the thalamus (with activations linearly correlated with the improvement of alertness) and such as the locus coeruleus (LC) [30,102]. At the same time, light inhibited brain areas promoting sleep such as the VLPO [103]. All these discoveries demonstrated that light could thus enhance mood by increasing alertness with several direct tracts from ipRGCs to wake and mood areas (Figure 1 and Figure 2). These findings pave the way for further research to better unravel the relationships between light, mood and alertness (Table 1). Among wakefulness pathways, the orexinergic system, plays a key role in modulating arousal, but also plays a key role in many other functions including mood. In this context, the following paragraph will synthetize new insights about the relationships between light, mood and the orexinergic pathways.

Orexin neurons play a key role in the regulation of appetite, affect, and reward system, as well as in the regulation of the sleep–wake cycle [104]. Orexin neurons activate wake promoting nuclei, by releasing norepinephrine via the Locus Coereleus (LC) [105]. In contrast, the orexin inhibition facilitates sleep initiation and maintenance [106]. In addition, orexin neurons have been postulated to have a role in depressive mood with evidence of abnormalities of their activity in both animal models of depression and in humans with MDD [107,108]. Indeed, it has been observed from pre-clinical studies a positive correlation between the amygdala orexin receptor expression and the manifestation of depressive behaviors [107]. Moreover, severe depressive behaviors were associated with decreased hippocampal orexin expression in the amygdala [108]. In humans, abnormalities in the activity of orexin neurons have been also observed in MDD. For instance, one study evaluating cerebrospinal fluid levels of the A orexin (i.e., which is one of the two subtypes of orexin) found a decrease in the circadian amplitude variation in patients with MDD compared with healthy controls [109]. Orexin neurons also mediate the mood effects of light [107,108] by receiving direct retinal inputs [33] and indirect retinal inputs from other hypothalamic nuclei including the SCN [110,111]. Orexin neurons innervate the dorsal raphe nuclei (DRN) implicated in mood regulation, as reported in nocturnal rats [112] and also in diurnal rats [113]. A study from Adidharma et al. demonstrated the role of orexin neurons in mediating the effects of light on the mood-regulating monoaminergic areas [114]. They showed that orexin neurons are part of the tract through which light information is transmitted to the serotonergic nucleus within the dorsal raphe (DRN) [115]. They observed that rats housed in winter-like lighting conditions then presented depression-like behaviors, supporting their potential use as an animal model of SAD [115,116,117,118]. Then, to better explore the neural pathways mediating the effects of light on mood, an acute light pulse was administered in the early subjective day to rats housed in constant darkness. This light pulse was intended to mimic the acute effect of light therapy, which is generally carried out during early daytime [119,120]. Using c-Fos marker, the brain responses were examined, in three interconnected brain regions/cell populations that could potentially be involved in mediating the effects of light on mood regulation: the principal circadian clock in the SCN [121], the wakefulness-promoting orexin neurons [106], and the largest serotonergic nucleus within the dorsal raphe (DRN) [115,122]. Their study showed first that light exposure during subjective daytime increases immediately neural activity in orexinergic cells and in the DRN but not in the SCN [114]. Furthermore, blocking the orexinergic signaling using an orexin receptor-1 (OXR1) antagonist inhibits the light-induced neural activity in the DRN suggesting that light induces excitatory responses in the serotoninergic DRN through activating orexinergic pathways. Thus, their results support the idea that the activation of the wakefulness promoting orexinergic system underlies the therapeutically effects of light on SAD. Their study focused on the DRN, but orexin neurons project to and directly regulate all monoaminergic neurons that are involved in mood regulation [123]. Therefore, the orexinergic pathway seems to be very a key player to mediate the effects of light on monoaminergic system that ultimately contribute to light-dependent mood regulation (Figure 1). Therefore, in the following section, a closer look on the effects of light on monoaminergic systems is proposed.

It has been well established that light influences serotonin (5-HT) turnover: the rate of production of serotonin by the brain is the lowest in winter and rapidly increases with longer daylength during summer [124]. Light also influences serotonin transporter binding levels in the brain, which increase in autumn and winter compared to spring and summer [125].

The serotonergic system and the influence of light on the serotonergic system have been extensively investigated in all mood disorders, including SAD, unipolar depression and BD [126]. In addition, it has been reported higher seasonal serotonin transporter (5-HTT) fluctuations in SAD, when compared to healthy controls, in all brain regions, including mood regulation areas such as the anterior cingulate (ACC) and prefrontal cortex (PFC). It has been also observed that 5-HTT density and affinity were significantly decreased in the ACC and PFC after 2 weeks of LT and correlated with the decrease of depressive symptoms [127]. All these findings suggest that (i) in SAD, light impacts seasonal fluctuations of 5-HT and 5-HTT levels, and that (ii) 5-HTT plays a central role in the direct mechanisms of the antidepressant effect of light [75]. In other words, increasing central 5-HT availability seems to be an important effect of LT in SAD. How LT acts to modulate 5HTT affinity for serotonin is however still unclear. Furthermore, serotonin has a strong and bidirectional relationship with the biological clock [128]. Indeed, serotonin modulates the SCN response to light [129], and light induces indirect signaling, from the SCN to the raphe median nucleus [73]. Another possible hypothesis of the antidepressant effect of light through the modulation of 5-HTT levels is that light may induce direct signaling (i.e., independently of the SCN) between the retina and the serotoninergic system such as a direct action on the raphe dorsal nucleus (Figure 1 and Figure 2).

Dopamine also regulates sleep–wake rhythms [130] and is known to contribute to the adaptation to the light/dark cycles in retina photoreceptors [131]. Dopamine may also play a role in the pathophysiology of SAD. In SAD, retinal alterations such as a lower rod retinal sensitivity and lower cone maximal amplitude assessed by the electroretinogram (ERG) are apparent during the depressive phase in autumn/winter, and are normalized after 4 weeks of LT or during the summer [132]. These findings suggest that these ERG changes regarding cone and rod functions are state markers for SAD [132]. These results raise questions about the underlying roles of the dopamine in SAD since multiple dopamine-dependent physiological mechanisms act through cone and rod circuits [131]. Other studies have also suggested the contribution of the cerebral dopaminergic system in SAD, including the decreased dopamine transporter (DAT) availability observed in untreated symptomatic depressed SAD patients [133]. LT appears to stimulate the dopamine production in the retina [134]. In this later work, authors suggest that the winter-induced retinal dopamine deficiency may cause SAD and that light reverses this syndrome. Finally, a recent preclinical study indicates that variations in the levels of hypothalamic DAT may act to model seasonal changes in a BD model [135] and suggest that light may influence DAT binding levels in the brain. Actually, their data reveal the potential contribution of the photoperiod-induced neuroplasticity within an identified circuit of the hypothalamus, linked with reduced DAT function, underlying switching between states in BD. Mice with reduced DAT expression exhibited hypersensitivity to summer-like and winter-like photoperiods, including more extreme mania-relevant and depression relevant behaviors [135]. Further clinical studies are warranted in the future to confirm if DAT could be considered as an important central biomarker of SAD and/or seasonal BD, and could thus play a role in the direct and SCN-independent mechanisms of light (Figure 1).

Mood disorders such as MDD or BD are characterized by altered structural or functional changes in areas involved in emotional processing, such as the amygdala, the hypothalamus, and the hippocampus [136]. Responses to emotional stimuli can have a large impact on mood [136]. LT is an effective treatment for mood disorders and so several authors assessed whether light affect mood through the modulation of brain areas involved in emotional processing. A first study of Vandewalle et al. assessed whether light can acutely influence normal brain emotional processing [100]. Their study showed that amygdala was one of the emotional brain area acutely affected by changes in ambient light [94]. This is not surprising as amygdala is known to be a core component of the emotional brain [136,137] that receives sparse direct projections from ipRGCs [33]. They also showed that blue light could induce the activation of three others cores components of the emotional brain: the left hippocampus, left thalamus and the brainstem locus coeruleus. The same team then explored whether light could directly modulate emotional brain responses to emotional stimuli in healthy subject by using functional MRI (fMRI) and a validated auditory emotional task [100]. They observed that blue light increased responses to emotional stimuli in the voice area of the temporal cortex and in the hippocampus, two other areas both implicated in emotional processing [100]. They also observed during blue illumination that emotional processing was associated with an enhanced functional coupling between the voice areas, amygdala and hypothalamus [100]. Their findings suggest that ambient blue light strengthens emotional brain reactivity and might promote accurate and contrasting responses to emotional signals, which could enhance efficient mood regulation processes. These results have important implications for the understanding of the mechanisms by which changes in lighting environment enhance mood [28,119]. More recently, Milosavljevic et al. demonstrated in mice that light has immediate effects on diverse brain structures implicated in mood, including emotional processing areas, such as the amygdala and the nucleus accumbens [95]. They also highlighted the key role of melanopsin-expressing retinal ganglion cells to emotional brain area responses to light. Finally, two new retino–brain pathways implicated also in emotion regulation has been recently discovered and will be detailed in the two following sections. These findings taken together confirmed that light might induce mood changes the refereeeehrough emotional processing pathways [9,138] (Figure 1 and Figure 2).

A key recent work reported that another SCN-independent pathway might mediate light-induced mood changes through emotional processing pathways [138]. In this study, Fernandez et al., found that a subset of ipRGCs (predominantly M1) in mice directly innervated the perihabenular nucleus (PHb) a region of the dorsal thalamus, known to be implicated in emotion regulation. They interestingly showed that an irregular light stimulation led to depressive-like behaviors associated with changes in the thalamic PHb. PHb neurons show a robust rhythmic expression of the clock gene PER2 under normal light conditions. In this study, when mice were housed under the T7 cycle (i.e., 3.5 h of light followed by 3.5 h of dark), the PER2 rhythmicity was abolished in the PHb [139,140]. Thus, these findings revealed that irregular light/dark cycles perturb the biological clock in the PHb through an ipRGCs-dependent circuit, then inducing depressive behaviors. The PHb is integrated into a loop of limbic thalamic nuclei and is connected to thalamo–frontocortical loop [141]: PHb neurons project to several mood regulating centers with collateral projections to the ventro-median prefrontal vortex (vmPFC), the dorsal and ventral striatum. The prefrontal cortex (PFC) is known to be fundamental to mood regulation and has been consistently implicated in MDD by imaging studies of patients [142,143,144] and using animal models of mood disorders [145]. The dorsomedial striatum participates to a thalamo–frontocortical loop, which is involved in affective emotional processing. The ventral striatum, the third target of PHb neurons, and mostly the nucleus accumbens (NAc) included in the ventral striatum, has been extensively implicated in mood and depression [146]. Furthermore, the study of Fernandez et al. demonstrated that the PHb is both necessary and sufficient for driving the effects of light on affective behavior. Indeed, they reported that inhibiting PHb neurons blocked the mood alterations triggered by abnormal light cycles and they reported that chronic activation exclusively of projecting PHb neurons induced significant alterations in mood. Chronic exposure to light at night in humans may cause similar neuronal changes leading to mood alterations.

Interestingly, Huang et al. recently observed that light influences also depressive-like behaviors in mice through a pathway linking the retina and the lateral habenula (LHb) [9]. The LHb from the epithalamus regulates information from the limbic system to the midbrain monoaminergic centers [147]. Stress factors can activate the LHb and induce depressive-like behaviors [147,148,149,150,151], whereas the inhibition of the LHb improves depressive symptoms [152]. In addition, recent pre-clinical studies have suggested that light can influence neuronal activity in the LHb, leading to depressive-like behaviors [38,153]. Indeed, in the Huang et al. study, they identified a retinal-brain pathway connecting the retina and LHb in the mouse [9]. In the retina, a subset of M4-type retinal ganglion cells (RGCs) innervates GABA neurons in the ventral lateral geniculate nucleus and intergeniculate leaflet (vLGN/IGL) [9], which in turn inhibit the neural activity of the LHb. They demonstrated that the activation of the retina–vLGN/IGL-LHb pathway is required for the antidepressant effects of BLT. Moreover, these works also showed that either the inhibition of LHb-projecting vLGN/IGL neurons or the activation of LHb postsynaptic neurons abolish the antidepressant effects of LT [9]. These works reported an absolutely new anatomically distinct neuronal circuit involved in light mediated mood deficits.