To best represent human disease, and explore relevant pathogenesis and novel treatment approaches, it is crucial to take full advantage of the available mouse models of spontaneous or inducible lupus-like disease. However, previous studies prefer using genetically susceptible animals to investigate spontaneous NPSLE[24]. Neuropsychiatric symptoms in the inducible wild-type models of lupus are less studied. Here, we systematically evaluated whether PIL mice are possible to be used as an environmental-related inducible model of NPSLE. Consistent with previous reports[18, 20, 21], we verified PIL mice presenting several peripheral lupus-like manifestations (Fig. 2). We further revealed some brain dysfunctions and neuropathological changes: a) olfactory dysfunction and an anxiety- and depression-like phenotype. b) cytokine overproduction in the brain tissues. c) BBB leakage. d) IgG deposition in the choroid plexus and lateral ventricular wall, and co-localization of IgG with microglia. e) changes of morphology and density of microglia and astrocytes in the hippocampus. f) lipofuscin deposition in the cortical and hippocampal neurons. Our results highlight the potential application of PIL mice for exploring the pathogenesis of environmental-related NPSLE.
Many NPSLE patients usually have olfactory dysfunction[25]. Spontaneous lupus-prone MRL/lpr mice also displayed a constellation of behavioral deficits dependent on olfaction[26, 27]. In our study, PIL mice lost their interest in biological odorants, decreased sensitivity to repellents (Fig. 3A). And, these olfactory dysfunctions appeared earlier than other behavioral deficits, suggesting that olfactory neural system is highly sensitive to immune attack. Thus, olfactory function test might be an effective approach for early diagnosis of NPSLE. Additionally, affective deficits are frequent neuropsychiatric events in NPSLE[28], which impact on patients’ health-related quality of life. Similarly, numerous genetically predisposed mouse models of NPSLE also exhibit affective disorders. For example, the MRL/lpr strain presents depression-like behavior and deficits in cognitive function without anxiety-like behavior[8]; NZB/W F1 mice exhibit congenital abnormalities, anxiety-like behavior and decreased locomotor activity[12] and B6.Nba2 mice show a strong anxiety phenotype and a mild depression phenotype[29]. Recently, several neuropsychiatric manifestations are also reported in PIL mice, including learning and memory disturbance[17] and decreased spontaneous activities[18]. In this study, we observed a severe and consistent anxiety- and depression- like phenotype in PIL mice from month 4 to 8 (Fig. 3B-D). In line with previous findings[30, 31], our results confirmed an association between olfactory dysfunction and depression-like behavior. However, no significant result was found in novel object recognition test, social novelty preference test, rotarod test and PPI test (Fig. 4). It should be noted that the MRL background of MRL/lpr mice itself may be involved in the severity of neurobehavioral abnormalities[24]. Whereas, our PIL mice, without specific genetic background, are more suitable to investigate the neuroimmune mechanism of NPSLE.
Several observations have suggested that cytokine dysregulation may contribute to depression-like behavior of NPSLE. First, increased levels of cytokines, especially IL-6, IL-8, IL-1, TNF-α and interferon-α (IFN-α), are found in the cerebrospinal fluid of NPSLE patients[32–35]. Second, the early dysregulation of cytokine production concords with the onset of symptoms of depressive-like behavior in MRL/lpr mice[36] and in other rodent strains[37, 38]. Third, anhedonia and other behavioral indices of depressive-like behavior in mice can be replicated by exogenous cytokines, such as IL-6 and TNFα[36, 39], and are prevented by knockout of their receptors[39, 40]. In this study, we found substantial increase of cytokine expression (IL-1β, IL-6, TNF-α and IL-10) in the brain of PIL mice at as early as month 1 (Fig. 5A). Because cytokine overexpression occurred earlier than behavioral deficits, it may be one of the initiating factors in the depressive-like behavior of PIL mice. The increase of cytokine levels in the brain might be attributable to an interaction between the peripheral and central immune systems. Pristane, known as a membrane-activating compound, can induce apoptosis in peripheral tissues, and produces sufficient autoantigen substrates for immune intolerance, which can lead to overproduction of cytokines and development of lupus-like autoimmunity[41]. The increased peripheral inflammatory cytokines can enter into the brain directly[42] or disrupt the integrity of BBB[43, 44] by acting on vascular endothelial cells[45–47]. In this study, we detected significant extravasation of evens blue dye in the brain tissues suggesting a leakage of BBB in PIL mice (Fig. 5B). Meanwhile, the innate immune cells in the brain, such as glia cells, can be activated and produce more inflammatory cytokines. Previous studies on the models with systemic inflammation have revealed an increase of glia activation even under the condition of intact BBB[48]. Furthermore, we found IgG deposition in the choroid plexus and lateral ventricular wall (Fig. 6). A significant increase of IgG deposition has also been observed in the hippocampus zone of PIL mice[19]. IgG deposition in the brain may occur following the impairment of blood ventricular barrier and choroid plexus vascular barrier caused by cytokines in PIL mice. IgG deposition can promote endothelial damage, microglial activation and inflammatory mediator production, resulting in a positive feedback loop to disrupt the balance of immunity homeostasis[49, 50]. Furthermore, we found an indication of phagocytosis of IgG by microglia (co-localization of IgG with microglia), suggesting that microglia may play a neuroprotective role in PIL mice.
Microglia are innate immune cells in the brain, connecting the nervous system with the immune system. Several studies have suggested that microglia are involved in the development of neuropsychiatric symptoms of NPSLE. First, microglia can be activated by cytokines, such as TNF-α, and inhibited by downregulating markers of microglia activation in MRL/lpr mice[51]. Second, microglia may initially migrate to BBB to protect its integrity, and then transform into a reactive phenotype to destroy BBB and trigger neuroinflammation in NPSLE model mice[52]. Third, antagonism of microglia activation can significantly attenuate spatial memory deficit and depression-like behavior in MRL/lpr mice[53]. In line with these, we found microglial activation during the initial stage of the disease in PIL mice, indicated by an increase of microglia density and specific morphological changes (Fig. 7A, C). It has been suggested that activated microglia may play a neuroprotective role during the early stage of neurodegenerative disorders[54]. Microglia could respond to the stimulation of neurotransmitters, such as adenosine triphosphate (ATP), glutamate and histamine, released by damaged neurons, and consequently enhance phagocytic activity to clear unwanted debris, protein aggregates and soluble antigens, thus reduce their damaging effects on neurons[55]. However, we also found the density of microglia declined sharply in the hippocampus of PIL mice from month 4 to 8. The sustained immune attack during the protracted process of SLE may cause activation-mediated apoptosis of microglia[56]. The loss of microglia may severely hamper their capacity for combating immune challenge and repairing tissues, resulting in neural damages and behavioral deficits. It is noted that recent study has demonstrated that the neuronal degeneration or remodeling induced by the interaction between activated microglia and neurons may contribute to the cognitive dysfunction[57]. One limitation of our study was lack of the examination of neuronal degeneration markers, such as neuronal complexity, length of dendrites and the number of spines in the brain of PIL mice. This issue is required to be fully investigated in the future. Together with previous results, our findings suggest that the dynamic change of microglia is an important involvement of the neuropathology of NPSLE, and regulating microglia may be a promising therapeutic strategy for NPSLE. In contrast, astrocytes were persistently activated throughout the observed course of the disease in PIL mice (Fig. 7B, D). Astrocytes generally interact with both neural and non-neural cells and exhibit dynamic activities crucial for neural circuit function, neurological function and behavior[58]. In neural disorders, astrocytes can be activated and produce nerve growth factors and immune mediators, such as IL-1 and nitric oxide (NO), in response to brain inflammation or injury[59]. Histopathological investigations of NPSLE patients’ brains confirm the widespread presence of activated astrocytes along with microglia in the heterogeneous pathological changes[60]. As astrocytes are an important component of BBB, activated astrocytes are the typical hallmark of BBB dysfunction[61]. In addition to the physiological role of astrocytes in synaptic refinement, activated astrocytes have also been implicated in pathological synapse loss and dysfunction following injury or nervous system degeneration in adults[62]. Overall, astrocyte activation may contribute to the neural dysfunction in PIL mice, however the detailed mechanisms warrant further studies.
Another important finding of our study was that lipofuscin deposited in the neurons of both cortex and hippocampus in PIL mice (Fig. 8). Lipofuscin, known as age pigments, are autofluorescent lipopigments formed by lipids, metals and misfolded proteins, which are especially abundant in neural cells. Lipofuscin within the brain increase not only with age but also with pathological processes such as neuronal dysfunction and a repertoire of cellular alterations, including oxidative stress, proteasome, lysosomal and mitochondrial dysfunctions[63–67]. Recent evidence has suggested that lipofuscin may represent a risk factor or driver for different neurodegenerative disorders[68]. Furthermore, senescent neural cells have been found accumulating in the hippocampal of MRL/lpr SLE model mice with depressive behavior[69]. The increase of lipofuscin deposition in PIL mice may be attributable to the overexpression of cytokines in the brain, which can induce the synthesis of oxidative stress products. Accumulated oxidative stress products within neurons may orchestrate inflammation, disrupt the metabolisms of lipid and metals, and then promote the generation of lipofuscin. Loss of neuroprotective effects from microglia may accelerate this pathological process during the progressive stage of SLE. Even though, we neither observed significant changes of neuron density nor signs of neural apoptosis in the cortex and hippocampus. Thus, lipofuscin deposition is not a consequence of increased neuronal cell death and neural dysfunctions in PIL mice remain reversible.