Microglia pre-activation and neurodegeneration precipitate neuroinflammation without exacerbating tissue injury in experimental autoimmune encephalomyelitis

Multiple sclerosis (MS) is a chronic neuroinflammatory disease, which in the majority of patients starts with a relapsing/remitting disease course [32]. In this stage of the disease, new focal lesions are formed mainly within the white matter and are characterized by an influx of T and B cells into the brain associated with blood-brain barrier damage, primary demyelination and reactive astrocytic scarring [25]. After ten to fifteen years, the relapsing/remitting disease course converts into a phase of steady progression (secondary progressive MS). Some patients skip the relapsing disease stage and already start with steady progression (primary progressive MS) [32]. The pathological substrate of progressive MS is the slow expansion of pre-existing lesions in the white matter and the cerebral cortex as well as diffuse neurodegeneration in the entire normal-appearing white and grey matter [26]. While the pathology of relapsing/remitting MS is partially reproduced in models of experimental autoimmune encephalomyelitis (EAE), no experimental animal model has yet been described that reflects the pathological changes in the progressive stage of MS [24]. Recent studies on autopsy tissue from patients with progressive MS suggest that a cascade of events, including microglia activation, oxidative injury, mitochondrial damage and the resulting energy deficiency (virtual or histotoxic hypoxia), plays a major role in driving demyelination and neurodegeneration in the progressive stage of the disease [8, 9, 34].

Since progressive MS develops in an age-dependent manner [46, 60], it has been suggested that age-related changes in the brain contribute to the propagation of the neurodegenerative process in the affected patients. In humans, brain aging is associated with chronic microglia activation [39], iron accumulation in myelin and oligodendrocytes [12], a pro-oxidative state of the brain [45], mitochondrial alterations as well as progressive myelin damage and neurodegeneration [51]. Thus, inflammatory MS lesions in aging patients develop on the background of a pre-injured CNS, which is not reflected in rodent EAE models. We therefore hypothesized that neuroinflammation induced on a background of pre-existing inflammatory and neurodegenerative conditions may be amplified and even convert from an acute self-limiting monophasic disease into chronic progressive disease like in the progressive stage of MS.

To test this hypothesis, we made use of a rat model (LEWzizi) that originally descended from the zitter (zi/zi) rat, a spontaneous attractin (Atrn) mutant [22, 42] presenting with a variety of neuropathological features that resemble key aspects of MS pathology, such as neurodegeneration [52, 53], hypomyelination [18], microgliosis [16, 49], extensive iron accumulation [49] and dysregulated anti-oxidative systems [11]. Here, we first characterized the newly established LEWzizi rat model and further investigated whether pre-existing microglia activation, hypomyelination and axonal damage amplify neuroinflammation and tissue injury in passive MBP-EAE. We observed a minor increase of clinical disease; however, the combination of a pre-injured CNS environment with the induction of EAE did not lead to an exacerbation of oligodendrocyte or axonal pathology. Interestingly, we observed a topographic shift of inflammation spreading from the spinal cord to forebrain regions, which were pronouncedly pre-affected by zitter pathology.

Since we intended to induce EAE in the zitter rat model, we first had to transfer it to an immunologically compatible background. For this, we crossed zitter rats (Sprague-Dawley background) with wild-type Lewis animals resulting in LEWzizi (LEW.SD-Atrn^zi/zi) rats. They exhibited the same phenotypic traits as described for the original zitter model [42] such as moderate generalized body tremor, unsteady gait and age-dependent moderate progressive flaccid paresis of the hind limbs (data not shown).

In the present study, we aimed to investigate whether pre-existing microglia activation, iron accumulation and neurodegeneration, which are conditions seen in human brain aging [58] and particularly in the normal-appearing white matter of patients with progressive multiple sclerosis [34], amplify experimentally induced neuroinflammation and tissue injury in the course of passive EAE. Additionally, we intended to test, whether acute monophasic T cell-mediated EAE is transformed into a chronic progressive course, when occurring on such a pre-injured background. To this end, we crossed the outbred zitter rat model to inbred Lewis rats, which are highly susceptible to the induction of EAE, thereby generating so-called LEWzizi rats. Like zitter rats [14, 16, 18, 19, 49], LEWzizi rats present with massive microgliosis accompanied by a general pro-inflammatory environment and with hypomyelination associated with aberrant myelin sheaths, reduced numbers of myelinated fibres, decreased oligodendrocyte densities and the presence of myelin degradation products within phagocytes. Similarly to zitter rats [14, 52, 53, 55, 56], we observed neurodegenerative processes, as shown by the accumulation of APP within neuronal spheroids and endbulbs, in naïve LEWzizi brains and spinal cords.

During normal aging in humans, iron accumulates in oligodendrocytes and axons [12]. In disease conditions, excessive iron accumulation can pose a major problem. When its levels exceed endogenous storage capacities or when it is liberated due to cellular stress and cell death, iron can potentiate oxidative stress via the Fenton reaction [59]. A general pitfall of rodent models of human diseases is their very low iron load; thus, a crucial co-factor for the study of neuroinflammation or neurodegeneration is absent in rodent-based experiments [48]. This can be circumvented by studying LEWzizi rats, which, similarly to zitter rats [49], present with age-dependently increasing, abnormally high iron levels within axonal tracts, oligodendrocytes (brain only) and microglia (both brain and spinal cord). Moreover, they show signs of oxidative stress, which had previously been comprehensively described for zitter rats as well [11, 14, 36, 54, 55]. Here, we used a histopathological approach and observed high numbers of LEWzizi microglia expressing iNOS. However, we could not detect any parenchymal p22phox expression in the whole CNS, in line with previous reports involving normal and inflamed rodent brains [48].

Taken together, LEWzizi rats represent a suitable model to study the consequences of experimentally induced neuroinflammation on a background of microglia activation, oxidative injury, neurodegeneration and iron accumulation. Induction of EAE by passive transfer of MBP-specific CD4⁺ T cells in both Lewis and LEWzizi rats led to typical monophasic EAE. Disease scores were slightly, but significantly higher in LEWzizi compared with Lewis rats, which, however, may have resulted from an additive effect of LEWzizi- and EAE-related clinical signs [42]. Clinical appearance of EAE manifests through lesions in the spinal cord and medulla oblongata, while lesions in the forebrain do not impact clinical scores. In LEWzizi rats, passive EAE started slightly, but significantly earlier than in Lewis controls. We suspect that the early start of disease resulted from the increased pro-inflammatory milieu in LEWzizi spinal cords. Initiation of EAE after induction of focal injuries reportedly reduced the threshold for lesion formation in previous studies [29].

Our neuropathological analyses clearly showed that T cell infiltration patterns significantly differed between wild-type Lewis and LEWzizi rats. Similarly to Lewis rats, T cell infiltrates were observed in lumbar spinal cords, the predilection site for classical MBP-EAE [1], of LEWzizi rats, although they were less abundant. On the contrary, T cells were detected at sites in the CNS, which are usually not targeted in Lewis controls, such as the deep brain stem nuclei including the mesencephalon. We believe that this difference in the topography of T cell infiltration is due to the underlying pro-inflammatory and neurodegenerative LEWzizi pathology, since T cell re-distribution specifically targeted pre-damaged areas. Similar re-distribution phenomena of EAE lesions have already been described earlier: Focal injuries to rat brains induced by burns, cryogenic insults, chemicals or anoxic injuries prior to active sensitization with a CNS antigen or passive transfer of CNS-reactive T cells led to the formation of EAE lesions in the pre-damaged forebrain areas, which were barely targeted in control rats [7, 21, 23, 27, 28, 30, 31, 40]. Additionally, it has been shown in mice that induction of forebrain neurodegeneration by a cuprizone-supplemented diet followed by active or passive MOG-EAE resulted in immune cell recruitment to areas of the brain parenchyma that are usually not targeted in naïve animals [4, 44, 47]. Interestingly, in contrast to the brain, LEWzizi-specific pre-damage in the spinal cord did not lead to enhanced T cell recruitment into this particular CNS area. Instead, significantly reduced numbers of T cell cuffs and total CD3⁺ immunoreactivity were observed upon disease induction by CD4⁺ T cells.

The role of microglia in EAE has been studied previously in transgenic mouse models. Stalling of microglia proliferation massively dampened clinical severity and neuropathology [15]. Restricting microglia activation abolished clinical disease and strongly diminished neuroinflammation and tissue destruction [10]. Microglia ablation via Csf1R inhibition at the onset of disease massively dampened clinical severity and greatly reduced EAE pathology [38]. Reduction of microglia reactivity by dipyridamole attenuated clinical as well as histological EAE scores [50]. Similarly, application of minocycline or microglia/macrophage inhibitory factor TKP attenuated EAE [6, 41]. Contrary, other previous studies have suggested that microglia pre-activation does not interfere with autoimmune disease induced by CD4⁺ T cells. Ablation of microglia in a CX3CR1-transgenic mouse model leads to a rapid repopulation of microglia from remaining progenitor cells resulting in type-1-interferon-pathway-related microglia activation associated with neurodegeneration [43]. In that study, induction of EAE at the peak of microglia activation did not change T cell-mediated inflammation, neurodegeneration or demyelination. Similarly, induction of EAE in animals with active cuprizone lesions resulted in the precipitation of inflammatory lesions at the sites of cuprizone-induced demyelination, but neither cuprizone-related demyelination and neurodegeneration nor EAE-related neurodegeneration were altered [44]. Here in our study, we found that microglia pre-activation in LEWzizi rats did not exacerbate tissue injury in CD4⁺ T cell-induced EAE. Although naïve LEWzizi animals initially had high numbers of activated microglia, the levels of CD68⁺ phagocytes (microglia and peripheral macrophages) were not increased at the peak of EAE compared with Lewis rats. Likewise, the number of Iba⁺ phagocytes, which was strongly elevated in naïve LEWzizi rats, was either equal or lower at the peak of EAE compared with Lewis animals. The phagocyte activation markers iNOS and p22phox as well as APP-positive neuronal spheroids and endbulbs (representing neuronal injury and damage, respectively) followed a rather similar pattern.

Taken together, we could collect in our study interesting and unexpected data that were in contrast to our initial key hypotheses. Importantly, we observed that tissue damage induced by CD4⁺ T cell-mediated inflammation was not aggravated, when induced on the LEWzizi-specific background of microglia pre-activation, cellular iron accumulation, diffuse myelin damage and neurodegeneration. Also, EAE did not convert into a chronic progressive inflammatory disease. In view of previously reported data, these findings were rather unexpected. For example, tissue injury in progressive MS is described to depend on chronic inflammation comprising T and B cells and occurs on the background of microglia activation, iron accumulation and neurodegeneration [34], conditions similar to those in our LEWzizi model. A possible explanation for this discrepancy may reside in differences in the inflammatory process between MS and EAE. Most EAE models, including the one used in our experiments, are driven by MHC class II-restricted CD4⁺ T cells [24]. However, in MS and particularly in patients with active progressive disease, the inflammatory reaction mainly consists of tissue-resident CD8⁺ memory T cells and B cells [33, 57]. Thus, future studies will be necessary to determine the interaction of these inflammatory cell populations with microglia activation and tissue damage.