Multifactorial White Matter Damage in the Acute Phase and Pre-Existing Conditions May Drive Cognitive Dysfunction after SARS-CoV-2 Infection: Neuropathology-Based Evidence

Background: There is an urgent need to better understand the mechanisms underlying acute and long-term neurological symptoms after COVID-19. Neuropathological studies can contribute to a better understanding of some of these mechanisms. Methods: We conducted a detailed postmortem neuropathological analysis of 32 patients who died due to COVID-19 during 2020 and 2021 in Austria. Results: All cases showed diffuse white matter damage with a diffuse microglial activation of a variable severity, including one case of hemorrhagic leukoencephalopathy. Some cases revealed mild inflammatory changes, including olfactory neuritis (25%), nodular brainstem encephalitis (31%), and cranial nerve neuritis (6%), which were similar to those observed in non-COVID-19 severely ill patients. One previously immunosuppressed patient developed acute herpes simplex encephalitis. Acute vascular pathologies (acute infarcts 22%, vascular thrombosis 12%, diffuse hypoxic–ischemic brain damage 40%) and pre-existing small vessel diseases (34%) were frequent findings. Moreover, silent neurodegenerative pathologies in elderly persons were common (AD neuropathologic changes 32%, age-related neuronal and glial tau pathologies 22%, Lewy bodies 9%, argyrophilic grain disease 12.5%, TDP43 pathology 6%). Conclusions: Our results support some previous neuropathological findings of apparently multifactorial and most likely indirect brain damage in the context of SARS-CoV-2 infection rather than virus-specific damage, and they are in line with the recent experimental data on SARS-CoV-2-related diffuse white matter damage, microglial activation, and cytokine release.


Introduction
The COVID-19 pandemic is placing a severe burden on socioeconomic, physical, and mental health. The mid-and long-term consequences are expected to be challenging for the coming years. Neurological symptoms and complications have been a focus of attention for both acute-phase symptoms (e.g., headache, hyposmia, cerebrovascular disease with ischemic and hemorrhagic lesions, symptoms associated with encephalitis mimics, seizures, and cranial and peripheral nerve damage, including Guillain-Barré syndrome) and the sequel of a SARS-CoV-2 infection, termed "Long-or Post-COVID-19" syndrome that includes various symptoms such as fatigue, memory, and attention problems ("brain fog"), potentially progressive cognitive decline [1][2][3] and autonomic dysfunction.
We therefore performed a postmortem neuropathological assessment of a series of Austrian patients, who died of/with COVID-19 infection, with the aim of identifying underlying brain, spinal, and neuromuscular pathologies, and to discuss potential substrate(s) of mid-and long-term cognitive alterations as have been reported for post-COVID-19 patients [18][19][20][21].

Materials and Methods
The study cohort included 32 patients (19 male, 13 female), aged 21 to 97 years at death (mean 65.8, median 66.5), who died during the pandemic in 2020 and 2021. Inclusion criteria were having had a positive PCR for SARS-CoV-2 before death and an autopsy Viruses 2023, 15, 908 3 of 25 including brain extraction. No exclusion criteria were applied. Details on the premorbid conditions of the patients and cause of death are summarized in Table 1.
Brains were obtained at autopsy at several Austrian Pathology Departments (Klinik Favoriten and Klinik Ottakring in Vienna, University Hospital Graz, University Hospital Innsbruck), at the Department for Forensic Medicine in Vienna, and by Dr. Denk (unaffiliated forensic medicine expert). Ethical approval for the use of human postmortem brain tissue for research purposes was obtained from the Medical University of Vienna (EK1454/2018). The study was performed in accordance with the World Medical Association Declaration of Helsinki.
For the neuropathological examination, we assessed formalin-fixed and paraffinembedded brain tissues from multiple brain areas, including olfactory bulb and olfactory tract, frontal, temporal, parietal, and occipital cortex, anterior and posterior basal ganglia, amygdala, anterior, and posterior hippocampus, posterior hypothalamus, including the corpora mamillaria, thalamus, and supra-and infratentorial white matter, midbrain, pons, medulla oblongata, cervical spinal cord, cerebellar vermis, and hemispheres with white matter and dentate nucleus. In a subgroup of patients, skeletal muscle and the peripheral nerve were also obtained. As muscles were obtained at autopsy from different departments, only formalin-fixed muscle tissue was available for the study. Muscle type differentiation was performed by immunohistochemistry, with antibodies directed against slow (type I) and fast (type II) myosin, acknowledging that some fibers co-express both.
Five-micrometer-thick sections were stained with hematoxylin-eosin, luxol fast blue (LFB), and, in selected cases and brain areas, Elastica van Gieson combined stain. Immunohistochemistry was performed in selected brain areas using a panel of primary antibodies (see Supplementary Table S1) on a DAKO autostainer (Dako, Glostrup, Denmark). For the visualization of an antigen-antibody reaction, we applied the Dako Envision Kit and diaminobenzidine as chromogen.
A total of 32 cases had a PCR-confirmed SARS-CoV-2 infection either in vivo (30 cases) or postmortem (2 cases), and two thirds had a clinical diagnosis of COVID-19 pneumonia that required intensive care unit treatment (Table 1); no cases with "Post-COVID-19" were studied. We also assessed three additional brains of patients who died in 2009/2010 due to influenza virus (H1N1) infection and compared the distribution of T-and B-cells, microglial reactivity, and additional neuropathological changes.
The assessment of the inflammatory reaction (anti-CD8, anti-CD20, and anti-HLA-DR) was graded in each brain area separately for perivascular spaces and within the parenchyma in a semiquantitative manner, as follows: 0 = absent (no cellular infiltrates); 1 = mild (<5 cells in 1 mm 2 ); 2 = moderate (5-10 cells/mm 2 ); 3 = abundant (>10 cells/mm 2 ). This scale refers to the density of the labeled cells for CD8 and CD20, and for HLA-DR, in addition to the density, we assessed the change in the morphology: in the perivascular space: 1 = mild (<50 cells in 1 mm 2 , rod to delicate ramified morphology); 2 = moderate (50-100 cells/mm 2 , moderate enlargement with coarse branching); and 3 = abundant (>50-100 cells/mm 2 , broad amoeboid morphology). In the parenchyma: 1 = mild (<200 cells/mm 2 , delicate ramified morphology); 2 = moderate (200-500 cells/mm 2 , moderate enlargement with broad branching); and 3 = abundant (>200-500 cells/mm 2 , broad amoeboid morphology). A comparison between patients who died from COVID-19 and patients who died from other conditions who additionally tested positive for SARS-CoV-2 was performed.  The evaluation of white matter changes was based on HE, Luxol Fast Blue, and HLA-DR stains and it was semi-quantitatively assessed in the regions of interest (see below) as: 0 = absent (no obvious vacuoles; no white matter pallor and homogeneous intense blue myelin staining in LFB), + = mild (<5 vacuoles/mm 2 , patchy and smooth discoloration in LFB stain), ++ = moderate (5-10 vacuoles/mm 2 , partly confluent areas of white matter discoloration in LFB stain), and +++ = severe (>10 vacuoles/mm 2 , diffuse confluent reduction in blue intensity in LFB). On HE and LFB, we assessed the degree of pallor and vacuolation, on HLA-DR the density and morphology of microglia in the white matter of the main lobes, frontal, temporal, parietal, and occipital, corpus callosum, and cerebellum, and integrated them in an overall score of "white matter damage". CD8, CD20, and HLA-DR were assessed for their perivascular and parenchymal compartment separately. As these patients were in the acute phase of disease and most of them were in the intensive care unit, a detailed correlation with neurological symptoms was not possible.
In one patient with concomitant herpes simplex virus type 1 infection, we performed an ultrastructural analysis of the viral particles with a Zeiss electron microscope. For this, a small fragment of formalin-fixed temporal cortex was embedded in Epon resin, cut ultrathin, and contrasted with uranyl acetate.

Results
A summary of the results is presented in Tables 2-5 and Supplementary Tables S1 and  S2, as well as in Figures 1-3.

Global Findings
All cases (case #1 to case #32) showed signs of diffuse edema with pericellular and perivascular tissue rarefaction. This was observed at different degrees of severity ranging from mild to severe. There was a spectrum of pathologies that ranged from a complete absence of focal lesions to acute ischemic infarcts, acute hemorrhages, single vascular thromboses, mild olfactory and/or cranial nerve neuritis, or nodular brainstem encephalitis. No signs of obvious vascular or meningeal inflammation were identified (see Tables 2-5). All cases showed a variable degree of microglial activation, which was predominantly and diffusely affecting the white matter.

Inflammatory Changes (Figure 1)
A: Olfactory system: eight patients showed mild CD8-positive T lymphocyte-dominated inflammatory infiltrates along the olfactory bulb and tract (cases #2, #3, #5, #6, #7, #12, #21, #22). In one patient, mild infiltrates were also noted within the entorhinal area adjacent to the hippocampus and amygdala (case #1). These infiltrates were accompanied by relatively marked microglial activation in both grey and white matter. All other cases showed diffuse, unspecific microglial activation but no obvious inflammatory nodules ( Figure 1A,D,G). In H1N1-infected cases, olfactory bulb was unfortunately not available for analysis, but entorhinal areas also showed relatively prominent microglial activation in 2 of 3 cases (cases #1 and #2), one of them with small microglial nodules ( Table 2, case #2).
B: (Micronodular) brain stem encephalitis: seven cases showed mild (cases #2, #6, #12, #16, #19, #20, #21, #32) and three cases moderate (cases #3, #4, #5) inflammatory nodules in the midbrain, pons, and/or medulla oblongata. The motor nucleus of the vagal nerve and its emergence from the brainstem was involved in most cases. The cellular infiltrates were mainly composed by CD8-positive T lymphocytes and microglia/macrophages ( Figure 1B-D). No CD20-or CD79A-positive B-cells were identified in any brain region in the parenchyma, only isolated intravascular or isolated perivascular cells were seen in single cases. The distribution of pathology was comparable to that observed in three brains of H1N1-infected patients ( Table 2). The spinal cord was analyzed in 10 cases, and no signs of myelitis were identified. No signs of meningeal inflammation were detected in any case.
We additionally analyzed the peripheral nerves (n. medianus) of 10 patients. We found no signs of neuritis, perineuritis, endotheliitis, vasculitis, or of demyelination suggestive of GBS/CIDP. Only one elderly patient showed a moderate axonal neuropathy. Spinal ganglia were not available.
Moreover, we investigated the formalin-fixed skeletal muscle of the forearm of 10 cases and found no signs of myositis, endotheliitis, vasculitis, necrotizing myopathy, or metabolic damage. Only two patients showed a tendency to type 2 fiber atrophy by immunohistochemistry (cases #17, #29), which could be attributed to the prolonged immobilization at the intensive care unit. D: Co-occurring infections: one patient (case #32) with prior immunosuppression due to treated breast cancer presented an acute herpes simplex encephalitis (HSV1) in the context of severe respiratory distress due to bilateral SARS-CoV-2 pneumonia ( Figure 3C1-C4).
Although there were no major differences in respect to inflammatory changes between patients who died of SARS-CoV2-related pneumonia and those with COVID-19 who died of other conditions, patients with longer intensive care and septic state had, in general, more perivascular CD8-positive cells and diffuse microglial activation than patients with shorter disease duration and acute death, caused by myocardial complications or the pulmonary embolism. We could not identify a specific immunoreactivity for SARS-CoV-2 antigens.      Figure 2) A: Hemorrhagic leukoencephalopathy: one patient (case #4) showed multiple subacute white matter hemorrhages that were distributed throughout the supratentorial and infratentorial white matter and affected all cerebral lobes, the basal ganglia, brainstem, and cerebellum. The lesions showed a peripheral rim of hemosiderin pigment and central red blood cells in different stages of degradation (Figure 2A-C). These lesions were mostly associated with white matter vessels, but without signs of vasculitis, fibrinoid necrosis of the vascular wall, or thrombosis. B: Leukoencephalitis: in one patient with olfactory neuritis and brain stem encephalitis, we further identified the mild foci of CD8-positive lymphocytes diffusely distributed within the white matter ( Figure 2D-F, case #2 Table 2); however, these were identified without signs of selective demyelination.
C: Diffuse white matter damage: this was a nearly constant feature in all cases and ranged from subtle changes in the LFB stain to the formation of relatively prominent white matter vacuoles ( Figure 2D-I), likely representing intramyelinic oedema. These changes were more prominent in posterior, parieto-occipital areas. Moreover, there was a diffuse, white matter predominant microglial activation with clear perivascular accentuation ( Figure 2E,F,I, Table 3), which also involved the corpus callosum and the anterior commissure, and was associated with a perivascular accumulation of CD8-positive lymphocytes. No obvious phagocytosis of perivascular myelin such as in acute demyelinating encephalomyelitis (ADEM) was observed. Moreover, we could not identify demyelinating plaques in the white matter or cortical demyelination. No prominent white matter gliosis was detected. Axons were comparatively well preserved. Figure 3A, Tables 4 and 5) A: Vascular thrombosis: four cases: venous and arterial thrombosis was identified in a patient who also had a sinus vein thrombosis in the context of a generalized thrombotic event (case #9). One patient showed small arterial thrombosis in a meningeal vessel (case #19, Figure 3A2) and another a fibrin clot in the dural sinus (case #25). In case #26, we observed hemorrhagic lesions around a centrally located fresh venous thrombus in the frontal cortex. In the temporo-basal cortex, in an area of an apparently old traumatic injury, we found an occluded vessel with bleeding and hemosiderin deposits ( Figure 3A1). The thrombotic events were associated with acute ischemic or hemorrhagic lesions.

Vascular Pathology (
We frequently found stagnant vessels with abundant entrapped leukocytes throughout most of the cases of our series, without the direct attachment of CD8-positive T-cells to endothelial cells, thrombosis, the disruption of the endothelial cell layer, and no signs of endotheliitis or vasculitis. This phenomenon is frequently observed in postmortem brains in the context of diffuse stagnation and sepsis, independent of the cause. The adherence of some neutrophils on the vessel wall should therefore not be considered a specific inflammation of the endothelium. B: Diffuse hypoxic-ischemic damage: this type of damage was observed in several patients, either in relation to the venous or arterial thrombosis or in the context of a cardiac dysfunction (Table 5). Moreover, most patients suffered from severe pneumonia, a frequent cause of brain hypoxia. Hippocampus CA4 and CA1 sectors, thalamic neurons, cerebellar Purkinje cells, and dentate nucleus neurons appeared to be most vulnerable. Areas of acute or subacute cortical laminar necrosis were also identified (cases #6, #12, #13, #20, #25).
C: Lewy body pathology: alpha-synuclein immunoreactive Lewy bodies and Lewy neurites were identified in three patients in early brainstem or olfactory stages (cases #4, #10, #11), mainly affecting the dorsal motor nucleus of the vagal nerve, the olfactory bulb and tract, and the olfactory area around the amygdala. As these patients had not been clinically diagnosed as having Parkinson's disease, they most likely represent an "incidental" Lewy body pathology ( Figure 3B4). D: TDP43 proteinopathy: two elderly patients (88 and 86 years, case #1 and #26) showed pTDP-43 protein aggregates in the limbic system representing LATE (limbic agerelated TDP43 encephalopathy), with neither hippocampal sclerosis nor signs of motor neuron disease or extensive fronto-temporal involvement. E: Prion disease: not a single case of (co-)incidental Creutzfeldt-Jakob disease or other type of prion disease was observed. (Table 2) The three patients who died of acute respiratory distress syndrome after H1N1 infection had similar neuropathological changes to those described here for SARS-CoV-2 (see also Table 2). Case #1 (f, 57) and Case #2 (m, 57) showed a disseminated microangiopathy with fibrinoid necrosis of several vessels with disseminated microhemorrhages in white matter hemorrhagic leukoencephalopathy, associated with moderate hypoxic-ischemic damage and diffuse edema, the latter also present in case #3. Moreover, disseminated fungal microabscesses (fungal pneumonia and sepsis, case #2) and a fungal septic necrotic focus in the occipital white matter (septic shock, case #1) were identified. Both cases also had brainstem microglial nodules and case #3 diffuse and brainstem predominant microglial activation. Patient #2 also showed a mild metabolic encephalopathy with increased Alzheimer type II astrocytes. No viral antigens could be demonstrated by PCR in any of the three cases in brain tissue.

H1N1-Infected Cases
age-related TDP43 encephalopathy), with neither hippocampal sclerosis nor signs of motor neuron disease or extensive fronto-temporal involvement. E: Prion disease: not a single case of (co-)incidental Creutzfeldt-Jakob disease or other type of prion disease was observed.

Other
One patient suffered from a traumatic subdural hemorrhage and severe brain contusion that led to death (case #29); their death was not causally linked to COVID-19. Acute traumatic brain injury 1 (3%)

Discussion
We present the neuropathological findings in a series of SARS-CoV-2-infected patients who died during the pandemic in 2020-2021 in Austria; two thirds of these patients had severe COVID-19 and received ICU treatment. We observed that one common neuropathological feature is the development of a diffuse white matter damage with intramyelinic edema and prominent diffuse and perivascular-accentuated microglial activation associated with the mild accumulation of perivascular CD8-positive T-cells and comparatively less axonal damage. The intensity of these alterations varied between cases, but it was globally moderate to severe and also included single cases with prominent vacuolar leukoencephalopathy, partly reminiscent of posterior reversible encephalopathy syndrome (PRES-like) [22], and one case of a micro-hemorrhagic HURST-like leukoencephalopathy [23]. The latter was also observed in one of three patients with H1N1 influenza infection from 2009 to 2010 and could be also related to disseminated intravascular coagulation, among other factors.
These changes seemed to be unrelated to the presence of acute/focal lesions and were accentuated in the parieto-occipital regions. Diffuse white matter damage is a non-specific finding and may be caused by several afflictions. Aside from inherited conditions causing, for example, hereditary leukodystrophies or CADASIL, acquired toxic or metabolic derangements may also alter the myelin sheath. We propose that a (severe) SARS-CoV-2 infection as in most of our cases may induce myelin damage not by the virus itself, but by a combination of an excessive cytokine release by the immune system and a metabolic/hydroelectrolytic derangement caused by the functional damage of vital organs and/or toxic damage by poly-medication (e.g., anticoagulation during extracorporeal membrane oxygenation).
During the acute phase of the disease, around 34% of patients experienced impaired memory, concentration, or attention [24]. Some recent studies have examined the immune cell profile in the CSF of patients with "Neuro-COVID-19" by single-cell sequencing [25,26] and found signs of local immune overactivation with a broad clonal T-cell expansion and reduced interferon response [25]. Another recent study performed a multidimensional characterization of immune mediators in the CSF and plasma of patients belonging to different Neuro-COVID-19 severity classes and identified a distinctive set of CSF and plasma mediators associated with blood-brain barrier (BBB) impairment, elevated microglia activation markers, and a polyclonal B-cell response targeting self-antigens and non-self-antigens underlying post-acute COVID-19 syndrome [27]. Altered BBB biomarkers have been also described in infected patients with neurological complications by other authors [28]. These results from other studies support the existence of immune-mediated mechanisms in severe COVID-19 that may contribute to neurological alterations, and that there might be a compromised antiviral response [25,26,29]. While one group did not observe an expansion of B-cells in CSF, findings that are also in line with our postmortem study where a B-cell component was nearly absent, others did. These differences between previous studies may have resulted from the analysis at different time points of the disease. Yet, another study analyzed several cyto/chemokines in the CSF and/or sera of patients with COVID-19 and neurological symptoms and found predominantly an alteration of the neurovascular unit rather than a prominent COVID-19-related neuroinflammatory profile [30]. Furthermore, a recent study performed in mice and humans observed a white matter-selective microglia reactivity after SARS-CoV-2 infection. The authors also observed a persistently impaired hippocampal neurogenesis, decreased oligodendrocyte counts, and signs of myelin loss in the context of increased CSF cytokines/chemokine levels, including CCL11 [31]. These findings were also detected after H1N1 influenza infection. These previous experimental findings do also support our postmortem findings of diffuse white matter microglial activation and likely intramyelinic edema.
In the post-illness stage, 15-80% of patients complained of cognitive dysfunction as part of a long-term clinical manifestation called post-COVID-19 syndrome [32]. This syndrome encompasses symptoms that persist after about 12 weeks of infection and include fatigue and concentration, attention, and memory deficits, among others. As white matter plays an important role in cognition [33], it might well be that acute and dif- fuse, probably multifactorial, white matter damage contributes to the development of post-COVID-19 cognitive dysfunction. In addition to the possible consequences of altered cytokine/chemokine levels, acute vascular pathologies, including thrombotic events, hypoxic/ischemic/hemorrhagic lesions in strategic brain areas, and/or the presence of pre-existing small vessel disease with/without leukoencephalopathy in patients with cardiovascular risk factors (altogether roughly observed in about one third of our cohort) may contribute to short-, mid-, or even long-term memory/attention problems in post-COVID-19 syndrome [34][35][36][37]. Additionally, important are underlying "preclinical" neurodegenerative morbidities in elderly people (e.g., Alzheimer type changes, limbic TDP-43 proteinopathy, premotor Lewy body pathology, argyrophilic grain pathology) might act as additional factors. These frequently develop in strategic limbic and brainstem nuclei (in our study, in about two thirds of patients, including patients in their late 50s/early 60s) and may also contribute to cognitive deficits after SARS-CoV-2 infection.
The reversibility of these changes will probably depend on the severity of the myelin damage and axonal integrity as well as on the host-dependent immune system, cognitive reserve, and several yet unknown factors. Of note, cognitive dysfunction has been also repeatedly reported after critical illness [38].
As we learnt from previous pandemics, such as the one caused by H1N1 virus in 1915-1930, long-term neurological consequences might appear even decades after the infection. A good example is linked to postencephalitic parkinsonism, a neurodegenerative tauopathy that developed years after influenza infection and an episode of encephalitis lethargica. In those cases, however, no viral antigens have yet been demonstrated in the brains [39]. It has been therefore postulated that an immunological process would be more likely the trigger of neurodegeneration rather than the virus per se. Therefore, it will be of utmost importance to perform future neuropathological studies in patients who survive COVID-19 and die years later of another condition, to investigate whether there are any sequelae, including unusual neurodegenerative features. It will be also particularly important to perform well-designed and controlled morphometric studies.
We observed a mild inflammatory CD8-positive T-cell infiltration of the bulbus/tractus olfactorius in eight cases. In addition, we identified a mild to moderate micronodular brainstem encephalitis in ten patients. In two of those, a cranial nerve inflammation was also identified. It is a matter of debate whether these changes are directly related to the viral infection and contribute to some of the symptoms, including the prominent brainstem dysfunction and respiratory problems the patients present, or whether this is a consequence of severe, end-stage diseases [40][41][42][43]. As in some other studies [44,45], we could not identify a specific signal for SARS-CoV-2 antigens by conventional immunohistochemistry, but this does not exclude the presence of viral particles. Some authors have detected small viral sequences by RT-PCR and RNAseq in a small fraction of patients [46], but it is still unclear whether they represent true CNS infection. Further studies with other, more sensitive modalities are currently in progress. The micronodular brain stem encephalitis that we observed in about one third of the SARS-CoV-2-infected patients was comparable to that seen in the three H1N1infected patients of our study, and might therefore not necessarily be a direct consequence of a SARS-CoV-2 infection itself. Microglial activation in the postmortem human brain is a frequent, non-specific finding and can be observed in a vast range of conditions, particularly in the context of critical illness, where an interplay of hypoxic-systemic-metabolic-infectious conditions generally contribute to death [17,47,48]. An alternative cause of these findings could be a reactivation of a latent viral infection of another type. This is frequently observed in autopsy series in the context of sepsis or severe terminal disease [49] and is not unique to SARS-CoV-2 infection. Interestingly, a single patient with chemotherapy-related immunosuppression developed a fulminant necrotizing herpes simplex virus encephalitis in the context of SARS and bilateral pneumonia. Moreover, two patients had a non-CNS reactivation of herpes simplex virus infection, two patients of Epstein-Barr virus, one patient suffered from aspergillosis, and one from mucormycosis. These co-morbid conditions might have contributed to patients' immune response in the brain and underscores the importance of ruling out potential co-occurring infections. Per contra, patients were critically ill and received several medications, including antiviral, antibiotic, antimycotic, and anti-inflammatory drugs, including corticosteroids, which likely altered the inflammatory milieu in the brain, influencing the postmortem neuropathological findings, which is another note of caution for the interpretation of our results.
The limitations of the study include its retrospective nature, the analysis of postmortem brain tissue of a clinically heterogeneous patients group with different pre-existing conditions and risk factors, different clinical evolutions and treatment strategies, and the lack of patients with "post-COVID-19" syndrome.
In summary, our results support some previous neuropathological findings of an apparently multifactorial brain damage in the context of SARS-CoV-2 infection and are in line with recent experimental data on diffuse white matter damage, microglial activation, and cytokine release. At least, the alpha and/or delta SARS-CoV-2 variant that affected most of the patients of our study seem not to be what is classically considered a "neuro(no)tropic virus" in a strict sense [26,46,50,51]. Such neurotropism combines neuroinvasiveness, neurovirulence, and neuroreplication, the direct infection of neural cells/neurons by the virus and its replication there, causing an acute polioencephalitis-associated variable Tcell-mediated inflammation and neuronal loss, as known for well-established neurotropic viruses, e.g., flaviviruses [52]. In contrast, we believe that the prominent immunological reaction induced by the SARS-CoV-2 virus and subsequent cytokine boost, potential immunological mimicry of SARS-CoV-2 antibodies with neuronal/glial antigens [53], the septic condition, and the dysfunction of vital organs with subsequent hypoxic brain damage, mitochondrial/metabolic/hydroelectrolytic derangements, are all likely contributors to a diffuse myelin damage, as already observed in other viral diseases where a pathogenetic network co-orchestrates virus-specific as well as -unspecific factors, e.g., in HIV infection [54]. Such alterations may decompensate or accelerate pre-existing "silent" brain conditions, in particular, underlying subtle neurodegenerative pathologies, chronic vascular pathologies, and latent viral infections, that might alter synaptic and neurotransmitter functions in the short-, mid-, and long-term.
While those who died are not necessarily representative of the majority of people who survive and develop long COVID-19/post COVID-19 and cognitive sequelae, the observed white matter damage in postmortem studies might be an important contributing factor to the long-term effects of COVID-19 infection.
Whether new virus variants will change or broaden the spectrum of neurological complications remains unclear, but it is likely. More recent COVID-19 variants that are likely to have a lower mortality may perhaps have less neuropathological impact. Further interdisciplinary studies are essential, and neuropathological expertise and brain banks can provide a valuable resource of tissue-based information to further elucidate the neurological, and particularly the cognitive consequences/sequelae of a severe infection, with multiorgan involvement, as exemplified by SARS-CoV-2.
Supplementary Materials: The following supporting information can be downloaded at: https://www. mdpi.com/article/10.3390/v15040908/s1, Table S1: List of primary antibodies; Table S2: Semiquantitative assessment and mapping of inflammatory infiltrates across the different brain regions.