Autoantibodies Affecting Cell Signaling Proteins and Their Symptoms in Long COVID Patients

Long/Post-COVID Syndrome (LC/PCS) is characterized by persisting or novel clinical symptoms after an acute infection with the severe acute respiratory syndrome coronavirus 2 (SARS-CoV2) virus, causing Coronavirus Disease 2019 (COVID-19) [1]. LC prevalence varies widely, being dependent on i.a. risk factors and vaccination status. [2,3]. As many as 65 million people were estimated to be affected with LC worldwide [4]. Consequently, LC has the potential to impact the quality of life of each patient, health care systems and economy [5].

The diverse and varying LC symptoms argue for different molecular pathways, which mediated LC pathogenesis. Several patients suffer at clinical symptoms, caused by specific restrictions and dysfunctions of distinct organ systems (e.g., heart, lung) [5,6,7,8] being objectively visible on, e.g., X-ray, while some patients complain about dominantly psychological impairments after COVID-19 [9]. Additionally, patients report of post-exertional malaise (PEM), neurological impairments, including cognitive symptoms (e.g., impairment of memory issues, attention deficits, brain fog), disrupted sleep patterns, loss of hair, taste or smell, postural orthostatic tachycardia syndrome (POTS) and dysautonomia [10,11,12,13,14,15,16,17,18,19,20]. The precise underlying molecular mechanisms of the diverse clinical phenotypes in the context of LC are elusive. Researchers propose that next to viral persistence, a hyperstimulation of the immune system or molecular mimicry might induce the generation of autoantibodies (AAb) targeting a variety of receptors, involved in the regulation of the autonomic nervous system, immune modulation and vascular response [21,22,23,24,25]. Recent data suggest an involvement of AAb targeting G protein-coupled receptors (GPCR-AAb) in LC pathogenesis. Interestingly, a subgroup of GPCR-AAb with functional activity (GPCR-fAAb) were observed in sera of patients with PCS [26]. In contrast to well-known AAb, which primarily trigger immune responses, destroying their targets and/or associated tissues, GPCR-fAAb do not consistently lead to apoptosis or inflammation in cells, tissues or organs. Instead, these 'functional autoantibodies' initiate uncontrolled signaling cascades via hyperactivation of their target receptors with consecutive prolonged desensibilization of them [27]. This activation is hypothesized to lack regulatory mechanisms, such as receptor downregulation, thus inducing a pathological cellular cascade [28].

Previously these GPCR-fAAb were observed in sera of patients with different disorders, sharing a vascular component and no causal therapeutic option. Research data showed that Trypanosoma cruzi, the pathogen responsible for Chagas' Cardiomyopathy, triggers a functional autoimmune response against the human β1adrenergic receptor (β1-fAAb) in the cardiovascular system, mediated by molecular mimicry [29]. Additionally, a seropositivity of β2-fAAb was also observed in sera of patients with myocarditis, hypertension, POTS, glaucoma, allergic asthma and chronic fatigue syndrome (CFS) [27,30,31,32,33,34,35,36]. Of interest, the coexistence of β2-fAAb and muscarinergic M2 receptor (M2)-fAAb has also been observed in sera of patients with POTS, dysautonomia and PCS, respectively [26,37]. It can be hypothesized that different GPCR-fAAb pattern might trigger different clinical phenotypes. Thus, it was the aim of the present study to investigate the phenotype of patients with LC and a seropositivity of GPCR-fAAb considering the different GCPR-fAAb patterns in comparison to controls.

GPCR-fAAb seropositive patients with LC showed different patterns of GPCR-fAAb (Figure 1). Most patients with LC showed a seropositivity of β2-fAAb (92.8%), M2-fAAb (87.1%), AT1-fAAb (85.6%), and MAS-fAAb (85.6%). Seropositivities of α1-fAAb (16.0%), Nociceptin-fAAb (14.4%), and ET-A-fAAb (2.6%) were a less common feature in the study cohort. There was no gender effect across the different GPCR-fAAbs (Figure 1A).

The frequency of the self-reported symptoms in this study cohort is visualized in Figure 2. 'Fatigue/PEM' (88.7%), 'lack of concentration (88.7%)' and 'POTS (53.6%)' were the most common ones, reported by the patients with LC. The frequencies of the other self-reported symptoms were 'dizziness' (47.4%), 'hair loss' (29.4%) and 'deterioration of pre-existing disorders' (10.8%, especially neurological disorders, 2.1%). A gender effect was observed for clinical symptom 'lack of concentration' (p = 0.0133, CI: [1.25, 6.89]; OR = 2.94), 'hair loss' (p = 2.23 × 10⁻⁶, CI: [3.14, 15.88]; OR = 7.06) and 'deterioration of pre-existing disorders' (p-value = 0.049854; CI: [1.00, 9.88]; OR = 3.14).

A decision tree could be created for the clinical symptoms, 'hair loss', 'dizziness', and 'POTS', yet not for the other clinical symptoms (Figure 3A–C).

These decision trees suggest, which biological markers (antibodies, age, gender) are most influential for each symptom in the dataset. The clinical symptom 'hair loss' is primarily separated by gender, then age. The clinical symptom 'dizziness' hinges almost entirely on age splits, whereas the clinical symptom 'POTS' uses a mix of GPCR-fAAb (β2-fAAb, M2-fAAb and α1-fAAb) and age.

Additionally, the clustered heatmap (Figure 4) displays the pairwise correlations among all the studied variables, including gender and age. Hierarchical clustering dendrograms are placed along the top and left edges of the heatmap. These dendrograms group variables based on the similarity of their correlation profiles: those that tend to be positively or negatively correlated with the same sets of variables cluster more closely. Diagonal cells (variable vs. itself) are shown in bright red, representing perfect correlations of 1.0. By examining the color blocks and the branch lengths in each dendrogram, it is possible to visually identify variable groupings ('lack of concentration' and 'fatigue/PEM') that share similarities, as well as pairs or clusters of variables that exhibit notably high or low correlations. Among the GPCR-fAAb, β2-fAAb and M2-fAAb also form a tight cluster, indicating a strong positive correlation and possibly shared mechanistic roles or occurrence patterns.

Logistic regression output (including all GPCR-fAAb in the model) showed a significant relation between β2-fAAb and the clinical symptom 'dizziness' (p-value = 0.015, CI: [1.45, 31.59]). The positive estimate (1.91) indicates that individuals with a seropositivity of β2-fAAb are more likely to experience this symptom. After adjusting for confounding variables, the presence of the specific predictor variable (β2-fAAb) is significantly associated with a 6.77 times higher odds of experiencing the outcome ('dizziness') compared to its absence. In addition, a significant relation between the symptom β2-fAAb and the symptom 'lack of concentration' (p-value = 0.014, CI: [1.50, 37.07], estimate 2.0) and POTS (p-value = 0.0126, CI: [1.52, 32.17], estimate 1.94) were observed, respectively. After adjusting for confounding variables, the presence of the specific predictor variable (β2-fAAb) is significantly associated with 7.45 times higher odds of experiencing the outcome ('lack of concentration') and 6.99 for 'POTS' compared to its absence. A seropositivity of ET-A-fAAb was significantly related to the symptom 'deterioration of pre-existing neurological disorders' (p-value = 0.044, CI: [0.08, 5.69]). The confidence interval suggests that this association is statistically significant, and the lower limit provides a lower bound for the strength of the association. Odds ratio was 16.67.

A subsequent analysis was conducted to evaluate the interaction between M2-fAAb and β2-fAAb, with fatigue/PEM designated as the outcome variable. The interaction term was not statistically significant, indicating no strong evidence that the effect of M2-fAAb on fatigue/PEM is modified by β2-fAAb levels. In a model restricted to the main effects of M2-fAAb and β2-fAAb, both β2-fAAb and gender were significantly associated with fatigue/PEM at the 5% significance level (p = 0.02 for each). Variance Inflation Factor (VIF) to assess multicollinearity for each predictor can be seen in Figure A1.

LC has a multifaceted clinical phenotype, characterized by various clinical symptoms, thus diverse molecular pathways can be assumed. One of these LC pathways seems to be mediated by a specific autoimmune pattern [26]. Classical autoimmunity is based on a loss of self-tolerance, which is based on each individuals' epigenetics and genetics, being modified by the external factors (e.g., life style, environment) [38,39]. The innate und adaptive immune response interact in these classical molecular mechanisms, generating autoantibodies and finally apoptosis [40]. In addition to this well-known autoimmune response, a different type of AAb was observed in Grave's disease and dilatative cardiomyopathy [41,42,43]. These AAb did not induce apoptosis. They can activate their targets and thus change the receptor mediated physiologically cellular response in a pathologically way, therefore called fAAb. Contrary to physiological agents these fAAb-mediated response is characterized by a prolonged desensibilization of the receptors and lack of receptor downregulation [27,44]. This special type of AAb was observed in systemic and local disorders previously, most of them involving the cardiovascular system. As a recent study described a seropositivity of GPCR-fAAb in patients with PCS, it was the aim of the present study to investigate a potential clustering of GPCR-fAAb and their association to specific clinical symptoms in patients with LC. A total of 194 sera of patients with LC were tested by the established cardiomyocyte-bioassay for analysis of the presence and functionality of the GPCR-fAAb [45]. A pattern of β2-fAAb, M2-fAAb, AT2-fAAb and MAS-fAAb was observed to be the most common ones in the study cohort. Especially, β2-fAAb, was observed to be linked to the clinical symptom 'dizziness', 'lack of concentration' and POTS. As β2-fAAb and M2-fAAb were observed to form a tight cluster, a strong positive correlation was assumed. In the statistical model restricted to the main effects of β2-fAAb and M2-fAAb, both β2-fAAb and gender were significantly associated with fatigue/PEM. Interestingly, the β2-fAAb does not modify the M2-fAAb response on that clinical symptom.

The present study identified a cluster of β2-fAAb, M2-fAAb, AT1-fAAb and MAS-fAAb as most common feature in this study cohort. Clustering of different GPCR-AAbs seems to be present in several disorders: Patients with POTS showed seropositivities of β1-fAAb, β2-fAAb, α1-fAAb, M2-fAAb, and AT1-fAAb [44,46,47,48]; distinct patterns including β2-fAAb and M2-fAAb were observed in Chagas cardiomyopathy, electrical cardiac abnormalities, complex regional pain syndrome and fatigue syndrome [35,49,50,51]. A combination of α1-fAAb and AT1-fAAb was reported in Thromboangiitis obliterans [52]. A slight age-trend was observed for Noci-, MAS-, M2- and AT1- fAAb, potentially arguing for the well-known feature in classical autoimmunity that dysregulation can be followed by autoimmunity in older ages [44]. In addition, recent data demonstrate that age is associated with increased natural autoantibody production, being further amplified in patients with severe COVID-19 [53].

Considering gender as cofactor, the clinical symptoms, 'hair loss', 'lack of concentration' and 'deterioration of pre-existing conditions' were significantly more common symptoms in female patients than in male patients, potentially due to hormonal factors [54,55].

The way and timepoint for generation of GPCR-fAAb is still elusive. Classical AAbs targeting Angiotensin II (AngII), Angiotensin Converting Enzyme 2 (ACE2) and MAS1 oncogene (MAS) receptors were frequently observed in COVID-19 patients, suggesting that the production of GPCR-fAAb may begin even at the time of acute COVID-19 infection [56,57,58]. The virus enters cells through the RAS system and changes their affinity for their ligands [59]. In addition, the high intensity of the immune response to viral infections may also contribute to the production of GPCR-fAAb, potentially by molecular mimicry. However, as the virus is eliminated by the immune response, GPCR-fAAb persist, building up a cyclical process involving continuous GPCR-fAAb production, receptor overstimulation, increased intra- and extracellular oxidative stress and ultimately microcirculatory dysfunction, consequently resulting in LC symptoms [56].

The most common GPCR-fAAb, observed in patients with LC were β2-fAAb, M2-fAAb, MAS-fAAb and AT1-fAAb. It can be assumed that the molecular pathways of that four GPCR-fAAb might interact: β2-fAAb showed a link to an impaired capillary microcirculation in local and systemic disorders [27,36,37,60,61]. In addition, AT-1 receptor mediated signals can trigger vasoactive processes via, e.g., vasoconstriction [26]. It has been shown that AT1-fAAb can increase intracellular reactive oxygen species (ROS) production in human vascular smooth muscle cells, probably due to increased NADPH oxidase activity [62]. Increased ROS can induce DNA damage, triggering the activation of a DNA repair enzyme called poly-ADP-ribose polymerase-1 (PARP1) in endothelial cells [63]. In addition, research data showed that the AT1-fAAb mediated effect on renal arteries is intensified in ischemic or inflammatory tissues, whereas only a moderate effect was observed on renal arteries with good oxygen supply [37]. Of interest, M2-fAAb triggers the production of prostaglandin E2 and nitric oxide synthase, and induce the expression of cyclooxygenase 2 and inducible NOS mRNA, consequently inducing proinflammatory signals [64]. Further on, M2-fAAb was observed to inhibit the L-type Ca²⁺ current, increase cyclic GMP and the outward potassium current [65]. MAS-fAAb can reduce the effects of angiotensin II [66,67].

Summarizing this, it can be assumed that GPCR-fAAb act as pathogenic agonistic AAbs by (I) inducing a permanent activation of their targets, which (II) lose their mechanisms for protection against overstimulation; these GPCR-fAAb-mediated effects were (III) enhanced in ischemic/inflammatory tissues, inducing a (IV) pathogenetic circulus of inflammation and ischemia (Figure 5).

The present study is not without limitations. The study was designed as cross-sectional study; thus, it would be of interest if a seropositivity of GPCR-fAAb would change over time, potentially also associated with the clinical LC-symptoms. In addition, as study participants were adults, a clinical study, investigating the presence of GPCR-fAAb in patients with LC, being younger than 18 years, would be of interest.

Patients with LC were diagnosed with a wide range of symptoms. Serological analysis yielded that a cluster of β2-fAAb, M2-fAAb, MAS-fAAb and AT1-fAAb was the most common one in LC-patients with fatigue/PEM as one major clinical symptom (Figure 6). As additionally, a strong positive correlation between β2-fAAb and M2-fAAb and an association of β2-fAAb and gender to the main self-reported clinical symptom fatigue/PEM were observed, we assume a critical impact of GPCR-fAAb on LC-pathogenesis.