Assessment of a combination of plasma anti-histone autoantibodies and PLA2/PE ratio as potential biomarkers to clinically predict autism spectrum disorders

Autism spectrum disorder (ASD) is a neurodevelopmental disorder characterized by deficiencies in social interaction and repetitive behaviors. Multiple studies have reported abnormal cell membrane composition and autoimmunity as known mechanisms associated with the etiopathogenesis of ASD. In this study, multiple regression and combined receiver operating characteristic (ROC) curve as statistic tools were done to clarify the relationship between phospholipase A2 and phosphatidylethanolamine (PE) ratio (PLA2/PE) as marker of lipid metabolism and membrane fluidity, and antihistone-autoantibodies as marker of autoimmunity in the etiopathology of ASD. Furthermore, the study intended to define the linear combination that maximizes the partial area under an ROC curve for a panel of markers. Forty five children with ASD and forty age- and sex-matched controls were enrolled in the study. Using ELISA, the levels of antihistone-autoantibodies, and PLA2 were measured in the plasma of both groups. PE was measured using HPLC. Statistical analyses using ROC curves and multiple and logistic regression models were performed. A notable rise in the area under the curve was detected using combined ROC curve models. Additionally, higher specificity and sensitivity of the combined markers were documented. The present study indicates that the measurement of the predictive value of selected biomarkers related to autoimmunity and lipid metabolism in children with ASD using a ROC curve analysis should lead to a better understanding of the pathophysiological mechanism of ASD and its link with metabolism. This information may enable the early diagnosis and intervention.

www.nature.com/scientificreports/ amino acid residue in proteins. Although histone proteins are conserved and weak immunogens, but they display strong immunogenicity after nitration. Rabbits challenged with peroxynitrite-modified histone prompt high titre antibodies, demonstrating that peroxynitrite modification created immunogenic epitopes. This might help to suggest that peroxynitrite-modified histones autoantibodies are contributed in the initiation and progression of autoimmune diseases among which is ASD 15 . Nadeem et al. 30 proved that BTBR mice as rodent model of ASD have increased lipid/protein oxidation products in cerebellum and neutrophils/CD3 + T cells, raised NADPH oxidase (NOX2) activity, and inducible nitric oxide synthase (iNOS) expression concomitant with decreased levels of antioxidant enzymes and glutathione. Histones have been shown to bind strongly to anionic phospholipids and are rapidly released during cell injury 31 . Hirai et al. 32 have demonstrated that phospholipids and cardiolipin can inhibit histone function. Pereira 33 has postulated an important role of anti-histone antibody in anti-cardiolipin reactions. They suggest the possibility that autoantibodies to histones or to histone-phospholipid complexes may contribute to anti-phospholipid activity. In case of disrupted BBB in autistic patients, the influx of anti-histone antibodies from blood to the brain and its binding to membrane phospholipids could lead to neuronal damage either by direct interaction of these complexes with neurons or by functional impairment due to their interaction with astrocytes, activation of endothelial cells and adherence to different brain cells 34 .
The ROC curve is the most commonly used graphical tool for evaluating the diagnostic value of a biomarker. In ROC curves, the area under the curve (AUC) is an effective way to summarize the overall diagnostic accuracy of the test. AUCs are effective and combined measure of sensitivity and specificity that measures the characteristic validity of a biomarker. It takes values from 0 to 1, where a value of 0 shows an inaccurate test and a value of 1 reflects a perfectly accurate test. In general, an AUC of 0.5 proposes no diagnostic or discrimination value of a biomarker, 0.7 to 0.8 is considered acceptable, 0.8 to 0.9 is considered excellent, and more than 0.9 is considered outstanding ability to diagnose patients with and without the disease or condition based on the ROC analysed biomarker 35 .
Due to the complexity of the brain and pathogenesis of ASD, a combination of several biomarkers could epitomize a more potent approach to diagnose this disorder 36,37 .
The ROC curve is the most common graphical tool for assessing the diagnostic power of a biomarker. For example, our most recent study showed that a combination of nine biomarkers and 5 ratios distinguished autistic children from healthy controls with a high AUC (area under the ROC curve) value 37 .
In this study, it was very interesting to test the possibility to find ROC combination models between twenty four previously published biomarkers presenting different signalling pathways in the plasma of the same group of patients with ASD. Logistic regression was performed between the twenty four variables with all possible permutations, and two models of combined ROC that have significant increase in the AUC of one or more variable were considered Supplementary Table S1. First model was published as combined Alpha-Synuclein, cyclooxygenase-2 and prostaglandins-EP2 receptors as neuroinflammatory biomarkers of ASD 38 . The second model is that presented in the current study showing combination between anti-histone autoantibodies with PLA2/PE relative ratio as a perfect biomarker of pro-inflammation and membrane phospholipid alteration in ASD. Understanding the mechanism behind the combination of anti-histone autoantibodies and PLA2/PE might help to highlight their roles in the etiology of this disorder.

Methods
Participants. The study protocol was approved by the ethics committee of medical Collage, King Saud University according to the most recent Declaration of Helsinki (Edinburgh, 2000). Two groups of study participants were enrolled in the study comprising of 40 patients with ASD and 40 age and gender matched healthy control. The inclusion period ranged from May 2019-December 2019. Informed consents were obtained from all participants and signed by their parents. Both groups were recruited through the ART Center (Autism Research & Treatment Center) clinic. The diagnosis of ASD was confirmed in all study subjects using the Autism Diagnostic Interview-Revised(ADI-R) and the Autism Diagnostic Observation Schedule (ADOS) and 3DI (Developmental, dimensional diagnostic interview) protocols. The ages of autistic children involved in the study were 5.5 ± 2.2 years old. All were simplex cases (i.e. family has one affected individual). All are negative for fragile x syndrome gene. They have a well characterised autistic phenotypes with mild-severe recorded values between 30-45.5 Childhood Autism Rating Scale (CARS), moderate-severe Social Responsiveness Scale (SRS) (70-˃ 76), and a well characterized Short Sensory Profile. Forty age-and sex-matched control participants, aged (5.3 ± 2.3) years old were recruited from the Pediatric Clinic at the same hospital. Participants were excluded from the study if they had fragile X syndrome, serious neurological disorders, or any other medical conditions. All participants were screened by way of parental interview for existing or historical physical illness. All patients and controls included in the study were on similar but not identical diet and none of them were on any special gluten restricted diet.
Blood sampling. After overnight fasting, blood samples were collected from children with ASD and healthy controls by a skilled technician into 3-ml blood collection tubes containing EDTA. Immediately after collection, blood was centrifuged at 4 °C at 3000g for 20 min. The plasma was separated, distributed into three 0.5 ml aliquots (to avoid several freeze-thaws cycles) and stored at − 80 °C until use 39  which was coupled with Apex M625 software (Autochrom, USA). As the nebulizing gaz, N2 was used at a flow rate of 4 l/ min, and a nebulizing temperature of 40 °C. The gain was set at 8 and 2.0 bar N2. A 125 × 4.0 mm Si-60 column with 5 μm particle diameter (Lichrosher) was used. The elution program was a linear gradient with 80:19.5:0.5 (V/V) chloroform: methanol: water: ammonia (NH3) at 22 min and the column was allowed to equilibrate until the next injection at 27 min. The injection volume was 50 μl. A liquid phase extraction procedure adapted from the method described by Bligh and Dyer 40 was used to extract the serum samples. Briefly, 50 μl of each sample was diluted with 750 μl deionized water and mixed well. Then 2 ml of methanol and 1 ml of chloroform were added to the sample and mixed well. Then the mixture was homogenized (Rotary mixture 34,526, Snijders) for 15 min. The mixture was centrifuged for 5 min at 4000 rpm, and 0.5 ml of supernatant was directly used in a 2 ml vial for HPLC analysis 39 . The HPLC instrument was programmed to accommodate 100 vials in one run. Assay of cPLA2 cPLA2 concentration was measured according to the manufacture's instruction using a competitive enzyme immunoassay technique a product of Amsbio, Blue Gene Company. The detection range of the product was 1.56-100 ng/ml.

Assay of antihistone-antibodies.
Antihistone-autoantibody was measured using ELISA kit, a product of Genway Biotech, San Diego, USA. The samples were run together, in parallel on the same run with the standards, and controls. After 30 min incubation at 25 °C, antibodies to highly purified human antihistone-specific antibodies, if present in the serum, would bind in the wells. Washing the wells to eliminate the unbound serum antibodies. Horseradish peroxidase enzyme conjugated anti-human IgG immunologically binds to the bound patient antibodies, forming a conjugate/antibody/ antigen complex. When substrate is washed in the presence of bound conjugate, it is hydrolyzed to form a blue colour. The addition of an acid stops the reaction, resulting in a yellow end product. The intensity of this yellow colour was measured photometrically at 450 nm. The amount of colour is directly proportional to the concentration of IgG antibodies present in the control and patients samples. To increase accuracy, all samples were analysed twice in two independent experiments to evaluate inter assay differences and to confirm reproducibility of the detected results 39 .
Statistical analyses. In this study, IBM SPSS software, version 22.0 (IBM Inc., Armonk, USA) was used to analyze the data. Normality of data for each group was done using Shapiro-Wilk Test, the results are presented as the Minimum, Maximum and Median, and Mann-Whitney Test for comparing two nonparametric groups was used for statistical evaluations, with P ≤ 0.05 considered a significant difference. Correlation between various nonparametric variables was done using Spearman rank correlation coefficient (R). For the combined ROC curves, odds ratios (ORs) obtained from logistic regression analyses describe associations of biomarkers with the clinical status. ROC curves were constructed for each logistic regression model. The area under the curve (ROC-AUC) was compared between each marker and marker combination using a nonparametric method. Similarly, the predictiveness curves, as a complement to ROC curves of the measured parameters, were drawn using a Biostat 16 computer program and the x-axis represents percentile rank of the biomarker, the y-axis represents the probability of identifying the disease, and the horizontal line is the prevalence of the disease. The predictiveness curve represents a spontaneous and graphical tool to compare the predictive power of any measured biomarker.
Ethics approval and consent to participate. All procedures performed were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. This work was approved by the ethics committee of King Khalid Hospital, King Saud University (Approval number: 11/2890/IRB).

Consent for publication.
All authors have read the manuscript and agreed for the submission. Table 1 demonstrates the demographic data of control and ASD participants. Table 2 and Fig. 1 demonstrate the normality of PLA2/PE and anti-histone in plasma of children with ASD compared to gender and age matching healthy controls. It can be easily noticed that while the PLA2/PE ratio was not normally distributed in control, it was normally distributed in ASD patients. Anti-histone was not normally distributed in both groups. Table 3 and Fig. 2 show the min., max., and median of PLA2/PE and anti-histone in plasma of patients with ASD compared to healthy controls. It is clear that while PLA2/PE is significantly higher in ASD patients, anti-histone is significantly lower compared to controls (P = 0.001). Table 4 and Fig. 3 demonstrate the negative correlations between the two measured variables with P of 0.023. Table 5 and Figs. 4 and 5 show the ROC and predectiveness curves of the two measured variables independently while Table 6 and Figs. 6 and 7 demonstrate the combined ROC curves showing the remarkable increase of the AUC and the much improvement of the predictiveness curve of the two combined markers.

Discussion
The present study has revealed the usefulness of logistic regression as a simple clinical method with great potential for assisting the diagnosis of ASD. The results of ROC analysis indicated that the combination of anti-histone autoantibodies and PLA2/PE produced remarkably higher AUCs, best sensitivity and specificity for diagnosis of ASD. This could help to understand and interpret the relationship between different recorded markers. www.nature.com/scientificreports/ A majority of diseases are triggered by imperfections in signalling pathways. The nature of these imperfections and how they are induced and interacted differs extremely. The notion of signalosome alteration and disease offers a frame for considering how defects in signalling pathways can result in disease. It is useful to separate these defects into phenotypic and genotypic remodelling of the signalosome. In case of phenotypic remodelling, the behaviour of the cells is changed leading to dysfunction and finally to disease. Clearly, there is an urgent need to understand how phenotypic changes in signalosome are interacting to clinically present disease states. Understanding the overlapping etiological mechanisms involved in a complex disorder such as ASD is a crucial challenge in recent biomedical research. Despite the increasing prevalence of ASD, their brain phenotypes remain unclear. Atypical levels of phospholipids, lipid mediators, glutamate/glutamine, creatine /phosphocreatine measured compounds suggest that neuron or glial density, mitochondrial energetic metabolism, and/or inflammation contribute to ASD neuropathology 41 . This might help to design better therapies 41,42 . Signalosomes are protein complexes that undergo assembling to form large supramolecular complexes that increase the local concentration and signalling activity of their individual components. It is well documented that conjugation of PE with selective proteins is involved in membrane closure, and autophagy as a catabolic process that is largely regulated by extracellular and intracellular signalling pathways and contributed to multiple neurological disorders among which is ASD [43][44][45][46] .
The evidence that the immune dysfunction expected plays a role in the etiology/pathophysiology of ASD is significant 47 . In the current study while the PLA 2 /PE ratio was dramatically increased level (P = 0.001), antihistone was significantly reduced (P = 0.001) in the plasma of autistic patients compared to healthy controls (Tables 2 and 3 and Figs. 1 and 2). As the main esterase participating in the synthesis of several inflammatory mediators, cPLA2 has been established to play vast roles in oxidative stress and autoimmune in the neurological disorders 48 . Children with autism are recognized to have elevated phospholipase A2 (PLA 2) activity compared to their matched healthy control 10,49 . It has been suggested that the instability noticed in fatty acid concentrations may be because of an increase in PLA 2 activity concomitant with much lower phospholipids possibly in connection with the high oxidative stress observed in autistic patients 50 .
The mechanisms that bring about the unusual functions of antihistone and how these impacts on neurological pathogenesis stay inadequately understood. In the present study, the increase of PLA2/PE can easily relate to the alteration of membrane phospholipids and multiple lipid mediators in ASD [51][52][53][54][55] . Phospholipids in the brain are rich with polyunsaturated fatty acids (PUFAs), and PLA2s are the key enzymes catalysing the hydrolysis of the PUFAs in the sn-2 position of the glycerol moiety. Omega 6 arachidonic acid (AA) as a substrate for cyclooxygenases and lipoxygenases, serves as the precursor for synthesis of eicosanoids and prostanoids that  57 and is the precursor for biosynthesis of neuroprotectin D (NPD1) and resolvins [58][59][60] .
In the present study, the logistic regression combination recorded between anti-histone as anti-nuclear antibodies with PLA2/PE demonstrating a much higher AUC value of 0.991, higher sensitivity, and specificity (Table 6 and Fig. 6), compared to the independent ROC-AUCs of each variable (0.77 & 0.900) ( Table 5 and Fig. 4) was interesting because it could help to highlight a defected signalosome in ASD which can be targeted as early intervention strategy.
Although research has focused on the identification of genetic abnormalities, emerging studies increasingly suggest that immune dysfunction as a possible risk factor contributing to the neurodevelopmental discrepancies observed in ASD. Edmiston et al. 61 , extensively studied the relationship between autoimmunity and autoantibodies and ASD. They highlighted the involvement of maternal anti-brain autoantibodies, which are supposed  www.nature.com/scientificreports/ to enter the fetal compartment during pregnancy, as a risk factor for developing ASD and are anticipated to contribute to early neurodevelopmental impairments in the developing fetus, later clinically presented as autistic phenotypes 52,62,63 . The remarkable reported lower anti-histone as nuclear autoantibodies in the plasma of patients with autism compared to healthy control (Table 3 and Fig. 2), could be concomitant with much higher level within the brain, a suggestion that can find support through considering the recorded negative spearman's correlation (Table 4 and Fig. 3). This can be attributed to the fact that maternal antibodies have greater access to the fetal brain due to increased permissiveness of the blood-brain barrier (BBB) during gestation 64 . Rossi et al. 65 through the use of immunohistochemistry can visualize not only the presence or absence of anti-brain autoantibodies,    www.nature.com/scientificreports/ but also the specific brain profiles to which they drag. They offer a convincing evidence for several brain-reactive antibodies in plasma from individuals with ASD. This might help to support the suggested influx of anti-histone autoantibodies from plasma to brain through the disrupted BBB as autistic phenotypic remodelling 66,67 .
The importance of combining ROC model between anti-histone and PLA2/PE (Fig. 6) was ascertained with the remarkable improvement of the predictiveness curve of the two combined variables (Fig. 7) relative to the predictiveness curves of each independent variable (Fig. 5). In order to understand the pathological effects of anti-histone in ASD, it is very interesting to know that binding of these autoantibodies to brain structures requires cell destruction, release of antigen and formation of immune complexes. Pereira et al. 31 suggest that anionic phospholipids among which are PE might be another target structure for these immune complexes. Although these phospholipids are known to be present in the inner part of the cell membrane, they may be uncovered under certain conditions such as activation of neuron and microglial cells as another autistic phenotype 9,68 . This phenomenon of phospholipid-autoantibodies complexing was recently ascertained by the work of Nalli et al. 69 which demonstrates that anti-phospholipid antibodies analysed by line immunoassay (LIA) can recognize different proteins epitopes depending on to which phospholipids the protein is binding. The obtained higher AUC of combined ROC for anti-histone and PLA2/PE (AUC of 0.991) can find great support in the previous work of Samuelsen et al. 70 which indicate high prognostic and scientific potential of cross-linked phospholipid-protein conjugates. Moreover, the finding of the present study can find more support in the previous work of Careaga et al. 71 showing an increased anti-phospholipid antibody levels in young children with ASD that is directly related   www.nature.com/scientificreports/ to the severity of the disorder. In statistics, negative correlation is a relationship between two variables in which one variable decreases as the other increases, and vice versa. A perfect negative correlation means the relationship that occurs between two variables is accurately opposite all of the time. Base on this fact the recorded negative correlation between PLA2/PE and anti-histone autoantibodies with a satisfactory value of − 0.3 could explain and support the remarkable increase of ROC AUCs from 0.900, and 0.77 for independent variables to 0.991 for the two combined variables. Up to the suggested influx of anti-histone autoantibodies from plasma to brain, much higher brain anti-histone autoantibodies, could be contributed in the impairment of membrane fluidity as etiological mechanism of ASD presented by altered PLA2/PE and thus combining both variable could have much higher diagnostic accuracy 72 . Up to the discussed data, the suggested relationship between anti-histone autoantibodies and PLA2/PE relative ratio in inducing neuronal damage is illustrated in Fig. 8.

Conclusion
This study highlighted the importance of anti-histone autoantibodies as marker of auto-immunity, and interrelated PLA2/ PE ratio as two potential new targets for understanding the mechanisms involved in the pathogenicity of ASD. Although there are reports on autoimmunity family history in ASD, the presence of only 50% (21/40) of the participants with family history of autoimmunity might affect the generalizability of the presented data which could be considered as limitation of this study. This preliminary findings support the importance of further study of the biological impact of autoantibodies and their association with other etiological mechanisms, behavioural, and cognitive impairments in children with ASD.

Data availability
Raw data can be available on request.
Received: 25 November 2021; Accepted: 27 July 2022 Figure 8. Illustration of the suggested relationship between combined anti-histone autoantibodies, and PLA2/ PE relative ratio in inducing neuronal damage. In response to neuroinflammation and oxidative stress, induced elevation of PLA2 as phospholipid hydrolysing enzyme can trigger the elevation of pro-inflammatory lipid mediators (Cysteinyl leukotrienes, and prostaglandins) as ASD phenotype. Suggested influx of anti-histone autoantibodies from plasma to brain through the disrupted BBB induces the formation of PE-anti-histone autoantibodies complex which increases anti-phospholipid antibody levels and greatly affects membrane fluidity.