Detection of Antibodies against the Acetylcholine Receptor in Patients with Myasthenia Gravis: A Comparison of Two Enzyme Immunoassays and a Fixed Cell-Based Assay

The detection of serum anti-acetylcholine receptor (AChR) antibodies is currently an important tool for diagnosing myasthenia gravis (MG) since they are present in about 85% of MG patients. Many serological tests are now available. Nevertheless, results from these tests can be different in some patients. The aim of this study is to compare the sensitivity of a commercially available fixed cell-based assay (F-CBA) to that of enzyme-linked immunosorbent assay (ELISA) kits for anti-AChR detection in patients with a diagnosis of MG. Overall, 143 patients with a confirmed MG diagnosis were included in the study. The detection and measurement of serum anti-AChR antibodies were performed by three analytical methods, namely, a competitive ELISA (cELISA), an indirect ELISA (iELISA), and an F-CBA, according to the manufacturers’ instructions. Anti-AChR antibody titers were positive in 94/143 (66%) using the cELISA, in 75/143 (52%) using the iELISA and in 61/143 (43%) using the F-CBA (adult and/or fetal). Method agreement, evaluated by concordant pairs and Cohen’s kappa, was as follows: cELISA-iELISA: 110/143 (77%), k = 0.53 (95%CI 0.40–0.66); cELISA-F-CBA: 108/143 (76%), k = 0.53 (95%CI 0.41–0.66); iELISA-F-CBA: 121/143 (85%), k = 0.70 (95%CI 0.57–0.80). Our findings show that the cELISA has better analytical performance than the iELISA and F-CBA. However, the iELISA and F-CBA show the highest concordance.


Introduction
Myasthenia gravis (MG) is an autoimmune neuromuscular disease characterized by autoantibodies targeting proteins in the neuromuscular junction of the skeletal muscles [1].
The clinical manifestations of the disease differ from mild and focal weakness to myasthenic crisis, an acute respiratory paralysis that requires intensive care [2][3][4]. MG symptoms can be limited to the eye muscles, commonly called ocular MG (OMG), or involve other skeletal muscle symptoms, leading to generalized MG (GMG).
Antibody testing is crucial to confirm the clinical suspicion of MG and guide the management of patients [5].
The most common antibody type in the sera of MG patients is against the nicotinic acetylcholine receptor (AChR), which consists of α, β, δ and ε subunits (adult-type AChR-ε) or α, β, δ and γ subunits (fetal-type AChR-γ). Antibodies against all five AChR subunits are detectable in about 85% of patients with GMG and about 50% of patients with OMG [5,6]. They also represent a useful serological biomarker for thymoma, which can be detected in 10-20% of MG patients [7].
About 5-10% of MG patients have antibodies against muscle-specific tyrosine kinase (MuSK) [8][9][10]. Anti-AChR and anti-MuSK are very specific, and, in practice, their detection in patients with suggestive symptoms confirms the diagnosis. When the clinical suspicion of MG arises, anti-AChR antibodies are first tested, followed by Anti-MuSK in AChRnegatives, according to the Italian recommendations for the diagnosis and treatment of myasthenia gravis [5]. Only in a few sporadic cases are both antibodies, i.e., anti-AChR and anti-MuSK, present in the same patient. In recent years, new antibody targets have been identified in seronegative MG patients, including antibodies against lipoprotein-receptorrelated protein 4 (LRP4), agrin, collagen, antistriational muscle (Kv1.4, titin and ryanodine receptors) and cortactin [11][12][13][14]. However, despite the progress achieved in serological testing, no antibodies can be detected in around 1-15% of MG patients [15]. This may be related to the low sensitivity of current testing methodologies.
Various analytical methods are available for serological analysis, including the radioimmunoprecipitation assay (RIPA), enzyme-linked immunosorbent assay (ELISA), dot-blot testing and a commercial biochip based on a fixed cell-based assay (F-CBA), which measures antibodies against AChR and MuSK simultaneously [16][17][18][19][20][21][22][23]. F-CBAs and live cell-based assays (L-CBA) are reported to have higher sensitivity compared to RIPAs or ELISAs. Notably, it has been reported that discordant results may be achieved for identical samples tested by different analytical methods.
In this study, we aimed to compare the performance of three analytical methods, namely, a competitive ELISA (cELISA), an indirect ELISA (iELISA), and an F-CBA, in detecting AChR antibodies in patients with MG.

Study Population
We performed an observational retrospective study at the University Hospital "P. Giaccone", Palermo, Italy, including 143 patients (66 males, 77 females, median age 61 years) with a confirmed diagnosis of MG according to the International Consensus Guidance for Management of MG [24]. We enrolled blood donors as age-and sex-matched healthy controls.
Clinical data, including MG crisis and the state of immunosuppressive treatment, were recorded by reviewing medical records. The presence of thymoma was investigated in all patients by means of computed tomography or magnetic resonance imaging scanning of the mediastinum.
MG patients were classified into five groups according to the Myasthenia Gravis Foundation of America (MGFA) clinical classification at the onset of myasthenic symptoms and at each follow-up [25].
The study was conducted in accordance with the ethical standards as formulated in the Helsinki Declaration and approved by the 'Palermo I' Ethical Committee (nr. 05/2021) on 19 May 2021.
For each subject enrolled, we collected blood samples in dry tubes to obtain sera. The latter was separated within 3 h after drawing and stored at −80 • C until analysis. All analyses were performed at the Institute of Clinical Biochemistry, Clinical Molecular Medicine and Clinical Laboratory Medicine, University of Palermo.

Anti-AChR Antibody Assays
Detection and measurement of serum anti-AChR antibodies were performed by three different assays (Table 1).  [18]. It is a non-isotopic assay based on the ability of AChR autoantibodies to compete with three different AChR monoclonal antibodies (MAbs 1-3) for binding sites on affinitypurified fetal and adult-type AChR. One MAb (MAb1) is coated onto ELISA plate wells, and the other two are labeled with biotin and used in the assay in the liquid phase. In the absence of serum AChR autoantibodies, a sandwich is formed among MAb1, the AChR and the two biotinylated MAbs, which are subsequently detected by the addition of streptavidin peroxidase, which is bound specifically to biotin. In the presence of serum AChR autoantibodies, the formation of the sandwich fails, and the amount of biotinylated MAbs is reduced. A higher concentration of serum AChR autoantibodies is associated with greater inhibition of MAb-biotin binding. The concentration of AChR autoantibodies is measured in nmol/L, and a raised value above the cut-off (0.5 nmol/L) is considered nearly 100% specific for MG. II. iELISA was performed using the commercially available Anti   Table 1 describes detailed performance characteristics of the three different assays.

Statistical Analysis
Statistical analysis was performed by R Language v.4.2.1 (R Foundation for Statistical Computing, Vienna, Austria), with additional packages including "dplyr", "ggplot2", "boot", "VCA" and "irr". Method comparison, using continuous values, was evaluated by non-parametric Passing-Bablok regression. Concordance was also evaluated as a percentage of concordant pairs and by Cohen's kappa with its 95% confidence interval. Analytical performances were evaluated by calculating sensitivity, specificity, positive predictive value and negative predictive value and by ROC curve analysis. AUCs were compared by the DeLong method.

Discussion
The detection of anti-AChR antibodies is currently an important tool for diagnosing MG since even very low titers of serum anti-AChR antibodies are sufficient to confirm the clinical suspicion [9]. Moreover, the gradually increasing titers of anti-AChR antibodies may be detected up to 2 years before the onset of typical MG symptoms [27].
RIPAs represent the gold standard for detecting anti-AChR antibodies due to their high specificity and sensitivity, reaching 99% and 85%, respectively [13]. Additionally, a RIPA is also a quantitative method, and, thus, the quantification of antibody levels could be helpful for patient monitoring. However, RIPAs have some limitations, mainly being the use of radioactive reagents. Over time, non-radioactive alternatives have been developed and commercialized. Among these, ELISAs are the most used. In the last decade, CBAs have also been introduced in MG diagnosis. CBA is a methodology based on the expression of high levels of antigen, i.e., AChR, in the membrane of cells, which can be live (L-CBA) or fixed (F-CBA). L-CBAs have been proven to be highly specific and sensitive, even more than RIPAs [28]. However, the use of L-CBAs in clinical practice is hampered because it requires expertise and cellculture facilities. The limitations of RIPAs and L-CBAs can be overcome by using an FCBA. Indeed, the latter is not based on radioactive material and is less technically

Discussion
The detection of anti-AChR antibodies is currently an important tool for diagnosing MG since even very low titers of serum anti-AChR antibodies are sufficient to confirm the clinical suspicion [9]. Moreover, the gradually increasing titers of anti-AChR antibodies may be detected up to 2 years before the onset of typical MG symptoms [27].
RIPAs represent the gold standard for detecting anti-AChR antibodies due to their high specificity and sensitivity, reaching 99% and 85%, respectively [13]. Additionally, a RIPA is also a quantitative method, and, thus, the quantification of antibody levels could be helpful for patient monitoring. However, RIPAs have some limitations, mainly being the use of radioactive reagents. Over time, non-radioactive alternatives have been developed and commercialized. Among these, ELISAs are the most used. In the last decade, CBAs have also been introduced in MG diagnosis. CBA is a methodology based on the expression of high levels of antigen, i.e., AChR, in the membrane of cells, which can be live (L-CBA) or fixed (F-CBA). L-CBAs have been proven to be highly specific and sensitive, even more than RIPAs [28]. However, the use of L-CBAs in clinical practice is hampered because it requires expertise and cell-culture facilities. The limitations of RIPAs and L-CBAs can be overcome by using an FCBA. Indeed, the latter is not based on radioactive material and is less technically demanding and time-consuming than an L-CBA. Thus, similarly to an ELISA, it could be easily implemented in clinical practice.
In this study, we first compared the analytical performance of an F-CBA with two ELISA assays in a population of patients with MG. The main findings of our study can be summarized as follows: (i) the cELISA detected the most AChR antibodies in comparison to both the iELISA and F-CBA; (ii) the iELISA and F-CBA had the highest concordance; (iii) the comparison of continuous titers between the iELISA and cELISA did not provide additional information given the different magnitude and the extension of the results (different limit of the analytical range). Indeed, a slope of 26% indicated that one method overall gave values that are one-third of another; (iv) the cELISA showed the highest analytical performances. Se and NPV increased by adding the iELISA (combined double strategy of cELISA+iELISA), respectively, from 66% to 71% and from 59% to 63%. No further increase was observed by adding iIF (combined triple strategy). Overall, our findings show that the cELISA has better analytical performance than the iELISA and F-CBA. An ELISA has some advantages over CBAs, both live and fixed. First, a CBA undergoes subjective interpretation, which is related to the expertise of the reader and could undermine its reproducibility. Additionally, a CBA provides only qualitative results. Finally, a cELISA does not require dilution.
To date, only a few studies have evaluated the performance of F-CBAs for anti-AChR detection. We previously demonstrated that the F-CBA for MG diagnosis did not accurately identify low anti-AChR and anti-MuSK levels, which were detected by an ELISA [26]. Mirian et al. showed that F-CBAs have similar specificity and higher sensitivity compared to RIPAs but lower performance than L-CBAs [22]. Also, Spagni et al. found that L-CBAs are more sensitive than F-CBAs [29].
Overall, the literature comparing live and fixed CBAs for detecting neural antibodies shows a decreased sensitivity for F-CBAs [30].
In our study, we compared three analytical methods that are commercially available and easy to introduce in clinical laboratories.
Notably, in our study, we found a percentage of seronegative MG patients, in accordance with the literature [31]. The seronegativity could result from low affinity or low levels of antibodies requiring more sensitive assays. Thus, the implementation of analytical methods with high sensitivity and specificity and that are easy to perform are still sought after.
The main limitation of our study is the lack of a comparison with a RIPA. Further studies on larger cohorts comparing the diagnostic performance of commercially available assays for anti-AChR antibody detection are required.