Abstract
Rationale
Sarcoidosis is a systemic granulomatous disorder associated with hypergammaglobulinemia and the presence of autoantibodies. The specific antigens initiating granulomatous inflammation in sarcoidosis are unknown, and there is no specific test available to diagnose sarcoidosis. To discover novel sarcoidosis antigens, we developed a high-throughput T7 phage display library derived from the sarcoidosis cDNA and identified numerous clones differentiating sarcoidosis from other respiratory diseases. After clone sequencing and a homology search, we identified two epitopes (cofilin μ and chain A) that specifically bind to serum IgGs of patients with sarcoidosis.
Objectives
To develop and validate an epitope-specific IgG-based immunoassay specific for sarcoidosis.
Methods
We chemically synthesized both immunoepitopes (cofilin μ and chain A) and generated rabbit polyclonal antibodies against both neoantigens. After extensive standardization, we developed a direct peptide ELISA and measured epitope-specific IgG in the sera of 386 subjects, including healthy control subjects (n = 100), three sarcoidosis cohorts (n = 186), pulmonary tuberculosis (n = 70), and lung cancer (n = 30).
Measurements and Main Results
To develop a model to classify sarcoidosis distinctly from other groups, data were analyzed using fivefold cross-validation when adjusting for confounders. The cofilin μ IgG model yielded a mean sensitivity, specificity, and positive and negative predictive value of 0.97, 0.9, 0.9, and 0.96, respectively. Those same measures for chain A IgG antibody were 0.9, 0.83, 0.84, and 0.9, respectively. Combining both biomarkers improved the area under the curve, sensitivity, specificity, and positive and negative predictive value.
Conclusions
These results provide a novel immunoassay for sarcoidosis. The discovery of two neoantigens facilitates the development of biospecific drug discovery and the sarcoidosis-specific model.
Keywords: sarcoidosis, antigen, chain A, cofilin, T7 phage display
At a Glance Commentary
Scientific Knowledge on the Subject
Sarcoidosis is considered as an antigen-driven pathology; yet, the antigens remain unknown.
What This Study Adds to the Field
We generated a T7 phage display library derived from sarcoidosis tissues and identified two novel epitopes that differentially bind to the IgGs of sarcoidosis sera. Using these novel epitopes, we developed a novel direct peptide ELISA immunoassay for sarcoidosis.
Sarcoidosis is a multisystem granulomatous disease of unknown etiology characterized by aberrant immune function that includes noncaseating granuloma, cutaneous anergy, hypergammaglobulinemia, increased local and circulating inflammatory cytokines with a shift toward T helper type 1 (Th1), and, in some cases, increased production of autoantibodies (1–6). A heightened humoral immune response in sarcoidosis manifests as an increase in circulating immunoglobulins and the presence of antibodies with reactivity toward multiple antigens and self-antigens (2, 4, 7).
Pulmonary tuberculosis (TB) caused by Mycobacterium tuberculosis and sarcoidosis share several clinical, radiological, and histological features, making differential diagnosis difficult (8, 9). In fact, numerous studies using proteomics, genomics, transcriptomics, and high-throughput technology suggested similarities in molecular signatures in sarcoidosis, TB, and cancer (10). The current sarcoidosis markers angiotensin-converting enzyme (ACE), lysozyme, and soluble IL-2R have poor sensitivity and specificity and do not differentiate sarcoidosis from TB and even malignancies (11, 12).
We hypothesized that sarcoidosis is an immune disorder triggered by a group of specific antigens. To test this, we previously developed a custom-made complex T7 phage display library derived from sarcoidosis cDNA isolated from BAL and peripheral blood cells to identify novel sarcoidosis antigens (3, 13). After biopanning and selection of sarcoidosis-specific antigen clones, an antigen microarray platform was constructed and immunoscreened with sera from patients with sarcoidosis, healthy control subjects, and patients with TB (3, 13). Recently, we identified a panel of immunoepitopes (antigen clones) that differentially bind to IgGs of sarcoidosis sera (3, 13). After sequencing the antigen clone (ID number P197-BP4-922) (false discovery rate [FDR], 8.10 × 10−9) and using the National Center for Biotechnology Information Basic Local Alignment Search Tool, we identified that this clone has some homology to cofilin-1, isoform CRA_a (gene name CFL1). Cofilin is an essential cellular protein and acts as a regulator of actin filament and a key regulator of cellular dynamics, including cellular movement, phagocytosis, and cell signaling (14). Another significant antigen clone (ID number P-197-BP4-921) (FDR, 2.7 × 10−8) had homology to chain A, Human Metap1 (gene name 4FLI_A). Methionine aminopeptidase (MetAP) is a metalloprotease cleaving the methionine initiator from nascent peptides (15, 16). MetAP plays critical roles in protein biogenesis, protein folding, processing, and targeting nascent protein into specific organelles (3, 13, 15–17). The aberrant function of this peptidase leads to misfolded proteins, causing cellular stress. Both identified epitopes were peptides encoded by inserted gene fragments and were out-of-frame peptides (3, 13) and had some differences from the native proteins (cofilin-1 and MetAP1); therefore, we considered them as neoantigens.
In this study, we used both novel neoantigens to develop an ELISA-based immunoassay for sarcoidosis diagnosis and determined whether immunoreactivities to these antigens can be considered valid predictors of sarcoidosis by measuring IgG immunoreactivity from a large and diversified sarcoidosis population (three independent cohorts), healthy control subjects, and subjects with TB and lung cancer. We calculated the performance of epitope-specific IgGs in terms of specificity, sensitivity, accuracy, positive predictive value (PPV), and negative predictive value (NPV).
Methods
Chemicals and Antibodies
ELISA plates and 3,3′,5,5′-tetramethylbenzidine substrate mixture were purchased from R&D Systems; goat anti-human IgG horseradish peroxidase–linked antibody was purchased from Abcam; and blocking buffer was purchased from Bio-Rad Laboratories.
Patient Selection
The Committee for Investigations Involving Human Subjects at Wayne State University approved the protocol for obtaining blood by phlebotomy from human subjects (IRB protocol 055208MP4E and 1904002164). All study subjects signed a written informed consent. Sarcoidosis diagnosis was based on the American Thoracic Society/European Respiratory Society/World Association of Sarcoidosis and other Granulomatous Disorders statement (1, 18). Sera were collected from healthy volunteers (n = 100); three sarcoidosis cohorts (n = 186) from Detroit (n = 98), San Francisco (n = 48), and the ACCESS (Case Control Etiologic Study of Sarcoidosis) cohort (19) (n = 40); lung cancer (LC; n = 30); and pulmonary TB (n = 70) from the Foundation for Innovative New Diagnostics (Geneva, Switzerland). All patients with TB had smear-positive sputum.
Direct ELISA Development for Detection of IgG against Cofilin μ and Chain A Peptides
Cofilin μ and chain A peptides were chemically synthesized by GenScript. A microplate (R&D Systems) was coated with cofilin μ or chain A peptide at the concentration of 500 ng/ml (100 μl/well) overnight at room temperature (RT). Plates were washed three times with washing buffer (PBS with 0.05% Tween-20) and blocked at RT for 1 hour with blocking buffer and then were washed three times with washing buffer. Serum samples (1:500 in blocking buffer) in duplicates were added onto wells and incubated at RT for 2 hours. The plates were washed three times and incubated with goat anti-human IgG horseradish peroxidase–linked secondary antibody at RT (90 min). The plate was washed, 100 μl/well of 3,3′,5,5′-tetramethylbenzidine substrate mixture was added (20 min), and the reaction was stopped by adding 50 μl of 2 N H2SO4. The optical density (OD) was measured at 450 nm using a microplate reader (BioTek Instruments).
Generation of Rabbit Polyclonal Antibody against Cofilin μ and Chain A Peptides
The rabbit polyclonal anti-cofilin μ or chain A antibody was custom generated by GenScript. Briefly, 0.2 mg of keyhole limpet hemocyanin–conjugated peptides with complete Freund’s adjuvant was administered subcutaneously per rabbit (New Zealand strain). Three booster doses of peptides were subcutaneously administered on Days 14, 28, and 42. The immunized serum was collected, and the polyclonal anti-cofilin μ or chain A IgGs were purified by the antigen affinity purification method.
Statistical Analyses, Sample Size Calculation, and Power Analysis
Power and sample size were calculated using G*Power (20). We estimated the effect size of 0.5–0.6 with 90% power and an FDR of 1–5%. On the basis of this calculation, the required sample size for sarcoidosis was 93 and that for healthy control subjects was 48. To reach higher power to near 100% and be able to cross-validate (CV), we increased the sample size and included control subjects (n = 100) and subjects with sarcoidosis (n = 186). We added 100 samples of disease control: TB (n = 70) and LC (n = 30). ELISA data were analyzed using the programming language R. A two-tailed Mann-Whitney test was applied to determine the significant differences of peptide specific IgGs among groups. The significance scores (P values) were corrected for multiple comparisons using the FDR method (21). The values of specific IgGs against chain A and cofilin μ were used to develop classification models and the receiver operating characteristic (ROC) curve to predict sarcoidosis while adjusting for confounders (sex, race, and smoking) and by stepwise selection. Final classification models were derived algorithmically with stepwise selection by the Akaike information criterion (22).
The models were validated following fivefold CV. The data were split into five sets, and each set was given an opportunity to be the test set. We assessed the ability to correctly identify sarcoidosis in each model on the basis of sensitivity, specificity, PPV, NPV, and accuracy.
Results
Novel Neoantigen Selection and Development of Cofilin μ Peptide ELISA
To identify the elusive sarcoidosis antigen(s), we created a custom-made T7 phage antigen display library through selective biopanning and antigen cloning and a high-throughput protein microarray followed by immunoscreening (3, 13, 23). We identified several antigen clones that specifically bind to IgGs present in sarcoidosis sera and immunoreactivities differentiate sarcoidosis from various groups, including healthy controls and TB (3, 13, 23). Two high-performance antigen clones (epitopes) were selected for ELISA development. One antigen clone with amino acid (AA) sequence of SACLQSLRTQLLTWALVGDVGQP is partly homologous to AAs 39–45 on the β2 strand of human cofilin-1 and partly homologous to flagellin (see Figure E1 in the online supplement). Using the Predicting Antigenic Peptides tool, we found that one of the antigenic epitopes of cofilin-1 has homology to our identified epitope (24). Because of the differences of this epitope with both native proteins (cofilin and flagellin), we considered this peptide a neoantigen and labeled it as cofilin μ. Cofilin-1 interacts with the actin cytoskeleton and is involved in actin depolymerization and remodeling (25). We selected cofilin μ as a novel immunoepitope to develop a peptide-based ELISA for detection of cofilin μ–specific IgGs. Cofilin μ peptide was chemically synthesized followed by standardization of peptide concentration, serum dilution, incubation parameters, and detection antibody. We determined the avidity as the relative strength of binding between IgG antibody to the cofilin μ epitope. A total of 386 study subjects were composed of healthy control subjects (n = 100) and three sarcoidosis cohorts: Detroit (n = 98), San Francisco (SF; n = 48), and the ACCESS cohort (n = 40). Two diseased control cohorts (TB, n = 70; and LC, n = 30) were added to exclude a nonspecific IgG reaction to infection or tumor antigens. The characteristics of the study groups are shown in Table 1. The clinical features among sarcoidosis cohorts are shown in Table E1. We found significant differences in race, sex, and treatment but no significant differences in Scadding chest radiograph stages. Figure 1A shows the OD values (mean ± SD) of cofilin μ–specific IgG in sera among sarcoidosis cohorts (1.5 ± 0.7) and healthy control subjects (0.4 ± 0.4). There were significant differences in the OD values between sarcoidosis cohorts and healthy control subjects (P = 4.7 × 10−34). We generated a custom-made rabbit polyclonal antibody against cofilin μ (anti-cofilin μ IgG) to develop a standard curve (Figure 1B). The concentration of cofilin μ–specific IgG was calculated on the basis of a standard curve. The mean values of cofilin μ–specific IgG across the sarcoid cohorts were 152 μg/ml (Detroit), 80 μg/ml (SF), and 121 μg/ml (ACCESS). Figure 1C shows the distribution of cofilin μ–specific IgGs. Cofilin μ IgG concentrations with respect to race, sex, and radiological stages are shown in Figures E2A–E2C. To assess if the IgG immunoreactivity against cofilin μ is sarcoidosis specific, we determined the concentrations of cofilin μ–specific IgGs in the sera of LC and TB. There were significant differences in the concentrations of IgGs between sarcoidosis cohorts and healthy control subjects (P = 1.5 × 10−36), TB (P = 3.2 × 10−32), and LC (P = 1.6 × 10−18) (Figure 1D). These data clearly show that despite the differences among cohorts, sarcoidosis sera collectively exhibited significantly higher concentrations of anti-cofilin μ IgG antibodies.
Table 1.
Subject Demographics
| Subjects with Sarcoidosis | Control Subjects | Subjects with TB | Subjects with LC | |
|---|---|---|---|---|
| (n = 186) | (n = 100) | (n = 70) | (n = 30) | |
| Age, yr, mean ± SD | 48 ± 11 | 40 ± 13 | 35 ± 12 | 62 ± 11 |
| BMI, kg/m2, mean ± SD | 30 ± 7 | 27 ± 5 | 28 ± 6 | 28 ± 9 |
| Sex, n (%) | ||||
| Female | 119 (64) | 66 (66) | 19 (27) | 18 (60) |
| Race, n (%) | ||||
| Black | 117 (63) | 42 (42) | 16 (23) | 0 (0) |
| White | 58 (31) | 46 (46) | 0 (0) | 30 (100) |
| Hispanic | 7 (4) | 14 (14) | 0 (0) | 0 (0) |
| Asian | 4 (2) | 12 (12) | 54 (77) | 0 (0) |
| Smoker, n (%) | 73 (39) | 19 (19) | 32 (46) | 25 (83) |
| Organ involvement, n (%) | ||||
| Pulmonary involvement | 169 (91) | NA | NA | NA |
| Extrapulmonary involvement | 132 (71) | NA | NA | NA |
| Skin | 61 (33) | NA | NA | NA |
| Eye | 41 (22) | NA | NA | NA |
| Heart | 18 (9) | NA | NA | NA |
| CNS | 16 (8) | NA | NA | NA |
| Other | 24 (13) | NA | NA | NA |
| Radiograph stages, n (%) | ||||
| Stage 0 | 14 (8) | NA | NA | NA |
| Stage 1 | 21 (11) | NA | NA | NA |
| Stage 2 | 76 (41) | NA | NA | NA |
| Stage 3 | 39 (20) | NA | NA | NA |
| Stage 4 | 37 (18) | NA | NA | NA |
| Medication | ||||
| Corticosteroids | 71 (37) | NA | NA | NA |
| Methotrexate | 28 (15) | NA | NA | NA |
| Azathioprine | 8 (4) | NA | NA | NA |
| Hydroxychloroquine | 7 (3) | NA | NA | NA |
| Infliximab | 3 (1.6) | NA | NA | NA |
| Mycophenolate mofetil | 4 (2) | NA | NA | NA |
Definition of abbreviations: BMI = body mass index; CNS = central nervous system; LC = lung cancer; NA = not applicable; TB = tuberculosis.
Figure 1.
Patients with sarcoidosis exhibit increased levels of anti-cofilin μ–specific IgGs. (A) Cofilin μ–specific IgGs were measured via standardized cofilin μ peptide ELISA in the sera of subjects with sarcoidosis (n = 186) and healthy control subjects (n = 100). All measurements were duplicates and standardized in our laboratory. The optical density (OD) value at 450 nm was a measure of cofilin μ–specific IgGs. The OD values (mean ± SD) of cofilin μ–specific IgG in sera of patients with sarcoidosis was 1.5 ± 0.7, whereas the healthy control subjects had a mean OD of 0.4 ± 0.4. There were significant differences in the ODs between sarcoidosis cohorts and healthy control subjects (P = 4.7 × 10−34). The horizontal lines represent the mean OD values of healthy subjects (black line) and subjects with sarcoidosis (red line). (B) Polyclonal anti-cofilin μ IgG antibodies were generated by immunizing rabbits with cofilin μ peptide. After three immunizations with cofilin μ peptide, the immunization achieved a high concentration of polyclonal IgG antibody that was purified. The polyclonal rabbit anti-cofilin μ IgG was used to obtain a standard curve for cofilin μ peptide ELISA. Anti-cofilin μ IgG standard curve showing OD at 450 nm versus anti-cofilin μ IgG concentration (ng/ml). (C) Sarcoidosis cohorts displayed significantly higher concentrations of serum anti-cofilin μ IgGs (μg/ml). The mean value of cofilin μ–specific IgGs in healthy control subjects was 27 ± 31 μg/ml, and the mean values in different sarcoid cohorts were 152 ± 61 μg/ml (Detroit), 81 ± 38 μg/ml (San Francisco [SF]), and 122 ± 46 μg/ml (ACCESS). The horizontal lines represent the mean anti-cofilin μ IgG values of healthy subjects (black line), the Detroit cohort (purple line), the SF cohort (green line), and the ACCESS cohort (blue line). There was a significant difference between the healthy control subjects and the Detroit (P = 1.8 × 10−27), SF (P = 4.0 × 10−12), and ACCESS (P = 8.9 × 10−16) cohorts. Within the sarcoidosis cohorts, the differences were as follows: Detroit versus SF (P = 1.87 × 10−6) and ACCESS (P = 7.3 × 10−3) and ACCESS versus SF (P = 2.6 × 10−5). (D) Patients with sarcoidosis displayed significantly higher concentrations of serum anti-cofilin μ IgGs than control subjects. The values (mean ± SD) of anti-cofilin μ IgGs in sarcoidosis were 127 ± 60 μg/ml, 23 ± 12 μg/ml in healthy control subjects, 18 ± 17 μg/ml in tuberculosis (TB), and 5 ± 2 μg/ml in lung cancer (LC). The horizontal lines represent the mean anti-cofilin μ IgG values of healthy control subjects (black line), subjects with sarcoidosis (red line), and subjects with TB or LC (gray line). There was a significant difference between the subjects with sarcoidosis and healthy control subjects (P = 1.5 × 10−36), between subjects with sarcoidosis and those with TB (P = 3.2 × 10−32), and between subjects with sarcoidosis and those with LC (P = 1.6 × 10−18).
Cofilin μ Antibody-based Classifications and ROC Curve
To develop a model to predict sarcoidosis from the other groups, we generated ROC curves using the values of cofilin μ–specific IgG and performed fivefold CV on the entire dataset while correcting for confounders. The fivefold CV with cofilin μ yielded mean values for sensitivity (0.95 ± 0.02), specificity (0.89 ± 0.04), accuracy (0.92 ± 0.03), NPV (0.97 ± 0.02), and PPV (0.9 ± 0.04) (Figure 2). Table 2 displays the performance of anti-cofilin μ IgG as a sarcoidosis immunoassay. These results show that ELISA-based detection of serum IgG against cofilin μ has high sensitivity, specificity, and accuracy to distinguish subjects with sarcoidosis from other groups, even TB.
Figure 2.
Classification and receiver operating characteristic (ROC) curve of cofilin μ. ROC curves for anti-cofilin μ IgG were generated by using fivefold cross-validation (CV). CV models were obtained from the entire dataset by stepwise inclusion of confounders (sex, race, and smoking) and selecting confounders to minimize the Akaike information criterion to obtain the optimal model. Cofilin μ ROC curve yielded areas under the curve (AUC) of 0.98 (black), 0.96 (red), 0.98 (green), 0.93 (blue), and 0.96 (light blue).
Table 2.
Cofilin μ Performance and Cross-Validation
| Cross-Validation | AUC | Sensitivity | Specificity | Accuracy | PPV | NPV |
|---|---|---|---|---|---|---|
| CV1 | 0.988 | 0.972 | 0.948 | 0.960 | 0.945 | 0.973 |
| CV2 | 0.961 | 0.944 | 0.923 | 0.933 | 0.918 | 0.947 |
| CV3 | 0.989 | 0.972 | 0.897 | 0.933 | 0.897 | 0.972 |
| CV4 | 0.939 | 0.916 | 0.846 | 0.880 | 0.846 | 0.916 |
| CV5 | 0.960 | 0.972 | 0.871 | 0.920 | 0.875 | 0.971 |
| Mean ± SD | 0.967 ± 0.02 | 0.956 ± 0.02 | 0.897 ± 0.04 | 0.925 ± 0.03 | 0.896 ± 0.04 | 0.956 ± 0.02 |
Definition of abbreviations: AUC = area under the curve; CV = cross-validation; NPV = negative predictive value; PPV = positive predictive value.
Development of IgG-based Chain A Peptide ELISA
In two different immununoscreening sets, a high-performance antigen clone with the sequence AGISRELVDKLAAALE was identified with selective immunoreactivity (P = 2.70 × 10−8) to IgG present in sarcoidosis sera (3, 13). This sequence has partial homology to human chain A MetAP1 on the AA sequence of ELVDKLAAALE spanning AAs 310–320, and the N-terminal of the peptide (AGISRE) has similarity with cilia- and flagella-associated protein (Figure E3). We considered this peptide as a neoantigen because of the AA differences in the native MetAP1 protein. By applying semiempirical analysis (24), we determined MetAP1 antigenicity and compared it with the identified epitope (24), and we found a homology of the discovered epitope (ELVDKLAAALE) with one of the antigenic determinants on the native MetAP1 protein (FEFELVDKLAAALEHH spanning 307–322 AAs) (Figure E3). We selected this epitope to develop an ELISA and chemically synthesized the chain A peptide followed by standardization of ELISA and evaluated the sera of the same study cohorts for the presence of chain A–specific IgGs. The OD value (mean ± SD) of chain A–specific IgGs in sera of patients with sarcoidosis was 1.3 ± 0.5, whereas in sera of healthy control subjects, it was 0.4 ± 0.2 (Figure 3A). There were significant differences in the OD values between subjects with sarcoidosis and healthy control subjects (P = 7.5 × 10−34). We generated a specific antibody against chain A by immunization of rabbits to standardize chain A–specific IgG measurements. The purified rabbit polyclonal anti-chain A IgG served to develop the standard curve (Figure 3B), based on which the concentration of chain A–specific IgGs were quantified. The mean values of chain A–specific IgGs in the study cohorts are shown in Figure 3C. Chain A IgG levels with respect to race, sex, and chest radiograph stages of disease are shown in Figures E4A–E4C. Furthermore, we determined the concentrations of chain A–specific IgGs in the sera of patients with TB and patients with LC. There were significant differences in the concentrations of IgGs between sarcoidosis cohorts and healthy control subjects (P = 1.8 × 10−33), TB (P = 7.2 × 10−16), and LC (P = 1.1 × 10−15) (Figure 3D). These results show that the novel chain A peptide ELISA can be used to identify subjects with sarcoidosis on the basis of the presence of specific serum IgG antibodies against the chain A epitope.
Figure 3.
Patients with sarcoidosis exhibit increased levels of anti-chain A–specific IgGs. (A) Chain A–specific IgGs were measured via standardized chain A peptide ELISA in the sera obtained from subjects with sarcoidosis (n = 186) and healthy control subjects (n = 100). The optical density (OD) value at 450 nm was a measure of chain A–specific IgGs. The OD values (mean ± SD) of chain A–specific IgGs in patients with sarcoidosis was 1.3 ± 0.5, whereas in healthy control subjects, it was 0.4 ± 0.2. There were significant differences in the ODs between sarcoidosis cohorts and healthy control subjects (P = 7.5 × 10−34). The horizontal lines represent the mean OD values of healthy subjects (black line) and subjects with sarcoidosis (red line). (B) Polyclonal anti-chain A IgGs were generated by immunizing the rabbits with chain A peptide. After three immunizations of rabbits with chain A peptide, the immunization achieved a high concentration of polyclonal IgG antibody that was purified to develop a standard curve for chain A peptide ELISA. A standard curve showing OD at 450 nm versus anti-chain A IgG concentration (ng/ml) is presented. (C) Sarcoidosis cohorts displayed significantly higher concentrations of serum anti-chain A IgGs. The mean value of chain A–specific IgGs in healthy control subjects was 6 ± 4 μg/ml, and the mean values in different sarcoid cohorts were 22 ± 11 μg/ml (Detroit), 26 ± 6 μg/ml (San Francisco [SF]), and 17 ± 6 μg/ml (ACCESS). The horizontal lines represent the mean anti-chain A IgG values of healthy subjects (black line), the Detroit cohort (purple line), the SF cohort (green line), and the ACCESS cohort (blue line). There was a significant difference between the healthy control group and the Detroit cohort (P = 1.34 × 10−22), the SF cohort (P = 1.26 × 10−22), and the ACCESS cohort (P = 5.82 × 10−16). Among the sarcoidosis cohort, differences were as follows: Detroit versus SF (P = 0.06) and ACCESS (P = 9.29 × 10−3), ACCESS versus SF (P = 1.62 × 10−5). (D) Sarcoidosis cohorts displayed significantly higher concentrations of serum anti-chain A IgGs than control subjects. The values (mean ± SD) of anti-chain A IgGs were 22 ± 9 μg/ml (sarcoidosis), 6 ± 4 μg/ml (healthy control subjects), 11 ± 4 μg/ml (tuberculosis [TB]), and 5 ± 3 μg/ml (lung cancer [LC]). The horizontal lines represent the mean anti-cofilin μ IgG values of healthy control subjects (black line), subjects with sarcoidosis (red line), and subjects with TB or LC (gray line). There was a significant difference between the subjects with sarcoidosis and healthy control subjects (P = 1.8 × 10−33), between subjects with sarcoidosis and subjects with TB (P = 7.2 × 10−16), and between subjects with sarcoidosis and subjects with LC (P = 1.1 × 10−15).
Chain A Antibody-based Classifications and ROC Curve
To develop a model to predict sarcoidosis from all other groups, we generated ROC curves by integrating stepwise confounders and fivefold CV on the entire chain A dataset (Figure 4). The fivefold CV yielded mean ± SD values for sensitivity (0.98 ± 0.02), specificity (0.88 ± 0.04), accuracy (0.92 ± 0.02), NPV (0.98 ± 0.02), and PPV (0.89 ± 0.04). Table 3 elucidates the performance of the anti-chain A IgG immunoassay, indicating that ELISA-based detection of serum IgG against chain A has high accuracy to distinguish subjects with sarcoidosis from healthy subject and subjects with LC or TB.
Figure 4.
Classifications and receiver operating characteristic (ROC) curve of chain A. ROC curves for anti-chain A IgGs were generated by using fivefold cross-validation (CV). CV models were obtained from the entire dataset by integrating confounders (sex, race, and smoking) in the best classification model. Chain A ROC curve yielded areas under the curve (AUC) of 0.97 (black), 0.94 (red), 0.95 (green), 0.91 (blue), and 0.94 (light blue).
Table 3.
Chain A Performance and Cross-Validation
| Cross-Validation | AUC | Sensitivity | Specificity | Accuracy | PPV | NPV |
|---|---|---|---|---|---|---|
| CV1 | 0.970 | 0.916 | 0.923 | 0.920 | 0.916 | 0.923 |
| CV2 | 0.943 | 0.861 | 0.923 | 0.893 | 0.911 | 0.878 |
| CV3 | 0.955 | 0.861 | 0.846 | 0.853 | 0.837 | 0.868 |
| CV4 | 0.918 | 0.916 | 0.769 | 0.840 | 0.785 | 0.909 |
| CV5 | 0.942 | 0.944 | 0.717 | 0.826 | 0.755 | 0.933 |
| Mean ± SD | 0.946 ± 0.01 | 0.9 ± 0.04 | 0.835 ± 0.09 | 0.866 ± 0.04 | 0.841 ± 0.07 | 0.902 ± 0.03 |
Definition of abbreviations: AUC = area under the curve; CV = cross-validation; NPV = negative predictive value; PPV = positive predictive value.
Improvement of Classification Accuracy by Combining Biomarkers and Their Association with Common Sarcoidosis Markers
Recent studies indicate that a combination of multiple biomarkers will increase the robustness of classification (26). We determined the collinearity of two markers via regression analysis and found a moderately strong association between cofilin μ and chain A antibodies among all study groups (r = 0.751) and the sarcoidosis cohort (r = 0.59) (Figure E5). Therefore, we evaluated if combining both IgGs would improve the classification accuracy of the model. The ROC curves were generated using fivefold CV on the entire dataset by combining both markers while adjusting for confounders (Figure 5). The fivefold CV with combined markers yielded mean ± SD values for sensitivity (0.98 ± 0.02), specificity (0.88 ± 0.04), accuracy (0.92 ± 0.02), NPV (0.98 ± 0.02), and PPV (0.89 ± 0.04). Table 4 depicts the performance of combined cofilin μ and chain A antibodies in each CV. The results indicate improvement of ROC when combing both markers.
Figure 5.
Classification and receiver operating characteristic (ROC) for combined epitopes (cofilin μ and chain A). Cross-validation (CV) models were obtained from the entire dataset by stepwise inclusion of confounders (sex, race, and smoking) to achieve the optimal model by minimizing the Akaike information criterion. The combined ROC curve for cofilin μ and chain A yielded areas under the curve (AUC) of 0.99 (black), 0.98 (red), 0.99 (green), 0.93 (blue), and 0.97 (light blue).
Table 4.
Performance of Combined Biomarkers
| Cross-Validation | AUC | Sensitivity | Specificity | Accuracy | PPV | NPV |
|---|---|---|---|---|---|---|
| CV1 | 0.994 | 0.972 | 0.923 | 0.946 | 0.921 | 0.973 |
| CV2 | 0.986 | 0.944 | 0.923 | 0.933 | 0.918 | 0.947 |
| CV3 | 0.995 | 1.000 | 0.897 | 0.946 | 0.900 | 1.000 |
| CV4 | 0.938 | 0.972 | 0.820 | 0.893 | 0.833 | 0.969 |
| CV5 | 0.970 | 1.000 | 0.846 | 0.920 | 0.857 | 1.000 |
| Mean ± SD | 0.977 ± 0.02 | 0.977 ± 0.02 | 0.882 ± 0.04 | 0.928 ± 0.02 | 0.886 ± 0.04 | 0.978 ± 0.02 |
Definition of abbreviations: AUC = area under the curve; CV = cross-validation; NPV = negative predictive value; PPV = positive predictive value.
Current sarcoidosis markers (lysozyme, ACE, soluble IL-2R, and C-reactive protein [CRP]) are measurements of indirect immunological response of antigen-presenting cells and T cells to inflammatory triggers and are considered nonspecific. We asked if these markers have any association with antigen-specific IgGs (cofilin μ and chain A antibodies). We found a significant correlation between the cofilin μ and chain A IgG and other sarcoidosis markers (Table 5).
Table 5.
Correlation of Serum Markers
| Serum IgG | Lysozyme (n = 98) |
ACE (n = 146) |
sIL-2R (n = 186) |
CRP (n = 132) |
||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| rho | P Value | CI | rho | P Value | CI | rho | P Value | CI | rho | P Value | CI | |
| Cofilin-μ IgG (μg/ml) | 0.396* | <0.001 | 0.214–0.552 | 0.348* | <0.001 | 0.208–0.475 | 0.275* | <0.001 | 0.157–0.386 | 0.473* | <0.001 | 0.349–0.581 |
| Chain A IgG (μg/ml) | 0.393* | <0.001 | 0.211–0.549 | 0.198* | 0.008 | 0.048–0.340 | 0.300* | <0.001 | 0.183–0.409 | 0.345* | <0.001 | 0.207–0.469 |
Definition of abbreviations: ACE = angiotensin-converting enzyme; CI = confidence interval; CRP = C-reactive protein.
Nonparametric Spearman’s correlation is significant at the 0.01 level (two tailed).
Discussion
Sarcoidosis is characterized by granuloma formation affecting multiple organs associated with local cellular responses and cytokine secretions (4, 6). The prevailing immunological aberrancy in sarcoidosis is described as an antigen-driven pathology associated with Th1 cell–mediated immunity (4, 6, 27). Despite the well-documented active humoral immunity, less effort has been made toward identifying epitope-specific IgGs in sarcoidosis and their utility as a tool to differentiate sarcoidosis from infectious granulomatous diseases (2, 7). This is partly due to the lack of insight into the inciting sarcoidosis antigen(s) and the complexity and diversity of the human antibody repertoire (28). We postulated that antigen-specific humoral responses might be important drivers in sarcoidosis development and that these epitope-specific IgGs can differentiate sarcoidosis from other respiratory diseases. Our custom-made complex T7 phage antigen display library identified a panel of classifier antigens that differentially bind to IgG of sarcoidosis sera as compared with control subjects and subjects with TB (3, 13). Because we validated the significance of two high-performance immunoepitopes in three independent immunoscreenings (3, 13), both epitopes were selected for ELISA development. To ensure successful ELISA development, we undertook crucial steps to achieve robustness, including sequence alignment, antigenicity determination, peptide design and synthesis, rabbit immunization, standardization, and validation (Figure 6). The immunoreactivity of sarcoidosis against both epitopes was validated in three independent sarcoidosis cohorts, and their classification accuracy was assessed by including clinically relevant subjects with TB and LC. Sarcoidosis sera collectively exhibit increased concentrations of serum IgG against both epitopes compared with TB and LC, although there were some differences in the concentrations of specific IgG against cofilin μ among sarcoidosis cohorts. For instance, the Detroit cohort had a higher anti-cofilin μ IgG concentration that might be due to the higher percentage of Black population, untreated subjects, and multiorgan involvement. Despite heterogeneities of sarcoidosis cohorts, ROC classifications based on epitope-specific antibodies while adjusting for confounders achieved a robust sensitivity, specificity, and accuracy. A similar approach led to the discovery of transketolase epitopes and development of an IgG-based assay differentiating TB infection from sarcoidosis and latent TB infection (29). These data suggest that epitope-specific antibodies are uniquely precise in differentiating even very similar conditions.
Figure 6.
Schematic representation of the discovery and development of the highly specific and sensitive peptide ELISA for detecting cofilin μ– or chain A–specific IgG antibodies in serum. Flowchart depicting the steps involved in the discovery of sarcoid-specific clones, using the T7 phage display library with subtractive panning and immunoscreening, the identification of cofilin μ and chain A epitopes, and the development of cofilin μ and chain A peptide ELISA. MΦ = macrophages; PBMC = peripheral blood mononuclear cell; ROC = receiver operating characteristic; TB = tuberculosis.
Currently, the diagnosis of sarcoidosis rests on compatible clinical and radiological findings and histopathological evidence of noncaseating granuloma on tissue biopsy and negative microbiological studies, including acid-fast bacilli smear and cultures (1, 5). Although tissue diagnosis in pulmonary sarcoidosis is relatively easy, tissue biopsies are quite challenging in less accessible organs such as the eye, brain, spinal cord, or heart. There is an unmet need for noninvasive and more accurate sarcoidosis biomarkers (5). The specificity and sensitivity of serological biomarkers such as ACE concentrations or soluble IL-2RA in sarcoidosis is between 62% and 86%, respectively (30–33). Most reported biomarkers, including CRP, serum lysozyme, and soluble IL-2RA are elevated in various inflammatory conditions, infectious diseases, and cancers (11, 34, 35), because they are released in response to various antigenic triggers. Despite being nonspecific, they provide information about immune activation of antigen-presenting cells and T cells in sarcoidosis. We found that cofilin μ and chain A IgG antibodies correlated well with CRP and serum lysozyme and ACE, whereas chain A antibody had better correlation with soluble IL-2R. We propose that cofilin μ and chain A IgG antibodies could be included in the evaluation of sarcoidosis because they assess humoral responses by detecting epitope-specific IgGs.
Cofilin/actin depolymerization factor, the actin regulatory protein, has a vital function in cells by severing filaments by increasing the number of filament ends from which polymerization and depolymerization can occur (36). The cofilin pathway is important for actin dynamics, cell movement, cytoskeleton, distribution of receptors, synapse formation, and immune cell migration, such as antigen presentation and phagocytosis (37). Although an increased cofilin-1 protein in LC has been reported (38), no study evaluated serum antibody against cofilin-1. In our study, we detected a lower concentration of anti-cofilin μ IgG in subjects with LC. To assess the cross-reactivity of anti-cofilin μ antibody against cofilin-1, we designed two other antigenic epitopes of human cofilin-1 and found no cross-reactivity and a lower affinity of sarcoidosis IgGs against both epitopes (data not shown), suggesting specificity of sarcoidosis IgG affinity toward the cofilin μ epitope. This supports the finding that even one AA alteration on an epitope-binding site modulates the antibody affinity (39). The N-terminal AA sequence of cofilin μ has similarity with bacterial flagellar biosynthesis protein (40). Bacterial flagella are helical filaments and are important for movement, pathogenesis, and chemotaxis of various microorganisms (40, 41). Bacterial flagellin is recognized by the host TLR5 receptor (42). Various microbial constituents have been associated with the development of sarcoidosis (43). Because of the elevated IgG concentrations against this unique epitope in sarcoidosis, we speculated that this is a novel immunoepitope developed through the splice product, or hybrid of cofilin and flagella evoking antigenic responses in sarcoidosis. The presence of IgG antibody against cofilin μ in sarcoidosis may contribute to alteration of cytoskeleton dynamics involved in various pathological functions. Further studies need to elucidate the role of this antibody in clinical features of sarcoidosis.
Another immunoepitope had some homology with chain A of MetAP1. MetAP removes initiator N-terminal methionine from nascent peptide (44). MetAP1 knockout cells exhibit impaired cell proliferation and altered cell morphology. Perturbation of MetAP1 is implicated in cytoskeleton organization, lipid metabolism, endothelial cell proliferation, and angiogenesis (16, 44). We found that patients with sarcoidosis exhibit upregulation of chain A gene and protein expression (data not shown). The antigenicity of chain A clone is 1.34 as determined by the OptimumAntigen tool, indicating a highly immunogenic epitope that empirically was confirmed by high antibody titers in an immunized rabbit. We considered the chain A epitope as a neoantigen because of N-terminal AA differences from the human MetAP1 protein. Predominantly, neoantigens have been associated with cancer or paraneoplastic syndromes; however, recently, neoantigens have been described in various immune-mediated diseases (45–47). Our results support the concept of neoantigen-driven immune responses in the development of systemic sarcoidosis. We considered both peptides as neoantigens because there were differences in the N-terminus from the native proteins. Our study provides several novelties in the discovery of the potential antigenic trigger(s) in sarcoidosis. Because sarcoidosis IgG differentially binds to these epitopes, these epitopes may represent antigenic determinants of sarcoidosis. A critical barrier to the development of a sarcoidosis disease model is the unknown sarcoidosis-specific antigen(s). The discovery of these neoantigens may facilitate the development of an animal model of sarcoidosis.
No previous study evaluated the presence of IgG-specific antibodies against cofilin or MetAP1 in sarcoidosis or any other respiratory diseases. Immunoglobulins are antigen recognition molecules produced by B cells, whereas membrane-bound immunoglobulins on B cells serve as the B-cell receptor for antigens. Epitope discovery is crucial for understanding the molecular basis of immunity. Epitope recognition by the B-cell receptor initiates a cascade of signaling leading to antigen internalization, processing and presentation to T cells, and secretion of soluble immunoglobulins (48). In addition, once the antigen-binding site on variable regions is known, epitope-specific IgGs represent a direct drug target. Immunoglobulins in the form of IgG promote phagolysosomal fusion and opsonization and engagement with FcγR (49, 50). However, IgG autoantibodies can cause cytotoxicity and promote neutrophil recruitment through the formation of immune complexes or direct complement activation with potential targeted organ damage. Further studies are needed to evaluate the antigenic potential of cofilin μ and chain A peptides in the activation of immune cells in sarcoidosis.
Supplemental Materials
Footnotes
Supported by National Heart, Lung, and Blood Institute grants R01HL113508 (L.S.) and R21 HL148089 (L.S.).
Wayne State University has filed for comprehensive worldwide patent protection related to this technology through the U.S. Patent and Trademark Office, patent no. WO2016141347A2.
Author Contributions: C.P. conducted the experiments and contributed to the standardization of ELISA. J.T. contributed to the development and standardization of ELISA and to writing the manuscript. L.L.K. contributed to sample collection and writing the manuscript. B.S. and K.S. analyzed the data and performed the statistical analysis. L.S. conceived and designed the study, participated in all areas of the research, and oversaw the ELISA development, data analysis, and writing of the manuscript.
A data supplement for this article is available via the Supplements tab at the top of the online article.
Originally Published in Press as DOI: 10.1164/rccm.202306-1054OC on February 22, 2024
Author disclosures are available with the text of this article at www.atsjournals.org.
References
- 1. Crouser ED, Maier LA, Wilson KC, Bonham CA, Morgenthau AS, Patterson KC, et al. Diagnosis and detection of sarcoidosis. An official American Thoracic Society clinical practice guideline. Am J Respir Crit Care Med . 2020;201:e26–e51. doi: 10.1164/rccm.202002-0251ST. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Häggmark A, Hamsten C, Wiklundh E, Lindskog C, Mattsson C, Andersson E, et al. Proteomic profiling reveals autoimmune targets in sarcoidosis. Am J Respir Crit Care Med . 2015;191:574–583. doi: 10.1164/rccm.201407-1341OC. [DOI] [PubMed] [Google Scholar]
- 3. Talwar H, Rosati R, Li J, Kissner D, Ghosh S, Fernández-Madrid FF, et al. Development of a T7 phage display library to detect sarcoidosis and tuberculosis by a panel of novel antigens. EBioMedicine . 2015;2:341–350. doi: 10.1016/j.ebiom.2015.03.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Weeratunga P, Moller DR, Ho L-P. Immune mechanisms of granuloma formation in sarcoidosis and tuberculosis. J Clin Invest . 2024;134:e175264. doi: 10.1172/JCI175264. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Maier LA, Crouser ED, Martin WJ, II, Eu J. Executive summary of the NHLBI workshop report: leveraging current scientific advancements to understand sarcoidosis variability and improve outcomes. Ann Am Thorac Soc . 2017;14(Suppl 6):S415–S420. doi: 10.1513/AnnalsATS.201707-563OT. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Talreja J, Talwar H, Bauerfeld C, Grossman LI, Zhang K, Tranchida P, et al. HIF-1α regulates IL-1β and IL-17 in sarcoidosis. eLife . 2019;8:e44519. doi: 10.7554/eLife.44519. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Hunninghake GW, Crystal RG. Mechanisms of hypergammaglobulinemia in pulmonary sarcoidosis. Site of increased antibody production and role of T lymphocytes. J Clin Invest . 1981;67:86–92. doi: 10.1172/JCI110036. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Hajizadeh R, Sato H, Carlisle J, Nadaf MT, Evans W, Shepherd BE, et al. Mycobacterium tuberculosis antigen 85A induces Th-1 immune responses in systemic sarcoidosis. J Clin Immunol . 2007;27:445–454. doi: 10.1007/s10875-007-9080-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Oswald-Richter K, Sato H, Hajizadeh R, Shepherd BE, Sidney J, Sette A, et al. Mycobacterial ESAT-6 and katG are recognized by sarcoidosis CD4+ T cells when presented by the American sarcoidosis susceptibility allele, DRB1*1101. J Clin Immunol . 2010;30:157–166. doi: 10.1007/s10875-009-9311-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Maertzdorf J, Weiner J, III, Mollenkopf HJ, Bauer T, Prasse A, Müller-Quernheim J, et al. TBornotTB Network Common patterns and disease-related signatures in tuberculosis and sarcoidosis. Proc Natl Acad Sci USA . 2012;109:7853–7858. doi: 10.1073/pnas.1121072109. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Vorselaars AD, van Moorsel CH, Zanen P, Ruven HJ, Claessen AM, van Velzen-Blad H, et al. ACE and sIL-2R correlate with lung function improvement in sarcoidosis during methotrexate therapy. Respir Med . 2015;109:279–285. doi: 10.1016/j.rmed.2014.11.009. [DOI] [PubMed] [Google Scholar]
- 12. Buccheri G, Marino P, Preatoni A, Ferrigno D, Moroni GA. Soluble interleukin 2 receptor in lung cancer. An indirect marker of tumor activity? Chest . 1991;99:1433–1437. doi: 10.1378/chest.99.6.1433. [DOI] [PubMed] [Google Scholar]
- 13. Hanoudi SN, Talwar H, Draghici S, Samavati L. Autoantibodies against cytoskeletons and lysosomal trafficking discriminate sarcoidosis from healthy controls, tuberculosis and lung cancers. Mol Biomed . 2022;3:3. doi: 10.1186/s43556-021-00064-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Bamburg JR, Bernstein BW. Roles of ADF/cofilin in actin polymerization and beyond. F1000 Biol Rep . 2010;2:62. doi: 10.3410/B2-62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Giglione C, Vallon O, Meinnel T. Control of protein life-span by N-terminal methionine excision. EMBO J . 2003;22:13–23. doi: 10.1093/emboj/cdg007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Jonckheere V, Fijałkowska D, Van Damme P. Omics assisted N-terminal proteoform and protein expression profiling on methionine aminopeptidase 1 (MetAP1) deletion. Mol Cell Proteomics . 2018;17:694–708. doi: 10.1074/mcp.RA117.000360. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Zhang F, Bhat S, Gabelli SB, Chen X, Miller MS, Nacev BA, et al. Pyridinylquinazolines selectively inhibit human methionine aminopeptidase-1 in cells. J Med Chem . 2013;56:3996–4016. doi: 10.1021/jm400227z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Judson MA, Costabel U, Drent M, Wells A, Maier L, Koth L, et al. The WASOG Sarcoidosis Organ Assessment Instrument: an update of a previous clinical tool. Sarcoidosis Vasc Diffuse Lung Dis . 2014;31:19–27. [PubMed] [Google Scholar]
- 19. Baughman RP, Teirstein AS, Judson MA, Rossman MD, Yeager H, Jr, Bresnitz EA, et al. Case Control Etiologic Study of Sarcoidosis (ACCESS) research group Clinical characteristics of patients in a case control study of sarcoidosis. Am J Respir Crit Care Med . 2001;164:1885–1889. doi: 10.1164/ajrccm.164.10.2104046. [DOI] [PubMed] [Google Scholar]
- 20. Faul F, Erdfelder E, Lang A-G, Buchner A. G*Power 3: a flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behav Res Methods . 2007;39:175–191. doi: 10.3758/bf03193146. [DOI] [PubMed] [Google Scholar]
- 21. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple hypothesis testing. J R Stat Soc Series B Stat Methodol . 1995;57:289–300. [Google Scholar]
- 22. Portet S. A primer on model selection using the Akaike information criterion. Infect Dis Model . 2020;5:111–128. doi: 10.1016/j.idm.2019.12.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Talwar H, Hanoudi SN, Draghici S, Samavati L. Novel T7 phage display library detects classifiers for active Mycobacterium tuberculosis infection. Viruses . 2018;10:375. doi: 10.3390/v10070375. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Kolaskar AS, Tongaonkar PC. A semi-empirical method for prediction of antigenic determinants on protein antigens. FEBS Lett . 1990;276:172–174. doi: 10.1016/0014-5793(90)80535-q. [DOI] [PubMed] [Google Scholar]
- 25. Pope BJ, Zierler-Gould KM, Kühne R, Weeds AG, Ball LJ. Solution structure of human cofilin: actin binding, pH sensitivity, and relationship to actin-depolymerizing factor. J Biol Chem . 2004;279:4840–4848. doi: 10.1074/jbc.M310148200. [DOI] [PubMed] [Google Scholar]
- 26. Mazzara S, Rossi RL, Grifantini R, Donizetti S, Abrignani S, Bombaci M. CombiROC: an interactive web tool for selecting accurate marker combinations of omics data. Sci Rep . 2017;7:45477. doi: 10.1038/srep45477. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Talreja J, Talwar H, Ahmad N, Rastogi R, Samavati L. Dual inhibition of Rip2 and IRAK1/4 regulates IL-1β and IL-6 in sarcoidosis alveolar macrophages and peripheral blood mononuclear cells. J Immunol . 2016;197:1368–1378. doi: 10.4049/jimmunol.1600258. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Rees AR. Understanding the human antibody repertoire. MAbs . 2020;12:1729683. doi: 10.1080/19420862.2020.1729683. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Talreja J, Peng C, Nguyen T-M, Draghici S, Samavati L. Discovery of novel transketolase epitopes and the development of IgG-based tuberculosis serodiagnostics. Microbiol Spectr . 2023;11:e0337722. doi: 10.1128/spectrum.03377-22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Eurelings LEM, Miedema JR, Dalm VASH, van Daele PLA, van Hagen PM, van Laar JAM, et al. Sensitivity and specificity of serum soluble interleukin-2 receptor for diagnosing sarcoidosis in a population of patients suspected of sarcoidosis. PLoS One . 2019;14:e0223897. doi: 10.1371/journal.pone.0223897. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Damoiseaux J. The IL-2 - IL-2 receptor pathway in health and disease: the role of the soluble IL-2 receptor. Clin Immunol . 2020;218:108515. doi: 10.1016/j.clim.2020.108515. [DOI] [PubMed] [Google Scholar]
- 32. Hartung H-P, Hughes RA, Taylor WA, Heininger K, Reiners K, Toyka KV. T cell activation in Guillain-Barré syndrome and in MS: elevated serum levels of soluble IL-2 receptors. Neurology . 1990;40:215–218. doi: 10.1212/wnl.40.2.215. [DOI] [PubMed] [Google Scholar]
- 33. Symons JA, Wood NC, Di Giovine FS, Duff GW. Soluble IL-2 receptor in rheumatoid arthritis. Correlation with disease activity, IL-1 and IL-2 inhibition. J Immunol . 1988;141:2612–2618. [PubMed] [Google Scholar]
- 34. Miyoshi S, Hamada H, Kadowaki T, Hamaguchi N, Ito R, Irifune K, et al. Comparative evaluation of serum markers in pulmonary sarcoidosis. Chest . 2010;137:1391–1397. doi: 10.1378/chest.09-1975. [DOI] [PubMed] [Google Scholar]
- 35. Kuen D-S, Kim B-S, Chung Y. IL-17-producing cells in tumor immunity: friends or foes? Immune Netw . 2020;20:e6. doi: 10.4110/in.2020.20.e6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Pfaendtner J, De La Cruz EM, Voth GA. Actin filament remodeling by actin depolymerization factor/cofilin. Proc Natl Acad Sci USA . 2010;107:7299–7304. doi: 10.1073/pnas.0911675107. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Jönsson F, Gurniak CB, Fleischer B, Kirfel G, Witke W. Immunological responses and actin dynamics in macrophages are controlled by N-cofilin but are independent from ADF. PLoS One . 2012;7:e36034. doi: 10.1371/journal.pone.0036034. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Müller CB, de Barros RL, Castro MA, Lopes FM, Meurer RT, Roehe A, et al. Validation of cofilin-1 as a biomarker in non-small cell lung cancer: application of quantitative method in a retrospective cohort. J Cancer Res Clin Oncol . 2011;137:1309–1316. doi: 10.1007/s00432-011-1001-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Grevys A, Frick R, Mester S, Flem-Karlsen K, Nilsen J, Foss S, et al. Antibody variable sequences have a pronounced effect on cellular transport and plasma half-life. iScience . 2022;25:103746. doi: 10.1016/j.isci.2022.103746. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Liu R, Ochman H. Stepwise formation of the bacterial flagellar system. Proc Natl Acad Sci USA . 2007;104:7116–7121. doi: 10.1073/pnas.0700266104. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Yamaguchi T, Makino F, Miyata T, Minamino T, Kato T, Namba K. Structure of the molecular bushing of the bacterial flagellar motor. Nat Commun . 2021;12:4469. doi: 10.1038/s41467-021-24715-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Feuillet V, Medjane S, Mondor I, Demaria O, Pagni PP, Galán JE, et al. Involvement of Toll-like receptor 5 in the recognition of flagellated bacteria. Proc Natl Acad Sci USA . 2006;103:12487–12492. doi: 10.1073/pnas.0605200103. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Locke LW, Schlesinger LS, Crouser ED. Current sarcoidosis models and the importance of focusing on the granuloma. Front Immunol . 2020;11:1719. doi: 10.3389/fimmu.2020.01719. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44. Hu X, Addlagatta A, Lu J, Matthews BW, Liu JO. Elucidation of the function of type 1 human methionine aminopeptidase during cell cycle progression. Proc Natl Acad Sci USA . 2006;103:18148–18153. doi: 10.1073/pnas.0608389103. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45. Mölzer C, Liu Y-H, Muckersie E, Klaska IP, Cornall R, Wilson HM, et al. Colitis in a transgenic mouse model of autoimmune uveitis may be induced by neoantigen presentation in the bowel. Sci Rep . 2023;13:1256. doi: 10.1038/s41598-022-27018-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46. Luengo-Gil G, Conesa-Zamora P. Potential utility of induced translocation of engineered bacteria as a therapeutic agent for mounting a personalized neoantigen-based tumor immune response. Glob Chall . 2021;6:2100051. doi: 10.1002/gch2.202100051. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47. Casciola-Rosen L, Nagaraju K, Plotz P, Wang K, Levine S, Gabrielson E, et al. Enhanced autoantigen expression in regenerating muscle cells in idiopathic inflammatory myopathy. J Exp Med . 2005;201:591–601. doi: 10.1084/jem.20041367. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48. Treanor B. B-cell receptor: from resting state to activate. Immunology . 2012;136:21–27. doi: 10.1111/j.1365-2567.2012.03564.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49. Takai T. Roles of Fc receptors in autoimmunity. Nat Rev Immunol . 2002;2:580–592. doi: 10.1038/nri856. [DOI] [PubMed] [Google Scholar]
- 50. Wang H, Coligan JE, Morse HC., III Emerging functions of natural IgM and its Fc receptor FCMR in immune homeostasis. Front Immunol . 2016;7:99. doi: 10.3389/fimmu.2016.00099. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.