Detection of Antigen 38-kDa in Cerebrospinal Fluid using Recombinant Antibody as a Novel Diagnostic Test for Tuberculous Meningitis
Badrul Munir1, Dewi Santosaningsih2, Dwi Yuni Nur Hidayati3, Tri Yudani Mardining Raras4,
Tommy Alfandy Nazwar5, Sumarno Reto Prawiro2, Shelby Amrus Ernanda1,
Fahimma Fahimma1
1Neurology Department, Faculty of Medicine, Brawijaya University –
Dr. Saiful Anwar General Hospital, Malang 65111, Indonesia.
2Department of Clinical Microbiology, Faculty of Medicine, Brawijaya University, Malang 65145, Indonesia.
3Department of Clinical Microbiology, Dr. Saiful Anwar General Hospital, Malang 65111, Indonesia.
4Department of Biochemistry and Molecular Biology,
Faculty of Medicine, Brawijaya University, Malang 65145, Indonesia.
5Department of Neurosurgery, Faculty of Medicine, Brawijaya University –
Dr. Saiful Anwar General Hospital, Malang 65145, Indonesia.
*Corresponding Author E-mail: badroel2007@ub.ac.id
ABSTRACT:
The diagnosis of tuberculous meningitis (TBM) continues to pose significant difficulties owing to the inherent constraints of presently available diagnostic techniques. This study aimed to evaluate the diagnostic potential of Antigen 38-kDa (Ag38) detection in cerebrospinal fluid (CSF) using recombinant antibodies. This cross-sectional study enrolled 36 participants at Saiful Anwar General Hospital, Indonesia: 12 culture-confirmed TBM patients, 12 clinically-diagnosed TBM patients, and 12 non-TBM controls. CSF samples were analyzed using ELISA with recombinant Ag38 antibodies as primary antibodies and both IgM and IgG as secondary antibodies. Diagnostic performance was assessed through sensitivity, specificity, and ROC curve analysis. As a result, using IgG as secondary antibody, CSF Ag38 levels were significantly higher in culture-confirmed TBM (0.589 ± 0.147) and clinically-diagnosed TBM (0.579 ± 0.158) compared to controls (0.402 ± 0.036) (P = 0.004). Using an optimal cutoff threshold of 0.444, IgG-based detection demonstrated 83.3% sensitivity and 91.7% specificity, achieving a positive predictive value of 90.1% and a negative predictive value of 84.6%. ROC analysis demonstrated excellent diagnostic performance with AUC of 0.854 (P = 0.004). Detection of Ag38 in CSF using recombinant antibodies, particularly with IgG as secondary antibody, shows promising diagnostic potential for TBM. This method could serve as a valuable addition to the current diagnostic arsenal for TBM.
KEYWORDS: Antigen 38-kDa, Cerebrospinal fluid, Diagnosis, ELISA, Tuberculous meningitis.
INTRODUCTION:
Tuberculosis meningitis (TBM) is a form of severe extrapulmonary infection caused by Mycobacterium tuberculosis (Mtb) with high mortality and morbidity rates. In developed countries, more than 50% of TB ME patients have permanent neurological sequelae, and more than two-thirds of patients experience disability complications. TB ME has a fairly high mortality rate, reaching 19.3-21.5%. In Indonesia, previous research at Saiful Anwar Hospital, East Java reported that the incidence of TBM mortality reached 42.9%1,2.
The high mortality and disability rates are related to the difficulty in establishing a diagnosis of TBM and the many complications that arise, as well as some tuberculosis drugs that have difficulty penetrating the blood-brain barrier2. Although the World Health Organization (WHO) has recommended the use of the Xpert gene for the diagnosis of TBM since 2013, meta-analysis data found that the sensitivity was only 79.5% and the predictive value was only 84%, even in endemic areas 1 in 6 TBM patients examined with Xpert was not detected3.
Antigen 38 (Ag38), also referred to as PstS-1, is a lipoprotein secreted by Mtb that is capable of eliciting immune responses from both B and T lymphocytes. This protein is composed of 373 amino acids containing signal peptides which are then released onto the surface of the Mtb membrane. It is secreted via the ATP-binding cassette (ABC) transport system and accumulates within the cell wall, thereby facilitating a robust immune response and activating the adaptive immune system. Ag38 is also known to be associated with active tuberculosis disease and is a glycosylated protein that can be found both intracellularly and secreted in extracellular supernatants of Mtb bacteria5,6.
Serological testing utilizing specific antigens is regarded as a promising diagnostic approach due to its simplicity, affordability, and suitability for implementation in resource-limited settings, including developing countries. Several studies have used antibodies for the diagnosis of TBM and several antigens have been identified and are promising to be evaluated as a diagnostic modality for ME TB, including antigen 5, p32 antigen, 88-kDa antigen, 30-kDa antigen, cord factor, and ESAT-67. However, studies on Ag38 in central nervous system TB infection are still limited. In fact, this antigen plays a very important role in the process of TB infection outside the CNS.
This study aims to evaluate the diagnostic potential of Ag38 as a biomarker in TBM by enabling precise identification of the target antigen within cerebrospinal fluid (CSF) using recombinant antibody-based immunoassays. We hypothesize that Ag38 is detectable in the CSF of patients with TBM and may offer superior sensitivity and specificity compared to conventional diagnostic approaches.
MATERIALS AND METHODS:
A cross-sectional design was utilized for this study. The research was carried out Dr. Saiful Anwar General Hospital, located in East Java, Indonesia, during the period between July and December 2022. We enrolled three groups of participants: (1) culture-confirmed TBM patients, (2) clinically-diagnosed TBM patients, and (3) non-TBM controls. The inclusion criteria were age 18-60 years, availability of CSF samples, and neuroimaging results. Patients who had received anti-tuberculosis treatment, underwent trepanation surgery, or had incomplete data were excluded. The institutional ethics review board of Dr. Saiful Anwar General Hospital approved this investigation, and written informed consent was obtained from all subjects before their participation in the study.
CSF samples were collected through lumbar puncture following standard aseptic procedures. The samples were centrifuged four times for 30 minutes to separate the serum. For Ag38 detection, we used recombinant Ag38 antibodies produced in our previous study8 as primary antibodies, while commercial Goat anti-Mouse IgM Secondary Antibody, HRP (Thermofisher, Catalog # PA1-84383) and Goat anti-Mouse IgG (H+L) Secondary Antibody, HRP (Thermofisher, Catalog # 31430) were used as secondary antibodies.
All ELISA testing was conducted in strict adherence to standardized protocol guidelines. The assay protocol involved pipetting 50μL of calibration standards into appropriate wells (excluding blank controls), followed by administration of 50μL of biotin-labeled detection antibody. Plate sealing and thermal incubation were then performed at 37°C for precisely 45minutes. After completing four buffer wash cycles, 100μL of HRP-labeled conjugate was dispensed into all experimental wells (omitting blanks), then subjected to 37°C incubation for exactly 30minutes. Subsequent to five additional buffer wash cycles, 90 microliters of chromogenic substrate were dispensed into each well and incubated under light-protected conditions at 37°C for exactly 15minutes. The enzymatic catalysis was subsequently quenched through the addition of 50μL stop reagent. Subsequently, absorbance readings were obtained at a wavelength of 405 nm utilizing a microplate spectrophotometer.
The data's distributional properties were evaluated using the Shapiro-Wilk test for normality, and Levene's test was applied to determine the homogeneity of variances. Parametric data exhibiting a normal distribution and equal variances were analyzed using one-way ANOVA, followed by post-hoc testing employing the Least Significant Difference (LSD) method. Non-parametric datasets or those exhibiting variance heterogeneity were evaluated using the Kruskal-Wallis test, with subsequent pairwise comparisons conducted via the Mann-Whitney U test. Optimal cut-off values were established by maximizing both sensitivity and specificity. The assay's diagnostic accuracy parameters, comprising sensitivity, specificity, along with positive and negative predictive values (PPV and NPV), were calculated. Statistical significance was established at a probability level of p< 0.05. All data analyses were conducted utilizing the SPSS statistical software suite (IBM Corp., Armonk, NY).
RESULT:
Patient Characteristic:
A total of 36 participants were enrolled in this study, consisting of 12 culture-confirmed TBM patients, 12 clinically-diagnosed TBM patients, and 12 non-TBM controls. The mean age was 32.29±16.33 years in the culture-confirmed TBM group, 40.75±13.25 years in the clinically-diagnosed TBM group, and 29.67±9.50 years in the control group. Male predominance was observed in both TBM groups (66.7%), while the control group had an equal gender distribution. See table 1 for the patients’ characteristics in detail.
Table 1: Demographic and Clinical Characteristics of Study Participants
|
Characteristics |
Definitive TBM (n=12) |
Clinical TBM (n=12) |
Non-TBM (n=12) |
|
Demographics |
|||
|
Age, years (mean ± SD) |
32.29 ± 16.33 |
40.75 ± 13.25 |
29.67 ± 9.50 |
|
Male sex, n (%) |
8 (66.7) |
8 (66.7) |
6 (50.0) |
|
Clinical Features, n (%) |
|||
|
Decreased consciousness |
11 (91.7) |
6 (50.0) |
7 (58.3) |
|
Headache |
12 (100.0) |
8 (66.7) |
8 (66.7) |
|
Fever |
11 (91.7) |
11 (91.7) |
6 (50.0) |
|
Meningeal signs |
12 (100.0) |
6 (50.0) |
4 (33.3) |
|
Seizures |
3 (25.0) |
4 (33.3) |
4 (33.3) |
|
Cranial nerve palsy |
3 (25.0) |
4 (33.3) |
4 (33.3) |
|
Motor deficit |
7 (58.3) |
5 (41.7) |
4 (33.3) |
|
Clinical Parameters |
|||
|
GCS at admission (mean ± SD) |
9.8 ± 3.5 |
12.9 ± 2.8 |
11.4 ± 5.0 |
|
Temperature, °C (mean ± SD) |
37.66 ± 1.16 |
37.4 ± 1.02 |
36.57 ± 0.45 |
|
Sodium, mmol/L (mean ± SD) |
128.6 ± 3.7 |
133.5 ± 6.7 |
132.4 ± 7.5 |
|
Disease Severity, n (%) |
|||
|
Mild |
6 (50.0) |
9 (75.0) |
NA |
|
Moderate |
0 (0.0) |
3 (25.0) |
NA |
|
Severe |
6 (50.0) |
0 (0.0) |
NA |
|
Complications and Comorbidities, n (%) |
|||
|
Hydrocephalus |
7 (58.3) |
2 (16.7) |
2 (16.7) |
|
Pulmonary TB |
11 (91.7) |
5 (41.7) |
2 (16.7) |
|
Diabetes mellitus |
1 (8.3) |
0 (0.0) |
2 (16.7) |
|
HIV |
1 (8.3) |
2 (16.7) |
1 (8.3) |
|
Mortality, n (%) |
6 (50.0) |
1 (8.3) |
5 (41.7) |
TBM, tuberculous meningitis; SD, standard deviation; GCS, Glasgow Coma Scale; NA, not applicable; TB, tuberculosis; HIV, human immunodeficiency virus.
Table 2: CSF Ag38 Levels and Statistical Analysis Results
|
Detection Method |
Study Groups |
Mean Absorbance ± SD |
P-value* |
Cut-off |
Sensitivity (%) |
Specificity (%) |
PPV (%) |
NPV (%) |
AUC |
|
Ag38 (IgM) |
Culture-confirmed TBM |
0.197 ± 0.052 |
0.009 |
0.156 |
75 |
75 |
75 |
75 |
0.854 |
|
Clinical TBM |
0.240 ± 0.186 |
||||||||
|
Non-TBM |
0.137 ± 0.022 |
||||||||
|
Ag38 (IgG) |
Culture-confirmed TBM |
0.589 ± 0.147 |
0.004 |
0.444 |
83.3 |
91.7 |
90.1 |
84.6 |
0.854 |
|
Clinical TBM |
0.579 ± 0.158 |
||||||||
|
Non-TBM |
0.402 ± 0.036 |
||||||||
AUC, area under the curve; CSF, cerebrospinal fluid; NPV, negative predictive value PPV, positive predictive value; SD, standard deviation, TBM, tuberculous meningitis. Cut-off values were determined using maximum sensitivity and specificity method. Post-hoc analysis showed significant differences between both TBM groups and non-TBM controls for IgM (P = 0.005 and P = 0.014) and IgG (P = 0.003 and P = 0.006) detection methods.
*Kruskal-Wallis test
CSF Ag38 Levels:
Diagnostic Performance:
The optimal cut-off value for Ag38 detection using IgM as secondary antibody was 0.156, yielding a sensitivity of 75% and specificity of 75%. Using IgG as secondary antibody, the optimal cut-off value was 0.444, resulting in a sensitivity of 83.3% and specificity of 91.7%. The positive and negative predictive values for IgG detection were 90.1% and 84.6%, respectively. ROC curve analysis demonstrated excellent diagnostic performance with AUC values of 0.854 for both IgM and IgG detection methods (P = 0.003 and P = 0.004, respectively) (Figure 1).
Figure 1. Receiver Operating Characteristic (ROC) curves for Antigen 38-kDa detection in cerebrospinal fluid using IgM and IgG secondary antibodies.
The blue line represents IgM-based detection (AUC = 0.854, P = 0.003) and the pink line represents IgG-based detection (AUC = 0.854, P = 0.004). The green diagonal line serves as the reference line. Both methods demonstrated excellent diagnostic performance, with IgG-based detection showing optimal sensitivity (83.3%) and specificity (91.7%) at a cut-off value of 0.444, while IgM-based detection showed 75% sensitivity and 75% specificity at a cut-off value of 0.156
DISCUSSION:
This study demonstrates that Ag38 (PstS-1) can be detected in CSF of TBM patients with high diagnostic accuracy. Using IgG as secondary antibody, our method achieved 83.3% sensitivity and 91.7% specificity, showing better performance compared to conventional diagnostic methods such as Ziehl-Neelsen staining, which has been reported to have only 10-20% sensitivity despite 100% specificity7.
The increased levels of Ag38 detected in the CSF of TBM patients can be attributed to its role as a lipoprotein secreted by Mtb. Ag38 is produced and secreted through the ABC transport system and accumulates in the cell wall, enabling strong immune response induction and triggering adaptive immune responses4. During TBM infection, Ag38 may be released into the CSF through several possible mechanisms, including cell lysis during bacterial death, secretion through type V secretion systems, proteolytic cleavage, or host immune responses9,10.
Our findings suggest that IgG-based detection performs better than IgM for CSF Ag38 detection. This is particularly notable given that WHO's currently recommended diagnostic tool, GeneXpert, has been reported to have sensitivity of only 79.5%, with one in six TBM patients in endemic areas showing false-negative results3. The superior performance of IgG-based detection might be attributed to the high specificity of Ag38 as a serological marker, as previously suggested by studies in pulmonary tuberculosis5,6.
The observed differences in Ag38 levels between culture-confirmed and clinically-diagnosed TBM patients could be influenced by various factors, including strain variations, host immune responses, and disease manifestations. Proteomic analyses have shown significant variations in protein abundance among Mycobacterium tuberculosis strains, affecting virulence and pathogenicity11. Furthermore, the equilibrium between tumor necrosis factor (TNF) and its soluble receptors is critical, as an elevated ratio of soluble TNF receptors to TNF suggests insufficient TNF production, potentially influencing protein expression and the course of disease progression12.
The diagnostic potential of Ag38 detection in CSF aligns with previous studies that found Ag38 to be a promising biomarker in other forms of tuberculosis. A study on childhood TBM reported that Ag38 detection provided 84% sensitivity (95% CI: 77, 89) and 89% specificity (95% CI: 83, 93)13. Our results demonstrate comparable or slightly better performance, particularly when using IgG as the secondary antibody.
Although the findings of our method are encouraging, several limitations should be acknowledged. Firstly, this investigation was conducted as a single-center study with a modest sample size. Second, genetic sequencing of the PstS1 gene was not performed for individual TBM cases, leaving the possibility of strain variations unaddressed. Finally, we used a single antigen approach, whereas a multi-antigen panel might provide more comprehensive immunological data for diagnostic purposes.
CONCLUSION:
Detection of Ag38 in CSF using recombinant antibodies demonstrates excellent diagnostic potential for TBM. The method showed remarkable efficacy when IgG was used as the secondary antibody, achieving a sensitivity of 83.3%, specificity of 91.7%, and positive and negative predictive values of 90.1% and 84.6%, respectively. These findings suggest that CSF Ag38 detection could serve as a valuable diagnostic tool for TBM, potentially complementing existing diagnostic methods. Further multicenter studies with larger sample sizes are warranted to validate these findings and explore the potential of combining Ag38 detection with other biomarkers for improved TBM diagnosis.
CONFLICT OF INTEREST:
The authors have no conflicts of interest regarding this investigation.
REFERENCES:
1. Munir B, Rianawati S, Sutanto H, Setyawan UA. Mortality of central nervous system infection adult patients in indonesia (Descriptive study in a tertier referral hospital). J Neurol Sci. 2017; 381: 414.
2. Nguyen DT, Agarwal S, Graviss EA. Trends of tuberculosis meningitis and associated mortality in Texas, 2010-2017, a large population-based analysis. PLoS One. 2019; 14(2): e0212729.
3. Bahr NC, Marais S, Caws M, Van Crevel R, Wilkinson RJ, Tyagi JS, et al. GeneXpert MTB/Rif to diagnose tuberculous meningitis: perhaps the first test but not the last. Clin Infect Dis. 2016; 62(9): 1133–5.
4. Bando-Campos G, Juárez-López D, Román-González SA, Castillo-Rodal AI, Olvera C, López-Vidal Y, et al. Recombinant O-mannosylated protein production (PstS-1) from Mycobacterium tuberculosis in Pichia pastoris (Komagataella phaffii) as a tool to study tuberculosis infection. Microb Cell Factories. 2019; 18: 1–19.
5. Senol G, Erer OF, Yalcin YA, Coskun M, Gündüz A, Biçmen C, et al. Humoral immune response against 38-kDa and 16-kDa mycobacterial antigens in tuberculosis. Eur Respir J. 2006; 29(1): 143–8.
6. Singh AK, Malhotra HS, Garg RK, Jain A, Kumar N, Kohli N, et al. Paradoxical reaction in tuberculous meningitis: presentation, predictors and impact on prognosis. BMC Infect Dis. 2016; 16: 1–11.
7. Pukazhvanthen P, Anbarasu D, Basirudeen SAK, Raja A, Singh M. Assessing humoral immune response of 4 recombinant antigens for serodiagnosis of tuberculosis. Tuberculosis. 2014; 94(6): 622–33.
8. Munir B, Hidayati DYN, Nazwar TA, Raras TM, Prawiro SR. The Ag38-rec Mycobacterium tuberculosis antigen as a new candidate marker for the diagnostic of tuberculosis meningitis: In silico approach. Res J Pharm Technol. 2023; 16(11): 5289–95.
9. Pham T, Werneburg GT, Henderson NS, Thanassi DG, Delcour AH. Effect of chaperone-adhesin complex on plug release by the PapC usher. FEBS Lett. 2016; 590(14): 2172–9.
10. Hwang WH, Lee WK, Ryoo SW, Yoo KY, Tae GS. Expression, purification and improved antigenicity of the Mycobacterium tuberculosis PstS1 antigen for serodiagnosis. Protein Expr Purif. 2014; 95: 77–83.
11. Mehaffy C, Hess A, Prenni JE, Mathema B, Kreiswirth B, Dobos KM. Descriptive proteomic analysis shows protein variability between closely related clinical isolates of Mycobacterium tuberculosis. Proteomics. 2010; 10(10): 1966–84.
12. Rydberg J, Miörner H, Chandramuki A, Lantz M. Assessment of a possible imbalance between tumor necrosis factor (TNF) and soluble TNF receptor forms in tuberculous infection of the central nervous system. J Infect Dis. 1995; 172(1): 301–4.
13. Haldar S, Sankhyan N, Sharma N, Bansal A, Jain V, Gupta V, et al. Detection of Mycobacterium tuberculosis GlcB or HspX Antigens or devR DNA impacts the rapid diagnosis of tuberculous meningitis in children. 2012;
|
Received on 24.04.2025 Revised on 08.09.2025 Accepted on 17.11.2025 Published on 20.05.2026 Available online from May 25, 2026 Research J. Pharmacy and Technology. 2026;19(5):2128-2132. DOI: 10.52711/0974-360X.2026.00306 © RJPT All right reserved
|
|
|
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. Creative Commons License. |
|