Evaluation of antigen-specific immune responses for leprosy diagnosis in a hyperendemic area in China

Objective To evaluate antigen-specific immune responses for leprosy diagnosis in a hyperendemic area in China. Methods Eighty-three leprosy patients and 161 non-leprosy controls were enrolled from Hani-yi Autonomous Prefecture of Honghe, Yunnan Province, China. Leprosy patients were divided into multibacillary (MB, n = 38), paucibacillary (PB, n = 23), and post-multi-drug therapy (MDT, n = 22) groups. Controls were divided into the following groups: healthy household contacts (HHC, n = 119), tuberculosis (TB, n = 11), and endemic controls (EC, n = 31). The NDO-LID Rapid Test, M. leprae antigen-specific ELISA and antigen-specific IFN-γ secretion in a whole blood assay (WBA) were used to evaluate these subjects. Results The NDO-LID Rapid Test achieved higher positive response rates in MB than in PB patients[94.7%(36/38) vs 65.2%(15/23)], and these rates were higher than those observed by ELISA using anti-LID-1[92.1%(35/38) vs 52.2%(12/23)], anti-NDO-LID[92.1%(35/38) vs 47.8% (11/23)], and anti-ND-O-BSA[89.5%(34/38) vs 60.9%(14/23)]. However, the NDO-LID Rapid Test also showed a higher positive response rate in the EC group (33.3%,10/31), which was higher than the rates observed for anti-NDO-LID (12.9%,4/31) and anti-ND-O-BSA (16.1%,5/31). M. leprae antigen-specific ELISA demonstrated relatively high specificity (86.84–97.37%) but low sensitivity (15.97–72.73%) in discriminating between leprosy patients and non-leprosy controls by ROC curve analysis. In contrast, M. leprae antigen-specific IFN-γ secretion detection achieved higher positive response rates in PB than in MB patients (positive ratio of MB vs PB: 40% vs 56% for LID-1, 28.6% vs 47.8% for ML89, 31.4% vs 60.7% for ML2044, and 31.4 vs 47.8% for ML2028) and could distinguish MB from EC when stimulated with ML89(AUC = 0.6664) and PB fromTB when stimulated with ML2044 and ML2028(AUC = 0.7549 and 0.7372, respectively). Conclusion The NDO-LID Rapid Test and M. leprae antigen-specific ELISA are useful tools to assist in the diagnosis of leprosy patients, especially MB patients, although the former had higher sensitivity but lower specificity than the latter. M. leprae antigen-specific IFN-γ release assessed by WBA has diagnostic value for distinguishing PB from TB but not for distinguishing PB from HHC or EC. Screening novel M. leprae-specific antigens, combining different M. leprae antigens and a multi-cytokine analyte model may be needed for more effective diagnosis of leprosy.

Introduction Leprosy, a chronic disease caused by Mycobacteriumleprae (M. leprae) infection, has a wide range of clinical outcomes correlated with the host's immune response to the bacilli [1,2].
Current leprosy control strategies rely on diagnosing the disease as early as possible, followed by prompt treatment with multi-drug therapy (MDT) [1]. The implementation of World Health Organization (WHO) MDT for widespread, worldwide treatment has drastically reduced registered leprosy cases from the approximately 12 million reported in 1985 to fewer than 250,000 reported in 2006 [3]. Currently, leprosy is mainly diagnosed by clinicians using defined criteria, slit-skin smears and biopsies [4]. However, as the prevalence of the disease decreases, clinical expertise is diminishing, leading to extended delays between the onset of clinical signs and the diagnosis and consequent sustained transmission of M. leprae [5].
Leprosy patients are predominantly diagnosed by the appearance of disease signs, but they can also be characterized by the physical and histological attributes of skin or nerve lesions or by their immune response to crude or recombinant M. leprae antigens [6,7,8,9]. It has been demonstrated that the immune response to crude or recombinant M. leprae antigens is helpful for detecting multibacillary (MB) leprosy patients by their antibody response [6], for the diagnosis of paucibacillary (PB) patients by antigen-specific CMI [7], and for monitoring the effectiveness of MDT in MB and PB leprosy patients by the antibody response and antigen-specific CMI, respectively [8]. The M. leprae antigens used for ELISA in this study were Leprosy IDRI diagnostic-1 (LID-1), a fusion protein developed by fusing the ML0405 and ML2331 genes [9,10];NDO-LID, a conjugate of LID-1 with natural octyl disaccharide (NDO) [11];and ND-O-BSA, a synthetic PGL-I derivative. The NDO-LID Rapid Test in lateral flow-based format has been developed using NDO-LID. The single tetravalent 89-kDa fusion protein (ML89), designated LEP-F1, consists of the ML2028, ML2055 and ML2380 antigens [12]. A list of accession numbers/ID numbers for genes and proteins included in the NCBI search and mentioned in the text is shown inTable 1.
The purpose of this study was to evaluate the diagnostic value of three antigen-specific immune diagnostic tests, namely, the NDO-LID Rapid Test(antibody response), an antigenspecific enzyme linked immunosorbent assay (ELISA)(anti-LID-1, anti-NDO-LID, and anti-ND-O-BSA)(antibody response), and antigen-specific IFN-γ secretion in a whole blood assay (WBA) (stimulated by LID-1, ML89, ML2044 and ML2028)(antigen-specific CMI) for diagnosing leprosy in a hyperendemic area in China.

Ethics statement
This study was approved by the Medical Ethics Committee of Beijing Friendship Hospital, Capital Medical University, Beijing, P.R. China. Written informed consent was obtained from all adult participants, and all parents or guardians of child participants provided informed consent on their behalf. All of the procedures in the study involving human participants were performed 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.

Study area and subjects
Eighty-three leprosy patients, who were referred to the Honghe Prefecture Skin Disease Prevention and Cure Institute in Honghe Autonomous Prefecture, Yunnan Province, were included in the study. Leprosy diagnosis was established based on clinical signs and symptoms, skin smears, skin biopsy, and neuro-physiologic examinations. The leprosy patients were classified into five groups based on the Ridley and Jopling [13] classification: tuberculoid (TT), borderline-tuberculoid (BT), borderline-borderline (BB), borderline-lepromatous (BL), and lepromatous (LL) groups. For data analysis in this study, leprosy patients were also classified into three groups: PB and MB, according to the WHO operational classification [14] during MDT, or post-MDT. One hundred and sixty-one controls from the same endemic region were included as non-leprosy controls. The controls were further classified into three groups: healthy household contacts (HHC), tuberculosis (TB), and endemic controls (EC).

NDO-LID Rapid Test
Antigen-specific antibody detection by NDO-LID was performed as previously described [15]. Serum antibodies were measured by the NDO-LID rapid diagnostic test (RDT; procured from Orange Life, Rio de Janeiro, Brazil). Briefly, NDO-LID RDT was performed by first adding undiluted serum (10 μl) into the sample well within the test cassette, followed by the addition of running buffer (100 μl). Samples migrated through the cassette such that interactions with the test and/or control lines were revealed as red colored lines within the reading window. Tests were valid if the control line was observed. A positive result was defined by the presence of the test line. Visual results were interpreted after 20 minutes by two independent readers and scored subjectively as (±/+/++/+ + +), with faint (±) or no test line considered a negative result.  . The IFN-γ ELISA employed had a detection limit of 2 pg/ml, and a threshold for positive responses was arbitrarily selected at 50pg/ml according to a previous study [15].

Statistical analysis
Statistical analysis was performed primarily with GraphPad Prism software version 5.0 (GraphPad Software Inc., San Diego, CA, USA). The nonparametric Mann-Whitney U test was used to analyze differences between two groups. The Kruskal-Wallis test was used to analyze differences among three or more groups. Probability (p) values less than 0.05 were considered significant. The diagnostic utility of individual M. leprae antigen-specific responses for leprosy disease, including sensitivity, specificity, Youden's index, and area under the receiver operator characteristic curve (AUC), were ascertained by receiver operator characteristics (ROC) curve analysis. The concordance between results was determined by kappa values (κ), and p values were calculated (Statistical Package for the Social Sciences (SPSS) version 16.0).

Study area
The study was undertaken mainly in counties in Honghe

Basic characteristics of leprosy patients and controls
Eighty-three leprosy cases[MB, n = 38; PB, n = 23; and MDT, n = 22] and 161 controls [HHC, n = 119; TB, n = 11; and EC, n = 31] from the same endemic region were included. The basic information for each study group is summarized in Table 2.

Comparing the consistency of NDO-LID Rapid Test, M. leprae antigenspecific ELISA and WBA by kappa test
A kappa test analyzes for the agreement of results collected from various test formats. When a kappa test was performed between the NDO-LID Rapid Test and M. leprae antigen-specific    (Table 4). This finding indicates that the two tests showed high consistency for the diagnosis of MB leprosy patients.

Discriminating between leprosy patients and controls with M. leprae antigen-specific ELISA
For all three M. leprae antigens (LID-1, NDO-LID and ND-O-BSA ELISA), the OD values showed significant differences for MB vs the PB, post-MDT, HHC, TB or EC groups, and PB vs EC (Fig 1). Of note, NDO-LID was better than the other two antigens (LID-1 and ND-O-BSA) at discriminating PB leprosy patients from non-leprosy controls (Table 5). In addition, we evaluated the diagnostic ability of M. leprae antigen-specific ELISA using ROC curve analysis, AUC, sensitivity and specificity (Table 6) and demonstrated that this method had a relatively high specificity but low sensitivity.

Evaluation of IFN-γ as a potential diagnostic host biomarker for leprosy
We compared the analyte levels detected in M. leprae antigen-stimulated WBA supernatants in leprosy patients with the levels obtained from the non-leprosy control groups using the mean and standard deviation(SD) (Fig 2) and the median and range( Table 7). As described previously, newly diagnosed PB patients produce more IFN-γ than MB patients. We also evaluated the diagnostic potential of IFN-γ by ROC curve analysis and AUC. IFN-γ levels were significantly different in (1) Table 8).

Discussion
Widespread application of MDT therapy has led to major advances in leprosy control, with sharp declines in prevalence rates in the vast majority of countries over the last 20 years [15,17]. However, the disease remains a public health concern in many regions. In 2010, China reported 1324 new cases of leprosy to the WHO [18]. The majority of cases in China came from the ethnically diverse, mountainous, and underdeveloped southwest provinces of Yunnan, Guizhou, and Sichuan [19,20]. Honghe Autonomous Prefecture inYunan is considered a highly endemic area for leprosy in China. We enrolled 83 leprosy patients and 161 controls from this endemic region in this study to evaluate the ability of several diagnostic tests to correctly diagnose different categories of leprosy patients. We found that the NDO-LID Rapid Test and M. leprae antigen-specific ELISA was useful to diagnose leprosy patients in hyperendemic areas of leprosy disease, especially MB patients. The former method provides a pointof-care measurement of antibodies and had higher sensitivity but lower specificity than the latter. M. leprae antigen-specific IFN-γ secretion in WBA has diagnostic value for distinguishing PB from TB but not for distinguishing PB and HHC or EC. M. leprae-specific antigen tests have been developed as useful tools to diagnose leprosy. LID-1, NDO-LID, and ND-O-BSA(also named PGL-I), as representative M. leprae-specific antigens, have been widely evaluated as leprosy diagnostics in the hyper-endemic regions of Brazil [20][21][22][23], Colombia, the Philippines [24], and China [15] and have been demonstrated to be excellent tools for detecting MB leprosy patients in a simple and highly quantitative manner [24], predicting patients susceptible to developing leprosy type 2 reactions (T2R) [23], and distinguishing leprosy from other confounding dermatoses [18].  Table 5. P values of M. leprae antigen-specific ELISA between the leprosy group and control groupsfrom the Kruskal-Wallis test(among three groups) and the Mann-Whitney U test(between two groups).

Immune diagnosis in leprosy of China
The NDO-LID Rapid Testwas compared with M. leprae antigen-specific ELISA and demonstrated a high degree of sensitivity but significant differences in specificity for leprosy diagnosis [25]. Therefore, this test is an effective tool for screening and identifying individuals at high risk who might benefit from regular monitoring [26]. Our previous study showed that confirmation was achieved in 95% of MB leprosy patients with the NDO-LID Rapid Test, and a high degree of agreement was observed with LID-1, NDO-LID, and ND-O-BSA ELISA [15]. In addition, 63.6% of PB leprosy patients were confirmed, and the NDO-LID Rapid Test had a higher detection rate in PB leprosy patients than LID-1, ND-O-BSA, and NDO-LID ELISA [15]. In this study, we enlarged the sample size and obtained results similar to those of previous studies. These data indicate an improved capacity of the NDO-LID Rapid Test over M. leprae ELISA for detecting the disease. However, the test also suffers from higher positive responses in the EC group than did NDO-LID and LID-1 ELISA. This implies that the NDO-LID Rapid Table 6. ROC of the diagnostic potential of OD detected in M. leprae antigen-specific ELISA for leprosy diagnosis. All of the analytes that showed significant differences (p<0.05) between leprosy and non-leprosy controls according to receiver operating characteristic (ROC) analysis are shown. Sensitivity and specificity were selected based on Youden's index. AUC = Area under the receiver operator characteristic curve.

M. leprae Antigens Subgroups of Cases p value AUC Sensitivity (%) Specificity (%) Youden's Index
Test was more sensitive but less specific than M. leprae antigen-specific ELISA(anti-LID-1 and anti-NDO-LID) for discriminating the leprosy patient group fromthen on-leprosy EC group. Despite the relatively low specificity, the NDO-LID Rapid Test, as a low-tech, robust assay, can still be applied in resource-poor settings to measure the immune response to M. leprae infection and can be used as a tool for leprosy screening in combination with good specificity confirmation tests, which will lead to timely treatment and reduced transmission [27]. We also evaluated the capacity of M. leprae antigen-specific ELISA to discriminate between the leprosy and control groups. All three M. leprae antigens (LID-1, NDO-LID and ND-O-BSA) were able to discriminate the MB group from all other leprosy and non-leprosy groups and the PB leprosy group from the non-leprosy EC groups, whereas only NDO-LID was able to discriminate the PB leprosy group from the non-leprosy HHC group. This indicates that all three M. leprae antigens have potential and specific value for research and medical applications. As described before, the M. leprae antigen-specific ELISA had lower sensitivity but better specificity than the NDO-LID Rapid Test. ELISA detection of specific antibodies may be preferred for confirming diagnoses, differentiating leprosy from other dermatological conditions, and performing follow-up studies for leprosy HHC and indeterminate leprosy, which are very early signs of the disease that are often missed by family members and medical personnel in the endemic area [27]. were used in these studies [20,[28][29][30][31]. IFN-γ and CXCL10 were evaluated as potential diagnostic host markers for PB leprosy patients in the hyper-endemic regions of Brazil [20,[28][29][30][31] and China [15]. Newly diagnosed PB patients produced more IFN-γ than MB patients [28][29][30][31], and IFN-γ was helpful in the differential diagnosis of leprosy from other confounding dermatoses [2]. CXCL10 discriminated PB from EC only in ML0276+LID-1 WBA; however, CXCL10 could not discriminate active disease (PB) from HHC individuals [28]. In this study, we also demonstrated that for new cases, PB patients produced more IFN-γ than MB patients; however, IFN-γ did not discriminate active disease (PB) from HHC or EC individuals.
In this study, we investigated the accuracy of IFN-γ as a host marker detected in supernatants after stimulation of whole blood with M. leprae-specific antigens (LID-1, ML89, ML2044, and ML2028) in an overnight culture assay and compared the IFN-γ marker levels in leprosy and non-leprosy control groups. Although IFN-γ can be useful as a host biomarker that contributes to a diagnostic signature of MB vs EC and that distinguishes PB vs TB groups, there was no evidence that IFN-γ was able to discriminate between the PB and HHC or EC groups. To screen novel M. leprae-specific antigens, combining different M. leprae antigens and facilitating a multi-cytokine analyte model may achieve improved diagnostic potential.
This study is limited by the small sample size, especially in the PB group. Antigen-specific immune responses have had limited diagnostic ability for leprosy disease and until recently have only been used for seroepidemiological investigation in hyperendemic areas of leprosy disease or in patients clinically suspected of having leprosy disease. However, the results should be interpreted with caution. Only a very limited number of M. leprae-specific antigens(LID-1, ML89, ML2044, and ML2028) and only one potential diagnostic host biomarker (IFN-γ) were tested for leprosy diagnosis in this study. Future studies should focus on additional M. lepraespecific antigens as well as additional host biomarkers. Median levels of analytes (pg/ml) and ranges (minimum and maximum) showing accuracies at discriminating between leprosy and controls in overnight culture supernatants for all of the study participants.  All of the analytes that showed significant differences (p<0.05) between leprosy and uninfected controls according to ROC analysis are shown. Sensitivity and specificity were selected based on Youden's index. AUC = Area under the receiver operator characteristics curve. https://doi.org/10.1371/journal.pntd.0006777.t008 In conclusion, the NDO-LID Rapid Test and M. leprae antigen-specific ELISA were helpful for diagnosing leprosy in hyperendemic areas of leprosy disease, especially for MB patients. The former had higher sensitivity but lower specificity than the latter. Although IFN-γ has been widely studied as a potential biomarker for PB leprosy patients, more research is needed to identify feasible M. leprae-specific antigens and other appropriate host biomarkers to improve its diagnostic value in PB patients in future studies.