Diagnostic performance of culture filtered protein 10‐specific perforin in pediatric patients with active tuberculosis

Abstract Background Mycobacterium tuberculosis (Mtb)‐specific perforin were significantly increased in patients with tuberculosis. This study aims to evaluate the diagnosis value of Mtb‐specific perforin in pediatric patients with tuberculosis. Methods Diagnostic performance of perforin levels induced by 6‐kDa early secreted antigen target (ESAT6) or culture filtered protein 10 (CFP10) were evaluated in eighty‐six samples from children participants by receiver operating characteristic curve analysis. Flow cytometry was used to detect the expression of perforin and INF‐γ of CD4+, CD8+ T cells in response to CFP10 stimulation. Results After ex vivo stimulation, levels of ESAT6/CFP10‐specific perforin in LTBI patients were significantly higher than active TB (ATB) patients, non‐tuberculosis infection (non‐TB), and health control (HC) individuals. The diagnostic efficacy of CFP10‐specific perforin for TB diagnosis was significantly higher than ESAT6‐specific perforin and T‐SPOT assay, and when 0.74 ng/mL was taken as the cutoff value, the sensitivity, specificity, and accuracy were 97.83%, 87.5%, and 93.02%. CFP10‐specific perforin in both CD4+ and CD8+ T cells were significantly higher in ATB patients compared to HCs and further increased in LTBI patients. However, INF‐γ was mainly secreted by CD4+ T cells and showed no significant difference between LTBI and ATB patients. In addition, CFP10‐specific perforin can effectively distinguish between ATB and LTBI with the cutoff value of 1.80 ng/mL. Sensitivity and specificity were 88.46% and 85.62%, respectively. Conclusions CFP10‐specific perforin may be used as a novel cellular immunity‐based diagnostic marker of pediatric patients with tuberculosis, and with the potential for discriminating ATB from LTBI.


| INTRODUC TI ON
Tuberculosis (TB) is one of the top 10 causes of death, and the leading cause from a single infectious pathogen (Mycobacterium tuberculosis, Mtb), ranking above HIV/AIDS. Pediatric patients account for an increasing proportion of TB cases in both high-resource and low-resource countries, with an estimated 1.1 million new cases in 2018. 1 It is estimated that most active TB (ATB) cases originate from an initial latent TB infection (LTBI), a state of persistent immune response to stimulation by Mtb antigens without evidence of clinical manifestations of ATB. 2 In contrast to adults, pediatric patients with LTBI are more likely to progress to ATB within the first year of primary infection 3 and the efficacy of currently available treatments ranges from 60% to 90%. 2 Thus, accurate and early identification of LTBI has become one of the key strategies to reduce TB incidence. 4,5 Clinical manifestations of TB, especially in infancy, are often non-specific and similar with many common childhood diseases.
Microbiological diagnosis of TB in pediatric patients remains a challenge as sputum is difficult to obtain and usually pauci-bacillary, which resulted a poor sensitivity (<15%) of sputum smear microscopy in pediatric TB, even with optimized methods such as centrifugation of samples and use of fluorescent microscopy. 6 While culture of Mtb is more sensitive than smear microscopy, bacteriological confirmation of pediatric TB is still less than 40%. 7,8 Automated real-time nucleic acid amplification tests (Xpert MTB/RIF) are only positive in a proportion of children who have been clinically diagnosed with tuberculosis. [9][10][11] Due to the difficulty in collecting satisfactory sputum samples from children, pediatric research has focused on finding new biomarkers in non-sputum-based samples for the detection of Mtb infection, including tuberculin skin test (TST) and interferon-gamma release assays (IGRAs). 12 Meta-analyses revealed that both TST and IGRAs had higher specificities (84%-98%) but low sensitivities (50%-80%) in pediatric patients, and were unable to distinguish between ATB and LTBI. [13][14][15] Importantly, TST are typically unresponsive during the first several weeks after infection, especially in infants and in cases of very severe or miliary ATB. 13 Therefore, there is an urgent need to explore experimental methods for rapid diagnosis of tuberculosis in pediatric patients, especially for the differentiation of two different clinical status, LTBI and ATB.
Many studies have tried to use the combinations of different immunodiagnostic biomarkers such as EGF and TGF-α; IL-1RA, IP-10, and VEGF; IL-2 and INF-γ; IL-15 and MCP-1; TNF-α, IL-12p40, and IL-17; EGF, MIP-1β, sCD40L, and IL-1α to improve the sensitivity of TB diagnosis. 16,17 In addition, Comella-del-Barrio et al 18 recently reported that the combination of INF-γ, IP-10, ferritin, and 25-hydroxyvitamin D may have the potentiality to discriminate ATB from LTBI in pediatric patients. Although the diagnostic accuracy of these methods still need to be confirmed in multicenter studies, it is undeniable that soluble factors of PBMCs induced by Mtb-specific antigen in vitro may be new markers with the potential for be used as new biomarkers for effectively diagnosis of pediatric patients with TB.
Host mainly depends on cellular immunity that mediated by CD4 + and CD8 + T lymphocytes to defense against Mtb infection. The cytotoxic molecules secreted by CD8 + T cells, such as granzyme and perforin, are directly involved in immune defense against Mtb 19 and usually suppressed in granulomas within the tuberculosis lesions. 20 In a multiple cohort to identify stage-specific host responses to Mtb infection, Roshni et al 21 reported that perforin-expressing cells were significantly higher in LTBI than in normal and ATB patients, and were associated with the progression of tuberculosis. It is worthy to note that, unlike IFN-γ mainly secreted by CD4 + T cells, perforin is not only secreted by CD8 + T cells, but also by CD4 + T cells. Laetitia et al 22 reported that perforin-expressing CD4 + T cells induced by Mtb-specific antigens (Heparin-binding hemagglutinin, HBHA) were found in most LTBI subjects and significantly higher than which in ATB patients. These results suggest that perforin may be a novel immune-responsive biomarker for diagnosis of Mtb infection and the differentiation of LTBI and ATB. Therefore, we compared perforin levels induced by Mtb-specific antigens in PBMCs of pediatric patients with TB in this study, and further evaluated their diagnostic potential for the differentiation of ATB and LTBI.

| Study design and patients
This study was carried out from January 12 to August 17, 2019, in Wuhan Children's Hospital and Wuhan Jinyintan Hospital (a hospital specializing in infectious diseases, Wuhan, China). Eighty-six children (≤14 years old) were classified into the following four groups: healthy controls (HC); non-TB controls; LTBI; and ATB, including pulmonary tuberculosis (PTB) and extrapulmonary tuberculosis (EPTB).
The main characteristics of participants were shown in Table 1.

| PBMCs isolation and stimulation
2-4 mL venous blood of all participants was collected into heparin lithium-anticoagulant tubes, and PBMCs were separated by Ficoll-Hypaque density gradient centrifugation. Then, 100 μL of fresh PBMCs (2.5 × 10 5 /well) was seeded in 96-well plates (BD company) and were stimulated with 50 μL of ESAT6, CFP10, or AIM-V medium (ESAT6 and CFP10 reagents were the same as TSPOT.TB kit, provided by Shanghai Fosun Medical Technology Co., Ltd.). The plates were incubated at 37°C for 16-20 hours. After incubation, the supernatant was collected after centrifugation at 1301 g, and stored at −80°C for use.

| T-SPOT assay
The T-SPOT assay was done according to the manufacture's clinical protocol (Shanghai Fosun Medical Technology Co., Ltd.). Then, 100 μL/well diluted biotinylated antibody (1:100) was added after 3 times washing and incubated for 1 hour at room temperature.

| ELISA for determination of perforin
Next, 100 μL/well of diluted Streptavidin-HRP antibody (1:100) was added after 3 times washing and incubated for 30 minutes at room temperature. Lastly, the absorbance was measured by a microplate reader (Molecular Devices) at OD 450 nm following TMB stain. The standard curve was drawn according to the standard samples; then, the concentration of perforin in the culture supernatant was calculated. Mtb-specific perforin was defined by the following formula: Mtb-specific perforin = perforin ESAT6 or CFP10 stimulation − perforin background . Perforin background represented the perforin levels in mediumstimulated PBMCs.

| Flow cytometry analysis
PBMCs (5 × 10 6 /well) were stimulated with CFP10 in AIM-V medium in 6-well flat bottom tissue culture plate at 37°C for 24 hours.

| Statistical analysis
Data were analyzed using GraphPad Prism 6.0 (GraphPad) and MedCalc V19.1. Statistical significance was determined by the Mann-Whitney U test. Proportions for categorical variables were compared using the chi-square test, although the Fisher exact test was used when the data were limited. Receiver operating characteristic (ROC) analysis was performed to determine the best cutoff levels of Mtb-specific perforin in the discrimination between ATB and LTBI. Area under the curve (AUC), sensitivity, and specificity were reported, as well as the 95% confidence intervals (CI). The optimal cutoff values were chosen as when Youden's index (YI = sensitivity +specificity − 1) was maximum. Statistical significance was determined as P < .05.

| Levels of total perforin in PBMCs stimulated by ESAT6/CFP10
To evaluate the potential of perforin for tuberculosis diagnosis, we firstly combined ATB and LTBI as TB group, non-TB, and  Figure 1C). These data suggest that although levels of total perforin induced by ESAT6 or CFP10 in PBMCs were significantly increased in TB patients, total perforin levels may not have the potential to distinguish ATB from LTBI.

| Levels of Mtb-specific perforin in PBMCs stimulated by ESAT6/CFP10
Then, we further evaluated the difference of Mtb-specific perforin   Figure 2C). These data suggest that ESAT6/CFP10-specific perforin in PBMCs not only distinguish TB from non-TB/HC, but may also with the potential to discriminate between ATB and LTBI.

| Comparison the diagnostic performance of total and Mtb-specific perforin for TB diagnosis
Receiver operating characteristic analysis was then performed to determine the optimal cutoff values of total/Mtb-specific perforin for the diagnosis of Mtb infection (Figure 3). When the cut- Differences between AUCs were analyzed using DeLong test.
The AUC of CFP10-specific perforins was significantly higher than that of total perforin induced by CFP10 and ESAT6, as well

F I G U R E 3
The receiver operating characteristic (ROC) curves (plotting sensitivity versus 1-specificity) of total/Mtb-specific perforin to discriminate patients with TB from control. The ROC analysis was performed for total/Mtb-specific perforin to determine cutoff levels for the diagnosis of TB disease

| Diagnostic performance of CFP10-specific perforin for the discrimination of ATB and LTBI
The AUC of ESAT6-specific perforin was 0.81 (95% CI: 0.68-0.94), with the sensitivity and specificity of 84.62% and 75.00%, when the cutoff value was 0.94 ng/mL ( Figure 5), while the AUC of CFP10specific perforin was 0.95 (95% CI: 0.89-1.00), with the sensitivity and specificity of 88.46% and 85.62%, when the cutoff value was 1.80 ng/mL. The AUC of CFP10-specific perforin was slightly higher than ESAT6 as analyzed by Delong test (P = .0591). However, these assays have limited sensitivity in children, particularly in those under 5 years of age. In high-income countries, the sensitivity of QFT-GIT and T-SPOT.TB is 79% and 67%, respectively, while in lowincome countries, the sensitivity is only 57% and 61%. 15 In addition, has been confirmed to be more sensitive for detecting Mtb infection compared to QFT-GIT. 29 In our study, we found the sensitivity of Mtb-specific perforin induced by ESAT6 or CFP10 were both significantly higher than that of IGRAs in previous meta-analysis. 15 In addition, the sensitivity, specificity, negative predictive value, and accuracy of CFP10-specific perforin for the diagnosis of TB in pediatric patients were significantly better than that of T-SPOT, which may be due to perforin can be secreted from both CD8 + T and CD4 + T cells.

| D ISCUSS I ON
To clarify the reason why diagnostic performance of CFP10-specific perforin in pediatric patients were significantly better than T-SPOT, we further compared the percentage of IFN-γ or perforin producing CD4 + and CD8 + T cells. We found that perforin induced in both CD4 + T cells and CD8 + T cells from ATB patients was higher than that from healthy In this study, total levels of perforin in response to ESAT6/CFP10 peptide stimulation in both ATB group and LTBI group were significantly higher than that of HC group and non-TB group, but there was no significant difference between ATB and LTBI group, which may be caused by the release of perforin at different baseline levels in four groups. According to the methodology conducted by Wang Feng et al, 33 they calculated Mtb-specific TNF-α by subtracting the background level of TNF-α secreted by the unstimulated PBMCs, we found that diagnostic performance of Mtb-specific perforin was substantially improved compared to total perforin for the diagnosis of TB disease and also discriminates between ATB and LTBI in this study. Interestingly, the diagnostic performance of CFP10-specific perforin was significantly better than ESAT6, which may be due to the fact that ESAT6 protein can directly bind to TLR2 or beta-2-microglobulin (β-2M) via the C-terminal six amino acid residues, activate TBK1 and IRF3, thus inhibiting the antigen presentation of host immune cells and the release of inflammatory factors, while CFP10 has no immunosuppressive function. 34,35 Overall, the data presented in this study indicated that CFP10specific perforin is a novel cellular immunity-based diagnostic marker of pediatric patients with tuberculosis, and it demonstrated high diagnostic potential in discriminating ATB and LTBI. The limitation of this study is that the small sample size comes from two hospitals, which may lead to sensitivity and specificity bias and our findings need to be further confirmed in a large, multicenter study of pediatric patients with tuberculosis.

E TH I C A L A PPROVA L
This study was approved by the ethical committee of Wuhan Children's Hospital (WHCH 2019016).

F I G U R E 5
Receiver operating characteristic (ROC) for Mtbspecific perforin as a classifier to distinguish between ATB and LTBI. The ROC analysis was performed for Mtb-specific perforin to determine cutoff levels in distinguishing between ATB and LTBI

I N FO R M ED CO N S ENT
Written consents were obtained from all participants or their guardians.