Differential expression of CD64 in patients with Mycobacterium tuberculosis infection: A potential biomarker for clinical diagnosis and prognosis

Abstract To evaluate the clinical utility of neutrophil (n)CD64 index to diagnose pulmonary tuberculosis (PTB) and extrapulmonary TB (ePTB) and to predict the outcome of Mycobacterium tuberculosis infection. We recruited 189 patients with active TB and 140 controls and measured the differential expression of nCD64 index using flow cytometry. The receiver operating characteristics (ROC) curve analysis was performed to estimate the diagnostic performance of the nCD64 index and T‐SPOT.TB assay for the diagnosis of TB. Furthermore, we analysed whether the nCD64 index in patients with TB was correlated with inflammatory indicators. Finally, we assessed the prognosis of patients by following the dynamic changes of the nCD64 index once a week. The nCD64 index was significantly higher in active TB group (PTB and ePTB), than in the anti‐TB and healthy controls (HC) groups. The sensitivity and specificity of nCD64 index for the differential diagnosis of PTB and pneumonia (PN) patients were 68.33% and 77.55%, respectively. The sensitivity and specificity of nCD64 index for the diagnosis of tuberculous meningitis (TBM) were 53.85% and 100%, respectively. Furthermore, there was a weak correlation between the nCD64 index and inflammatory indicators. More importantly, with the improvement in patient condition, the nCD64 index started to decline in the first week of anti‐TB therapy and significantly decreased at 4 weeks after treatment. Our study demonstrated that the CD64 assay is a rapid, non‐invasive and stable method for clinical application, and the nCD64 index can serve as a potential biomarker for the diagnosis and prognosis of TB.


| INTRODUC TI ON
Tuberculosis (TB), caused by Mycobacterium tuberculosis infection, is a major global health concern. There is a lack of specific biomarkers for the diagnosis and prognosis of patients with active TB, especially in patients with extrapulmonary TB (ePTB), whose lesions occur in the meninges, pleura, lymph nodes or bones, except the lungs. Hence, there is a need for novel methods or markers to diagnose pulmonary TB (PTB) and ePTB and to predict the outcome of M. tuberculosis infection.
CD64, one of the immunoglobulin (Ig)G Fc receptors (Fcγ RI), is constitutively expressed on monocytes and macrophages but hardly expressed on lymphocytes. 1 Under normal conditions, CD64 is expressed on resting neutrophils at a negligible level, but it increases by 10-fold at 4-6 hours after inflammatory response or stimulation by proinflammatory cytokines, such as granulocyte colony-stimulating factor (G-CSF) and interferon (IFN)-γ, and thus plays an important role in the immune response to infection. [2][3][4] Several studies have reported that neutrophil CD64 (nCD64) can be used as an early diagnostic biomarker in patients with sepsis, especially bacterial infection. [5][6][7][8][9][10][11] However, only a few studies have evaluated the diagnostic and prognostic utility of nCD64 in patients with M. tuberculosis infection, especially in patients with ePTB.
Therefore, our study was designed to address the above concerns. In the present study, we first analysed the differential expression of nCD64 in patients with PTB, ePTB, long-term anti-TB treatment (anti-TB) and healthy controls (HC); monocytes CD64 (mCD64) and lymphocytes CD64 (lymCD64) were used as the positive and negative controls, respectively. We then evaluated the clinical utility of the nCD64 index in the diagnosis of TB and finally evaluated serial changes in the nCD64 index during anti-TB therapy in patients with active TB.

| Participants and diagnosis
In this study, we recruited 329 individuals, comprising patients with other ePTB (n = 12); and the remaining received anti-TB therapy for more than one month at enrolment (n = 39). Patients with PTB and ePTB were followed up once a week from enrolment, and the flow chart is shown in Figure 1.
The diagnosis of TB was categorized as (a) confirmed diagnosis, with at least one of the following: acid-fast bacilli (AFB) smear or M. tuberculosis culture positive in any specimen (sputum, pleural effusion, cerebral spinal fluid or blood), tuberculous granuloma

| Measurement and calculation of the nCD64 index
Ethylenediaminetetraacetic acid (EDTA) anticoagulated peripheral blood samples (50 μL) were collected from the enrolled individuals, stored at room temperature (<24 hours) or 4°C (<48 hours) and then processed. 12 The monoclonal antibodies of CD45-PerCP, CD14-FITC and CD45-PE were purchased from BD Bioscience (San Jose, CA, USA). CD64 expression was measured by flow cytometry using FACSCanto Ⅱ (BD Bioscience, San Jose, CA, USA). The assay is simple and rapid, and it is based on whole-blood cell-surface staining.
The pan-leucocyte marker CD45 and monocyte marker CD14 were used to gate different WBC types according to their distribution in the forward side scatter (FSC) plot ( Figure 2C,D). Furthermore, the expression of CD64 on WBC is presented as median fluorescence  Given that CD64 is highly and stably expressed on the surface of monocytes and hardly expressed on lymphocytes, they can be used as the positive (mCD64 MFI) and negative (lymCD64 MFI) internal control, respectively ( Figure S1). The CD64 index was calculated using the following formula (

| Stability of the nCD64 index in stored blood samples
For the convenience of clinical application, sometimes specimens have to be stored when they cannot be processed in time. Therefore, it is important that the nCD64 index value is stable in stored samples. To test this issue, we randomly selected 10 blood samples from patients with high or low nCD64 expression (nCD64 index < 1, n = 2; 1 ≤ nCD64 index < 10, n = 3; 10 ≤ nCD64 index < 50, n = 2; nCD64 index ≥ 50, n = 3), stored these samples at room temperature and determined the nCD64 index at 0 hour and after 24, 48, 72 and 96 hours. instructions. The positive results were interpreted as previously described. 13

| Statistical analysis
Data were analysed using GraphPad Prism 8 (GraphPad, Inc., San Diego, CA, USA). Continuous variables were compared between independent groups using Mann-Whitney test (two groups) and Kruskal-Wallis test (multiple groups) followed by Dunn's post hoc test for multiple comparisons. Categorical variables were compared using the chi-square test or Fisher's exact test, as appropriate. The follow-up paired data were analysed using the paired Wilcoxon rank test. The receiver operating characteristics (ROC) curve analysis was performed to estimate the diagnostic performance of the nCD64 index and T-SPOT.TB assay for the diagnosis of TB. The association between common inflammatory markers (WBC count, neutrophil percentage, ESR, CRP, PCT and ferritin) and nCD64 index was assessed with Spearman correlation. The results with a P value of <.05 were considered significant.

| Individuals' characteristics
A total of 189 active patients with TB (PTB, ePTB and anti-TB) were enrolled in this study; 41 out of the 60 (68.33%) and 33 out of the 90 (36.67%) patients in the PTB and ePTB groups were confirmed diagnosis cases, respectively. The anti-TB therapy was administered for no more than two weeks in both PTB and ePTB groups at enrolment. The characteristics are summarized in Table 1. There were no significant differences in the clinical characteristics except for sex among the enrolled individuals; only 25% of the HC were men. Five out of the 189 (2.65%) patients died after enrolment.

| nCD64 index in patients with active M. tuberculosis infection
The nCD64 index was significantly higher in the active TB group  Figure 2F,G). In particular, the nCD64 index in the ePTB group was almost four times higher than that in the PTB group (P = .0110).
Furthermore, the nCD64 index was the highest in patients with disseminated TB (Table 1). There was no difference between the anti-TB and HC groups.  Figure 3B).

F I G U R E 3
The difference of nCD64 index between clinical and confirmed diagnosis in patients with (A) PTB or (B) ePTB. The boxes show medians and interquartile ranges, whereas the whiskers indicate the 10th and 90th percentiles. *P < .05. ePTB, extrapulmonary tuberculosis; ns, not significant; PTB, pulmonary tuberculosis

| Utility of the nCD64 index in the diagnosis of PTB
To estimate the performance of the nCD64 index in the diagnosis of PTB, we enrolled 49 patients with non-tuberculous PN, and among them, patients with severe PN were excluded (severe PN patients highly expressed nCD64, data were not shown) and collected their clinical characteristics (Table S1). As shown in  (Table 2).

| Utility of the nCD64 index in the diagnosis of TBM
To date, the differential diagnosis of CNS infection lacks specific biomarkers. We sought to understand whether the nCD64 index could be used to distinguish between TBM and non-tuberculous CNS infection (N-TBM). Therefore, we recruited 64 cases of N-TBM patients and collected their clinical documents (Table   S2) TB assay (Table 3).

| Dynamic determinations of the nCD64 index and prognosis in patients with active M. tuberculosis infection
We further divided all the patients into three groups based on the baseline nCD64 index values, as shown in Figure 6A,B. Among all TB patients, 44.67% had the nCD64 index of <10, in which 61.19% and 38.81% were from PTB and ePTB group, respectively. However, the patients with PTB accounted for less than one-fourth of those in the 10 ≤ nCD64 index < 50 (24.00%) and nCD64 index ≥ 50 groups (21.21%) ( Figure 6B). Furthermore, as we expected, the percentage of ePTB patients in the above two groups with high nCD64 index exceeded 75% ( Figure 6B). This suggested that high nCD64 index values might be associated with the dissemination of TB. We had data of at least one follow-up of 80 (53.33%) patients with active TB. According to the efficiency of treatment, the 80 patients were divided into improvement (n = 67) and poor improvement (n = 13, including 4 deaths) groups. Notably, 97.06%, 78.79% and 61.54% of the nCD64 index < 10, 10 ≤ nCD64 index < 50 and nCD64 index ≥ 50 patients, respectively, were included in the improvement group (P = .0077); moreover, the proportion of patients with high nCD64 index values significantly differed between the improvement and poor improvement groups ( Figure 6C,D). The nCD64 index value started to significantly decrease at 1 week of anti-TB treatment in the improvement group, reaching up to 10, and it evidently decreased to below 10 at 8 weeks of treatment (P < .0001, Figure 6E). However, in the poor improvement group, there were no significant differences in the nCD64 index at the baseline and 1, 2 and 3 week(s) (all P > .05, Figure 6F).

| Qualification of stability of the nCD64 index in stored blood samples
As shown in Figure 7A, we evaluated the stability of the nCD64 index based on the delta value and observed that the absolute delta value percentage was within 20% and the coefficient of variation (CV) was less than 10% when the blood samples were stored for <24 hours at room temperature ( Figure 7B). Thus, we recommend that the blood samples should not be stored for more than 48 hours at room temperature (delta value, 30%; CV, 15%), and they can be stored at 4°C for a short period when necessary.

| D ISCUSS I ON
In this study, we investigated the differential expression of nCD64 in patients with PTB and ePTB and observed that patients with M. tuberculosis infection highly expressed nCD64 when compared with the anti-TB patients and HC. Furthermore, the nCD64 index was considerably higher in patients with ePTB than in those with PTB, and more importantly, it was elevated up to 10-fold in patients with disseminated TB, indicating that there might be an association between the nCD64 index and TB dissemination. Our results showed there was a significant difference in sex among these groups, which might be due to the higher incidence of tuberculosis in men than in women, but it is reported that sex had almost no effect on the number of granulocytes, monocytes and CD64 expression 14 ; therefore, we thought the differences in sex  and HIV-negative participants, respectively. Therefore, it can be used to distinguish active TB from LTBI. 16 Additionally, the protein expression of CD64 on monocytes increased compared with that in PBMC samples in TB patients. 15 In contrast, another whole-blood  Our study had some limitations. First, the number of participants recruited in this study was relatively small, and larger number of individuals should be evaluated in the future. Second, we aimed to distinguish patients with PTB from those with non-tuberculous lung infection; thus, we did not estimate the utility of the nCD64 index for the differential diagnosis of PTB and LTBI. Third, we did not explore the CD64-associated mechanism involved in M. tuberculosis infection, and it should be explored in the future. Despite these limitations, our study presented robust data indicating that the nCD64 index could serve as a potential biomarker for the clinical diagnosis and prognosis of patients with TB.

TA B L E 2 Comparison of accuracy of T-SPOT.TB and nCD64 index for the diagnosis of pulmonary tuberculosis
In conclusion, our study demonstrated that the nCD64 index can be a potential biomarker to distinguish between PTB and PN patients and to identify patients with TBM from those with CNS infection, with the higher specificity and PPV, which were superior to those of the T-SPOT.TB assay. Moreover, this fast, non-invasive and stable method is convenient for clinical application. More importantly, the nCD64 index can reflect the responses to anti-TB treatment and serve as a therapeutic monitor.

ACK N OWLED G EM ENTS
We gratefully acknowledge the financial and technical support of

CO N FLI C T O F I NTE R E S T
All authors declare that there is no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data are available from the corresponding author upon reasonable request.