The diagnosis of tuberculosis remains challenging in individuals with difficulty in providing good quality sputum samples such as children. Host biosignatures of inflammatory markers may be valuable in such cases, especially if they are based on more easily obtainable samples such as saliva. To explore the potential of saliva as an alternative sample in tuberculosis diagnostic/biomarker investigations, we evaluated the levels of 33 host markers in saliva samples from individuals presenting with pulmonary tuberculosis symptoms and compared them to those obtained in serum. Of the 38 individuals included in the study, tuberculosis disease was confirmed in 11 (28.9%) by sputum culture. In both the tuberculosis cases and noncases, the levels of most markers were above the minimum detectable limit in both sample types, but there was no consistent pattern regarding the ratio of markers in serum/saliva. Fractalkine, IL-17, IL-6, IL-9, MIP-1
Tuberculosis (TB) remains a global health problem. An estimated 8.7 million new cases and 1.4 million deaths resulted from the disease in 2011 [
Saliva is primarily secreted through the parotid, submandibular, and sublingual glands. It is composed of 98% water and contains other substances including electrolytes, mucus, antibacterial compounds, and various enzymes [
There has recently been an interest in exploring saliva for potentially useful inflammatory biomarkers [
Individuals suspected of having pulmonary TB disease were recruited from the Fisantekraal community in the outskirts of Cape Town, South Africa, as part of the ongoing EDCTP-funded African European Tuberculosis Consortium (AE-TBC) study (
All participants presented at the rural health care facility with symptoms suggestive of pulmonary TB disease. Briefly, participants presented with persistent cough lasting ≥2 weeks and at least one of the following: fever, malaise, recent weight loss, night sweats, knowledge of close contact with a TB patient, haemoptysis, chest pain, or loss of appetite. Participants were eligible for the study if they were 18 years or older, willing to give written informed consent for participation in the study, including for HIV testing. Patients were excluded from the study if they had not been residing in the study area for more than 3 months, were severely anaemic (HB < 10 g/L), were on anti-TB treatment, had received anti-TB treatment in the previous 90 days, or were on quinolone or aminoglycoside antibiotics in the past 60 days. At enrolment, a case report form was completed for each participant before blood and saliva samples along with other samples, including urine and sputum as required for the main study, were collected. The study was approved by the Health Research Ethics Committee of the University of Stellenbosch (Reference no. N10/08/274) and the City of Cape Town.
Blood was collected into 4 mL plain BD vacutainer tubes (BD Biosciences) and transported at ambient conditions to the laboratory. The tubes were then centrifuged at 2500 rpm for 10 minutes after which serum was harvested, aliquoted, and frozen (−80°C) until use. Saliva was collected from all participants into salivette tubes (Sarstedt, Nümbrecht, Germany), according to the instructions of the manufacturer. Saliva samples were then transported on ice (4°C) to the laboratory after which the salivette tubes were centrifuged for 2 minutes (1000 g) and the saliva was harvested, aliquoted into labeled tubes, and kept at −80°C until analysis.
Sputum samples collected from all participants were cultured by the MGIT method (BD Biosciences). Positive MGIT samples were examined for AFB using the Ziehl-Neelsen method, to check for contamination, after which PCR experiments were performed to confirm the isolation of
The levels of 33 host markers, including interferon (IFN)-
Differences in analyte levels between the TB patients and participants without TB disease or between the marker levels detected in saliva and serum levels were evaluated by the Mann-Whitney
Of the 38 participants included in this study, 27 (71%) were females. The mean age of all study participants was 38.0 ± 10.2. Of the 28 study participants with available Quantiferon In-Tube results, 67.9% were positive using the manufacturer’s recommended cut-off (≥0.35 IU/mL). Eight (21%) of the study participants were HIV infected (Table
Demographic and clinical characteristics of study participants.
All | TB cases | Non-TB case | |
| |||
Number of participants | 38 | 11 | 27 |
Mean age |
|
|
|
Males/females | 11/27 | 3/8 | 8/19 |
HIV positive |
8 (21.0) | 2 (18.2) | 6 (22.2) |
Quantiferon Positive |
19/28 | 5/6 | 14/22 |
We evaluated the levels of host markers above the minimum detectable concentration (MDC; obtained from the manufacturer’s package insert), in saliva and serum, and then compared the levels of the markers detected in saliva to those obtained in serum. The levels of five of the 33 markers evaluated (MMP-9, IP-10, MIP-1
Proportion of study participants with host markers above the minimum detectable concentration (MDC) and differences between saliva and serum.
Marker | MDC (pg/mL) | Saliva | Serum |
|
||
---|---|---|---|---|---|---|
% >MDC | Median (IQR) | % >MDC | Median (IQR) | |||
(A) Host markers more abundantly expressed in saliva | ||||||
IL-1 |
1.5 | 100.0 | 4618.9 (1956.3–10000.0) | 21.0 | 0.0 (0.0-0.0) | <0.0001 |
IL-1 |
0.7 | 95.0 | 24.6 (12.23–54.7) | 18.0 | 0.0 (0.0–0.2) | <0.0001 |
IL-2 | 0.4 | 97.0 | 6.5 (2.3–14.4) | 32.0 | 0.0 (0.0–0.6) | <0.0001 |
IL-5 | 0.1 | 32.0 | 0.0 (0.0–1.1) | 0.0 | 0.0 (0.0-0.0) | <0.0001 |
IL-7 | 1.0 | 45.0 | 0.0 (0.0–19.0) | 16.0 | 0.0 (0.0-0.0) | <0.0001 |
IL-8 | 0.3 | 100.0 | 145.2 (78.6–237.3) | 97.0 | 13.6 (6.2–27.7) | <0.0001 |
IL-10 | 0.3 | 39.5 | 0.0 (0.0–16.8) | 2.6 | 0.0 (0.0-0.0) | <0.0001 |
IL-12p70 | 0.9 | 89.0 | 9.8 (3.7–16.9) | 16.0 | 0.0 (0.0–0.3) | <0.0001 |
IL-13 | 0.3 | 92.0 | 20.7 (11.4–34.1) | 0.0 | 0.0 (0.0-0.0) | <0.0001 |
IL-15 | 0.6 | 45.0 | 0.0 (0.0–8.4) | 5.0 | 0.0 (0.0-0.0) | <0.0001 |
IL-17 | 0.4 | 97.0 | 13.0 (8.6–18.9) | 16.0 | 0.0 (0.0-0.0) | <0.0001 |
IFN- |
0.4 | 82.0 | 4.1 (0.6–10.4) | 42.0 | 0.0 (0.0–5.0) | <0.0001 |
G-CSF | 3.9 | 100.0 | 1348.0 (842.0–2263.2) | 97.0 | 90.7 (45.3–114.4) | <0.0001 |
GM-CSF | 2.3 | 100.0 | 100.5 (65.3–137.9) | 8.0 | 0.0 (0.0-0.0) | <0.0001 |
TGF- |
1.4 | 100.0 | 9.5 (6.7–16.6) | 92.0 | 6.9 (3.2–20.8) | 0.08 |
EGF | 5.3 | 100.0 | 5717.0 (3991.9–7964.4) | 97.0 | 98.6 (45.5–185.1) |
<0.0001 |
VEGF | 10.1 | 100.0 | 618.2 (457.3–802.6) | 95.0 | 303.5 (145.7–493.2) | <0.0001 |
Fractalkine | 7.6 | 97.0 | 451.8 (137.9–699.9) | 10.0 | 0.0 (0.0-0.0) | <0.0001 |
MMP-9 | 1.0 | 100.0 | 164631.4 (105484.3–348292.9) | 100.0 | 2673.0 (1795.9–3951.6) | <0.0001 |
| ||||||
(B) Markers more abundantly expressed in serum | ||||||
sIL-2R |
7.5 | 8.0 | 0.0 (0.0-0.0) | 29.0 | 0.0 (0.0–10.4) | 0.01 |
GRO | 11.4 | 97.0 | 132.5 (74.1–204.0) | 100.0 | 1209.1 (856.0–2099.6) | <0.0001 |
IP-10 | 1.3 | 100.0 | 102.6 (67.4–213.3) | 100.0 | 408.0 (307.4–710.0) | <0.0001 |
MIP-1 |
3.2 | 100.0 | 12.0 (8.4–17.0) | 100.0 | 47.7 (22.2–81.3) |
<0.0001 |
MCP-1 | 1.2 | 100.0 | 124.5 (29.5–204.0) | 100.0 | 473.4 (314.3–644.6) |
<0.0001 |
CRP | 0.0012* | 71.0 | 88.2 (0.0–232.0) | 100.0 | 27668.5 (9213.7–127253.2) | <0.0001 |
SAA | 0.21* | 50.0 | 119.5 (0.0–848.8) | 97.0 | 11408.9 (2519.1–95050.4) | <0.0001 |
SAP | 0.055* | 21.0 | 0.0 (0.0-0.0) | 100.0 | 46954.9 (37567.1–60894.8) | <0.0001 |
MMP-2 | 48 | 13.0 | 0.0 (0.0-0.0) | 100.0 | 1148.8 (971.1–1333.0) | <0.0001 |
sCD40L | 5.2 | 100.0 | 353.8 (166.9–779.2) | 100.0 | *705911 (307862–1039000) | <0.0001 |
| ||||||
(C) No difference in expression levels between serum and saliva | ||||||
IL-4 | 0.6 | 0.0 | 0.0 (0.0-0.0) | 0.0 | 0.0 (0.0-0.0) | — |
IL-6 | 0.4 | 45.0 | 0.0 (0.0–37.3) | 32.0 | 0.0 (0.0–11.5) | 0.12 |
IL-9 | 1.1 | 15.8 | 0.0 (0.0-0.0) | 5.3 | 0.0 (0.0-0.0) | 0.14 |
TNF- |
0.2 | 87.0 | 10.6 (6.9–20.6) | 95.0 | 11.1 (6.5–13.0) | 0.23 |
Median levels of biomarkers detected in saliva and serum samples from all study participants (
There were, on average, 6-fold higher levels of IFN-
Levels of host markers detected in saliva and serum samples from all study participants (
Levels of host markers detected in saliva and serum samples from all study participants (
When the levels of markers detected in saliva of TB cases were compared to the levels obtained in the saliva of the noncases, 8 of the 33 markers (IL-6, CRP, IL-9, IL-5, MIP-1
Utility of host markers detected in saliva in the diagnosis of TB disease.
Marker | TB disease | No TB disease |
|
AUC (95% CI) | Cut-off value | Sensitivity % (95% CI) | Specificity % (95% CI) |
---|---|---|---|---|---|---|---|
IL-6 | 37.3 (0.0–52.2) | 0.0 (0.0–13.2) | 0.019 | 0.72 (0.54–0.91) | >25.8 | 63.6 (30.8–89.0) | 81.5 (61.9–93.7) |
CRP | 246.5 (22.0–353.9) | 45.9 (0.0–122.0) | 0.024 | 0.74 (0.53–0.94) | >271.7 | 45.5 (16.8–76.6) | 92.6 (75.7–99.0) |
IL-9 | 0.0 (0.0–11.0) | 0.0 (0.0-0.0) | 0.027 | 0.65 (0.44–0.86) | >10.3 | 27.3 (6.0–60.9) | 96.3 (81.0–99.9) |
IL-5 | 0.9 (0.0–9.2) | 0.0 (0.0-0.0) | 0.033 | 0.68 (0.48–0.88) | >7.8 | 27.3 (6.0–61.0) | 96.3 (81.0–99.9) |
MIP-1 |
17.0 (11.3–22.2) | 11.3 (8.4–15.6) | 0.039 | 0.72 (0.54–0.90) | >18.7 | 45.5 (16.8–76.6) | 92.6 (75.7–99.1) |
Fractalkine | 772.9 (225.8–1148.3) | 338.2 (104.3–565.5) | 0.041 | 0.72 (0.52–0.91) | >912.2 | 36.4 (10.9–69.2) | 96.3 (81.0–99.9) |
IL-17 | 18.9 (7.6–37.0) | 12.6 (8.6–16.6) | 0.085 | 0.68 (0.46–0.90) | >29.0 | 45.5 (16.8–76.6) | 96.3 (81.0–99.9) |
VEGF | 457.3 (307.7–754.9) | 680.0 (512.4–802.6) | 0.085 | 0.68 (0.47–0.90) | <370.5 | 45.5 (16.8–76.6) | 92.6 (75.7–99.1) |
Median levels and interquartile ranges (in parenthesis) of the markers and abilities to discriminate between pulmonary TB cases (
Levels of markers detected in the saliva samples of pulmonary TB cases and individuals without TB disease and receiver operator characteristics (ROC) plots showing the accuracy of these markers in the diagnosis of TB disease. Error bars in the scatter-dot plots indicate the median analyte levels. Only markers for which the area under the ROC curve (AUC) was ≥0.70 are shown.
When the data obtained from saliva were fitted into general discriminant analysis (GDA) models, optimal prediction of TB or no TB disease was achieved when markers were used in combinations of five. A combination of IL-5, IL-6, IL-15, TNF-
Frequency of analytes in the top 20 GDA predictive models that most accurately classified study participants as TB disease or no TB. The columns represent the number of times each analyte was included in the top 20 discriminatory models. (a) Frequency of analytes in the models generated from the host markers detected in saliva, (b) frequency of analytes in models generated when the data obtained from saliva were combined with the data obtained from serum samples.
When serum marker levels obtained in TB cases were compared to the levels obtained in the noncases, significant differences were obtained for four markers (IL-6, IL-2, SAP, and SAA). The levels of IL-6, IL-2, and SAP were significantly higher in the TB cases (
Utility of host markers detected in serum in the diagnosis of TB disease.
Marker | TB cases | Non-TB cases |
|
AUC (95% CI) | Cut-off value | Sensitivity % (95% CI) | Specificity % (95% CI) |
---|---|---|---|---|---|---|---|
IL-6 | 11.5 (0.0–28.1) | 0.0 (0.0-0.0) | 0.01 | 0.73 (0.54–0.92) | >27.54 | 27.3 (6.0–61.0) | 96.3 (81.0–99.9) |
IL-2 | 0.6 (0.0–1.3) | 0.0 (0.0-0.0) | 0.01 | 0.73 (0.53–0.92) | >0.95 | 45.5 (16.8–76.6) | 92.6 (75.7–99.1) |
SAP | 60894.8 (45137.4–65623.2) | 42251.4 (36985.9–53804.8) | 0.03 | 0.72 (0.54–0.90) | >58914 | 54.6 (23.4–83.3) | 85.2 (66.3–95.8) |
SAA | 239.1 (0.0–848.8) | 6133.8 (2012.1–40070.2) | 0.05 | 0.70 (0.52–0.88) | >941894 | 18.2 (2.3–51.8) | 96.3 (81.0–99.9) |
Median levels (pg/mL) and interquartile ranges (in parenthesis) of the markers and abilities to discriminate between pulmonary TB cases (
When the diagnostic accuracy of the markers detected in serum was investigated by ROC curve analysis, the AUC for all four markers that showed significant differences (IL-2, IL-6, SAP, and SAP) was ≥0.70 (Table
Levels of markers detected in the serum samples of pulmonary TB cases and individuals without TB disease and receiver operator characteristics (ROC) plots showing the accuracy of these markers in the diagnosis of TB disease. Error bars in the scatter-dot plots indicate the median analyte levels. Only markers for which the area under the ROC curve (AUC) was ≥0.70 are shown.
When the data obtained from serum samples was fitted into GDA models, the prediction accuracy of the 5-marker serum analytes tended to be poorer than that obtained for the models generated on saliva data. A combination of IL-6, IL-12p70, G-CSF, MMP-9, and MIP-1
We evaluated the levels of 33 host immunological biomarkers in saliva and serum samples from individuals suspected of having pulmonary TB disease. The main finding of our study was the dissimilar expression of host markers that were detected in both sample types, with up to 6-fold higher levels of some markers expressed in saliva. We have therefore shown that saliva, a relatively easy-to-obtain sample type, may be a very valuable sample in TB biomarker discovery investigations. Interestingly, some of the markers detected in saliva including IL-6, CRP, MIP-1
All the markers evaluated in this study are inflammatory markers including cytokines, chemokines, growth factors, acute phase proteins, and matrix metalloproteinases that have been widely investigated and shown to play diverse roles in the pathogenesis of different diseases including TB. Of note, the levels of IL-5, IL-6, IL-9, IL-17, MIP-1
IL-2 (reviewed in [
One of the most noticeable observations of our study was the marked differences in the levels of the markers detected in saliva and serum and that higher levels of most of the markers were obtained in saliva. Most of the markers that were more abundantly expressed in serum were chemokines (IP-10, MIP-1
Although investigations on saliva in the TB biomarker field are limited, saliva has been widely investigated in other diseases, notably in leukaemia, oral cancer, oral lichen planus, and periodontitis amongst others [
This study was done as a pilot for a larger, on-going study and, as such, was limited by the small sample size and the absence of individuals with other lung infections. However, our observations will serve as proof-of-concept for more research in the field given the fact that diagnostic tests based on easily obtainable samples like saliva would revolutionize the diagnosis of TB disease, especially if such markers are incorporated into lateral flow devices. We did not investigate the influence of factors such as food or drink in-take prior to sampling on the levels of the salivary biomarkers. The potential influence of such factors may require investigation in further studies.
In conclusion, the data presented in this study indicates that there are many differences in the levels of host markers expressed in saliva in comparison to those of serum and some of the markers detected in both sample types have potential in the diagnosis of TB disease. Our findings indicate that saliva might be a better alternative to serum in TB biomarker discovery investigations. Our findings warrant further investigation in larger studies.
The authors declare that there is no conflict of interests regarding the publication of this paper.
The authors are grateful to the study participants, our study nurses especially Mrs Shirley Mcanda, and members of the SUN-Immunology Research Group for their various roles in ensuring that the data needed for this paper were available. This work was funded by the EDCTP through the African European Tuberculosis Consortium (AE-TBC, Grant no. IP_2009_32040) and the Trials of Excellence in Southern Africa (TESA, Project code CG_cb_07_41700), Principal Investigator: Professor Gerhard Walzl. NN Chegou was supported by Postdoctoral fellowships from The Claude Leon Foundation and the South African MRC.