Journal of Infection (2017) 75, 132e145

www.elsevierhealth.com/journals/jinf

Mycobacteria-specific cytokine responses as

correlates of treatment response in active
and latent tuberculosis
Vanessa Clifford a,b, Marc Tebruegge a,b,c, Christel Zufferey b,
Susie Germano b, Ben Forbes b, Lucy Cosentino e,
Emma McBryde d, Damon Eisen d, Roy Robins-Browne b,e,
Alan Street d, Justin Denholm d,e,f, Nigel Curtis a,b,*

a
Department of Paediatrics, The University of Melbourne, Parkville, VIC, Australia
b
Murdoch Children’s Research Institute, Royal Children’s Hospital Melbourne, Parkville, VIC, Australia
c
Academic Unit of Clinical and Experimental Medicine, Faculty of Medicine & Respiratory Biomedical
Research Unit & Global Health Research Institute, University of Southampton, United Kingdom
d
Victorian Infectious Diseases Service, Royal Melbourne Hospital, Parkville, VIC, Australia
e
Department of Microbiology and Immunology, University of Melbourne, VIC, Australia
f
Victorian Tuberculosis Program, Peter Doherty Institute, Parkville, VIC, Australia

Accepted 5 April 2017

Available online 5 May 2017

KEYWORDS Summary Objectives: A biomarker indicating successful tuberculosis (TB) therapy would
Tuberculosis; assist in determining appropriate length of treatment. This study aimed to determine changes
Cytokines; in mycobacteria-specific antigen-induced cytokine biomarkers in patients receiving therapy for
Biomarkers; latent or active TB, to identify biomarkers potentially correlating with treatment success.
Latent TB; Methods: A total of 33 adults with active TB and 36 with latent TB were followed longitudinally
Active TB; over therapy. Whole blood stimulation assays using mycobacteria-specific antigens (CFP-10,
Monitoring ESAT-6, PPD) were done on samples obtained at 0, 1, 3, 6 and 9 months. Cytokine responses
(IFN-g, IL-1ra, IL-2, IL-10, IL-13, IP-10, MIP-1b, and TNF-a) in supernatants were measured
by Luminex xMAP immunoassay.
Results: In active TB cases, median IL-1ra (with CFP-10 and with PPD stimulation), IP-10 (CFP-
10, ESAT-6), MIP-1b (ESAT-6, PPD), and TNF-a (ESAT-6) responses declined significantly over the
course of therapy. In latent TB cases, median IL-1ra (CFP-10, ESAT-6, PPD), IL-2 (CFP-10, ESAT-
6), and IP-10 (CFP-10, ESAT-6) responses declined significantly.

* Corresponding author. Department of Paediatrics, The University of Melbourne, Royal Children’s Hospital Melbourne, Flemington Road,
Parkville, VIC 3052, Australia. Fax: þ61 3 9345 4751.
E-mail address: nigel.curtis@rch.org.au (N. Curtis).

http://dx.doi.org/10.1016/j.jinf.2017.04.011
0163-4453/ª 2017 The British Infection Association. Published by Elsevier Ltd. All rights reserved.
Cytokine biomarkers for monitoring TB treatment
Conclusions: Mycobacteria-specific cytokine responses change significantly over the course of
therapy, and their kinetics in active TB differ from those observed in latent TB. In particular,
mycobacteria-specific IL-1ra responses are potential correlates of successful therapy in both
active and latent TB.
ª 2017 The British Infection Association. Published by Elsevier Ltd. All rights reserved.
Introduction
The new World Health Organization (WHO) targets for
reducing tuberculosis (TB) incidence and prevalence rely
heavily on identifying cases of active TB and providing
effective treatment.1 There is also increasing emphasis on
identifying and treating latent TB infection (LTBI) in high-
risk individuals.2 A key component of both these strategies
is ensuring adequate treatment. At present there is no reli-
able biomarker of treatment success.3e5
In active TB, therapeutic response is usually judged
clinically, in conjunction with laboratory measures such
as conversion of sputum smears from positive to negative,
correlating with reduction in mycobacterial load. Polymerase
chain reaction (PCR) based methods, including the Xpert MTB/
RIF assay that is being rolled out globally, cannot provide
any useful information about treatment response, as these
detect both viable and non-viable mycobacteria.6 A biomarker
that predicts treatment response and can guide decisions on
length of treatment would be a significant advance in
combating the ongoing TB pandemic. Several studies have
shown differences in mycobacteria-specific cytokine
responses between patients with active TB and individuals
with LTBI at presentation.7e10 In addition, data from a small
number of studies indicate that monitoring cytokine
responses, including TNF-a, IP-10, IL-2 and IL-10 responses,
may have some predictive ability in patients with active
TB.11e14 However, most of these studies were small, had other
significant limitations, and reported conflicting results.3
In LTBI, a biomarker of treatment success would also be
of great benefit. At present, there is no test to confirm
Mycobacterium tuberculosis (Mtb) eradication at comple-
tion of LTBI treatment. While interferon-g (IFN-g) release
assays (IGRAs) are useful for establishing the diagnosis of
LTBI, longitudinal studies have shown that IGRAs are not
suitable for monitoring treatment response.15,16 Only a mi-
nority of patients with an initially positive categorical IGRA
result convert to a negative result after completing LTBI
treatment.15,16 Furthermore, quantitative changes in IFN-
g responses in IGRAs following treatment are also inconsis-
tent. Finally, tuberculin skin tests (TSTs) are also not useful
for monitoring treatment response, as most patients remain
TST positive for several years after completion of LTBI
treatment.17
An important potential application of a biomarker of TB
treatment success is the ability to individualise duration of
therapy. In some patients, treatment could potentially be
shortened, whilst in patients at risk of treatment failure,
therapy could be prolonged. Individualised treatment has
the potential to both reduce associated costs substantially
and decrease the risk of drug-related adverse events.18
Our study aimed to determine changes in Mtb antigen-
induced cytokine biomarker responses in patients receiving
133
treatment for LTBI or active TB to identify potential
biomarkers correlating with treatment success.
Methods
Participants
Adult participants starting treatment for suspected LTBI or
active TB were recruited following informed consent. The
study was conducted at the Royal Melbourne Hospital (Victo-
ria, Australia), a large tertiary referral centre, over a 3-year
period (March 2012 to August 2014). These patients were part
of a larger cohort recruited into our study of diagnostic
biomarkers for TB.19 Patients receiving immunosuppressive
medication or with a primary or secondary immunodeficiency
were excluded, as were patients who had received anti-
tuberculous therapy for more than one week.
We recorded data on demographic and clinical charac-
teristics on a standardised case report form. Demographic
details included country and date of birth, TB exposure
history, BCG immunisation status and clinical symptoms. In
addition, past medical history, current medications, phys-
ical examination and radiological results were recorded.
Tuberculin skin tests and interferon-gamma release
assays
All participants had a TST (5 Tuberculin Units PPD; Tubersol,
Sanofi Pasteur, Toronto, Canada) done by a trained health-
care professional, with the exception of participants
who had microbiologically-confirmed active TB at recruit-
ment. A positive TST result was defined as induration
10 mm at 48e72 h. In addition, blood samples were
collected in sodium heparin tubes for whole blood stimula-
tion assays and for QuantiFERON-TB Gold-in-Tube (QFT-GIT)
assays. QFT-GIT assays were processed in a fully accredited
diagnostic laboratory, following the manufacturer’s
instructions.
Categorisation of participants and follow-up
Based on TST, QFT-GIT and microbiological results,
participants were classified into the following well-
defined diagnostic groups: Group A e active TB: defined
as microbiologically-proven TB based on culture and/or
PCR results, Group B e LTBI: defined as asymptomatic
patients with a positive TST, a positive IGRA result and
a normal chest X-ray.
Participants who did not fulfil the criteria for these two
distinct diagnostic categories were excluded from further
analyses. Participants with LTBI were treated with isoniazid
for 9 months. Participants with active TB were treated
134 V. Clifford et al.

with quadruple combination anti-mycobacterial therapy
as per WHO recommendations and local management
guidelines.20
Participants in Groups A and B were followed up over the
course of treatment at 1, 3, 6 and 9 months, with blood
collected for whole blood stimulation assays and QFT-GIT
assays at each time point.
Whole blood assays and cytokine analysis
Whole blood was stimulated with the Mtb-specific antigens
ESAT-6 (10 mg/ml; JPT Peptide Technologies, Berlin Ger-
many) and CFP-10 (10 mg/ml; JPT), and PPD (20 mg/ml;
RT50, Statens Serum Institut, Copenhagen, Denmark) or
left unstimulated (negative control). All stimulation assays
were done in the presence of co-stimulatory antibodies
anti-CD28 and anti-CD49 (both BD Biosciences, San Jose,
CA, USA). All samples were then incubated at 37  C for 19 h.
Supernatants were then harvested and immediately cry-
opreserved at 80  C for later analysis.
Cytokines were measured in batched analyses using Bio-
rad human cytokine kits (Biorad, Gladesville, Australia) us-
ing an xMAP Luminex 200 Bioanalyser, following the
manufacturer’s instructions, as described in our previous
report.9 Based on our previous experience that the concen-
trations of certain cytokines exceed the dynamic detection
range of the Biorad assays,9 each sample was analysed using
two separate plates: (i) undiluted samples for IFN-g, IL-1ra,
IL-2, IL-10, IL-13, and TNF-a, and (ii) 1:20 diluted samples
for IP-10 and MIP-1b.
confirmed active TB and 36 with confirmed LTBI. Forty-two
patients (19/33 (58%) with active TB and 23/36 (64%) with
LTBI) had blood tests obtained at the end of therapy at
either 6 or 9 months.
Both active TB and LTBI patients were predominantly of
Asian extraction, with a median age of 26 and 28 years,
respectively. The majority of participants were BCG-
vaccinated (Table 1).
Active TB
Of the patients with active TB, 23 had pulmonary disease
and 10 had extra-pulmonary disease. All patients with
active TB responded clinically to treatment, with full
resolution of symptoms. All 16 patients with pulmonary
TB who were initially sputum-smear positive converted to
negative sputum-smear results.
There were no significant differences in cytokine re-
sponses at baseline between patients with pulmonary TB
and extra-pulmonary TB, except for PPD-stimulated TNF-a
responses, which were significantly higher in patients with
extra-pulmonary TB (data not shown).
Latent TB infection
All patients with LTBI were treated with isoniazid for a 9-
month-period with documented compliance, and none
developed symptoms of active TB over the study period.

Table 1 Demographic characteristics of participants

Statistical analysis
included in the final cohort.
All cytokine concentrations were background-corrected Active TB LTBI
prior to analysis by subtracting the concentration measured (nZ33) (nZ36)
in the negative control sample. Cytokine responses be- Age e median (range) 28 (21e59) 26 (18e38)
tween diagnostic groups and over time were compared with in years
non-parametric tests: Mann Whitney U tests for two-group
Male 15 (45%) 17 (47%)
unpaired comparisons and Wilcoxon signed rank test for
two-group paired comparisons. Statistical analyses were BCG vaccinated
done using Prism V5 (GraphPad Software Inc; La Jolla, CA, Yes 20 (61%) 27 (75%)
USA) and Stata V14 (StatCorp, College Station, TX, USA). Unknown 8 (24%) 6 (17%)
No 5 (15%) 3 (8%)
Ethical approval Born overseas 31 (94%) 35 (97%)
Region of birth
The study was approved by the Melbourne Health Human Africa 5 (15%) 5 (14%)
Research Ethics Committee (HREC approval number Asia 24 (73%) 27 (75%)
2011.128), and conducted in accordance with Good Clinical Australasia 3 (9%) 2 (6%)
Practice and the principles of the Declaration of Helsinki. Europe 0 1 (3%)
North America 1 (3%) 1 (3%)
Results Time since arrival in 6 (0e30) 4 (0.5e30)
Australia e median
Patient population (range) in yearsa
Pulmonary TB 23 (70%) e
A total of 120 patients were recruited to the study (55 with Extra-pulmonary TB 10 (30%) e
suspected active TB, 65 with suspected LTBI). Of these, 38 Lymph node 6 (18%)
and 43 respectively, fulfilled the study criteria for active TB Osteoarticular 2 (6%)
and LTBI. The 69 of these patients who had at least two Intra-abdominal 2 (6%)
blood samples taken over the study period were included in a
Includes only individuals born overseas.
the final analysis, namely 33 patients with microbiologically-
 Cytokine biomarkers for monitoring TB treatment
Table 2A Median cytokine concentrations (and interquartile ranges) in pg/mL measured at 0, 1, 3, 6 and 9 months after start of treatment in participants with active TB
infection. The median concentration at each time point was compared to the median concentration at the start of treatment (0 months) using Wilcoxon signed rank tests.
Active TB
Number 0 1 mo 3 mo 6 mo 9 mo 0 vs 1 mo 0 vs 3 mo 0 vs 6 mo 0 vs 9 mo
(nZ33) (nZ26) (nZ24) (nZ16) (nZ13) Wilcoxon signed rank test p-value
IFN-g
CFP-10 402.3 154.9 107.4 46.2 85.6 0.10 0.01 0.17 0.31
[102.3,1325] [28.5,887.3] [17.2,600.3] [6.6,514.3] [0,534.2]
ESAT-6 116.5 105.2 41.8 55.5 26.6 0.10 0.05 0.08 0.02
[30.0,731.7] [19.3,727.1] [6.7,371.4] [16.1,132.9] [13.6,178.5]
PPD 3797 5415 5495 4062 4148 0.06 0.07 0.07 0.89
[1976,6395] [2707,13 562] [2692,1092] [1705,11 333] [1583,11 813]
IL-1ra
CFP-10 291.6 40.6 16.3 108.2 173.3 0.11 0.11 <0.01 0.05
[135.6,193.0] [366.3,193] [157.4,303.2] [452.6,57.7] [717.9,27.2]
ESAT-6 100.2 60.7 0.0 81.3 28.4 0.67 0.89 0.11 0.17
[188.8,641.0] [239.0,591.8] [128.1,99.4] [329.6,203.9] [1284,223]
PPD 2166 2001 1481 1197 1418 0.33 0.06 0.02 0.31
[1205,3945] [894,2422] [1028,2434] [722,2778] [775,2182]
IL-2
CFP-10 127.7 120.1 63.87 14.80 87.99 0.65 0.07 0.24 0.89
[34.6,319.3] [24.9,235.5] [10.8,288.1] [6.9,201.1] [5.7,204.4]
ESAT-6 59.7 88.1 25.7 28.8 14.5 0.49 0.07 0.31 0.91
[18.1,174.1] [15.8,238.0] [3.3,149.9] [3.4,77.6] [1.8,128.4]
PPD 719 1511 1138 819 872 0.01 0.03 0.06 0.03
[427,1587] [774,2797] [571,2309] [527,2167] [446,2527]
IL-10
CFP-10 2.2 8.0 8.9 6.4 10.5 0.02 0.12 0.98 0.74
[14.6,0.8] [26.5,1.7] [25.5,2.4] [29.3,3.3] [28.8,6.0]
ESAT-6 2.6 3.8 9.4 7.9 10.7 0.15 0.29 0.82 0.79
[15.1,0.6] [21.1,0.5] [20.5,1.9] [26.6,3.0] [17.8,5.4]
PPD 49.9 45.1 60.1 85.5 115.9 0.13 0.77 0.66 0.02
[25.7,81.7] [21.1,117.2] [35.3,165.1] [32.3,169.9] [43.5,186.2]
IL-13
CFP-10 3.4 0.5 1.0 0.04 0.1 0.72 0.004 0.34 0.35
[0.70,10.6] [0,9.7] [0.2,3.3] [0,7.9] [0.2,4.1]
ESAT-6 1.0 0.9 0.5 0.2 0.0 0.92 0.87 0.49 0.05
[0,2.5] [0,8.6] [0.2,2.6] [0,3.4] [1.2,1.8]
(continued on next page)

135
 136
Table 2A (continued )
Active TB
Number 0 1 mo 3 mo 6 mo 9 mo 0 vs 1 mo 0 vs 3 mo 0 vs 6 mo 0 vs 9 mo
(nZ33) (nZ26) (nZ24) (nZ16) (nZ13) Wilcoxon signed rank test p-value
PPD 60.6 89.5 99.3 145.1 99.2 0.05 0.04 <0.01 0.06
[20.2,123.5] [42.6,231.3] [50.4,161.6] [48.0,276.6] [65.1,229.9]
TNF-a
CFP-10 60.7 42.3 23.1 26.0 15.2 0.68 0.27 0.94 0.95
[20.9,281.2] [4.9,158.3] [12.4,83.2] [1.4,57.5] [13.0,166.7]
ESAT-6 33.0 20.6 15.3 9.7 24.3 0.49 0.03 0.49 0.45
[8.1,144.8] [0,160.1] [0,56.3] [3.2,34.2] [1.3,66.2]
PPD 1164 1723 1303 1071 2206 0.02 0.09 0.22 0.43
[714,1777] [611,4521] [630,3623] [425,4700] [569,3902]
IP-10
CFP-10 86 608 72 211 78 515 20 912 19 930 0.88 0.32 0.02 0.07
[53 635,182 618] [23 845,201 940] [12 719,146 075] [1462,99 814] [2973,10 159]
ESAT-6 44 611 60 541 27 738 25 423 20 852 0.80 0.09 0.10 0.03
[13 647,132 177] [15 974,156 824] [3476,73 217] [11 685,54 061] [9698,55 434]
PPD 65 078 60 849 64 807 77 206 43 078 0.78 0.23 0.45 0.59
[24 415,138 771] [35 469,128 710] [38 721,138 309] [32 753,149 418] [11 104,121 099]
MIP-1b
CFP-10 4885 3231 2554 1572 275.2 0.13 <0.01 0.31 0.11
[2049,9563] [566.4,7048] [853.6,4907] [1083,6764] [3489,5089]
ESAT-6 2766 2356 1568 2987 1255 0.19 0.10 0.90 0.59
[1271,7686] [367,8521] [986,5005] [88,5954] [2301,7676]
PPD 77 302 92 387 90 048 86 572 82 693 0.43 0.28 0.62 0.11
[28 347,127 667] [51 384,139 803] [40 604,154 383] [41 044,189 525] [44 321,188 686]

V. Clifford et al.
 Cytokine biomarkers for monitoring TB treatment
Table 2B Median cytokine concentrations (and interquartile ranges) in pg/mL measured at 0, 1, 3, 6 and 9 months after start of treatment in participants with latent TB
infection. The median concentration at each time point was compared to the median concentration at the start of treatment (0 months) using Wilcoxon signed rank tests.
LTBI
Number 0 1 mo 3 mo 6 mo 9 mo 0 vs 1 mo 0 vs 3 mo 0 vs 6 mo 0 vs 9 mo
(nZ36) (nZ24) (nZ28) (nZ19) (nZ8) Wilcoxon signed rank test p-value
IFN-g
CFP-10 151.4 239.8 167.0 198.9 328.6 0.40 0.40 0.98 0.65
[24.1,553.3] [24.1,730.6] [44.2,641.2] [23.5,420.7] [25.5,735.6]
ESAT-6 95.3 354.5 111.0 186.8 76.8 0.30 0.81 0.67 0.82
[17.9,350.4] [41.0,789.0] [29.8,503.4] [20.7,352.2] [0.0,786.7]
PPD 2604 5624 2925 4372 4043 0.62 0.28 0.09 0.84
[1817,12 880] [3226,11 608] [1074,5548] [1909,5762] [1109,6178]
IL-1ra
CFP-10 30.8 65.2 18.1 24.4 326.6 0.90 0.03 0.02 0.05
[124.2,337.4] [56.6,815.9] [310.7,392.5] [105.1,38.7] [742.6,130.4]
ESAT-6 148.9 164.0 126.2 39.1 191.3 0.72 0.08 0.01 0.05
[10.6,461.6] [3.8,959.1] [7.5,483.4] [27.5,371.6] [620.7,105.2]
PPD 1822 2011 1200 1170 1097 0.52 0.49 <0.01 0.02
[986,2849] [1030,3430] [729,2837] [217,1812] [398,1557]
IL-2
CFP-10 91.9 80.3 65.9 48.4 68.4 0.51 0.19 0.05 0.30
[21.6,160.2] [17.9,223.2] [18.8,228.8] [8.6,129.9] [17.0,179.3]
ESAT-6 52.0 104.2 57.9 47.4 33.6 0.72 0.38 0.02 0.30
[12.5,184.3] [10.8,236.7] [9.5,203.3] [2.4,132.8] [13.7,331.1]
PPD 799 1070 773 767 1019 0.68 0.26 0.15 1.00
[527,1262] [542,1756] [411,1219] [312,1485] [454,1304]
IL-10
CFP-10 0.6 0.0 1.5 1.5 24.6 0.62 0.40 0.01 0.04
[2.9,0.1] [4.7,0.0] [10.8,0.2] [16.9,0.4] [34.2,9.4]
ESAT-6 0.3 0.6 0.9 1.9 18.6 0.04 0.50 0.01 0.13
[2.1,0.4] [6.7,0.0] [4.3,0.1] [10.4,0] [38.1,4.0]
PPD 24.8 24.7 23.4 20.1 147.5 0.83 0.39 0.82 0.04
[13.9,46.9] [11.7,43.0] [12.3,102.3] [11.6,67.7] [68.3,221.2]
IL-13
CFP-10 0.5 0.7 0.4 0.0 0.6 0.37 0.38 0.08 0.43
[0.0,3.0] [0,1.6] [0.1,3.3] [0.8,0.3] [1.6,1.2]
ESAT-6 0.7 1.2 1.4 0.1 1.0 0.82 0.23 0.24 1.00
[0.0,2.8] [0,4.1] [0.1,4.5] [0.7,4.3] [0.9,4.9]
PPD 56.8 57.1 66.2 70.6 37.7 0.66 0.91 0.27 1.00
[32.1,116.4] [28.7,162.1] [28.7,115.0] [36.6,154.8] [18.4,190.5]

137
(continued on next page)
 138
Table 2B (continued )
LTBI
Number 0 1 mo 3 mo 6 mo 9 mo 0 vs 1 mo 0 vs 3 mo 0 vs 6 mo 0 vs 9 mo
(nZ36) (nZ24) (nZ28) (nZ19) (nZ8) Wilcoxon signed rank test p-value
TNF-a
CFP-10 16.2 16.1 12.1 13.8 14.8 0.20 0.39 0.86 0.43
[1.3,52.8] [2.7,119.3] [0.4,54.7] [0.0,96.0] [1.1,371.8]
ESAT-6 14.5 47.6 32.7 26.6 69.9 0.27 0.76 0.86 0.25
[2.9,48.1] [7.0,152.6] [7.4,173.4] [0.0,185.8] [44.3,616.4]
PPD 527 749 566 595 1178 0.70 0.23 0.24 0.11
[179,1859] [220,2495] [174,1253] [115,1199] [579,2093]
IP-10
CFP-10 63 276 55 199 44 236 32 624 21 558 0.42 0.73 0.70 0.01
[23 160,122 204] [22 829,134 048] [19 092,189 776] [10 696,176 619] [12 956,45 512]
ESAT-6 48 084 61 115 35 641 23 499 7671 0.48 0.46 0.89 0.01
[14 737,115 683] [16 911,165 086] [3688,137 190] [344,104 136] [375,56 327]
PPD 75 131 99 418 63 304 64 672 38 269 0.92 0.29 0.73 0.74
[35 129,108 301] [45 186,133 170] [27 798,138 159] [39 650,131 087] [11 995,159 625]
MIP-1b
CFP-10 2891 3559 1695 1623 1190 0.06 0.26 0.46 0.36
[861,5702] [1201,9584] [633,5013] [55,9767] [1373,4605]
ESAT-6 4164 5060 3007 3613 2142 0.74 0.67 0.67 0.65
[1524,6308] [2242,9437] [727,8269] [971,9986] [3533,10 870]
PPD 55 807 55 963 45 693 36 295 54 265 0.16 0.37 0.05 0.95
[31 677,99 073] [30 767,96 662] [25 416,76 406] [31 078,54 250] [32 617,95 894]

V. Clifford et al.
Cytokine biomarkers for monitoring TB treatment 139

Figure 1 Box-plot with Tukey whiskers showing Mtb antigen-induced cytokine responses after CFP-10 (A), ESAT-6 (B) and PPD
(C) stimulation at baseline and at 1, 3, 6 and 9 months after starting treatment in patients with microbiologically-confirmed active
TB. The horizontal lines represent the medians. Statistically significant p-value results are shown.
140
Figure 1
Stimulant responses
Cytokine responses at 0, 1, 3, 6 and 9 months of treatment in
both active TB and LTBI patients are detailed in Table 2. In
general, higher cytokine responses were seen after PPD
stimulation, compared to ESAT-6 or CFP-10 stimulation
(Table 2). CFP-10 responses were higher than ESAT-6 re-
sponses in active TB patients (Table 2A), but responses to
CFP-10 and ESAT-6 were similar in LTBI patients. ESAT-6 stim-
ulation induced higher levels of IL-1ra response than CFP-10
in LTBI patients (Table 2B). We observed only very small IL-10
and IL-13 responses to CFP-10 or ESAT-6 stimulation.
Cytokine responses in participants with active TB
CFP-10 stimulation
We observed a reduction in median cytokine responses over
the study period for IFN-g, IL-1ra, IP-10 and MIP-1b
(Table 2A & Fig. 1A). Compared to baseline (0 months),
changes were statistically significant for IL-1ra at 6 and 9
months, for IFN-g at 3 months, for MIP-1b at 3 months
and for IP-10 at 6 months. Median IL-2 and TNF-a responses
also declined over the study period, although these changes
were not statistically significant.
ESAT-6 stimulation
We observed a reduction in median cytokine responses for
IFN-g, IP-10 and TNF-a over the study period (Table 2A and
Fig. 1B). Changes in cytokine responses were statistically
significant for IFN-g at 9 months, TNF-a at 3 months and
for IP-10 at 9 months. Median IL-1ra and IL-2 responses
V. Clifford et al.
(continued)
also declined over the study period, although comparisons
did not reach statistical significance.
PPD stimulation
Median cytokine responses decreased over the study period
for IL-1ra, which was statistically significant at 6 months
(Table 2A and Fig. 1C). Compared to baseline, there was a
statistically significant increase in IL-2, IL-13 and TNF-a re-
sponses at 1 month. Furthermore, compared to baseline, IL-
10 responses and IL-13 responses were significantly higher
at 6 months.
Cytokine responses in participants with LTBI
CFP-10 stimulation
There was a trend towards a reduction in median cytokine
responses over the study period for IL-1ra, IL-2, IP-10 and
MIP-1b (Table 2B and Fig. 2A). Changes were statistically
significant for IL-1ra at 3, 6 and 9 months, for IL-2 at 6
months, and for IP-10 at 9 months.
ESAT-6 stimulation
A reduction in median cytokine responses over the study
period was observed for IL-1ra and IP-10, with statistically
significant changes for IL-1ra at 6 and 9 months and for IP-
10 at 9 months (Table 2B and Fig. 2B).
PPD stimulation
We observed a reduction in median cytokine responses over
the study period for IL-1ra and MIP-1b (Table 2B and
Fig. 2C). Differences were statistically significant for
Cytokine biomarkers for monitoring TB treatment 141

Figure 2 Box-plot with Tukey whiskers showing Mtb antigen-induced cytokine responses after CFP-10 (A), ESAT-6 (B) and PPD
(C) stimulation at baseline and at 1, 3, 6 and 9 months after starting treatment in patients with latent TB infection. The horizontal
lines represent the medians. Statistically significant p-value results are shown.
142
Figure 2
IL-1ra at 6 and 9 months and for MIP-1b at 6 months.
Compared to baseline, there was also a statistically signif-
icant increase in IL-10 responses at 9 months.
IL-1ra (interleukin-1 receptor antagonist)
responses in active TB and LTBI
We found statistically significant reductions in IL-1ra re-
sponses in patients with active TB at 6 months after both
PPD stimulation and CFP-10 stimulation (Table 2A/B and
Fig. 3A/B). A decline in median concentration after ESAT-
6 stimulation was also observed over the study period,
but this did not reach statistical significance.
Most active TB patients had detectable IL-1ra responses
at baseline: 32/33 patients had a response to PPD stimu-
lation, 24/33 patients to CFP-10, and 21/33 patients to
ESAT-6 stimulation. There were no differences in the
demographic characteristics, disease type or IGRA result
between those who did and did not have IL-1ra responses to
ESAT-6 or CFP-10, although median IFN-g response in the
QFT-GIT assay were lower in non-responders.
In four patients with active TB who had an increase in
IL-1ra response over the study period to at least one Mtb-
specific antigen, there were no distinctive clinical features.
In LTBI patients, a statistically significant decline in IL-1ra
responses over therapy was observed at 6 and 9 months with all
3 stimulants (CFP-10, ESAT-6, PPD). When results for patients
with samples collected at baseline and 6 months were
analysed, 18/19 patients had a decrease in IL-1ra responses.
One patient had increased PPD-stimulated IL-1ra responses
over the study period; the patient was fully compliant with
LTBI therapy and did not develop symptoms of active TB.
V. Clifford et al.
(continued)
Most LTBI patients had detectable IL-1ra responses at
baseline: 33/36 had a response to PPD stimulation, 27/36
to ESAT-6 stimulation and 23/36 to CFP-10 stimulation.
There were no differences in the demographic character-
istics, disease type, TST or IGRA results between those
who did and did not have IL-1ra responses to ESAT-6 or
CFP-10.
Discussion
Our study provides novel data on the longitudinal kinetics
of cytokine biomarkers in patients treated for active TB and
LTBI. We found that mycobacteria-specific cytokine re-
sponses change significantly over time, and that their
kinetics in patients with active TB differ from those
observed in patients with LTBI. In addition, we identified
mycobacteria-specific IL-1ra responses as a potential corre-
late of successful treatment.
Recent studies, including own, have shown that
mycobacteria-specific IL-1ra responses are promising TB
biomarkers that have the ability to discriminate between
LTBI and active TB.8,9,21,22 IL-1ra, a cytokine produced by
macrophages and monocytes, inhibits the action of IL-1, a
pro-inflammatory cytokine, by competitively blocking
IL-1RI receptors.23 IL-1ra secretion is thought to be crucial
for granuloma formation, an important mechanism for the
containment of Mtb.24 The critical role of a balanced
IL-1ra/IL-1 environment for TB control is highlighted by
the observation that patients receiving monoclonal IL-1ra
antibodies for rheumatological conditions have a signifi-
cantly increased risk of TB re-activation.25,26
Cytokine biomarkers for monitoring TB treatment 143

Figure 3 A) Mtb antigen-induced IL-1ra responses after CFP-10, ESAT-6 and PPD stimulation in the 16 individuals with active TB
who had samples measured at both 0 and 6 months (end of treatment). B) Mtb antigen-induced IL-1ra responses after CFP-10, ESAT-
6 and PPD stimulation in the 19 individuals with LTBI who had samples measured at both 0 and 6 months.

A cross-sectional study by Juffermans et al. found that
IL-1ra concentrations in unstimulated serum samples were
markedly elevated in patients with active TB at the time of
diagnosis, and that their IL-1ra concentrations were signif-
icantly lower after completion of therapy.27 A subsequent
study made similar observations, and additionally, reported
that plasma IL-1ra concentrations had a tendency to be
higher in patients with delayed response to therapy than
in those who responded well.28 Both studies provide evi-
dence that IL-1ra concentrations correlate with disease ac-
tivity in TB. However, as discussed in detail elsewhere, the
use of unstimulated serum samples for diagnosis is problem-
atic, as without preceding antigenic stimulation, raised
cytokine concentrations simply reflect inflammation, rather
than being specific for infection with a particular
pathogen.19
Our study shows that whole blood assays using
mycobacterial antigens, including the Mtb-specific pep-
tides ESAT-6 and CFP-10, can elicit mycobacteria-specific
IL-1ra responses. Remarkably, 97% of patients with active
TB and 92% of patients with LTBI had detectable IL-1ra
responses at baseline when samples were stimulated with
PPD. As expected, overall IL-1ra responses following
stimulation with the single peptide antigens ESAT-6 and
CFP-10 were considerably lower than those observed with
PPD, which contains more than 200 mycobacterial pep-
tides. In patients with active TB, IL-1ra responses after 6
months of treatment were considerably lower than at
baseline, irrespective of the antigenic stimulant used
(reaching statistical significance in CFP-10 and PPD stim-
ulated samples). In LTBI patients, compared to baseline,
there was a statistically significant reduction in IL-1ra
responses both at 6 and 9 months with all three
stimulants.
Importantly, the observed decrease in IL-1ra responses
is highly likely the result of a reduction in microbial
antigenic burden and resulting immunological changes,
rather than being the result of a direct effect of anti-
tuberculous antibiotics on anti-mycobacterial immune
responses. By use of an ex vivo model, we have previously
144
shown that although certain anti-mycobacterial antibiotics
have anti-inflammatory properties, responses to stimula-
tion with mycobacterial antigens are unaltered in their
presence.29
A small number of previous studies have suggested that
TNF-a might be a useful biomarker of response to
treatment in active TB.3,11,12,30,31 The majority of these
studies were small and results were conflicting, with
some studies reporting a decrease in TNF-a responses dur-
ing or after treatment compared to baseline, while others
reported no significant change.3 A study by Eum et al., in
which samples from adult TB patients were stimulated
with CFP-10 for 72 h, even reported an increase in TNF-
a responses.11 In our study, where samples were stimu-
lated for a shorter period of time, we observed a trend to-
wards lower ESAT-6- and CFP-10-induced TNF-a responses
over time in patients with active TB. However, no such
trend was observed in LTBI patients, a finding that is
consistent with several studies that have found that
TNF-a responses are lower in LTBI than active TB before
treatment.9,32,33
In our study, we found a statistically significant reduc-
tion in IFN-g responses after CFP-10 and ESAT-6 stimulation
in active TB, but not in LTBI patients. This finding accords
with the conclusions of our recent systematic review on
commercial IGRAs for monitoring anti-tuberculous therapy.
The review found that, whilst there was a statistically
significant reduction in IFN-g levels after therapy, the
degree of individual variation in results means that IGRAs
are not useful for monitoring anti-tuberculous treatment in
clinical practice.15
We found a significant increase in PPD-stimulated IL-10
responses 9 months after initiation of treatment in both
active and LTBI patients. Compared to baseline, at 9
months median IL-10 responses were 2-fold and 6-fold
higher in active TB and LTBI patients, respectively. This
finding is consistent with the study by Eum et al., who
found an increase in CFP-10 induced IL-10 responses (after
24 h of stimulation) during therapy in active TB patients.11
Various lines of evidence show that IL-10, which predomi-
nately has anti-inflammatory properties, is a key cytokine
in anti-mycobacterial immune responses in both animals
and humans.34
A key strength of our study was the careful selection of
patients with a definitive diagnosis and exclusion of cases
with diagnostic uncertainty (i.e. possible and probable
active TB). To our knowledge, it is also one of the first to
quantify changes in antigen-stimulated cytokine responses
(other than IFN-g) over the course of therapy in patients
with LTBI. A limitation of our study is that, for logistic
reasons, we were unable to obtain blood samples from all
patients at every time point. Despite this, the study was
sufficiently large to detect statistically significant changes
in cytokine responses over time.
In conclusion, we have identified several mycobacteria-
specific cytokine biomarkers that have the potential to be
exploited in future immunodiagnostic tests designed to
define successful anti-tuberculous therapy, of which IL-1ra
responses are the most promising candidates. Further
longitudinal studies are needed to confirm the role of these
biomarkers in different patient groups, including children
and patients with drug-resistant TB.
V. Clifford et al.
Funding
This study was supported by a grant from the John Burge
Trust. Vanessa Clifford was supported by an Australian Gov-
ernment Research Training Scholarship.
Conflicts of interest and financial disclosure
The authors do not have a commercial association that
might pose a conflict of interest.
Acknowledgements
The authors thank Ms Snezana Spillane and Ms Vi Dang
(Murdoch Childrens Research Institute) for their help with
the cytokine stimulation assays and harvesting of superna-
tants, Ms Elizabeth Matchett for assistance with blood
sample collection and Ms Paulette Manton (Victorian Infec-
tious Diseases Department of the Royal Melbourne Hospital)
for administrative assistance.
Appendix A. Supplementary data
Supplementary data related to this article can be found at
http://dx.doi.org/10.1016/j.jinf.2017.04.011.
References
1. Stop TB Partnership. Global plan to end TB: 2016-2020. United
Nations; 2015.
2. Wejse C. Tuberculosis elimination in the post Millennium Devel-
opment Goals era. Int J Infect Dis 2015;32:152e5.
3. Clifford V, Zufferey C, Street A, Denholm J, Tebruegge M,
Curtis N. Cytokines for monitoring anti-tuberculous therapy:
a systematic review. Tuberculosis (Edinb) 2015;95:217e28.
4. Chegou NN, Heyckendorf J, Walzl G, Lange C, Ruhwald M.
Beyond the IFN-gamma horizon: biomarkers for immunodiag-
nosis of infection with Mycobacterium tuberculosis. Eur Respir
J 2014;43:1472e86.
5. Tebruegge M, Ritz N, Curtis N, Shingadia D. Diagnostic tests for
childhood tuberculosis: past imperfect, present tense and
future perfect? Pediatr Infect Dis J 2015;34:1014e9.
6. Lawn SD, Mwaba P, Bates M, Piatek A, Alexander H, Marais BJ,
et al. Advances in tuberculosis diagnostics: the Xpert MTB/RIF
assay and future prospects for a point-of-care test. Lancet
Infect Dis 2013;13:349e61.
7. Won EJ, Choi JH, Cho YN, Jin HM, Kee HJ, Park YW, et al. Bio-
markers for discrimination between latent tuberculosis infec-
tion and active tuberculosis disease. J Infect 2017;74:
281e93. http://dx.doi.org/10.1016/j.jinf.2016.11.010.
8. Suzukawa M, Akashi S, Nagai H, Nagase H, Nakamura H,
Matsui H, et al. Combined analysis of IFN-gamma, IL-2, IL-5,
IL-10, IL-1RA and MCP-1 in QFT supernatant is useful for distin-
guishing active tuberculosis from latent infection. PLoS One
2016;11. e0152483.
9. Tebruegge M, Dutta B, Donath S, Ritz N, Forbes B, Camacho-
Badilla K, et al. Mycobacteria-specific cytokine responses detect
tuberculosis infection and distinguish latent from
active tuberculosis. Am J Respir Crit Care Med 2015;192:485e99.
10. Frahm M, Goswami ND, Owzar K, Hecker E, Mosher A,
Cadogan E, et al. Discriminating between latent and active
tuberculosis with multiple biomarker responses. Tuberculosis
2011;91:250e6.
Cytokine biomarkers for monitoring TB treatment
11. Eum S-Y, Lee Y-J, Min J-H, Kwak H-K, Hong M-S, Kong J-H, et al.
Association of antigen-stimulated release of tumor necrosis
factor-alpha in whole blood with response to chemotherapy
in patients with pulmonary multidrug-resistant tuberculosis.
Respiration 2010;80:275e84.
12. Bertholet S, Horne DJ, Laughlin EM, Savlov M, Tucakovic I,
Coler RN, et al. Effect of chemotherapy on whole-blood cyto-
kine responses to Mycobacterium tuberculosis antigens in a
small cohort of patients with pulmonary tuberculosis. Clin Vac-
cine Immunol 2011;18:1378e86.
13. Harari A, Rozot V, Enders FB, Perreau M, Stalder JM, Nicod LP,
et al. Dominant TNF-þ Mycobacterium tuberculosis-specific
CD4þ T cell responses discriminate between latent infection
and active disease. Nat Med 2011;17:372e6.
14. Djoba Siawaya JF, Beyers N, van Helden P, Walzl G. Differential
cytokine secretion and early treatment response in patients
with pulmonary tuberculosis. Clin Exp Immunol 2009;156:
69e77. http://dx.doi.org/10.1111/j.365-2249.009.03875.x.
Epub 2009 Jan 22.
15. Clifford V, He Y, Zufferey C, Connell T, Curtis N. Interferon
gamma release assays for monitoring the response to treat-
ment for tuberculosis: a systematic review. Tuberculosis (Ed-
inb) 2015;95:639e50.
16. Chiappini E, Fossi F, Bonsignori F, Sollai S, Galli L, de Martino M.
Utility of interferon- release assay results to monitor anti-
tubercular treatment in adults and children. Clin Ther 2012;
34:1041e8.
17. Havlir DV, van der Kuyp F, Duffy E, Marshall R, Hom D, Ellner JJ.
A 19-year follow-up of tuberculin reactors. Assessment of skin
test reactivity and in vitro lymphocyte responses. Chest 1991;
99:1172e6.
18. Heyckendorf J, Olaru ID, Ruhwald M, Lange C. Getting personal
perspectives on individualized treatment duration in
multidrug-resistant and extensively drug-resistant tubercu-
losis. Am J Respir Crit Care Med 2014;190:374e83.
19. Clifford V, Tebruegge M, Zufferey C, Germano S, Denholm J,
Street A, et al. Serum IP-10 in the diagnosis of latent and active
tuberculosis. J Infect 2015;71:696e8.
20. Department of Health & Human Services. Management, control
and prevention of tuberculosis: guidelines for health care pro-
viders. 2015. Melbourne, Victoria.
21. Ruhwald M, Bjerregaard-Andersen M, Rabna P, Eugen-Olsen J,
Ravn P. IP-10, MCP-1, MCP-2, MCP-3, and IL-1RA hold promise
as biomarkers for infection with M. tuberculosis in a whole
blood based T-cell assay. BMC Res Notes 2009;2:19.
22. Anbarasu D, Raja CP, Raja A. Multiplex analysis of cytokine-
s/chemokines as biomarkers that differentiate healthy con-
tacts from tuberculosis patients in high endemic settings.
Cytokine 2013;61:747e54. http://dx.doi.org/10.1016/j.
cyto.2012.12.031. Epub 3 Feb 8.
145
23. Dinarello CA. Interleukin-1 in the pathogenesis and treatment
of inflammatory diseases. Blood 2011;117:3720e32.
24. Ruth JH, Bienkowski M, Warmington KS, Lincoln PM, Kunkel SL,
Chensue SW. IL-1 receptor antagonist (IL-1ra) expression, func-
tion, and cytokine-mediated regulation during mycobacterial
and schistosomal antigen-elicited granuloma formation. J Im-
munol 1996;156:2503e9.
25. Brassard P, Kezouh A, Suissa S. Antirheumatic drugs and the
risk of tuberculosis. Clin Infect Dis 2006;43:717e22.
26. Settas LD, Tsimirikas G, Vosvotekas G, Triantafyllidou E,
Nicolaides P. Reactivation of pulmonary tuberculosis in a pa-
tient with rheumatoid arthritis during treatment with IL-1 re-
ceptor antagonists (anakinra). J Clin Rheumatol 2007;13:
219e20.
27. Juffermans NP, Verbon A, van Deventer SJ, van Deutekom H,
Speelman P, van der Poll T. Tumor necrosis factor and
interleukin-1 inhibitors as markers of disease activity of tuber-
culosis. Am J Respir Crit Care Med 1998;157:1328e31.
28. Lee JH, Chang JH. Changes of plasma interleukin-1 receptor
antagonist, interleukin-8 and other serologic markers during
chemotherapy in patients with active pulmonary tuberculosis.
Korean J Intern Med 2003;18:138e45.
29. Clifford V, Zufferey C, Germano S, Ryan N, Leslie D, Street A,
et al. The impact of anti-tuberculous antibiotics and cortico-
steroids on cytokine production in QuantiFERON-TB gold in
tube assays. Tuberculosis (Edinb) 2015;95:343e9.
30. Petruccioli E, Petrone L, Vanini V, Sampaolesi A, Gualano G,
Girardi E, et al. IFN/TNFalpha specific-cells and effector mem-
ory phenotype associate with active tuberculosis. J Infect
2013;66:475e86.
31. Kim CH, Choi KJ, Yoo SS, Lee SY, Won DI, Lim JO, et al. Compar-
ative analysis of whole-blood interferon-gamma and flow cytom-
etry assays for detecting post-treatment immune responses in
patients with active tuberculosis. Cytom B Clin Cytom 2014;
86:236e43.
32. Wang F, Hou H, Xu L, Jane M, Peng J, Lu Y, et al. Mycobacte-
rium tuberculosis-specific TNF-alpha is a potential biomarker
for the rapid diagnosis of active tuberculosis disease in Chinese
population. PLoS One 2013;8. e79431.
33. Sutherland JS, de Jong BC, Jeffries DJ, Adetifa IM, Ota MO. Pro-
duction of TNF-alpha, IL-12(p40) and IL-17 can discriminate be-
tween active TB disease and latent infection in a West African
cohort. PLoS One 2010;5. e12365.
34. Dorhoi A, Kaufmann SH. Perspectives on host adaptation in
response to Mycobacterium tuberculosis: modulation of inflam-
mation. Semin Immunol 2014;26:533e42.