180 R. Waters and M. Vordermeier
BCG, thereby affording an ability to differentiate
infected from non-infected or BCG-vaccinated
animals (Rhodes et al., 2014). In a systematic
meta-analysis on the diagnosis of latent
M. tuberculosis infection of humans, it was deter-
mined that use of IL-2 release assays provides
some benefit, particularly when combined with
IGRAs (Mamishi et al., 2014). Soluble IL-2
receptor-alpha is also released in PBMC culture
supernatants after mycobacterial antigen stimu-
lation in M. bovis-infected cattle, and may have
potential as a biomarker of infection (Nualláin
et al., 1997).
There is also considerable interest in use of
IP-10 (IFN-γ-induced protein 10 or CXCL10) as
a diagnostic biomarker of M. tuberculosis infec-
tion in humans as measured in sera, urine, or
antigen-stimulated cultures (Chegou et al.,
2014; Tonby et al., 2015). Similarly, IP-10 has
shown promise as a biomarker of M. bovis infec-
tion in cattle (Waters et al., 2012) and African
buffaloes (Syncerus caffer) (Goosen et al., 2015).
One potential concern with use of IP-10 as a
specific marker of infection is that this chemo-
kine is often produced in large quantities as a
result of inflammation or due to various infec-
tions resulting in high levels of IP-10 in sera;
thus, while IP-10 has shown promise for diag-
nostic purposes and to monitor antimicrobial
therapy with M. tuberculosis infection in humans
(Tonby et al., 2015), it may also be falsely ele-
vated in response to other infections or inflam-
matory conditions (Waters et al., 2012; Clifford
et al., 2015), thereby confounding interpretation
of agent-specific assays using IP-10 as a readout.
Antibodies for detection of numerous bovine
cytokines and chemokines are increasingly
being developed and becoming readily available
through multiple commercial companies. Thus,
further studies will be warranted to evaluate the
clinical and diagnostic potential of emerging
immune markers as related to bovine TB.
Multi-parameter readouts of CMI responses
have also shown diagnostic potential with both
bovine and human TB (Jones et al., 2010a;
Tebruegge et al., 2015). Using a multi-parameter
approach, Tebruegge et al. (2015) demonstrated
that combinations of TNF-α/IL-1Rα and TNF-α/
IL-10 correctly classified latent versus active TB
in 95.5% and 100% of cases in children, respec-
tively. While numerous cytokines were evalu-
ated and there was considerable overlap between
treatment groups for the majority of cytokines,
the use of IP-10, TNF-α and IL-2 achieved a
high level of accuracy in the distinction between
TB-infected versus non-infected humans in this
study (Tebruegge et al., 2015). With bovine TB,
Jones et al. (2010a) demonstrated that the com-
bined use of IL-1β, TNF-α, and IFN-γ in response
to ESAT-6/CFP10 increased the sensitivity of the
assay by 11% as compared to use of IFN-γ alone;
however, there was a concomitant decrease in
specificity by 14% (Jones et al., 2010a). Thus,
balancing gains in sensitivity must be balanced
by the potential for loss in specificity when using
multi-parameter approaches. Interestingly, in
the Jones et al. (2010a) study, applying only
IFN-γ and IL-1β in parallel increased the sensi-
tivity of the assay by 5% without a loss in speci-
ficity. Multi-parameter approaches may be
particularly useful for ‘test-and-removal’ appli-
cations in which identification of all infected
animals is a priority over specificity of the assay,
as use of multiple parameters may increase the
odds over detecting a response to a single param-
eter. However, the overall costs of multiple
parameter test applications will have to be
balanced against their benefit in control
programmes. A potential pitfall of the multi-
parameter approach is that variables such as co-
infection with other pathogens (e.g.; parasites
[Flynn et al., 2009]), age of the animal (i.e. non-
specific production of cytokine by NK cells in
young animals [Olsen et al., 2005]), and stage of
the disease may be amplified by use of multiple
versus single readout tests.12.6.2 Whole blood assays and
‘in-tube’ strategiesWhole blood assays for the detection of IFN-γ
responses to TB have been used in cattle for over
25 years (i.e. Bovigam [Rothel et al., 1990]) and
humans for almost 20 years (i.e. Quantiferon
[Streeton et al., 1998]). A limitation of the initial
Quantiferon assay for humans was the require-
ment for shipment of blood samples to a labora-
tory for transfer of the blood to vessels (i.e. tubes,
microtiter plates, etc.) for antigen stimulation.
Thus, there was a requirement for easily acces-
sible satellite laboratories to process the sample
in a timely manner. The ‘in-tube’ approach for