Immunological Diagnosis 179
II (BoLA) binding prediction algorithms, which
are not available to date to cover MHC diversity
in cattle compared to the human or murine sys-
tems. We have used, with moderate success, a
method to predict human HLA binding proteins
(ProPred) (Vordermeier et al., 2003), or a BoLA
DRB3 structure-based prediction method
( Hepitom) (Jones et al., 2011) to predict bovine
promiscuously recognized peptides (i.e. peptides
recognized in the context of multiple BoLA class
II alleles). However, the accuracy of either
method, particularly in respect to the specificity
of detecting promiscuously recognized peptides
was not optimal, requiring impractically large
peptide sets to cover the whole proteome (Jones
et al., 2011). Nevertheless, better prediction
methods are under development and it is highly
likely that such genome-wide mapping studies
can soon be undertaken to define the M. bovis
T-cell antigenome in cattle.
12.5.4. Predicting specificityWhile predicting antigenicity or immunogenic-
ity is a challenge, so is the prediction of antigens
that are specific for M. bovis. The absence of a
particular protein within a genome (e.g. anti-
gens encoded on gene regions deleted from the
BCG genome compared to M. bovis) does not
guarantee specificity as the areas of cross-
reactivity can be harboured within short
stretches (<10–20 amino acid residues long)
within an antigen that can be shared with other,
otherwise undeleted antigens (Cockle et al.,
2002). The actual amino acid differences
between a peptide being ‘specific’ or ‘cross-
reactive’ can be minute and unpredictable, as
was shown for peptides from highly homologous
PE or PPE proteins (Vordermeier et al., 2012).
However, not surprisingly, the fewer identical or
homologous amino acids located within an epit-
opic region of a peptide, the higher the chance
of it being specific (Vordermeier et al., 2012).
This highlights not only the possibility of select-
ing individual specific peptides out of otherwise
cross-reactive proteins (Jones et al., 2010), but
also the need for continued ‘wet’ experimental
immunological investigation of cross-reactivity
and specificity.
12.6 Biomarkers: Promising
Candidates and Alternative
Approaches12.6 Emerging biomarkers for CMI-based
assays beyond IFN-gNumerous biomarkers beyond IFN-γ are being
evaluated for use in the immunodiagnosis of
M. tuberculosis infection of humans (reviewed by
Chegou et al., 2014). As mentioned in the immu-
nopathogenesis section (section 12.3), several
cytokines and chemokines have been identified
as biomarkers of M. bovis infection as well as
surrogates of protection in response to vaccina-
tion in cattle. Of these, Th-17-associated cyto-
kines are attractive as multiple studies have
implicated cytokines from this T-cell subset in
both protective responses elicited by vaccination
and/or responses associated with lesion severity
after M. bovis infection in cattle (Vordermeier
et al., 2009; Blanco et al., 2011, 2013; Aranday-
Cortes et al., 2012; Bhuju et al., 2012; Rizzi et al.,
2012; Shu et al., 2014; Waters et al., 2015a).
Indeed, using RNA-seq followed by RT-qPCR, we
have recently demonstrated that Th17-related
cytokine genes (i.e. IL-17A, IL-17F, IL-22, IL-19,
and IL-27) follow similar kinetics and levels as
compared to IFN-γ in the response to BCG vacci-
nation and subsequent infection with virulent
M. bovis (Waters et al., 2015a). Using ELISA to
measure protein, the IL-17A and IFN-γ responses
were highly correlated and exhibited similar
diagnostic capacity. Also, reduced IL-17A
responses by BCG vaccinates (i.e. exhibiting sig-
nificant protection upon necropsy) at 2.5 weeks
after M. bovis challenge correlated with
reduced disease burden. Thus, measure of Th17-
associated cytokines may also be useful both as a
biomarker of infection and as a surrogate of
protection in the immune response to bovine TB.
Other biomarkers for measure of CMI
responses (e.g. IL-2, IP-10, IL-1β, TNF-α, and
nitric oxide) have also been evaluated for diag-
nostic applications in cattle, although only in
proof of principle studies (Waters et al., 2003,
2012; Jones et al., 2010a; Rhodes et al., 2014).
Rhodes et al. (2014) demonstrated that IL-2
responses to ESAT-6/CFP10 or PPD are detect-
able in whole blood assays after infection with
virulent M. bovis but not after vaccination with