Bovine tuberculosis

Mycobacterium bovis Molecular Typing and Surveillance 65

multiple animals from the same cattle herd, with
distinct types seen in separate herds, consistent
with the independent evolution of regional vari-
ants (Whipple et al., 1997). Molecular types in
many wildlife species were indistinguishable
from those in captive elk populations, indicating
an epidemiological link (Whipple et al., 1997).
More recently, spoligotyping data have con-
firmed that most sources of M. bovis infection in
humans and cattle in California were Mexican
cattle (Rodwell et al., 2010). A discriminating
MLVA test was configured and showed remark-
able congruence with epidemiological data
when tested on selected USA isolates (Martinez
et al., 2008). Molecular typing data were
exploited to assess the extent of, and risk factors
for, M. bovis infections in humans in the USA
2006–2013 (Scott et al., 2016).
Early application of molecular epidemiol-
ogy to M. bovis in South America used the RFLP
technique (Fisanotti, 1998) and suggested that
regional variants were associated with Argen-
tina, Paraguay and Brazil (Fisanotti, 1998).
Spoligo- and RFLP typing confirmed these origi-
nal observations (Zumarraga et al., 1999). Spoli-
gotyping, RFLP and MLVA have subsequently
been applied to human and cattle populations in
Argentina, confirming that the most common
molecular type in cattle was also observed in 2%
of human pulmonary tuberculosis cases
(Etchechoury et al., 2010). Several human cases
exhibited identical molecular types that were
distinct from cattle isolates, perhaps indicative of
human to human transmission (Etchechoury
et al., 2010). More recently, MLVA and IS 6110
RFLP typing confirmed M. bovis/M. avium hom-
inissuis co-infection in pigs (Barandiaran et al.,
2014). In Brazil, an optimized MLVA panel
alongside spoligotyping increased the discrimi-
nation of M. bovis (Parreiras et al., 2012), and
has been used to infer multiple introductions
into single herds (Figueiredo et al., 2012).

5.5.4 Oceania

In Australia, REA and RFLP produced informa-
tive molecular epidemiological data in the early
1990s (Cousins et al., 1993). PFGE was used to
differentiate strains from different regions and to
infer transmission within different hosts and

between farms (Feizabadi et al., 1996). However,

Australia’s declaration of Officially Bovine

Tuberculosis Free (Lamoureux et al., 2012)

status in December 1997 (More et al., 2015) pre-

dated much of the more recent development of

molecular epidemiological tools.

Des Collins from New Zealand is quite

rightly recognized as having pioneered the

molecular epidemiology of M. bovis (Collins and

De Lisle, 1984), initially with the REA tech-

nique. Molecular epidemiology has remained a

valued and cost-effective tool in the country’s

bovine TB eradication scheme (Livingstone et al.,

2015a, b), helping to direct potentially costly

control interventions by determining whether

new breakdowns are more likely the result of

transmission from domestic livestock or wildlife

(Ryan et al., 2006; Price-Carter et al., 2011;

Buddle et al., 2015).

5.6 Local Findings from Northern

Ireland and Lessons Learned

Here we discuss findings from Northern Ireland

in a little more detail. While M. bovis molecular

typing provides useful information in the out-

break investigation setting, its optimal use is

likely to be in describing population-level epide-

miological effects (cattle movements, cattle–

cattle transmission, wildlife–cattle transmission,

relative virulence, etc.).

Northern Ireland has implemented herd-

level MLVA surveillance since 2003 (Skuce et al.,

2010) and animal-level MLVA surveillance since

2010. Significantly, the spatial distribution of
      M. bovis MLVA types sampled was not random;
      highly significant geographical localization of
      M. bovis genotypes was clearly evident and sug-
      gested that sources tended to be local and that
      TB was a locally driven epidemic. Each MLVA
      type could be considered as responsible for its
      own micro-epidemic and the observed geo-
      graphical localization implies that the epidemic
      comprises a group of local micro-epidemics
      (Smith et al., 2003).
      MLVA types in purchased cattle were
      slightly more dispersed (less clustered) than
      those from home-bred cattle. However, on
      occasions MLVA types were clearly translocated
      significant distances from their normal