THE SEAL TUBERCULOSIS AGENT, MYCOBACTERIUM PINNIPEDII, INFECTS DOMESTIC CATTLE IN NEW ZEALAND: EPIDEMIOLOGIC FACTORS AND DNA STRAIN TYPING

The New Zealand fur seal, Arctocephalus forsteri, has been under legal protection since 1916, and hunting is permitted only for research purposes (New Zealand Department of Conservation [NZDC], Marine Conservation Unit 2006). Breeding colonies are found principally on the South Island, Stewart Island, the Chatham Islands, and the sub-Antarctic islands (Crawley and Wilson 1976). The estimated fur seal population is approximately 100,000 in Australia and another 100,000 in and around New Zealand and neighboring islands (Goldsworthy et al. 2013), which is significantly higher than the estimate of 35,000 from 1976 (NZDC 2012). Pups are born from late November to early January; after which, the females mate again. After weaning at 9–10 mo, the pups congregate in pods and often stay in protected inlets. They suckle their mother periodically and gradually venture out to sea as they mature. Opportunities for contact with domesticated cattle (Bos primigenius) can occur either with pups and mothers near the rookeries or with bachelor males venturing into other areas. In general, fur seals prefer rocky outcrops but can also be seen on beaches or pastures near the sea (Crawley et al. 1976).

Seals have been shown to harbor numerous zoonotic diseases, including tuberculosis (Woods et al. 1995), brucellosis (Lynch et al. 2011), leptospirosis (Colegrove et al. 2005), herpes virus, Toxoplasma gondii, and Dirofilaria immitis (Aguirre et al. 2007). Tuberculosis (TB) in the New Zealand fur seal is caused by Mycobacterium pinnipedii, a member of the Mycobacterium tuberculosis complex (Cousins et al. 2003). Other members of this complex occasionally found in New Zealand are M. tuberculosis in humans and Mycobacterium bovis in cattle and wildlife. Mycobacterium pinnipedii has been commonly found in fur seals and sea lions (Otaria flavescens) in Argentina, Australia, and New Zealand. The first article describing TB in seals assumed the disease was caused by M. bovis, which is the common cause of TB in cattle (Cousins et al. 1990). Later work by Cousins et al. (2003) proved that this strain differed from M. bovis and was designated as M. pinnipedii based on spoligotyping and other DNA tests.

Despite domestic cattle and red deer (Cervus elaphus) herd period prevalence (the sum of the point prevalence rate at the beginning of a specified time period plus the cumulative incidence rate for the remainder of the specified time period) of TB falling below 0.2 in 2011 (TB Free New Zealand 2012), New Zealand has a continuing problem with bovine TB in cattle and farmed deer (primarily Cervus elaphus) herds because of a reservoir of M. bovis in wildlife, particularly the brushtail possum (Trichosurus vulpecula) population. The bovine TB control and eradication program has made use of DNA typing of M. bovis as a part of its program for more than 25 yr. All new cases of bovine TB are routinely characterized by DNA typing to determine relationships to previous outbreaks and because strains of bovine TB in wildlife are geographically distinct. This has been of use in verifying the origin of infections. Until recently, restriction endonuclease analysis (REA) was used in New Zealand for typing M. bovis (Collins and de Lisle 1985) because it provided the most valuable epidemiologic information. Currently, typing of M. bovis isolates is done using variable number tandem repeats and the direct repeat locus (VNTR/DR). Although the VNTR/DR method is not as discriminatory as REA, it has the advantage of providing a faster result, is more easily analyzed, and can be reported numerically (Price-Cater et al. 2011).

Seals and other pinnipeds are the natural hosts for M. pinnipedii. Although M. pinnipedii has been transmitted from pinnipeds to humans working at marine zoos (Forshaw and Phelps 1991; Thompson et al. 1993; Kiers et al. 2008), there has been only one documented report of natural transmission from seals to other nonhuman mammals, a Bactrian camel (Camelus bactrianus) and Malayan tapir (Tapirus indicus) in zoos (Moser et al. 2008). We document seven cases of M. pinnipedii infections in cattle. We detail the location of the farms, case presentation, histology, and VNTR/DR typing. In addition, we collected data from a limited survey of 55 neighboring farms that did not experience TB outbreaks by telephone survey to determine possible risk factors for M. pinnipedii infection in cattle.

The cases of M. pinnipedii infection were identified from 1991 to 2011 during routine investigations as part of the bovine TB control program of cattle herds by TB Free New Zealand Ltd. They were identified through routine herd testing or when slaughtered through registered slaughter premises and sampled for TB detection. Upon slaughter, relevant lymph nodes (mandibular, parotid, retropharyngeal, apical, bronchial, mediastinal, supramammary, popliteal, ileocecal, and mesenteric) are checked for tuberculous, and gross lesions of the lungs, liver, kidneys, and gut are noted. Breeding herds in New Zealand were tested annually, biennially, or triennially, depending on the risk of becoming infected with M. bovis from a terrestrial wildlife source. An exception to this policy was for herds located on the Chatham Islands, 750 km east of the South Island. They were declared an accredited (free) area for M. bovis in 1992. The Chatham Islands have no commercial meat or milk processing facilities, and no herd testing has been done there since they were declared free of M. bovis. Most of the animals from the Chatham Islands are shipped and slaughtered on the main two New Zealand islands. All cattle that are slaughtered through a registered slaughter premise, for human consumption, are examined for evidence of bovine TB, and samples are taken from suspicious lesions, for histologic examination, bacteriologic identification, and DNA strain typing.

Tissues were fixed in 4% neutral-buffered formalin, dehydrated with ethanol, embedded in paraffin, sectioned at 30 µm, placed on glass slides, and stained with H&E. A board-certified pathologist (Gribbles Veterinary Pathology Ltd., Christchurch, Hamilton, or Dunedin, New Zealand) examined the histologic sections of the lesions and diagnosed a result of negative, suspicious, or typical for TB, as well as whether or not acid-fast organisms were found. Typical TB is equivalent to granulomatous inflammation with central caseation and mineralization. The necrotic tissue is usually surrounded by a mixture of epithelioid cells, lymphoid cells, and occasional Langhans giant cells. Suspicious lesions show some of the above characteristics, and TB cannot be ruled out.

The procedure used for isolating mycobacteria involved homogenization of tissues using Ten Broek grinders, decontamination with NaOH, and inoculation of the treated samples into Bactec 12B (Becton Dickinson, Franklin Lakes, New Jersey, USA) culture vials. Microbial growth was detected by the release of CO₂, and mycobacteria were initially identified by acid-fast tests as viewed under light microscopy. Preliminary identification of mycobacterial isolates was based on use of a DNA probe specific for the M. tuberculosis complex (Accuprobe, Gen-Probe, San Diego, California, USA). Mycobacterial isolates were examined by REA using the method of Collins and de Lisle (1985) and were confirmed as M. pinnipedii by PCR using M. pinnipedii–specific multiplex for the determination of the absence (168-base pair [bp] product, M. pinnipedii) or presence (293-bp product) of the DNA region of difference (RD2^seal) as described by Warren et al. (2006).

The method for DNA typing was the combination of nine VNTR assays and two PCR direct repeat assays (Price-Carter et al. 2011). The M. pinnipedii isolates from the bovine cases were compared with the reference database of M. bovis isolates that had been subjected to the same assays.

Each isolate was defined by a string of 11 integers (the VNTR/DR profile), corresponding to the number of repeats found at the 11 VNTR/DR loci used in the analysis. The seven unique M. pinnipedii VNTR profiles were analyzed using the MIRU-VNTRplus web-based application (Alix-Beguec et al. 2008; Weniger et al. 2010;). Clustering analysis used a distance matrix generated using the most conservative coefficient—a categorical distance—and the dendrogram was constructed with the unweighted pair-group method with arithmetic mean. The categorical distance scores the number of markers with a different allele (number of repeats) divided by the total number of VNTR/DR assays used. Missing data were ignored for the distance calculations. The dendrogram was drawn using the program TreeDyn (Chevenet et al. 2006).

Case herds were those where the M. pinnipedii–infected cattle originated. Control herds were neighbors to the case herds. A telephone survey was conducted, which consisted of a number of questions (see Supplementary Material) relating to the property location with respect to the sea, access of seals to the property or cows to the beach, herd size, and sightings of seals. All of those questions were also asked of owners of the infected herds. Descriptive statistics (tables, averages, and frequencies) and χ² analyses were performed after questionnaire data were entered into a Microsoft (Redmond, Washington, USA) Excel® spreadsheet and uploaded into NCSS^© (Kaysville, Utah, USA) 2000 statistical software (Hintze 2001).

Seven cases of M. pinnipedii infection in beef cattle were identified between 1991 and 2011 (Table 1). Two cases were identified following a positive tuberculin skin test using 0.1 mL of purified bovine tuberculin injected intradermally in the caudal tail fold (Lepper et al. 1977) and compulsory slaughter. The remaining five cases were not skin tested but had macroscopic lesions resembling TB identified at routine meat inspection. All the cases presented at slaughter with a single lesion in a lymph node, which macroscopically had caseous necrosis and resembled those caused by M. bovis. Histologic changes were consistent with those described in the “Materials and Methods” section. We noted if acid-fast staining organisms were either absent or present. These changes were classified as either suspicious or typical of those found in cases of TB in cattle caused by M. bovis. No histologic features were identified that enabled the M. pinnipedii lesions to be distinguished from those caused by M. bovis.

In all cases, the initial, presumptive identification of the mycobacterium isolated was M. bovis based on cording of acid-fast organisms in liquid cultures and a positive reaction to the M. tuberculosis complex DNA probe. The DNA typing using REA revealed banding patterns the same or similar to those found with M. pinnipedii and were clearly different from the patterns found with M. bovis. The identity of the isolates was confirmed by M. pinnipedii–specific PCR.

As with REA, the VNTR/DR results showed clear differences between the M. pinnipedii and M. bovis isolates, which is illustrated in the dendrogram (Fig. 1 and Supplementary Material). Although the M. pinnipedii isolates were not identical, they are sufficiently different from the bovine isolates to have a separate branch of the dendrogram. Several loci were more highly repeated in M. pinnipedii than they were in M. bovis isolates. For example, there are normally three or four copies of the QUB18 repeat in New Zealand M. bovis isolates, whereas there are 12 or 13 in the M. pinnipedii isolates; there are one to four copies of the QUB26 repeat in New Zealand M. bovis isolates, whereas there are seven copies in M. pinnipedii isolates, and there are three to six copies of ETRD in New Zealand M. bovis isolates, but seven copies in M. pinnipedii isolates. The QUB3232 copy number is varied in both M. bovis and M. pinnipedii, but tends to be larger in the M. pinnipedii isolates, with most isolates harboring more than 16 copies. There is also considerable difference in the size of the product generated with DR1 primers; two, three, five, or six inserts in the DR1 region are amplified from M. bovis isolates, whereas there are consistently 11 inserts in the DR1 region amplified from M. pinnipedii isolates.

Following an initial presumptive diagnosis of infection with M. bovis, all herds with M. pinnipedii infection were placed under movement restriction. A new herd test was scheduled for all animals in the herd, and tracing of all animals into and out of the herd during the last 12 mo. was performed. Epidemiologic investigations revealed no links to possible sources of M. bovis infection but showed three of the seven cases were associated with possible contact with seals on coastal farms on the main two islands of New Zealand; the other four cases were animals most likely to have been infected on the Chatham Islands and shipped to the South Island of New Zealand for finishing and slaughter. This was deducted because they were finishing beef cattle, which had been imported to mainland New Zealand (based on herd of origin ear tags and shipping documents) just before slaughter. Five of the cases were from different farms, whereas the remaining two cases were from the same farm on the Chatham Islands with an interinfection interval of 6 yr. They were presumed to be due to separate incidences of contact with seals because, during the 6-yr interval, a large proportion of the herd had been slaughtered annually without any sign of disease. Because that herd was located on Pitt Island in the Chatham Islands, no herd testing was done. Furthermore, the VNTR/DR typing of the two isolates from that herd was identical, and a within-herd infection could not be conclusively ruled out.

Of an initial list of 91 control herds, 55 owners were contacted and completed the telephone survey. There was no statistical difference between control and case herds for the following categories: distance to sea, herd size, herd type, or farming practice. Some of those parameters were difficult to assess because of the low number of cases. There was a significant difference between the groups on whether or not seals had been sighted on the property. Seals had been seen on four of the six farms with M. pinnipedii infection, compared with 10 of 55 control farms (OR = 9, 95% CI = 1.4–56.1). On all case farms, cattle had access to seals by grazing areas adjacent to beaches or areas with direct connections to the ocean via creeks.

Presence of seals on farms was based on photographic evidence or observations of multiple witnesses. In all case farms, cattle had access to seals by grazing areas adjacent to beaches or areas that were in direct connection with the ocean via creeks. When comparing farms on the Chatham Islands to those in mainland New Zealand, cattle from the islands were more likely to have direct access to the seals via beach or waterways (OR = 3.6, 95% CI = 1.1–11.4).

We document the first recognized cases, to our knowledge, of M. pinnipedii infection in cattle and the likely transmission of infection from wildlife to domestic animals. Limited postmortem investigations of fur seals and sea lions from New Zealand, including its sub-Antarctic islands, have revealed M. pinnipedii is prevalent and widespread in pinnipeds (Lenting pers. comm.). In marked contrast, M. pinnipedii has been isolated from only seven cattle between 1991 and 2011. In all cases of M. pinnipedii in cattle, only one animal in the herd was found to be infected. Although two cases were recorded in one herd, the interval between the cases was 6 yr. and was most likely due to separate contacts with infected seals. The M. pinnipedii infections were identified either by tuberculin skin test or routine meat inspection. Many cases of TB in cattle in New Zealand tend to be dormant, walled-off lesions, where there is only one animal found in the herd, but occasionally, there are numerous animals affected, either through a recent wildlife infection or by reactivation of a dormant infection causing within-herd spread by aerosol or through the milk to calves (New Zealand Animal Health Board unpubl.). There is little danger of human infections through meat and milk because any animals with generalized TB lesions are discarded (slaughter protocol; New Zealand Animal Health Board 2010) and should an infected cow shed any type of TB in the milk, that would be killed by routine pasteurization (Lobue et al. 2010). However, there is some danger of farmers or seal handlers becoming infected by close contact with infected animals (Thompson et al. 1993). As the infection in cattle with M. pinnipedii seems to be limited to one animal in all of the present cases, the danger of human infection from cattle appears to be low. The close genetic relatedness of M. pinnipedii and M. bovis would preclude any of the currently available antemortem tests from distinguishing between infections with these organisms.

In six of seven cases, the lesions were identified in the thoracic lymph nodes, indicating a likely aerosol transmission pathway. For successful aerosol transmission, there would need to be a close association between pinnipeds and cattle. Aerosol transmission has also been proposed for the transfer of M. bovis between brushtail possums and cattle. Experimental studies provide evidence for close contact between moribund possums and cattle, providing the opportunity for aerosol spread of M. bovis (Paterson and Morris 1995).

The lack of multiple cases of M. pinnipedii within a herd suggests that cow-to-cow transmission does not occur or is rare. The finding of M. pinnipedii infection in beef animals was probably a reflection of those animals being more extensively farmed and more often given free range as opposed to dairy cattle, which are more intensively farmed on well-fenced and improved pastures. This is particularly the situation on the Chatham Islands, where unrestricted access to beaches was more common than on mainland New Zealand. No cattle census is taken on the Chatham Islands, but reliable estimates based on slaughter numbers would put it no higher than 8,000 adult cattle (there are no slaughter facilities on the islands, and shipping records provided by the two sole stock agents show fewer than 6,000 animals per year being shipped to the mainland for slaughter). Based on a 20% replacement rate and a 95% calving rate, that would bring the estimated adult breeding cows to fewer than 8,000, which would make the incidence of M. pinnipedii in the Chatham Islands cattle substantially greater than mainland New Zealand coastal areas, where more than 100 times that number graze, and only three cases have been recorded between 1991 and 2011. The VNTR/DR typing showed some variation in the M. pinnipedii isolates, with only two being identical, and those originated from the same farm on one of the islands in the Chatham group. However, all isolates were easily distinguishable from M. bovis isolates found in New Zealand cattle.

Although there have only been seven cases of M. pinnipedii infection recorded in cattle, new cases become increasingly more important as New Zealand enters the eradication phase of its TB program. Misidentification of these cases as M. bovis could lead to unnecessary and inappropriate control measures. In addition to the bovine cases observed in New Zealand, M. pinnipedii has also been documented in other countries: in a human in a marine park, and in a Bactrian camel and a Malayan tapir in zoos. The bovine infections, as well those associated with zoologic parks, are new examples of the natural spread of a pathogen from wildlife to domestic animals and humans.

Supplementary material for this article is online at http://dx.doi.org/10.7589/2013-09-237.

We thank Wendi Roe and Paul Livingstone for reading the manuscript. We acknowledge the technical support and financial contributions of TB Free New Zealand Ltd., Gribbles Veterinary Pathology and AgResearch. Finally, we acknowledge the farmer participation in the survey and their financial support toward the control and eradication of M. bovis infection in domestic and feral animals in New Zealand.