Detection and Molecular Characterization of Mycobacterium microti Isolates in Wild Boar from Northern Italy

Abstract

Approximately 23,000 hunter-harvested wild boars from the pre-Alpine area of northern Italy were examined for tuberculosis over a 9-year period (2003 to 2011). Retropharyngeal and mandibular lymph nodes from the wild boars were examined grossly, and 1,151 of the lymph nodes were analyzed in our laboratory by histology (728 samples) and culture isolation (819 samples). Mycobacterium tuberculosis complex (MTBC)-specific PCR (1,142 samples) was used for molecular-level detection in tissue samples, as was a gyrB restriction fragment length polymorphism (RFLP) assay (322 samples). Lesions compatible with tuberculosis and indistinguishable from those described in cases of Mycobacterium bovis infection had been observed since 2003. Mycobacterium microti was identified directly in 256 tissue samples by the adopted molecular approaches. However, only 26 M. microti strains were obtained by culture isolation due to the well-known difficulties in isolating this slow-growing mycobacterium. During 2006, a prevalence study was performed in two provinces of the area, and the diffusion of M. microti was calculated to be 5.8% (95% confidence intervals surrounding the estimated prevalences [CI_P95%], 3.94 to 7.68%). Over the following years (2007 to 2011), the presence of M. microti appeared to be stable. All isolates were genotyped by spoligotyping and exact tandem repeat analysis (ETR types A to F). In addition to the typical vole type (SB0118), a new spoligotype lacking the 43 spacers was found. Spoligotyping was also applied directly to tissue samples, and a geographical cluster distribution of the two spoligotypes was observed. This is the first report studying the diffusion and genetic variability of M. microti in wild boar.

INTRODUCTION

Mycobacterium microti belongs to the Mycobacterium tuberculosis complex (MTBC), which also includes Mycobacterium tuberculosis, Mycobacterium bovis, Mycobacterium caprae, Mycobacterium africanum I and II, Mycobacterium canetti, and Mycobacterium pinnipedii. These species are closely related and difficult to distinguish by biochemical properties alone.

M. microti was first described in the 1930s as the causative agent of an epizoonotic disease in the wild English field vole (Microtus agrestis) (1). M. microti can infect a variety of mammalian species, including wood mice (Apodemus sylvaticus), bank voles (Clethrionomys glareolus), and shrews (Sorex araneus) (2–5). Occasionally, it was also found in llamas, cats, pigs, cows, and dogs (3, 6–13). Recently, the widespread host range of M. microti seemed to be increasing, with reports of an infected cat population in England (9, 14, 15) and with accounts of new hosts, such as squirrel monkeys (16), meerkats (17), and South American camelids (18). In humans, infection by M. microti is infrequent, and to date only 27 cases have been reported in immunosuppressed and/or immunocompetent patients (19–23).

Identification of M. microti by traditional methods is difficult since the typical curved cell morphology may be lost upon subculturing (6). Moreover, it grows very slowly on solid medium, and it shows variable biochemical properties for pyrazinamidase and urease activities as well as niacin accumulation (22, 24). Therefore, M. microti's prevalence, geographical distribution, and host range have probably been underestimated.

At present, the identification and characterization of M. microti can easily be determined by molecular methods. Spoligotyping (spacer oligotyping) was successfully applied to distinguish M. microti from the other members of the MTBC (9). Several genotypes of M. microti have been recognized by spoligotyping. The most common are the llama type (SB0112) and the vole type (SB0118); both types are involved in animal and human infections (3, 19–23).

Additionally, informative chromosomal deletions, such as RD1^mic and MiD1 were identified as characteristic of M. microti strains (25). In particular, the deletion of MiD1 is responsible for the vole spoligotype, which only contains two spacers.

The wild boar harbors many important infectious agents that are transmissible to livestock and other animal species, including humans (26). In particular, the presence of MTBC in the wild boar has been reported in several European countries, including Italy (27–34).

However, the role of the wild boar in the epidemiology of bovine tuberculosis (bTB) is still under discussion and can vary among regions. At present, the wild boar is considered a reservoir for tuberculosis in Iberia (29, 35), whereas in Italy, it is considered a spillover host (27, 36). Recently, a study conducted in Sicily of free-ranging domestic black pigs suggested a role in the maintenance of bTB in this ecosystem (37).

Additionally, the geographical range and population densities of the Eurasian wild boar are currently increasing in Europe (38), including Italy (39), reaching previously unrecorded levels. To control the population size, hunting has increased, resulting in a rise of wild boar meat consumption (40). The awareness of a greater risk of interspecies disease transmission and zoonosis led to the development of wildlife health-monitoring programs in many countries.

In northern Italy, wildlife infections and parasitic diseases have been monitored since 1998. As a result of these examinations, the presence of Mycobacterium strains in wild boar was reported (27, 36, 41). Both M. bovis and M. microti were isolated from wild boars in this area, and recently, M. microti was also detected using direct real-time PCR in tissue samples from two wild boars from a neighboring area in Switzerland (34).

Our goal was to assess the diffusion of M. microti in the wild boar population in northern Italy and to provide baseline data for future investigations. Through this study, we show the critical points for detecting M. microti and have developed a specific diagnostic procedure. Additionally, we molecularly characterize M. microti strains circulating in the area of study, providing new insights into the genomic features of this poorly studied mycobacterium.

MATERIALS AND METHODS

Study area and sample collection.

This study was carried out on a wild boar population of the Central Alps in northern Italy and in particular in the provinces of Brescia (BS), Bergamo (BG), Como (CO), and Varese (VA) (Fig. 1).

FIG 1.

Geographical distribution of spoligotype profiles detected in wild boar tissue samples from the pre-Alpine area of northern Italy.

To detect tuberculosis infections in hunted wild boar, we conducted a monitoring program from 2003 to 2011. In the early stage, from 2003 to 2005, we collected retropharyngeal and mandibular lymph nodes (n = 242), both with TB-like lesions and without visible lesions (NVLs), from samples from BS and BG. On the basis of the results obtained during this period, the “preliminary study,” we carried out two field studies.

(i) Study 1.

To evaluate the presence of M. microti in wild boar, all of the hunted animals during 2006 (n = 602) coming from BS and BG were analyzed. This study was limited to two provinces because it was not financially feasible to analyze the number of hunted wild boars from the whole area (2,500 to 5,000/year).

(ii) Study 2.

During the period from 2007 to 2011, only wild boars (307) with TB compatible lesions were analyzed. In this study, samples came from all four provinces, and the culture incubation time was prolonged to 18 weeks.

Postmortem examination was performed on approximately 23,000 wild boars by trained vets. Retropharyngeal and/or mandibular lymph nodes coming from wild boars with TB-like lesions (n = 444) and without visible lesions (n = 707), were submitted to other diagnostic assays. Most of these samples were submitted to PCR, while histology and culture isolation were performed on only some of them (Table 1). Histology was performed on a lymph node from each animal. Culture isolation and IS6110 PCR were carried out using samples of pooled lymph nodes from each animal. When the PCR analysis was positive, the gyrB restriction fragment length polymorphism (RFLP) assay was performed for confirmation and to identify the MTBC members present in the sample. Isolates were further characterized by biochemical methods and by molecular assays, such as spoligotyping and genotyping by exact tandem repeat analysis from A to F (ETR-A to -F).

TABLE 1.

Summary of diagnostic tests performed during the period from 2003 through 2011

Diagnostic test	No. of resultsa
	Preliminary study (n= 242)	Study 1 (n = 602)	Study 2 (n= 307)	Total (n = 1,151)
Culture isolation	240	273	306	819
Histology	84	356	288	728
IS6110 PCR	234	602	306	1,142
gyrB RFLP	74	51	197	322

^aThe number of wild boars (n) is shown in parentheses for each group.

Postmortem examination.

Macroscopic inspection was performed on retropharyngeal and mandibular lymph nodes to detect nodular lesions.

Nodular lesions, such as caseous, necrotic, and calcified nodules, and abscesses were considered potential tuberculosis lesions. Lymph nodes with nodules ranging from 1 to 3 cm in diameter as well as NVLs were divided into two sections. One section was submitted for cultural isolation and molecular assays, and the other was fixed in 10% neutral buffered formalin for histological investigation. Lymph nodes with nodules of less than 1 cm in diameter were wholly submitted for microbiological assays.

Histology.

Portions of formalin-fixed lymph nodes were paraffin embedded and cut into three serial sections 3 μm thick. The first section was stained with hematoxylin and eosin and the next two with Ziehl-Neelsen (ZN) stain, which is specific for acid-fast bacteria. Fifty fields of each section were examined by light microscopy (60×).

We identified three different types of granuloma: (i) specific granuloma, (ii) atypical, and (iii) club-forming granuloma. The specific granuloma corresponded to the typical TB lesion that had a necrotic core, surrounded by a mixed population of inflammatory cells (lymphocytes, epithelioid cells, macrophages, and multinucleated giant cells) and several degrees of fibroplasia. The lesions ranged from initial granulomas to strongly necrotic and calcified granulomas with intense fibroplasia. The atypical granuloma was composed of a central necrotic zone with variable neutrophilic infiltrations and mineralizations and a zone of foamy macrophages containing frequent multinucleated giants cells and fibroplasias. The club-forming granuloma was characterized by the presence in the middle of the so-called “sulfur granules.”

Additionally, the histological lesions were divided into five classes: class 0, negative (absence of granuloma); class 1, mycobacteriosis (specific granuloma and positive ZN staining); class 2, suspect mycobacteriosis (specific granuloma and negative ZN staining); class 3, atypical granuloma; and class 4, club-forming granuloma.

Bacteriology.

For each animal, lymph nodes were pooled and cultured. Approximately 3 to 5 g of tissue samples was homogenized with 3 ml of 1× phosphate-buffered saline (PBS [pH 6.8]) (30 mM phosphate buffer [pH 7.2], 150 mM NaCl, 2 mM EDTA), for 120 s in a Stomacher 80 laboratory blender. The homogenate was collected and decontaminated with an equal volume of 4% NaOH for 30 min at 37°C, neutralized with drops of 10% H₂SO₄, and centrifuged at 3,000 × g for 15 min at 4°C. The sediment was suspended in 1× PBS (pH 6.8) and cultured at 37°C onto solid medium—Löwenstein-Jensen (L-J) medium (Heipha, GmbH, Eppelheim, Germany) and Stonebrink (St) medium (Heipha, GmbH, Eppelheim, Germany) plus pyruvate—and the liquid medium modified Middlebrook 7H9 broth (BBL MGIT, Becton, Dickinson) in the Bactec MGIT 960 system (Becton, Dickinson and Company, Sparks, MD). Growth on solid medium was monitored every week, while the Bactec MGIT 960 system was used to check the liquid medium every day. In study 2, the incubation time for the isolation of M. microti was extended from 6 to 18 weeks. After that time, the absence of growth or characteristic colonies was interpreted as negative.

Identification of isolates. (i) Phenotypic and biochemical identification.

The isolates were evaluated for a panel of biochemical and cultural features by standard procedures (42, 43). The tests included an analysis of the growth at 43°C, pigment production, colony morphology, and biochemical tests, including nitrate reduction, niacin accumulation, arylsulfatase, and urease activities, and tolerance to thiophene-2-carboxylic hydrazide and pyrazinamidase.

(ii) Molecular identification.

Mycobacterial isolates were identified by a multiplex PCR assay described by Kulski et al. (44) and by detection of the RD1^mic region as described by Smith et al. (9). Additionally, the identification of M. microti by PCR-RFLP of the gyrB gene was performed both on isolates and on DNA from tissue samples. This analysis was carried out by the amplification of a 765-nucleotide (nt) fragment of the gyrB gene by using the MTUBf primer (45) and the gyrB Seq primer (46). Five microliters of a heat-inactivated bacterial suspension or DNA extracted from a lymph node was used in a 50-μl reaction mixture containing 0.1 mM each primer, 0.2 mM each deoxynucleoside triphosphate (dNTP), 1× KAPA2G GC buffer, and 0.25 μl KAPA2G robust HotStart DNA polymerase (KAPA Biosystems). The amplification profile was as follows: 15 min at 95°C, followed by 40 cycles of 60 s at 95°C, 60 s at 60°C, and 70 s at 72°C and a final elongation step for 7 min at 72°C. The PCRs were run using a GeneAmp 9700 PCR system (Applied Biosystems). The amplicons were analyzed by restriction with RsaI (Roche) (47). Restriction digests were separated in 2% agarose gel along with a 50-bp DNA ladder (Roche).

Nontuberculous mycobacteria were characterized using the Microseq 500 16S rRNA-based bacterial ID system (Applied Biosystems) following the manufacturer's recommendations. Sequencing was performed in an ABI Prism 3130 genetic analyzer (Applied Biosystems).

DNA extraction and IS6110 PCR.

A 200-μl aliquot of homogenized lymph node sample was submitted to mechanical lysis using 100 mg of glass beads (100 to 200 μm in diameter) in a Qiagen tissue lyser apparatus for 5 min at 30 Hz. Samples were digested with 25 μl of proteinase K (25 mg/ml) for 1 h at 65°C and purified using a Qiaamp DNA minikit (Qiagen) according to the manufacturer's instructions. DNA was eluted in a final volume of 200 μl of TE buffer (10 mM Tris-HCl, 1 mM EDTA [pH 8]).

For PCR analysis, the primers Ext-1 (5′CCCGGACAGGCCGAGTTT3′) and Int-1 (5′-CCCCATCGACCTACTACG-3′) were used to generate a fragment (209 bp) of the IS6110 region. The PCR was performed in a 50-μl reaction mixture containing 5 μl of DNA sample, 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl₂, 0.2 mM each dNTP, 0.4 mM primers, and 1.25 μl Taq Gold polymerase (Applied Biosystems). DNA samples, after 1 cycle at 95°C for 10 min, were amplified through 50 cycles of denaturation for 30 s at 95°C, primer annealing for 30 s at 58°C, and elongation for 45 s at 72°C. In the last cycle, the elongation step was extended for 10 min. Amplified products were detected by electrophoresis in a 2% agarose gel and 1× TBE buffer (100 mM Tris-base, 100 mM boric acid, 2 mM EDTA [pH 8.3]). For positive IS6110 PCR samples, differentiation of MTBC species was attempted by PCR-RFLP of the gyrB gene by using the DNA of tissue samples as described above.

Genotyping of M. microti isolates.

The molecular characterization of 26 M. microti isolates collected from 2003 to 2011 was endeavored by three different techniques: spoligotyping, sequencing of a portion of the direct repeat locus (DR), and variable-number tandem repeat (VNTR) typing. Spoligotyping was performed as described by Kamerbeek et al. (48). The spacer sequences contained in the DR locus were detected by hybridization onto a spoligotyping membrane (Isogen Bioscience BV, Maarssen, The Netherlands). For spoligotyping performed directly on tissue samples, the PCR product of the DR locus was analyzed by electrophoresis in a 2% agarose gel and 1× TBE buffer to detect the typical smeared pattern. Moreover, we quantified the MTBC DNA in the samples with the Artus M. tuberculosis PCR kit (Qiagen), according to the manufacturer's instructions. Sequencing of a portion of the DR region located 3′ of the IS6110 element was done following amplification with the primers IM_dir (5′-CTACTACGCTCAACGCCAG-3′) and IM_rev (5′-ATCGACTGCGACAACATCC-3′). The PCR product was 573 or 387 bp, depending on the presence or absence of spacers. The fragments obtained were sequenced using an ABI PRISM 3130 instrument (Applied Biosystems). VNTR typing was assessed by amplification of five loci according to Frothingham and Meeker-O'Connell (49), with modifications described by Boniotti et al. (50). Allelic diversity was calculated by the method of Selander et al. (51).

Statistical analysis.

A statistical analysis was performed with Excel 2010 (Microsoft). Data obtained from study 1 were used to evaluate the prevalence of M. microti infections. In this study, animals were considered infected when the IS6110 PCR and/or the gyrB RFLP was positive. The 95% confidence intervals surrounding the estimated prevalences (CI_P95%) were calculated using a binomial distribution.

RESULTS

Field studies. (i) Preliminary study.

In the early stage of our study (2003 to 2005), 242 samples from BG and BS were randomly monitored (Table 1). Lesions consistent with TB infection were macroscopically detected in 89 head lymph nodes. Using IS6110-based PCR, we found 61 MTBC-positive samples, and another 8 were detected among NVL samples. By gyrB RFLP analysis, applied directly to tissue samples, 52 were confirmed to be M. microti and 2 were M. bovis. By culture isolation, performed with protocols optimized for M. bovis growth (42), we isolated one M. microti strain and two M. bovis strains.

Despite the inefficiency of culture isolation, these preliminary data showed a considerable presence of M. microti in the wild boar population, which was much more prevalent than M. bovis. This underlined the importance of designing a specific survey to determine the prevalence of infection and to standardize a set of molecular methods to identify and characterize M. microti strains.

(ii) Study 1.

During the hunting season of 2006, we analyzed all of the hunted wild boars (n = 602) from two provinces located in northern Italy (BS and BG). Forty-eight samples showed tuberculous-like lesions, whereas 554 contained NVLs (Table 2). The samples underwent IS6110 PCR, and those that tested positive were analyzed using the gyrB RFLP assay to confirm and identify the presence of a member of the MTBC. Culture isolation was attempted, using a protocol optimized for M. bovis growth, on 273 samples (44 with macroscopic lesions and 229 with NVLs), but only six Mycobacterium spp. were isolated. In the same year, two M. microti strains and two M. bovis strains were isolated from two provinces (CO and VA) not included in study 1.

TABLE 2.

Tuberculous-like lesion and M. microti molecular detection in study 1

Type of lesion found at postmortem examination (n)a	No. of wild boars with:
	MTBC detection by IS6110 PCR		M. microti identification by gyrB RFLP (+)
	+	−
TBL (48)	24	24	22
NVL (554)	27	527	13
Total (602)	51	551	35

^aTBL, tuberculous-like lesions; NVL, no visible lesions.

Using IS6110 PCR, 24 positive results in samples with macroscopic lesions and 27 positive results in NVL samples were obtained. Using gyrB RFLP, the presence of M. microti was confirmed in 22 samples with macroscopic lesions and in 13 NVL samples. The presence of M. microti in the wild boar population of these regions is 8.5% (CI_P95%, 6.24 to 10.69%) based on PCR detection and 5.8% (CI_P95%, 3.94 to 7.68%) based on gyrB RFLP identification. Considering that 47% of the PCR-positive samples were detected in the 48 samples showing lesions, compared to the 53% detected in 554 samples without lesions, we decided to focus our subsequent efforts on samples with lesions, thereby significantly reducing the experimental costs.

(iii) Study 2.

In this survey, the area of study was extended to the whole sub-Alpine territory of the Lombardy region, but only lymph nodes with macroscopic TB-like lesions were analyzed. A total of 307 samples out of 15,922 were collected during the hunting seasons from 2007 to 2011 and were submitted for culture isolation and analyzed by the IS6110 PCR assay (Table 3).

TABLE 3.

Positive results of diagnostic tests in study 2^a

Yr	No. of positive results
	TB lesions	MTBC (IS6110 PCR)	M. microti (gyrB RFLP)	Culture isolates	M. microti isolates
2007	82	38	37	22	12
2008	58	44	41	1	0
2009	53	34	31	6	5
2010	54	38	25	5	3
2011	60	43	35	9	3
Total	307	197	169	43	23

^an = 307.

The number of animals with TB lesions was stable from 2008 to 2011, and the amount of MTBC-positive samples, 60 to 70%, was also stable.

Of the samples positive by IS6110 PCR, 85% were confirmed to be M. microti by gyrB RFLP analysis. However, only 23 were isolated using culturing protocols with a prolonged incubation time (18 weeks). The number of culturable isolates was still very low compared to the number of M. microti-positive samples detected by molecular techniques.

Pathology.

Postmortem examination was performed on retropharyngeal and mandibular lymph nodes from approximately 23,000 wild boars during the whole period of study (2003 to 2011). Gross lesions were present in 444 wild boars as yellowish-white nodules from 0.5 to 3 cm in diameter. The nodules were single, rarely multiple, and always surrounded by a capsule. The center of the lesion ranged from necrotic and caseous or necrotic and calcified to purulent. Sometimes, the necrosis was so extensive as to occupy almost the entire lymph node.

Histological examinations were performed on 728 samples (Tables 1 and 4). TB-specific granulomas were observed in 187 samples (classes 1 and 2) (Table 4 and Fig. 2). Lesions were mostly multifocal and appeared as strongly necrotic and calcified granulomas with intense fibroplasia. Scattered eosinophils and granulocytes were distributed close to the fibrous capsule. A total of 127 samples stained ZN positive. The acid-fast bacteria were rare and observed in low numbers (1 to 3) in necrosis. Most often, deposits of calcium fragments, intensely basophilic, interfered with the interpretation of the ZN staining. Rarely, acid-fast bacilli were observed in the cytoplasm of the multinucleated giant cells. Sixty TB granulomas stained ZN negative (class 2). In the latter group, four samples had no gross lesions, and the PCR analysis was positive for two samples. Seventy-seven samples were classified as club-forming granulomas (class 4). Most of them had an intensely eosinophilic precipitate in the center with a marked giant cell response. The precipitates formed a radiating appearance ending with clubbed projections. Eleven were observed in NVL lymph node samples. Seventy-seven samples were classified as atypical granulomas (class 3). The histological appearance was variable (Fig. 2). Large granulomas frequently had numerous foamy macrophages around the central area of mineralization and numerous multinucleated giant cells. The smaller granulomas had karyorrhectic centers surrounded by macrophages and histiocytes. Sometimes, only focal or multifocal microscopic mineralizations surrounded by macrophages were present. Seven atypical granulomas were observed in NVL lymph node samples, and the ZN staining was always negative. A total of 387 lymph nodes had no histological lesions, even though 86 had gross lesions and 30 were positive by IS6110 PCR analysis.

TABLE 4.

Histological results from 2003 to 2011^a

Lesion class	No. of wild boars with:
	Gross lesions (n = 405)			NVL (n = 323)
	No. of samples	MTBC (IS6110 PCR)	M. microti (gyrB RFLP)	No. of samples	MTBC (IS6110 PCR)	M. microti (gyrB RFLP)
0	86	30	21	301	14	6
1	127	120	110	0	0	0
2	56	50	43	4	2	2
3	70	52	46	7	0	0
4	66	13	7	11	0	0
Total	405	265	227	323	16	8

^an = 728.

FIG 2.

Histopathology in wild boar lymph nodes. Panels A to F show an NVL lymph node (A) and nodal lesions (B to F) stained by hematoxylin and eosin (A, B, C, E, and F) and Ziehl-Neelsen stain (D). (A) Class 2 granuloma (suspected mycobacteriosis), shown as a cluster of multinucleated Langhans type giant cells in the paracortex region (20×). (B) Class 2 granuloma. Shown are granulomas in the cortical area with a necrotic center and minimal calcification surrounded by epithelioid cells, multinucleated giant cells, and macrophages (20×). (C) Class 2 granuloma. Shown is an extensively necrotic and calcified granuloma surrounded by a wide connective capsule (20×). (D) Class 1 mycobacteriosis. Shown are rare acid-fast bacteria in necrotic debris (100×). (E) Class 3 atypical granuloma. Shown is a granuloma in the medullary zone with minimal necrosis, neutrophilic infiltration, foamy macrophages, and fibroplasia (20×). (F) Class 4 club-forming granuloma. Shown are colonies with radiating clubs surrounded by a granulomatous reaction with giant cells (20×).

Culture isolation and biochemical identification of strains.

Table 5 summarizes the isolates obtained during the period from 2003 to 2011. All of the M. microti as well as M. bovis isolates obtained during the preliminary study and study 1 were from samples with macroscopic lesions, while most of the M. avium complex (11/18) and M. nonchromogenicum (6/11) isolates were from NVL samples.

TABLE 5.

Isolates from wild boar from 2003 to 2011

Organism	No. of isolates from:
	Preliminary study	Study 1	Study 2	Total
M. microti	1	1 + 1a	23	26
M. bovis	2	2a		4
M. avium complex	12		6	18
M. nonchromogenicum	6	3	2	11
M. hiberniae	1			1
M. engabaekii		1		1
Nocardia	1			1
Hafnia alvei			1	1
Actinomadura		2	2	4
Micromonospora			2	2
Streptomyces			1	1
Total	23	10	37	70

^aFrom the CO and VA provinces, which were not included in study 1.

M. microti isolates took between 12 and 18 weeks to grow. Eleven isolates grew on the liquid Middlebrook 7H9 medium only, 10 grew on solid medium only, and 5 grew on both liquid and solid media.

After the primary isolation, M. microti strains were analyzed by phenotypic and biochemical tests (Table 6) (42, 43). Due to the slow growth and the low bacterial loads, subculturing was difficult to perform, and the biochemical results were often weak or negative. Almost all of the strains (25/26) showed urease activity, a characteristic shared by M. bovis and M. tuberculosis. Niacin accumulation, which discriminates M. microti from M. bovis, was present in 18 strains. None of the 26 strains showed nitrate reduction as had been reported by Magee and Ward (43), and only three strains showed pyrazinamidase activity (20, 43).

TABLE 6.

Phenotypic and biochemical characteristics of M. microti isolates in this study

Phenotypic or biochemical test	Result for isolate no.a:																										Expected result (reference)a
	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19	20	21	22	23	24	25	26
Niacin accumulation	++	+	+	+	+	+	−	+	+	−	+	+	−	−	−	−	+	+	+	−	+	+	+	+	++	−	+ (24, 43)
Nitrate reduction	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	+/− (24, 43)b
Urease	+	+	+	++	−	+	++	++	+	+	+	++	+	++	+	+	+	+	+	++	+	++	+	+	+	+	+ (24, 43)
Pyrazinamidase	+	-	+	++	−	−	−	−	−	−	−	−	−	−	ND	−	−	−	−	−	−	−	−	−	−	−	+/− (20, 43)c
TCHd tolerance	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	− (24, 43)
14-day arylsulfatase	−	−	−	−	−	−	−	−	−	ND	ND	−	−	−	ND	−	−	−	−	−	−	−	−	−	−	−	− (24, 43)
Growth at 43°C	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	− (24, 43)
Pigmentation	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	− (24, 43)

^a+, weak; ++ medium; −, negative; ND, not determined.

^bLevy-Frebault and Portaels (24) reported absence of nitrate reduction in M. microti strains, while Magee and Ward (43) reported that 11 to 89% of the strains showed nitrate reduction.

^cMagee and Ward (43) reported a positive reaction to pyrazinamidase, and van Soolingen et al. (20) reported 2 positive results among 8 strains.

^dThiophene-2-carboxylic hydrazide.

Molecular identification and genotyping of M. microti strains.

The 26 M. microti strains isolated during the period from 2003 to 2011 were identified by PCR-RFLP analysis of the gyrB gene. All of the strains also possessed a deleted RD1^mic region (Table 7).

TABLE 7.

Genotyping of M. microti strains and allelic diversity at each locus^a

Strain no.	Yr	Spoligotype	Diversity by ETR type:
			A	B	C	D	E	F
1	2005	SB0118	9	3	6	4	1	2.2
2	2006	SB0118	6	3	5	4	1	2.2
3	2007	SB0118	6	3	5	8	1	2.2
4	2007	SB0118	9	3	5	4	1	2.2
5	2007	SB0118	6	3	5	9	1	2.2
6	2007	SB0118	9	3	5	4	1	2.2
7	2007	SB0118	9	3	5	4	1	2.2
8	2007	SB0118	6	3	5	8	1	2.2
9	2007	SB0118	6	3	5	8	1	2.2
10	2007	SB0118	9	4	5	4	1
11	2007	SB0118	9	3	5	6	1	2.2
12	2007	SB0118	6	3	5	8	1	2.2
13	2007	SB0118	6	3	5	8	1	2.2
14	2007	SB000	9	3	5	6	1	2.2
15	2007	SB000	9	3	5	6	1	2.2
16	2009	SB0118	9	4	5	4	1	2.2
17	2009	SB0118	9	3	5	4	1	2.2
18	2009	SB0118	9	3	5	4	1	2.2
19	2009	SB000	9	3	5	6	1	2.2
20	2009	SB0118	10	3	5	7	1	2.2
21	2010	SB0118	10	3	5	7	1	2.2
22	2010	SB0118	6	3	5	8	1	2.2
23	2010	SB000	11	3	6	6	1	2.2
24	2011	SB0118	6	3	5	10	1	2.2
25	2011	SB0118	9	3	5	6	1	2.2
26	2011	SB0118	6	3	5	8	1	2.2

^aAll strains shown had the RD1^mic A chromosomal deletion. Total allelic diversity was as follows: spoligotyping, 0.28; ETR-A, 0.59; ETR-B, 0.11; ETR-C, 0.11; ETR-D, 0.74; ETR-E, 0; ETR-F, 0.

Spoligotyping was performed for all of the M. microti isolates. Out of 26 isolates, 22 showed the characteristic vole-type profile (SB0118), while the remaining 4 showed an absence of spacers. The genome sequence of this region confirmed these results. In the SB0118 isolates, we found the spacers 51, 52, 58, 66, 67, and 68, as reported by van Embden et al. (52), where 51 and 52 correspond to spacers 37 and 38 (48). No spacers were found in the other four isolates, indicating a larger MiD1 deletion in this region. This profile was named SB000 in this study and was registered in the Mbovis.org database (56). The international spoligotype name for the pattern called SB000 here is SB2277.

By VNTR analysis, 10 genotypes were identified, which when combined with the two spoligotypes, results in 11 total genotypes (Table 7). Due to the low number of strains analyzed, the allelic diversity value calculated for each locus is not reliable; however, a few differences are evident compared to M. bovis isolates. The ETR-E and -F loci were monomorphic, as already observed by Smith et al. (9) in M. microti strains from cats. Additionally, ETR-D showed an increased allelic diversity (0.74) compared to the M. bovis population from cattle in Italy (0.22) (50).

Geographical localization of genotypes.

Spoligotyping was performed directly on DNA from tissue samples when M. microti was identified by gyrB RFLP analysis; however, the spoligotype profiles were not always readable. To exclude PCR amplification problems, M. microti DNA was quantified by the Artus M. tuberculosis PCR kit (Qiagen). All of the SB000 samples showed an amount of target DNA comparable to that in the SB0118 samples (data not shown). We could assign spoligotypes to 118 samples, with 84 showing spoligotype SB0118, while 34 did not show any spacers. The frequencies of the two spoligotypes in tissue samples (SB0118, 84/118 [71%]; SB000, 34/118 [29%]) are consistent with those found among the isolates (SB0118, 19/23 [83%]; SB000, 4/23 [17%]). By georeferencing the spoligotype profiles obtained from tissue samples, we noticed a geographical clustering of genotypes (Fig. 1). The profile SB000 is predominant in the area near Garda Lake. SB000 was occasionally present also in BG and CO in 2007, but it was not detected in the following years. Conversely, SB0118 prevailed in CO, VA, and BG, and only occasionally appears in BS.

DISCUSSION

Our work shows a considerable and unexpected presence of M. microti in the wild boar population of the Central Alps in northern Italy, with a prevalence estimate of approximately 5.8%.

Since 2003, lesions compatible with tuberculosis infections have been observed in wild boar. At the macroscopic and microscopic levels, lesions present in retropharyngeal and mandibular lymph nodes were similar to those described for M. bovis infections, and no notable differences were observed in this study.

However, the hunted wild boars appeared to be in good condition, with no signs of clinical disease. Similarly to what happens in M. bovis infections, calcified granulomatous lesions predominated, indicating prolonged and chronic infection. Histological examinations and ZN staining showed a paucity of bacilli as well.

At the beginning of our study, culture isolation from visible TB lesions failed to identify the mycobacterium involved. Instead, an IS6110-based PCR analysis showed the presence of the MTBC in tissue samples. However, only by using a gyrB RFLP assay, which is able to discriminate the members of the MTBC, could we establish that the causative agent of the tuberculosis infection was M. microti. This molecular approach was extremely important to correctly identify the mycobacterium involved, which can be confused with M. bovis, and the assay is essential in detecting very-slow-growing and difficult-to-culture mycobacteria, such as M. microti, whose prevalence would otherwise be underestimated.

This microorganism is generally thought to be a disease agent of small mammals, but occasional isolations have been reported from other species, such as pigs, llamas, dogs, cattle, and humans (6, 7, 11, 12, 19–23).

Abundant recent literature reports the presence of MTBC in the Eurasian wild boar (Sus scrofa) in several European countries, including Italy (27–34, 36, 37).

M. bovis is the most frequently isolated agent, but occasionally M. microti is also isolated from wild boar (34, 36). In the area of study, the presence of M. bovis was marginal, with only four positive samples isolated in 2005 (2) and 2006 (2) from CO and VA. The genotyping of isolates was performed by spoligotyping and mycobacterial interspersed repetitive unit (MIRU)-VNTR analysis using the following 12 markers: ETR-A, -B, -C, -D, and -E; MIRU-26; and VNTR 2163a, 2163b, 4052, 3155, 1895, and 3232 (50). Three of the isolates had the same genetic profile (SB0120, 4-5-5-3-3/4-5-10-4-5-3-1-6), and the fourth had a similar genotype with a single variation at locus ETR-E. No related M. bovis strains have been detected in the same geographical area in cattle herds. It is noteworthy that the prevalence of bovine tuberculosis in cattle is very low in the study area. In fact, Lombardy has been designated an officially TB-free region since 2010, with prevalence below 0.1% from 2004. In other Italian regions, the situation is different. In Piemonte, Liguria (36), Sardegna, and Marche, M. bovis isolates from wild boar possessed genotypes found also in local cattle herds (unpublished results). Recent studies (9) reported that M. microti is endemic in certain areas of Great Britain in cats and that the geographic areas where M. bovis and M. microti were recovered from domestic cats do not overlap. The authors suggested that exposure to M. microti could contribute to protection against M. bovis, working as a natural vaccine, reducing the presence of bovine tuberculosis even in other wild and domestic animals. Even though in the study area the TB-free status was achieved thanks to the eradication campaign, we cannot exclude that the presence of M. microti in the wild boar could contribute to protection against M. bovis in the future.

Attention to wild boar health has intensified in the last few years, mainly as a consequence of a considerable increase in the population density and geographical range (39, 53). In particular, in the Central Italian pre-Alps, the wild ungulate population has grown as a result of changes in land use and game management strategies (39, 54). Consequently, hunting and the consumption of wild boar meat have increased, making the chances of human exposure to wild boar diseases greater. Skinning and evisceration procedures might lead to infection by direct contact with the organs and tissues of infected animals (55). Tuberculosis caused by M. microti is uncommon in humans, even if it is considered able to cause clinical illness in both immunocompetent and immunosuppressed patients (19, 20, 21, 22, 23). However, the prevalence of human infection may have been underestimated due to the difficulties in primary isolation and differentiation and because cases of pulmonary tuberculosis caused by M. microti and M. tuberculosis are clinically indistinguishable.

Additionally, livestock partially shares pastures with wild animals, and contact between wild boars and domestic animals may occasionally occur. Cattle might get in contact with M. microti-infected wild boar or other environmental sources and become false-positive skin reactors interfering with the bTB diagnosis.

Monitoring activities were implemented in Italy in 1998 with the cooperation of hunting associations and the veterinary official service. In the frame of this control program, meat inspection was performed on a total of approximately 23,000 wild boar head lymph nodes from hunted animals during the period from 2003 to 2011.

During study 1, the diffusion of M. microti infection in wild boar in two provinces of northern Italy was calculated by analyzing the hunted wild boars coming from BS and BG. The prevalences of M. microti were 8.5% (CI_P95%, 6.24 to 10.69) based on PCR detection and 5.8% (CI_P95%, 3.94 to 7.68) based on gyrB RFLP identification. The discrepancy between IS6110 PCR and gyrB RFLP analyses is probably due to the different sensitivities of the two techniques. The first amplifies a target (209 bp) present in multiple copies in the genome of M. microti, while the second amplifies a product more than three times longer (765 bp). For these reasons, the prevalence of M. microti in this wild boar sampling is likely to be closer to 8% than 6%.

Based on this study, the probability of detecting MTBC in samples with macroscopic lesions (50% [24/48]) was higher than that for NVL samples (4% [27/554]). Therefore, in the following studies, we analyzed only lymph nodes with visible lesions.

In study 2, the presence of M. microti in the Lombardia region over a 5-year period (2007 to 2011) was monitored. The presence of TB-like lesions and M. microti-positive samples was essentially stable (Table 3). To increase the efficiency of culture isolation, the culture growth incubation period was increased to 18 weeks. Despite this, only 23 of 169 M. microti-positive samples were isolated. It is possible that a different sample pretreatment, culture medium composition, and a combination of liquid and solid media could be optimized for more efficient M. microti growth.

Histological examination revealed the different evolutionary stages of the TB-specific lesions as well as of unspecific lesions (Fig. 1). Especially in NVL samples, the histological examination was important to discriminate unspecific granulomas, such as atypical granulomas and club-forming granulomas. Granulomas are frequently observed in wild boar, probably due to their feeding behaviors, such as rooting, which can induce microtraumas to the oral mucosa, facilitating the entry of environmental bacteria that can colonize the retropharyngeal and mandibular lymph nodes.

Histology was less efficient than PCR in detecting TB infections (Table 4). Since the culture isolation in a TB diagnostic laboratory takes priority, the most symptomatic portion of the sample is used for this purpose. The remaining portion may be less symptomatic, resulting in a disadvantage for the histological analysis. However, the combination of histopathology and PCR proved to be useful to reveal mixed infections (classes 3 and 4).

During the study period (2003 to 2011), a total of 26 M. microti strains were isolated and characterized by biochemical and molecular methods (Tables 6 and 7). Molecular characterization by spoligotyping discriminated two genotypes: the vole-type profile (SB0118) and a new, never described before, profile in which no spacers were found. This result, confirmed by sequencing of the DR region located at the 3′ end of the IS6110 element, indicates a rearrangement of this region. Further genomic studies will be needed to understand the evolutionary origin of this strain and its association with the wild boar host. All of the strains contained deletions of the RD1^mic region (Table 7), a common feature of M. microti strains. Genotyping by ETRA-F analysis identified 10 genotypes that, combined with the two spoligotypes, resulted in a total of 11 genotypes (Table 7). The ETR-E and -F loci were monomorphic, as already observed in the United Kingdom by Smith et al. (9) for M. microti strains from many sources.

Spoligotyping, performed directly on DNA from those tissue samples in which M. microti was confirmed by gyrB RFLP analysis, showed a geographical clustering of genotypes (Fig. 1). Geographical clustering of M. microti genotypes was also observed by Smith et al. (9) in the United Kingdom.

In wild boar, M. microti infections can be transmitted via direct contact with wild rodents and other small mammals favored by their behavior and eating habits (rooting and scavenging on carcasses). However, the high prevalence of M. microti in the wild boar population suggests an active role of the wild boar in maintaining M. microti in the environment, and on the basis of current knowledge, intraspecies transmission cannot be excluded.

The understanding of the wild boar's role in the epidemiological cycle of M. microti needs further investigation. In the meantime, constant monitoring is desirable since M. microti can interfere with the diagnosis of TB infections.