Canine leproid granuloma caused by a member of the Mycobacterium tuberculosis complex

Abstract

Canine leproid granuloma (CLG) is a chronic form of dermatitis that has been associated with nontuberculous mycobacterial infections in Africa, Oceania, the Americas, and Europe. We report here a case of CLG associated with a member of the Mycobacterium tuberculosis complex (MTBC), which could be of public health concern. An 8-y-old pet dog developed 0.5–1-cm diameter, raised, firm, nonpruritic, alopecic, painless skin nodules on the external aspects of both pinnae. Histologic examination revealed severe pyogranulomatous dermatitis with intracellular Ziehl-Neelsen–positive bacilli that were immunoreactive by immunohistochemistry using a polyclonal primary antibody that recognizes tuberculous and nontuberculous Mycobacterium species. DNA extracted from formalin-fixed, paraffin-embedded skin sections was tested by a Mycobacterium genus–specific nested PCR assay targeting the 16S rRNA gene. BLAST sequence analysis of 214-bp and 178-bp amplicons showed 99.5% identity with members of the MTBC; however, the agent could not be identified at the species level. Although CLG has been associated traditionally with nontuberculous mycobacterial infections, the role of Mycobacterium spp. within the MTBC as a cause of this condition, and the role of dogs with CLG as possible sources of MTBC to other animals and humans, should not be disregarded given its zoonotic potential.

Cutaneous nodular mycobacteriosis in dogs was first documented in the 1970s in southern Africa ¹³ and Australia. ¹¹ The condition was initially referred to as canine leprosy until the names “canine leproid granuloma” (CLG) and “canine leproid granuloma syndrome” were coined.

CLG is characterized by the formation of single or multiple, nonpruritic, indolent skin nodules, seen histologically chiefly as pyogranulomatous inflammation that may form distinct pyogranulomas. ² The nodules affect mostly the dermis and/or subcutaneous tissue of the ears, other areas of the head, and to a lesser extent the extremities, particularly but not exclusively in dogs with short haircoats. ⁸ Affected dogs do not usually have any clinical signs, and there is no involvement of regional lymph nodes or internal organs,^8,13 although large skin nodules may ulcerate when traumatized, sometimes leading to secondary infections. ⁷

Because variable numbers of acid-fast bacilli are present histologically in the pyogranulomatous inflammation, ² the disease was initially thought to be a cutaneous manifestation of tuberculosis, ¹³ in reference to diseases caused by Mycobacterium tuberculosis or Mycobacterium tuberculosis variant bovis. Mycobacterial culture in cases of CLG has been unsuccessful, even at mycobacterial reference laboratories.^8,13 Based on these observations, CLG was thought to be caused by saprophytic mycobacteria with fastidious growth requirements precluding isolation in culture media. ⁸

Molecular techniques have helped to identify mycobacteria in cases of CLG. Using 16S rRNA gene analyses, a potentially novel Mycobacterium sp. was identified in 16 of 43 Australian dogs with CLG, ⁶ suggesting that this agent was a likely cause in some cases. The agent was named Mycobacterium sp. Murphy (GenBank AF144747). It shares high sequence homology with Mycobacterium sp. strain IWGMT-90143 (GenBank X88906), considered a probable ribovar of M. simiae. ¹⁵ Mycobacterium sp. Murphy identified in cases of CLG also shares high nucleotide sequence identity with other nontuberculous mycobacteria such as M. tilburgii, M. interjectum, and M. genavense.^6,14 Mycobacterium sp. Murphy has also been identified in cases of CLG in various U.S. states (California, Florida, Hawaii), ⁴ Italy, ³ and Brazil, ¹⁰ suggesting that this agent is the most likely cause of this condition in different geographic locations. Phylogenetic analysis indicated that Mycobacterium sp. Murphy identified in cases of CLG clusters with other species of the M. simiae clade of nontuberculous mycobacteria, ³ represented by slow-growing mycobacteria of low pathogenicity.

We searched the scientific literature and retrieved no cases of CLG caused by mycobacteria of the tuberculosis complex (MTBC), which contains zoonotic pathogens. Here, we describe a case of CLG associated with a Mycobacterium sp. of the MTBC.

An urban pet dog from Colonia del Sacramento, Uruguay, was presented for examination to a local veterinary clinic in March 2021. The patient was an 8-y-old, ~40 kg, spayed female of the Cimarrón Uruguayo breed. Upon initial examination, 0.5–1-cm diameter, raised, firm, nonpruritic, alopecic, painless skin nodules were noted on the external aspect of one pinna; she was treated with hydrocortisone cream applied topically to the nodules. Over the course of ~1 mo, similar nodules developed on the skin of the contralateral pinna (Fig. 1) and the lateral aspect of the left thigh.

Figure 1.

Gross appearance of the cutaneous mycobacterial pyogranulomas in an 8-y-old dog. Multifocal-to-coalescing 0.5–1-cm diameter raised alopecic nodules in the skin of the right pinna.

On April 4, 2021, one of the ear nodules was biopsied and submitted to the veterinary diagnostic laboratory in 10% neutral-buffered formalin for routine histopathology. Histologic examination revealed severe pyogranulomatous dermatitis with a few intrahistiocytic Ziehl-Neelsen (ZN)–positive bacilli, typical of CLG (Fig. 2A, 2B).

Figure 2.

Microscopic images of the cutaneous mycobacterial pyogranulomas. A. Pyogranulomatous dermatitis. H&E. B. A multinucleate giant cell contains numerous intracytoplasmic, acid-fast bacilli. ZN stain. C. Scattered Mycobacterium spp. antigen immunoreactivity. IHC with rabbit polyclonal antibody against Mycobacterium spp. purified protein derivative. Hematoxylin counterstain.

Based on the suspicion of cutaneous mycobacteriosis, another skin biopsy was obtained from a pinna on May 17, 2021; results of an automated CBC performed on a whole blood sample obtained concurrently were within reference intervals. The biopsy was collected aseptically in a sterile container and submitted fresh to the laboratory. Upon arrival, half of the biopsy was fixed in 10% neutral-buffered formalin and processed for histopathology; microscopic lesions were similar to the ones described for the first biopsy, including the presence of occasional intrahistiocytic ZN-positive bacilli. The other half of the biopsy was frozen at −20°C until processed for DNA extraction (MagMAX CORE nucleic acid purification kit; Thermo Fisher), with the digestion protocol recommended by the manufacturer. After incubation with proteinase K, bead-beating was conducted in 0.1- and 0.5-mm lysis tubes (ZR BashingBead; Zymo Research) on a bead beater (BioPrep6; Allsheng) for 2 rounds of 40 s at 6 m/s. DNA was used as template for a TaqMan real-time PCR assay for M. tuberculosis variant bovis using primers and probe targeting the RD4 region ¹² and by a conventional PCR assay using primers targeting the ETR-D region present in different copy numbers in M. tuberculosis and M. tuberculosis variants bovis, africanum, and microti ⁵ (Table 1); results of both assays were negative.

Table 1.

Sample, primers, target region, and reaction conditions used for a Mycobacterium tuberculosis variant bovis TaqMan real-time PCR assay and a M. tuberculosis, M. tuberculosis variants bovis, africanum, and microti conventional PCR assay.

Agent	Sample*	Primers (probe)	Target region	Annealing temperature and time	Cycle	Amplicon size, bp	Reference
M. tuberculosis variant bovis	DNA extracted from skin	F-5′-GACGCCTTCCTAACCAGAAT-3′† R-5′-CTAAGATATCCGGTACGCCCG-3′† (5′-TACAAGCCGTAGTCGTGCAGAAGC-3′)‡	Region of difference 4 (RD4)	60°C/20 s	40	120	Sales et al. 12
M. tuberculosis, M. tuberculosis variant bovis, africanum, M. microti	DNA extracted from skin	F-5′-CAGGTCACAACGAGAGGAAGAGC-3′ R-5′-GCGGATCGGCCAGCGACTCCTC-3′	Exact tandem repeats (ETR-D)	59°C/30 s	35	300	Frothingham and Meeker-O’Connell 5

*

DNA was extracted from 2 subsamples of the same skin biopsy. The DNA concentration was assessed in each subsample (Invitrogen Qubit fluorometer; Thermo Fisher) resulting in 67.2 and 8.36 ng/µL.

†

Primers were modified to ensure that they would hybridize with the target region of M. tuberculosis variant bovis and not hybridize with the host genome.

‡

The probe was designed using Integrated DNA Technologies (IDT) internal ZEN and Iowa Black FQ quencher and 6-FAM as dye reporter.

A serial section of the formalin-fixed, paraffin-embedded (FFPE) block containing skin of the second biopsy was sectioned and processed using an immunoalkaline phosphatase technique (immunohistochemistry, IHC) with a primary rabbit polyclonal IgG antibody raised against M. tuberculosis purified protein derivative (anti–M. tuberculosis antibody ab905; Abcam), which reacts with tuberculous and nontuberculous mycobacteria. The Mach 4 alkaline phosphatase polymer kit (Biocare Medical) and permanent red chromogen (Cell Marque; MilliporeSigma, Aldrich) were used as the detection system; appropriate positive and negative controls were used. Positive immunoreactivity was identified multifocally in the dermal pyogranulomas (Fig. 2C) and in the positive control tissue, but not in the negative controls.

Additionally, DNA was extracted from the FFPE skin biopsy specimen (QIAamp ultra clean production pathogen kit; Qiagen) and tested by a Mycobacterium genus–specific nested PCR assay targeting the 16S rRNA gene ¹ (primer sequences unpublished). Positive PCR amplicons of 214 bp and 178 bp were identified by gel electrophoresis, extracted from the gel, and directly sequenced by Sanger sequencing. BLAST sequence analysis (https://blast.ncbi.nlm.nih.gov/Blast.cgi) of positive amplicons showed 99.5% identity with members of the MTBC, although the agent could not be identified at the species level.

Upon the initial histologic diagnosis of CLG with intracellular ZN-positive bacilli, the dog received antibiotic treatment for mycobacterial infections (clindamycin and erythromycin for 60 d), after which all of the skin nodules resolved. Prolonged antibiotic therapy, particularly combined therapy with rifampicin and clarithromycin, has been shown to be effective in the treatment of CLG ⁹ ; however, spontaneous recovery has also been documented. ⁸ Our patient shared the household with another dog and their owner; neither the other dog nor the owner developed skin lesions contemporaneously.

Our diagnostic investigation allowed for confirmation of CLG caused by a member of the MTBC, with no clinical evidence of extracutaneous involvement, and no hematologic abnormalities. The occurrence of this case at the beginning of autumn coincided with the weak seasonal trend in autumn and winter described for CLG in Australia. ⁸

Histologically, in most cases of CLG, lesions are paucibacillary, although the number of bacilli can be moderate to many in a small subset of cases. ² In our case, the number of ZN-positive bacilli observed histologically was low and had a multifocal distribution; several serial sections of the skin biopsies stained with ZN were required to detect the bacilli. This paucibacillary nature of the lesions may have negatively affected our ability to further identify the agent by molecular means.

Although we could not identify the causative agent at the species level, based on sequence homology, the agent in our case was very different from Mycobacterium sp. Murphy (91% homology) and other members of the M. simiae clade. Our case highlights the value of performing PCR to identify MTBC in FFPE tissues, which was key to the diagnosis. This approach has been used to diagnose MTBC infections in human patients. ¹ Interestingly, the PCR assays conducted on DNA extracted from the fresh-frozen biopsy in our case were negative for some Mycobacterium spp. of the MTBC, such as M. tuberculosis, M. tuberculosis variants bovis, africanum, and microti. This negative finding could indicate null or low numbers of mycobacteria in the fresh biopsies (below the limit of detection of the PCR assays used) or that the dog was infected with another Mycobacterium sp. within the MTBC (e.g., M. canettii, M. orygis, M. mungi, M. tuberculosis variants caprae and pinnipedii, “M. suricattae”, the dassie bacillus, or other).