Abstract
Osteoarthritis is a common cause of morbidity and mortality in geriatric gazelles. Propionibacterium australiense has been reported as a cause of systemic granulomas in cattle, but there are no descriptions of this bacteria infecting other species nor causing osteoarthritis, to our knowledge. An 8-y-old, castrated male, sand gazelle (Gazella leptoceros leptoceros) was managed for chronic, intermittent, progressive osteoarthritis of the right tarsus. Serial biopsies revealed pyogranulomatous dermatitis with intralesional bacteria. Serial diagnostic imaging identified osseous and soft tissue proliferation with draining tracts. Treatments over 1 y included broad-spectrum antibiotics, anti-inflammatories, joint debridement, and infusion with platelet-rich plasma and stem cells. Despite therapy, lameness persisted, azotemia developed, and subsequently, the animal was euthanized. On postmortem examination, the periarticular tissue of the right tarsus was markedly expanded by pyogranulomas and fibrosis. Histologically, the synovium, joint capsule, and overlying soft tissues were markedly expanded by pyogranulomas and numerous gram-positive and acid-fast–negative filamentous bacteria surrounded by Splendore–Hoeppli material. Within the joint, there was regionally extensive cartilage ulceration, osteonecrosis, osteolysis, and pannus formation. PCR assay of affected formalin-fixed, paraffin-embedded tissue amplified segments of 16S rRNA and β subunit of bacterial RNA polymerase (rpoB) genes with 99.7% and 95.6% identity to P. australiense. This bacterium should be considered a differential for chronic pyogranulomatous osteoarthritis in gazelles.
The sand gazelle (syn. Rhim gazelle, slender-horned gazelle; Gazella leptoceros leptoceros) is a terrestrial ungulate of Saharan desert climates, residing primarily in Algeria, Egypt, Libya, and Tunisia. 16 They are classified as “endangered” by the IUCN 16 with an estimated 300–600 mature individuals in their native range. The population continues to decline because of increased hunting, encroachment of territory by human activity, and drought.6,16 Acquisition of up-to-date knowledge of diseases impacting this species is critical to maintain effective conservation and propagation programs.
A retrospective study of 16 species of captive gazelle (n = 311 animals) identified trauma (32.5%), bronchopneumonia (11.3%), and maternal neglect (8.4%) as the most common contributors to morbidity leading to euthanasia. 3 In that study, osteoarthritis was the seventh most common cause for euthanasia, accounting for 3.9% of cases. Degenerative osteoarthritis (i.e., degenerative joint disease) is considered a common antemortem finding in geriatric gazelle species and a cause for euthanasia in some individuals when clinical management is not sufficient to alleviate clinical signs. 17 Bacterial osteoarthritis has been reported sporadically in non-domestic hoofstock, including sand and Dorcas gazelles.3,17,22,23 However, we found no reports of clinical, microbiologic, gross, and histologic findings in a case of bacterial osteoarthritis in a gazelle using Google, PubMed, and The Ohio State University library databases with search terms “gazelle”, “osteoarthritis”, “bacteria”, and “Propionibacterium”. Herein we present a case of chronic bacterial osteoarthritis, cellulitis, and synovitis in a gazelle attributable to Propionibacterium australiense infection that was managed by surgical intervention, medical management, and stem cell therapy, which, to our knowledge, has not been reported previously.
An 8-y-old, captive, castrated male, sand gazelle was managed for 5 y for intermittent lameness of the right hindlimb. At 3-y-old, the gazelle suffered a suspected traumatic injury to the lateral aspect of the right tarsus resulting in swelling and lameness; the injury was thought to be secondary to interspecific fighting. Clinical signs associated with this swelling resolved rapidly with stall rest and a 3-d course of oral NSAID therapy with flunixin meglumine (Banamine, 1.8 mg/kg/d PO; Merck). At 4-y-old, lameness recurred, and examination revealed an injury to the lateral claw of the right hindlimb that resulted in prolonged lameness and swelling of the ventrolateral tarsus. Radiographs of the distal limb identified mild abaxial rotation and dislocation of the proximal interphalangeal joint; the tarsus was radiographically unremarkable. Tarsal swelling at that time was presumed to be secondary to digital trauma. Lameness resolved within 1 mo with stall rest and oral NSAIDs (meloxicam, two 2–3-d courses of 0.5 mg/kg/d every other day PO followed by phenylbutazone, 3 doses at 12.0 mg/kg/d q24h PO) therapy, and the animal returned to normal function.
When the gazelle was 7-y-old, a cutaneous, raised, red, crusting, 5.0-cm nodule was noted on the right lateral hindlimb at the level of the tarsus. At that time, no lameness was noted, and the lesion was considered unrelated to prior trauma of the foot and tarsus. Histologic evaluation of a punch biopsy of the cutaneous nodule revealed diffuse pyogranulomatous inflammation with large bacterial colonies with associated Splendore–Hoeppli material. Notably, biopsy was limited to the skin and dermis. The patient was managed with antibiotics (12.0 mg/kg/d of injectable ceftiofur) and NSAIDs (0.28 mg/kg/d of IM meloxicam followed by a 3-d course of 0.35 mg/kg/d of oral meloxicam), resulting in mild reduction in swelling associated with the biopsy site; despite this therapy, the gazelle slowly developed progressive lameness. Stall rest and similar antibiotic and NSAID therapy was continued, resulting in initial mild clinical improvement. Computed tomography (CT) performed at this time indicated degenerative changes of the tarsal joints, a chip fracture of the talus, and soft tissue swelling suspected to represent edema, fibrosis, and synovial thickening. Three months following initial biopsy, there was continued lameness and increased right tarsal swelling. A second punch biopsy was performed from the right lateral hock, and histologic evaluation revealed pyogranulomatous dermatitis and intralesional bacteria similar to the initial biopsy.
Immediately following the second biopsy, the gazelle’s lameness worsened despite continued management with oral meloxicam (multiple 5–7-d courses of 0.70–1.0 mg/kg/d doses), injectable ceftiofur (multiple 12.0–12.5 mg/kg/d doses), and stall rest. Repeated CT evaluation identified progression of degenerative joint disease, characterized by tibial periosteal proliferation with bony fragments embedded within soft tissues. These findings raised concern for concurrent inflammation and/or septic osteoarthritis. The patient then underwent anesthesia for right tarsal debridement, platelet-rich plasma (PRP) injection, and bone marrow biopsy for collection of stem cells. During surgery, the tarsal joint was opened and debrided via an incisional approach, PRP injection was performed, and bone marrow was obtained from the right tuber coxae of the ilium for processing of stem cells. Intraoperatively, the synovium was noted to be grossly thickened and the joint fluid was opaque; the radiographically identified bone fragment was not found. Immediately following surgery, stem cell therapy was initiated intraarticularly in addition to injectable antibiotics (ceftiofur 12.5 mg/kg/d, tulathromycin 3.0 mg/kg/d) and oral meloxicam (1.0 mg/kg/d for multiple weeks); during this period and the 2 wk following surgery while these therapies were in place, there was a significant improvement in lameness.
However, 5 mo following surgery and stem cell treatment, lameness and tarsal swelling returned. Aspirate of synovial fluid from the tarsus at this time yielded 2.0 mL of yellow, turbulent material from the joint. Cytologic evaluation of the retrieved fluid revealed a highly cellular exudate primarily of degenerate neutrophils. Standard anaerobic culture via swab collection yielded growth of Actinomyces sp. as identified via MALDI-TOF Biotyper software (Bruker). Notably, speciation was not performed, and susceptibility testing was not performed as it was considered invalid for anaerobic bacteria by the diagnostic laboratory. Broad-spectrum injectable antibiotic therapy (ceftiofur and tulathromycin at doses noted above) and injectable meloxicam (at doses noted above) were continued over a 2-wk period, and resulted in transient improvement in lameness and swelling. However, 3 mo following bacterial culture, the right tarsus became markedly erythematous and swollen, exhibited decreased flexion, and had moderate crepitus on palpation. Radiographs and CT performed at that time identified severe degenerative joint disease and fusion of the distal intertarsal and tarsometatarsal joints accompanied by significant periarticular soft tissue swelling (interpreted as fluid accumulation vs. soft tissue proliferation) and a dorsolateral draining tract (Fig. 1). Serial serum biochemistry panels revealed development and subsequent exacerbation of azotemia over 2 d that was not present 3 mo prior (Table 1), and renal failure was suspected as a result of prolonged NSAID therapy. A component of pre-renal azotemia could not be ruled out given the lack of urine specific gravity, and with the persistence of azotemia despite fluid therapy (0.6 L and 1.0 L lactated Ringer solution boluses administered SQ on 2021.08.25 and 2021.08.26, respectively), renal contribution to azotemia was prioritized. The gazelle became hyporexic, and given the progression of azotemia and ongoing lameness, the animal was euthanized and submitted for postmortem evaluation.
Computed tomographs in a gazelle with chronic actinomycotic osteoarthritis:
Blood chemistry renal parameters over 3 mo in a sand gazelle.
1.0 L of subcutaneous lactated Ringer solution was administered on 2021.08.26.
On postmortem examination ~72 h following euthanasia, the right tarsus was markedly swollen with 10–15 alopecic, dark-brown, firm, crusted, 1.0–2.5-cm nodules on the medial and lateral aspects of the right tarsus. Nodules had dark-red, central ulcers surrounded by a moderate amount of dried blood. Upon sagittal sectioning, the underlying subcutaneous tissues were markedly expanded by pale-tan, firm, fibrous connective tissue that surrounded numerous, firm, 3–7-mm nodules that had a pale-tan, firm, fibrous rim and a core of soft, inspissated, yellow-tan, opaque exudate, suggestive of pyogranulomas (Fig. 2). The right tibiotarsal joint capsule was diffusely and markedly thickened by fibrosis. Along the ventral aspect of the trochlea of the talus was a 5-mm focus of cartilage loss with extension into underlying cortical bone, with dark-red, irregular margins. The remainder of the postmortem examination identified dark-red, pinpoint, depressed foci on the right renal cortical surface, interpreted as renal cortical infarcts with atrophy and fibrosis. Within the thoracic cavity was ~100 mL of red-tinged, clear, watery fluid, and the lungs released a moderate amount of clear, red-tinged foamy fluid (interpreted as edema secondary to agonal aspiration). All sections of lung floated when placed in formalin. Samples of the affected joint capsule with overlying subcutis and skin, kidney, and lung were collected and processed routinely for microscopic evaluation.
Postmortem gross, cytologic, and histologic features of chronic Propionibacterium australiense osteoarthritis with secondary renal amyloidosis in a gazelle.
Cytology was performed via swab preparation of the right tarsal joint exudate, which yielded a markedly cellular sample that consisted of 94% degenerate neutrophils, 5% large foamy macrophages, and <1% small lymphocytes and plasma cells (Fig. 3). Numerous beaded filamentous bacteria that occasionally formed chains and mats were noted within macrophages and extracellularly. Rare multinucleate giant cells were both surrounded by and contained intracytoplasmic red blood cells. Large mats of bacteria were occasionally surrounded by neutrophils and macrophages. Occasional spindle cells with a large, round nucleus, stippled chromatin, and 1–3 nucleoli were noted and suspected to be reactive synoviocytes (Fig. 3).
Histologically, synovium, joint capsule, and overlying subcutis and dermis of the right tarsal joint were markedly expanded and effaced by numerous pyogranulomas composed of an outer layer of abundant epithelioid macrophages, plump, reactive fibroblasts, and granulation tissue. In the core of the pyogranulomas were large aggregates of degenerate neutrophils surrounding large mats of 2-μm beaded filamentous bacilli. Bacteria were admixed with Splendore–Hoeppli material (Fig. 4). Multifocally throughout all layers of granulation tissue were coalescent aggregates of lymphoplasmacytic inflammation and large foci of hemorrhage. The synovial lining and surrounding associated tendon were ulcerated and replaced by fibrin admixed with lymphocytes, plasma cells, and fewer neutrophils.
On histologic evaluation of the tarsal bones, inflammation similar to that seen in soft tissues also was present in the joint space, with multifocal disruption, erosion, loss of articular cartilage, and pannus formation (Fig. 5). Underlying cortical bone was occasionally necrotic and/or osteolytic, with scalloped margins and osteoclasts in Howship lacunae. Bacteria noted within the affected tissues were identified by additional staining as gram-positive, Wright-Giemsa–positive, and modified acid-fast (Fite)-negative. Our morphologic diagnosis was severe, chronic, pyogranulomatous arthritis, synovitis, and cellulitis with fibrosis, cartilage erosion and pannus formation, osteolysis, and osteonecrosis with intralesional filamentous bacteria.
Histologically, the renal cortical interstitium was moderately expanded by mature lymphocytes and plasma cells with admixed fibrous tissue; similar fibrous stroma multifocally mildly expanded glomerular capsules and tubular basement membranes. Hemorrhage occasionally dissected through the affected interstitium. Few (<5) <1-mm cortical scars subsequent to infarction were present, with occasional synechiation and sclerotic glomeruli present within areas of cortical scarring. Hyalinized eosinophilic deposits within the interstitium (Fig. 6) were amyloid, based on congophilia and apple-green birefringence when viewed under polarized light (Fig. 7). Histologic evaluation of the lungs was unremarkable.
Given the lack of speciation obtained from the antemortem bacterial culture, we pursued molecular characterization of the organism via PCR and sequencing. DNA was extracted from scrolls of formalin-fixed, paraffin-embedded (FFPE) tissue from both the initial antemortem skin biopsy as well as affected skin and subcutaneous tissue collected postmortem for PCR amplification of partial 16S rRNA and rpoB genes. DNA was extracted from FFPE scrolls (QIAsymphony DSP DNA mini kit, QIAsymphony SP DNA extraction robot; Qiagen) after overnight incubation of the scrolls at 56°C in 250 μL of ATL tissue lysis buffer (Qiagen). Three PCRs targeting the 16S gene of Propionibacterium genera were used in our study, including 2 quantitative PCRs (qPCRs) and a conventional PCR (cPCR), to amplify a larger portion of the gene (Table 2). Amplification assays were performed (CFX96 real-time detection system, C1000 thermal cycler; Bio-Rad) for qPCR or cPCR. Amplification reactions contained 1× SYBR Green supermix (Bio-Rad) for qPCRs or 1× MyTaq HS mix (Bioline) for cPCRs, 5 µL of DNA template, and primers at a final concentration of 0.4 µM. Thermocycler conditions for all assays are as follows: 98°C for 3 min, followed by 40 cycles at 98°C for 15 s, 67°C or 63°C for 15 s, and 72°C for 20 s. Positive controls consisted of plasmids synthesized by Integrated DNA Technologies. Negative controls included molecular-grade water and uninfected mammalian genomic DNA. Isolated amplicons were sequenced via Sanger sequencing (Azenta), and contigs were created with overlapping 16S rRNA amplicon sequences. DNA sequences were compared to other DNA sequences deposited in GenBank using BLASTn (https://blast.ncbi.nlm.nih.gov/Blast.cgi). A 906-bp segment of 16S rRNA (GenBank OQ850307) from the biopsy specimen had 99.7% identity match to 3 Propionibacterium australiense strains (LR134442.1, KX298726.1, NR_025076.1), including 2 strains identified in granulomatous lesions of cattle. A 154-bp segment of 16S rRNA from the postmortem tissues had 100% identity to the sequences of both the antemortem biopsy specimen as well as the previously noted P. australiense strains. Further, a 342-bp segment of the rpoB gene (OQ859038) from the biopsy specimen had 95.6% identity match to 2 P. australiense rpoB sequences (LR134442.1, LC187297.1). Primers designed to selectively amplify Actinomyces species 16S rRNA DNA failed to produce discrete DNA amplicons, and any DNA product from these PCRs failed any sequencing efforts (no results were obtained with Sanger sequencing). Collectively, these results supported that the bacterial colonies seen in both antemortem and postmortem tissues were P. australiense.1,11
PCR primer sequences and conditions for amplification of Propionibacterium australiense 16S rRNA and rpoB genes.
NA = not applicable, melt curve analysis not performed for conventional PCR; Tm = melting temperature.
Primers were utilized for attempted Actinomyces-specific PCRs, and failed to produce a discrete amplicon for sequencing.
P. australiense is a filamentous, gram-positive, acid-fast–negative, anaerobic bacterium in the order Actinomycetales and family Propionibacteriaceae. This organism was first described in 2000 and has only previously been reported as a cause of multisystemic granulomatous disease in cattle from Australia.4,5,15 Our case implies that this organism is present beyond Australia and can infect exotic hoofstock. There have been no reports of bone or joint involvement in bovine cases; however, skin involvement is common.4,15 Given the complete clinical history in our case, particularly the initial characterization of skin lesions prior to overt lameness and joint involvement, we suspect that skin infection preceded infection of deeper joint tissues. In humans, Propionibacteriaceae, including Cutibacterium acnes (formerly Propionibacterium acnes), are common skin commensals that are associated with dermatitis, acne formation, and in some cases, deep-seated opportunistic infections of underlying tissues, particularly in medical devices and orthopedic implants. 9 Given this association with related organisms, it is possible that the bacterial infection in our case was opportunistic after an initial traumatic event to the tarsal skin, and over the long clinical course of this patient ultimately resulted in deeper bacterial osteoarthritis and synovitis. Interestingly, our case had an apparent traumatic tarsal injury 4 y prior to the onset of significant lameness. Although thought to be an unrelated injury, the suspected conspecific trauma may have predisposed to, and/or initiated, infection. Furthermore, if percutaneous infection occurred at the time of this injury, pyogranulomatous inflammation may have been walled off and inapparent. However, joint fluid was not analyzed at the time of the initial injury.
Based on the initial antemortem culture results and histologic and histochemical features of the organism, infection with Actinomyces sp. was originally considered as the most likely cause of the tarsal lesion in this gazelle. However, the sequence identity results found in multiple tissue samples are diagnostic for P. australiense infection.1,11 We have several theories that may explain these discordant results between culture and molecular data. First, the gazelle may have been coinfected with both bacteria, and the 2 tests preferentially identified separate organisms. To this point, it is possible that bacterial isolation and identification may have been impeded by concurrent antimicrobial therapy at the time of synovial fluid collection, further contributing to discordant results. Second, the genus identification obtained via MALDI-TOF may have been inaccurate. Accuracy rates in correct identification of Actinomyces and Propionibacterium organisms at the genus level have been reported at 81% and 91%, respectively 20 ; hence, misidentification is possible, particularly if a coinfection is present. A third possibility would be that the PCR results represent a contaminant in the tissue. However, this is unlikely given successful amplification of DNA of multiple genes from multiple tissue samples that represent both superficial and deep affected tissues obtained at different times in the course of disease. Fourth, it is possible that the organism infecting this gazelle is a novel species of bacteria, therefore yielding discordant results; however, this is also unlikely given that the levels of sequence identity for P. australiense found in both amplified bacterial genes is considered sufficient for speciation.1,11
The primary differential diagnosis in our case, Actinomyces, is in the same family as Propionibacterium (Actinomycetales) and has indistinguishable histologic and histochemical features from Propionibacterium. Actinomyces organisms have been implicated as a cause of osteomyelitis, specifically “lumpy jaw,” in exotic artiodactyls. 17 Microscopically, this lesion in bone is similar to our observation in the tarsus from our case, although the lesion has not been described previously affecting joints or bones of the limbs in exotic artiodactyls. Other bacteria implicated in osteomyelitis in exotic hoofstock include Fusobacterium necrophorum and Trueperella pyogenes. Although all artiodactyls are reported susceptible to F. necrophorum, springbok, impala, and roan antelope are especially susceptible.12,17 Gross lesions most commonly consist of necrotizing pododermatitis that can extend to deeper tissues with chronicity, causing cellulitis, arthritis, tenosynovitis, and osteomyelitis.12,17 Oral lesions can also manifest with this entity and are often mixed anaerobic infections, particularly with T. pyogenes. 17 Historical reports of osteoarthritis in several different exotic ungulates have been associated with isolated infection with T. pyogenes.14,19,24,25
Chronic or recurrent T. pyogenes infection has been associated historically with renal medullary amyloidosis in gazelles. 23 Renal medullary amyloidosis is prevalent in, and the second most common cause of death in, sand gazelles. We suspect that the amyloid deposits in our case were reactive amyloid A (AA amyloid) secondary to the chronic inflammation attributed to P. australiense infection. However, as we did not confirm AA versus AL amyloid, we cannot rule out renal amyloidosis as incidental and unrelated to chronic infection in our case.
Infectious arthritis is well-documented in domestic small ruminants; beyond F. necrophorum and T. pyogenes, other identified bacterial entities include Chlamydia pecorum, Erysipelothrix rhusiopathiae, Escherichia coli, Histophilus somni, Staphylococcus spp., Streptococcus spp., and Mycoplasma spp. 18 Bacterial arthritis may be caused by contamination of a joint via direct trauma, extension from peripheral infection, or iatrogenically by arthrocentesis. 18 Small ruminant lentiviruses including Visna-maedi virus of sheep and caprine arthritis encephalitis virus are well-documented viral causes of arthritis, with rare reported adaptation to exotic hoofstock populations.13,21
Typical treatment modalities for bacterial arthritis include appropriate antibiotic therapy following culture and susceptibility testing, joint lavage to remove inflammatory exudate, and NSAIDs to decrease inflammation and pain. 18 Similar treatment modalities were used in our case and further supplemented with joint debridement, PRP, and stem cell therapy. Use of PRP, which promotes chondrogenesis and cartilage regeneration, 10 has increased in treatment of joint disease in veterinary species over the last decade. 8 Further, mesenchymal stem cells (MSCs) have the capacity to differentiate into and proliferate as various cell lineages, have immunogenic properties, and are easily obtainable from peripheral blood or bone marrow. Intraarticular transplantation in cases of arthritis promotes MSCs to mobilize into the injured tissue and promote repair.2,7 Given the patient’s significant, albeit transient, improvement in clinical signs following use of PRP and stem cell therapy, these may serve as viable additional treatment modalities for bacterial osteoarthritis in exotic hoofstock. Notably, localized treatment modalities were only applied to the right tarsus, and it is unknown the extent to which the lateral claw injury may have contributed to and/or further exacerbated clinical signs.
The discordant results between molecular analysis and antemortem bacterial culture and the lack of speciation performed with the antemortem culture impeded full characterization of the bacteria in our case. Although cytologic and histologic evaluation of joint fluid and tissues can identify filamentous bacteria, our case underscores that several filamentous bacterial genera (e.g., Actinomyces, Propionibacterium, and likely other actinomycetes) may be indistinguishable by light microscopy, and require further laboratory testing (i.e., culture with speciation, or PCR) to classify.
Footnotes
Acknowledgements
We thank and acknowledge the diagnostic and treatment expertise contributed by the staff at the Columbus Zoo and Aquarium and the contribution of the staff of the OSU CVM Histology Laboratory. We also thank and acknowledge the diagnostic expertise of the North Carolina State University College of Veterinary Medicine Vector-Borne Disease Diagnostic Laboratory; specifically, we acknowledge the technical efforts of Charley V. Hall and Jaidyn F. Dillon.
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors declared that they received no financial support for their research and/or authorship of this article.
