Fatal non-traumatic gas gangrene caused by Clostridium perfringens type A in a Siberian Husky dog

Abstract

An 8-y-old, castrated male Siberian Husky dog was admitted to an emergency clinic with acute collapse and severe swelling of both forelimbs, ventral thorax, and axillary region. The clinical assessment, with laboratory tests and radiologic investigation, confirmed severe subcutaneous emphysema and multi-organ failure. The animal died while receiving emergency treatment. On postmortem examination, Clostridium perfringens was isolated from the subcutaneous fluid and the effusion from the thoracic and abdominal cavities. Relevant histopathology findings included subcutaneous emphysema and multi-organ perivascular and intravascular, intralesional myriad 0.5–3-µm gram-positive rod bacteria, with no associated inflammation. Whole-genome sequencing and phylogenetic analysis identified C. perfringens type A. Virulence genes detected included cpa (alpha toxin), cadA (v-toxin), colA (collagenase A), nagH (hyaluronidase), nanH, nanI, nanJ (sialidases), and pfoa (perfringolysin). These virulence genes have previously been reported to act synergistically with alpha toxin in C. perfringens–mediated gas gangrene.

Gas gangrene (clostridial myonecrosis) has been described as a highly lethal infection of soft tissue, caused by Clostridium species, which usually arises in traumatized tissue but can also arise spontaneously. ¹⁵ Clostridium species are gram-positive, spore-forming, ubiquitous obligate anaerobic bacilli bacteria, which are often associated with severe-to-fatal disease. ⁵ Clostridium perfringens has been reported more commonly in human than in animal cases of gas gangrene (previously referred as malignant edema).^8,10 In animals, gas gangrene has a higher incidence in ruminants ¹³ and horses,^1,16 with scattered reports in other species, including pigs, ¹⁷ cats, ²² and dogs.^20,23

Despite the numerous cases of gas gangrene reported in other mammals, there are no reports of generalized fatal gas gangrene in dogs, to our knowledge. We describe here a case of fulminant subcutaneous emphysema, and rapid death with no history of trauma that utilized pathology and microbiology methods to identify Clostridium perfringens as the potential disease-causing agent.

An 8-y-old, castrated male Siberian Husky was presented to a veterinary emergency clinic acutely obtunded in lateral recumbency and with severe pain. There was no history of access to toxins, and the diet was a combination of raw mincemeat, dry pet food, and occasional bones. The dog ate dinner at ~7 pm, went for a 5-km run at 8:30 pm, started yelping at 1 am, and collapsed within minutes. On admission at 2:15 am, the patient had pale mucous membranes, a temperature of 41.2°C, heart rate 168 bpm, mildly muffled heart sounds with normal respiratory rate and effort, and SpO₂ of 97%, with diffuse abdominal pain, subcutaneous emphysema of the proximal forelimbs, axilla, and ventral thorax. The results of in-house laboratory investigations revealed mild hemoconcentration (hematocrit 0.61 L/L [RI: 0.37–0.54 L/L], hemoglobin 192 g/L [RI: 122–184 g/L]), elevated glucose (10.7 mmol/L [RI: 3.9–8.0 mmol/L]), and elevation in liver enzyme activities (alanine aminotransferase 431 U/L [RI: 10–125 U/L]; gamma-glutamyl transferase 75 U/L [RI: 0–11 U/L]). Pain medication and fluid therapy were started, and the patient was transferred to imaging. Radiographs showed marked subcutaneous emphysema around the axillary region and ventral thorax, with no communication to the thoracic cavity or opening of the skin. While on the radiographic table, the patient’s condition deteriorated rapidly, and he started agonal breathing. Because of respiratory failure, the patient was intubated, CPR and adrenaline were administered. The combined therapy had no effect, and the patient died within 1.5 h of arrival. To determine the cause of death, the owners requested a postmortem examination. The animal was refrigerated by the pet crematorium prior to transport to our facility. The postmortem examination was conducted immediately upon arrival and performed within 36 h after the patient’s death.

The postmortem examination revealed advanced putrefaction and autolysis, with severe generalized subcutaneous crepitation and edema (Fig. 1A, 1B). The internal examination revealed emphysematous myositis, effusive peritonitis and pleuritis, entrapped air predominantly within the adipose tissue, gastric and intestinal wall, liver, kidneys, and all internal organs except for the brain, with an absence of associated inflammation. There were no signs of trauma or perforations of the skin or gastrointestinal tract. Samples from the main organ systems were processed routinely and sections stained with H&E. Brown–Hopps Gram stain was performed on all of the samples collected.

Figure 1.

Gross and histologic findings in an 8-y-old Siberian Husky dog with gas gangrene. A. The forelimb (hair clipped) is severely swollen; a longitudinal dissection shows the expanded subcutis with entrapped air (interpreted as gas bubbles). The head is to the right of the image. B. A close-up of the forelimb subcutis showing separation to the underlying fascia by clear red-to-purple-tinged, gelatinous-to-watery fluid containing myriad <0.5-mm diameter gas bubbles. C. Liver with severe hepatocellular vacuolation, dissociation of hepatic cords, and autolysis. H&E. Bar = 10 μm. D. Liver with myriad extracellular straight rod bacteria with blunt ends, 0.5 μm wide × 1–3 μm long. Brown–Hopps Gram stain. Bar = 20 μm. Inset: higher magnification of the bacteria. Brown–Hopps Gram stain.

Histopathology demonstrated a striking number of gram-positive 0.5–1.2 × 2–5-µm straight rod bacteria. These were observed in and around blood vessels, admixed with variably sized, up to 2-mm diameter, well-demarcated clear spaces, interpreted as gas bubbles, that compressed adjacent parenchyma with absent inflammatory cell infiltrates within the subcutaneous tissue, gastrointestinal tract, liver, kidneys, thyroid glands, spleen, and heart. Hepatic central veins were dilated, and hepatic cords were disrupted, with loss of differential staining and indistinct or absent nuclei (Fig. 1C, 1D).

Samples of subcutaneous tissue from the forelimbs, thoracic fluid, and peritoneal fluid were cultured at 37°C aerobically (sheep blood agar [SBA], Columbia agar plates, MacConkey agar3; Thermo Fisher) and anaerobically (SBA). The aerobic culture yielded a light (1+) mixed growth. Enterobacter cloacae, identified with a Microbact 24E system (Thermo Fisher), was the predominant growth. Anaerobic conditions were generated (Oxoid AnaeroGen 2.5-L sachets; Thermo Scientific), and jar and plates were incubated overnight at 37°C. The anerobic culture yielded a light growth of E. cloacae (identified as above) and a heavy (3+) grow of a large beta-hemolytic gram-positive bacilli, initially identified as a suspect C. perfringens in-house using phenotypic methods, including positive growth on C. perfringens selective agar (Perfringens agar base OPSP with selective supplements A & B; Thermo Fisher), and lipase-negative and lecithinase-positive results on egg yolk agar; egg yolk agar was manufactured in-house using blood agar base 2 with the addition of egg yolk emulsion (Thermo Fisher). Identification of C. perfringens was confirmed via matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (Bruker Biotyper) at the Department of Agriculture and Fisheries Biosecurity Sciences Laboratory, Coopers Plains, Queensland, Australia.

The bacterial isolate (B20-0404) was submitted to the Public Health Microbiology laboratory at Queensland Health Forensic and Scientific Services for whole-genome sequencing and phylogenetic analysis. Sequences of the isolate were generated (NextSeq 500 genome sequencing platform; Illumina). Sequences were assembled into contigs (Spades v.3.12; GenBank PRJNA781736).

Multilocus sequence typing (MLST) and ribosomal MLST were performed using the schemes hosted on PubMLST (https://pubmlst.org/organisms/clostridium-perfringens; https://pubmlst.org/rmlst/ [accessed 2020 Oct 20]). Toxinotype ⁶ and virulence gene detection were analyzed by in silico PCR using Geneious R11. Core genome MLST (cgMLST) analysis was performed on the assembled contigs using an ad hoc scheme developed in SepSphere+ (v.6.0.0; Ridom)

The isolate was identified as a C. perfringens type A, MLST sequence type (ST)102. Virulence genes detected included cpa (alpha toxin), cadA (v-toxin), colA (collagenase A), nagH (hyaluronidase), nanH, nanI, nanJ (sialidases), and pfoa (perfringolysin). The genes cpe (enterotoxin), cpb (beta toxin), cbp2 (beta 2 toxin), etx (epsilon toxin), iap (iota toxin), netA (putative toxin), netF (pore-forming toxin), and netG (putative toxin) were not detected. Analysis of this isolate’s genome sequences by ribosomal MLST against all C. perfringens sequences in PubMLST found the closest match to CP15 (accession NZ_CP019468), a chicken isolate from the United States. Like the isolate identified here, CP15 is ST102, and was the only ST102 isolate in the PubMLST isolate database.

The genome sequence for CP15, as well as 19 other publicly available C. perfringens sequences were downloaded from GenBank and analyzed by cgMLST along with this isolate (Suppl. Table 1). By cgMLST analysis, B20-0404 had the highest level of relatedness to CP15, compared to the other strains included in the analysis (Fig. 2).

Figure 2.

Neighboring-joining tree built using core genome MLST allele differences. The length of branches represent distance between groups as indicated by the scale bar. B20-0404 is highlighted in red.

The cpaA gene was detected from the C. perfringens isolate grown from our case, along with the pfoA gene. ¹⁹ The pfoA gene encodes for perfringolysin, a pore-forming toxin that has been reported to act synergistically with alpha toxin (CPA) in C. perfringens–mediated gas gangrene. ² Perfringolysin causes macrophage cytotoxicity in the early stages of infection, associated with thrombosis and a decrease in inflammation at later stages, ¹⁵ as do the sialidases encoded by nanI and nanH. ⁴ The toxins encoded by the netE, netF, and netG genes have been described in a case of canine disease, ⁷ but these genes were not detected in our case.

C. perfringens CPA, a necrotizing toxin produced by all strain types of C. perfringens, is involved in gas gangrene of domestic animals, and is the major toxin associated with the pathogenesis of gas gangrene in humans. ²⁵ Gas gangrene caused by C. perfringens type A has been reported mostly in ruminants, pigs, and horses. ¹⁵

CPA is a zinc-dependent phospholipase C that causes extensive local tissue destruction by degrading phosphatidylcholine and sphingomyelin components of the plasma membrane of cells, which affects host signaling and leads to tissue necrosis, thrombosis, and generally absence of leukocytes at the infection site, ⁷ as noted in our case. However, the role of CPA in animals is unknown, although it is likely an important virulence factor in animals with gas gangrene. ¹⁵

Notably in our case, there was no history of trauma or perforating lesions that could have led to the extensive emphysematous lesions. Another notable feature of our case was the extremely rapid clinical course, given that the animal was normal when exercising with the owner at 8:30 pm, but in extremis at 1 am, only 5 h later.

Fatal non-traumatic sepsis caused by C. perfringens was described in a Labrador retriever that died days after not responding to antimicrobial treatment for fever of unknown origin. ²⁰ A 10-mo-old Rottweiler died within 24 h of sudden swelling of the hindlimbs, with subsequent isolation of Clostridium septicum. ¹⁸ Other fatal cases reported in dogs have involved death as a result of hemorrhagic gastroenteritis that was caused by C. perfringens. ²¹

Cases of non-fatal myositis in dogs are unusual. A German Shepherd dog acquired C. perfringens following a contaminated intramuscular injection 2 d prior, developed acute non-weightbearing lameness and severe emphysematous myositis, but survived following extensive medical treatment. ²³

In people, reports of fatal non-traumatic gas gangrene caused by C. perfringens include findings of elevated body temperature, with signs of multi-organ failure including increased liver enzyme activities, and autopsy findings that include sponge-like gas collection appearance in solid organs, ²¹ similar to findings noted in our case. Most of the human cases associated with gas gangrene involve pre-existing conditions, such as hepatic cirrhosis ¹¹ and diabetes, ¹⁴ which can lead to emphysema of the limbs. ⁹

Reports of other bacterial species (non-clostridial bacteria) causing gas gangrene in humans have been reported, including Klebsiella pneumoniae and mixed infections. ²⁴ In addition, necrotizing fasciitis caused by Lancefield group A streptococci, which occurs in humans, or staphylococci are differential diagnoses for clostridial gas gangrene. ¹² However, in our case, there was a heavy growth of C. perfringens type A, and a light growth of Enterobacter spp. Enterobacter spp. have not been implicated in gas gangrene, to our knowledge, but it is not uncommon for other aerobic, and facultative anaerobic and anaerobic bacteria, to be mixed with the Clostridium spp. ³

To our knowledge, non-traumatic fatal gas gangrene (clostridial myositis) has not been reported previously in a dog that died suddenly after peracute collapse and development of subcutaneous emphysema. We speculate that C. perfringens type A was associated with the gas gangrene of this animal. The possibility of postmortem changes and clostridial overgrowth would not explain the fulminant clinical course and subcutaneous emphysema in our case. Exploration of similar cases, including detailed genetic analysis of potential causal organisms, is strongly warranted to definitively prove an association with C. perfringens type A as a cause of non-traumatic fatal gas gangrene in dogs.