Isolation and characterization of atypical Listeria monocytogenes associated with a canine urinary tract infection

Abstract

Listeria monocytogenes, a well-described cause of encephalitis and abortion in ruminants and of food-borne illness in humans, is rarely associated with disease in companion animals. A case of urinary tract infection associated with an atypical, weakly hemolytic L. monocytogenes strain is described in a diabetic dog. The serotype of the L. monocytogenes isolate was determined to be 1/2a (3a), with the multilocus genotyping pattern 2.72_1/2a. A nucleotide substitution (Gly145Asp) was detected at residue 145 in the promoter prfA region. This residue is within the critical helix-turn-helix motif of PrfA. The source of the L. monocytogenes strain remains unknown, and the dog recovered after a 4-week course of cephalexin (30 mg/kg orally twice daily).

A 12-year-old, 35-kg male neutered Labrador Retriever was referred to the North Carolina State University Veterinary Hospital (NCSU-VH; Raleigh, North Carolina), for evaluation of recently diagnosed diabetes mellitus. Approximately 3 weeks prior to presentation, the dog was presented to the referring veterinarian for facial swelling. He was administered prednisone (1 mg/kg/day for 3 days, then 0.5 mg/kg/day for 3 days, and then 0.5 mg/kg every other day for 3 days) for a potential allergic reaction to an insect bite. One week later, the owner reported that the dog became polydipsic and polyuric in spite of discontinued prednisone. Soon thereafter, the dog became lethargic, and displayed hindlimb weakness and diffuse muscle wasting. During that 1-week period, the dog lost 2.2 kg despite a normal to slightly reduced appetite. When evaluated by the referring veterinarian, 4 days prior to referral to NCSU-VH, a serum biochemistry panel revealed hyperglycemia (483 mg/dL, reference interval [RI]: 70–143 mg/dL) and increased alkaline phosphatase (ALP; 473 U/L, RI: 23–212 IU/L) and alanine aminotransferase (ALT; 134 U/L, RI: 10–100 IU/L). Glycosuria, ketonuria, and pyuria were noted on urinalysis. A urine cortisol:creatinine ratio performed at that time was increased at 197. The owner reported that the dog had a 1-year history of hair loss on the tail that had progressed to its hindlimbs.

At the time of presentation to the NCSU-VH, the dog was noted to be quiet and mildly dehydrated. The dog was normothermic (T = 37.9°C), had a normal heart rate (70 beats/min), and was breathing normally. The abnormalities detected by physical examination included a mildly pendulous abdomen with suspected hepatomegaly, muscle wasting on both hindlimbs, and dermatologic changes, including a dull hair coat, and thinning of hair coat along the dorsal tail, caudal thighs, and around the neck. Generalized scaling, papules, and pustules, as well as multiple drying epidermal collarettes were present on the trunk and ventral abdomen. Multifocal alopecia was present on the dog’s face.

As part of the patient’s initial assessment, a complete blood cell count (CBC), serum biochemistry, and urinalysis were performed. CBC abnormalities included neutrophilia (11.0 × 10⁹/L; RI: 2.8 - 9.1 × 10⁹/L), lymphopenia (0.41 × 10⁹/L; RI: 0.59 - 3.30 × 10⁹/L)), monocytosis (1.37 × 10⁹/L; RI: 0.08 – 0.85 × 10⁹/L), and thrombocytosis (520 × 10⁹/L; RI: 190 – 460 × 10⁹/L). Serum biochemistry values revealed increased ALP (448 U/L, RI: 16–140 U/L), increased ALT (100 U/L, RI: 12–54 U/L), hypochloremia (99 mmol/L, RI: 108–112 mmol/L), hyperkalemia (5.4 mmol/L, RI: 4.0–5.3 mmol/L), hyperglycemia (24 mmol/L; RI: 3.9 – 7.3 mmol/L), and hyperlipidemia (649 IU/L, RI: 12–147 IU/L). Fructosamine measurement (489 μmol/L, RI: 260–375 µmol/L) was consistent with prolonged hyperglycemia. Urinalysis revealed glycosuria (4+), ketonuria (2+), and trace bacteriuria with a urine specific gravity of 1.024. No abnormalities were present on thoracic radiographs. Abdominal ultrasound revealed bilateral adrenomegaly (left caudal pole 1.1 cm; right caudal pole 1.0 cm) with a hyperechoic nodule in the right cranial pole, hepatomegaly, and diffuse hepatopathy. The bladder wall was diffusely thickened, measuring up to 0.34 cm, with an irregularly marginated mucosa.

Specimens were collected from the pustules and urine by cystocentesis and submitted to the NCSU-VH Clinical Microbiology Laboratory for routine aerobic culture and susceptibility testing. Skin specimens were submitted on a swab, and the urine sample was submitted in transport medium. ^a Samples were plated onto Columbia agar with 5% sheep blood ^b and MacConkey ^b agar and incubated at 37°C for up to 48 h under 5% CO₂. Urine samples were plated with a calibrated 10-µL loop so culture results could be quantified. Isolated colonies from each sample type were presumptively identified by Gram stain and biochemical profiling. ^c Aerobic culture of the pustules yielded 2 colony types: 3+ (moderate growth) Staphylococcus pseudintermedius and 1+ (light growth) Bacillus species. The urine culture yielded >100,000 colony forming units (CFU)/mL of Listeria monocytogenes. Based on culture and sensitivity results from the skin and urine, the dog was treated for both infections with cephalexin (30 mg/kg orally twice daily). In addition, the pyoderma was treated with a chlorhexidine shampoo and a moisturizing spray. ^d Therapy for the diabetes was instituted with 0.25 IU/kg neutral protamine Hagedorn insulin ^e given subcutaneously twice daily.

The dog was seen 10 days after discharge for a recheck examination at the NCSU-VH. At that time, it was assessed to be persistently polyuric and polydipsic despite an improved energy level. A second urine culture, collected as before, yielded a negative aerobic culture. The pustules, papules, and epidermal collarettes had resolved. A 4-week course of cephalexin was continued. A urine culture performed 7 days after the end of the antibiotic course confirmed resolution of the infection. Thereafter, care of the dog was taken over by the referring veterinarian. Approximately 6 months after the initial presentation to our hospital, the patient’s diabetes mellitus was well controlled. His body weight had increased by 1.2 kg, and fructosamine measurement was 402 µmol/L (RI: 177–314 μmol/L). In spite of this improvement, the patient remained mildly polydipsic. A low-dose dexamethasone suppression test performed at that time was consistent with a diagnosis of pituitary-dependent hyperadrenocorticism.

Because of the unexpected finding of L. monocytogenes in the urine, the strain obtained from this dog was further characterized by serotyping, multilocus genotyping (MLGT), virulence testing using a Galleria mellonella insect model, and evaluated for resistance to heavy metal compounds and the quaternary ammonium disinfectant benzalkonium chloride. The strain was noted to have unusually reduced hemolytic activity on 5% Columbia sheep blood agar, as compared to typical L. monocytogenes. Zones of hemolytic activity could not be detected following 48-h growth at 37°C; however, when the colonies were removed, extremely weak hemolysis was observed. Using a previously described multiplex PCR, ⁶ the serotype of the L. monocytogenes isolate was determined to be 1/2a (3a). The MLGT type, determined by previously reported methods, ²⁰ was 2.72_1/2a. The insect G. mellonella has previously been validated as a model system to assess the virulence potential of Listeria species. ¹⁵ Briefly, Galleria larvae were injected with 10⁶ CFU/mL of the L. monocytogenes strain obtained from this patient, and the same concentration of a L. monocytogenes serotype 4b strain, previously isolated from canine blood, ¹⁷ as well as a reference strain from a human listeriosis outbreak. ³ The percentage of surviving larvae was calculated daily for 7 days to determine virulence. The strain obtained from this case was as pathogenic as a previously characterized strain implicated in canine listeriosis and exhibiting normal hemolytic activity. ¹⁷ Ninety to 95 percent of the larvae had died by day 5 postinoculation for all 3 tested strains (data not shown). When the strain was assessed for resistance to cadmium, arsenic, and benzalkonium chloride, ¹¹ no resistances were detected (data not shown).

The markedly reduced hemolytic activity of the strain obtained from this case prompted us to analyze the sequence of prfA, which encodes a major regulator of virulence genes, including the hemolysin gene hly (listeriolysin O). ¹² The dual prfA promoters previously described ⁸ were intact in this strain; however, a nucleotide substitution (glycine to aspartic acid) was detected at residue 145. This residue is within the critical helix-turn-helix motif of PrfA, and interestingly, an alternative substitution at this residue (Gly145Ser) has been found to result in a hyper-hemolytic phenotype and overexpression of hly and other virulence genes. ¹⁸ Further analysis is needed to determine whether the Gly145Asp substitution in this strain was indeed responsible for the reduction in hemolytic activity.

L. monocytogenes is a frequently documented pathogen in pregnant, elderly, and immunocompromised people and in ruminants; it is also occasionally implicated in illness in various other animal species. ⁵ However, there are few documented cases of canine listeriosis.^{4,10,13,14,17,19}

As with most reports of listeriosis in companion animals, the source of infection in this case remains unknown. L. monocytogenes contaminates food largely through contamination of the environment and equipment at food processing facilities. In 2014, whole genome sequences of 4 L. monocytogenes strains from pet food were reported; all 4 strains were serotype 1/2a and 1 strain was closely related in whole genome sequence content to one of the serotype 1/2a strains implicated in a major listeriosis outbreak involving cantaloupe in 2011. ² However, this latter strain from pet food had a premature stop codon in the key virulence regulator prfA, resulting in a truncated PrfA and also in complete absence of hemolytic activity ² and is thus presumed to be avirulent. A 2014 report documenting prevalence rates of L. monocytogenes from various types of dog food found that whereas 16% of raw or exotic foods were positive, none of the commercial dry or semimoist diets tested positive for the organism. ¹⁶ The dog in our study was reported to receive a commercial dry diet but the extent to which he may have consumed additional food items remains unknown. Ingestion of L. monocytogenes–harboring feces or tissues from other animals appears to be unlikely associated with transmission to this dog, but cannot be excluded. Though some reports from outside the United States have suggested high exposure rates in the canine population based on listeria serology,^1,9 fecal shedding of L. monocytogenes appears to be present in no more than 1% of the healthy canine population. ²¹ Moreover, the dog in our report was the sole pet in the household. Regardless of the L. monocytogenes source, age and impaired systemic immunity are likely risk factors for infection. These are known risk factors for human listeriosis, and were also noted in a canine listeriosis case that we recently described. ¹⁷ Impaired systemic immunity has been suggested as a risk factor in at least 1 other canine case report. ¹⁴ This is relevant because the dog reported herein was diagnosed with diabetes mellitus as well as hyperadrenocorticism, 2 endocrinopathies that are known to decrease systemic immunity.

It is interesting to note that the dog in our report did not display any signs suggestive of a lower urinary tract infection (UTI). This may be in part explained by the fact that dogs with hyperadrenocorticism often fail to generate an inflammatory response to UTIs and thus fail to develop the classic signs of pollakiuria and stranguria. ⁷ The lack of clinical signs may also be related to specific virulence attributes of the L. monocytogenes strain. Although we observed some pathogenic potential in the G. mellonella model, the strain appears to be weakly hemolytic. Strains with the 1/2a serotype, like the one reported herein, are represented at high frequencies among L. monocytogenes isolates from foods and food processing plants. ²⁰ More than 50% of serotype 1/2a strains may have attenuated virulence caused by the presence of a truncated form of the InlA protein that is responsible for host cell invasion. ²⁰ However, MLGT analysis indicated that the strain reported herein does not harbor any known InlA truncation mutations.

The antibiotic treatment decision in our case was based in part on the sensitivity of L. monocytogenes to β-lactams as well as the culture and sensitivity panel from one of the dog’s pustules. Although shorter antibiotic courses are usually recommended for treatment of UTIs, according to published guidelines, ²² patients with comorbidities such as diabetes mellitus (as was the case in this dog) should be treated for extended periods. Consequently, the dog received 1 full month of the antibiotic initially prescribed for the skin infection. Clearance of the infection was confirmed with a urine culture 1 week after the end of antibiotic therapy. Given the paucity of reports of listeriosis in companion animals, the zoonotic potential of these cases, either directly or via contamination of human food through contact with urine or feces, remains unknown. Nonetheless, caution should be exercised in these cases, and owners should be warned to limit their exposure to the pet’s stool and urine until there has been documentation of clearance of the infection.