Disseminated Protothecosis Associated With Diskospondylitis in a Dog

Introduction

Protothecosis is an uncommon disease caused by several species of saprophytic, achlorphyllous algae belonging to the genus Prototheca.^1,2 The causative organisms exist ubiquitously within environmental niches of most major continents and have been implicated in localized and systemic disease in both animals and people worldwide.^1–6 Dermatologic manifestations of infection predominate in most species and are often associated with a more favorable prognosis.^2–6 In contrast, more severe, life-threatening cases of disseminated protothecosis appear to occur with much greater frequency in dogs and are believed to be associated with immune compromisation.^2–8 Signs of dissemination are generally referable to the gastrointestinal, ocular, and neurologic systems; however, lesions involving other organ systems have also been reported.^1,3,4 In the present case report, the authors describe a case of disseminated protothecosis associated with diskospondylitis and possible concurrent candidiasis in a dog presenting with primary ocular and neurologic symptoms. To the authors' knowledge, this report represents the first documented case of protothecal diskospondylitis and possible coinfection with Candida in the veterinary literature.

Case Report

A 6 yr old female Labrador retriever weighing 34 kg was referred to the Internal Medicine service at Texas Gulf Coast Veterinary Specialists for sudden-onset blindness that had rapidly progressed to include difficulty walking and a head tilt. Three days prior to the dog’s presentation to the veterinary specialists, therapy with oral doxycycline (5.88 mg/kg q 12 hr) and prednisone (1.17 mg/kg q 12 hr) had been initiated after an evaluation with a veterinary ophthalmologist revealed chronic panophthalmitis oculus dexter and retinal detachment associated with hemorrhage oculus sinister (OS). Allowed to roam freely outdoors with access to two nearby ponds, the dog was not on a heartworm preventative and had historically been treated with repeated courses of corticosteroids and antibiotics for recurrent bouts of pyoderma and otitis externa. Having previously whelped a litter of 11 puppies several years earlier, the dog was last reported to experience a heat cycle approximately 3 mo prior to her presentation.

At the time of referral, the dog was nervous, blind, but otherwise mentally appropriate and responsive to stimuli. She was mildly overweight with a body condition score of 6 out of 9 and normal vital parameters. Preliminary systolic blood pressure, determined via Doppler, was markedly elevated (mean, 220 mm Hg). Ocular examination revealed bilateral mydriasis with absent menace, dazzle, and direct and indirect pupillary light reflexes. Intraocular pressures measured 3 mm Hg oculus dexter and 5 mm Hg OS and cursory fundoscopic examination noted bilateral retinal detachment. Neurologically, the dog exhibited signs referable to a central vestibular lesion as determined by a general tendency to veer to the left while walking, a mild left-sided head tilt, positional rotary nystagmus, and left-sided hemiparesis. Mild to moderate hyperpathia was also encountered on spinal palpation at the level of the first and second lumbar vertebrae. Otoscopic examination, thoracic auscultation, abdominal palpation, and digital rectal examination were all unremarkable. Based on those findings and the dog's clinical history, a systemic vasculopathy was initially suspected, and differential diagnoses commonly associated with systemic hypertension were considered most likely.

Initial diagnostics included a hemogram, serum biochemical profile, urinalysis, cage-side serology for Anaplasma spp., Borrelia burgdorferi, Ehrlichia spp., and Dirofilaria immitis^a, abdominal and thoracic radiographs, and an abdominal ultrasound. Hematology subsequently revealed leukocytosis (24,300/µL; reference range, 4000–15,500/µL), neutrophilia (21,384/µL; reference range, 2060–10,600/µL), and monocytosis (972/µL; reference range, 0–840/µL). Serum biochemical analysis showed elevations in both total protein (7.5 g/dL; reference range, 5–7.4 g/dL) and globulins (4.4 g/dL; reference range, 1.6–3.6g/dL) as well as a low amylase (240 IU/L; reference range, 290–1125 IU/L). Cage-side serology was negative for the four organisms assayed, and no significant abnormalities were reported on routine urinalysis. Thoracic and abdominal radiographs revealed a mild, diffuse, interstitial lung pattern and bony lysis with adjacent sclerosis of the vertebral endplates of the first and second lumbar vertebrae (Figure 1). Aside from mild, equivocal, small intestinal ileus, sonographic evaluation of the abdomen was unremarkable.

View Figure

• Download as Powerpoint
• Download Figure

With those new findings, a disseminated infectious etiology causing diskospondylitis was suspected, prompting the dog's hospitalization for parenteral antibiotic administration, blood pressure measurement, pain management, and further diagnostics. Prior to initiating any further therapy, additional laboratory diagnostics included blood and urine cultures, serum protein electrophoresis, serology for Brucella canis and Aspergillus spp., urine antigen testing for Blastomyces dermatitidis and Histoplasma capsulatum, and an extended vector-borne polymerase chain reaction panel^b. Serologic testing for Brucella and Aspergillus was performed by immunofluorescent antibody and agar gel immunodiffusion, respectively, whereas a quantitative sandwich enzyme immunoassay^c was used for detection of Blastomyces and Histoplasma. An echocardiogram to evaluate the integrity of the heart valves was also performed but failed to identify any significant pathology. Needle aspiration of the dog's spinal lesion for culture and cytology as well as a cerebrospinal fluid (CSF) tap and MRI were recommended but declined by the owner.

Over the next 3 days, the dog was maintained on IV fluids^d, parenteral clindamycin^e (11 mg/kg IV q 12 hr) and enrofloxacin^f (10 mg/kg IV q 24 hr), and the oral doxycycline that was previously prescribed. Oral tramadol^g (2.2 mg/kg q 8 hr) and enalapril^h (0.44 mg/kg q 24 hr) were also used to help manage the dog's spinal discomfort and hypertension, respectively. Within 24 hr of starting those medications, the dog's hypertension was noted to significantly improve, measuring on average 213 mm Hg on day 1, 161 mm Hg on day 2, 133 mm Hg on day 3, and 142 mm Hg on day 4. Because of the concurrent improvement in the dog’s discomfort, it was concluded that the high blood pressure readings initially recorded were most likely pain-induced. Brucellosis was ruled out on day 2 after the immunofluorescent antibody assay was reported as negative. Early on day 4, topical 1% prednisolone acetateⁱ (1 drop oculus uterque q 6 hr) was started after observing increased scleral injection bilaterally. Despite that new observation, a persistent lack of vision, and ongoing neurologic deficits, the dog was discharged later that same day at the owner's request with oral clindamycin, doxycycline, enrofloxacin, tramadol, and topical prednisolone acetate.

With the exception of the blood cultures, final reports on all outstanding diagnostics were provided within 2 days of the dog's discharge. Results of the remaining fungal immunoassays and the extended vector-borne polymerase chain reaction panel were negative for all organisms assayed. Serum protein electrophoresis revealed a polyclonal elevation in the β fraction (1.55 g/dL; reference range, 0.7–1.3g/dL) and a low normal γ fraction (0.64 g/dL; reference range, 0.5–1.3 g/dL). Urine culture yielded growth of a yeast (10,000–50,000 colony-forming units/mL) that was later identified as Candida zeylanoides based on results obtained from a single yeast identification system^j. On day 3 postdischarge, the possibility of disseminated candidiasis was more strongly suspected after a second veterinary ophthalmologist reported a yellow/tan subretinal exudate on fundoscopic examination of both eyes. That suspicion was bolstered when blood culture results were returned the following day indicating the isolation of yeast from broth media that was also identified as Candida spp. by the above-mentioned identification system. To confirm the diagnosis of candidiasis, a recommendation to obtain either a subretinal or spinal aspirate for culture and cytology was made. The following day, the dog was prescribed voriconazole^k (4 mg/mg per os q 12 hr) after obtaining three aspirates from the subretinal and vitreal spaces OS. Cytologic analyses of those aspirates using standard Wright's stain revealed marked suppurative inflammation and multiple degenerated fungal-like organisms that proved unidentifiable. Fungal cultures were performed but yielded no growth of identifiable pathogens.

Five days after starting the voriconazole, the dog's clinical status began deteriorating rapidly. Concerned about quality of life and the grave prognosis for recovery, the owner elected humane euthanasia. Approximately 6 hr after euthanasia, a sample of CSF was aseptically collected via a cisternal tap, and the body submitted for postmortem examination at the Texas Veterinary Medical Diagnostic Laboratory. Fluid and cytology of the CSF revealed marked elevations in both protein (327 mg/dL; reference range, ≤30 mg/dL) and WBCs (7325/µL; reference range, 0–5/µL) and was characterized by a mixed pleocytosis consisting predominantly of lymphoid and mononuclear cells. No infectious organisms were identifiable within the sample and no cultures were submitted.

At necropsy, pertinent gross observations included thickening of the subcutaneous fat overlying the dorsal aspect of the right hip and softening of the intervertebral disc between the first and second vertebral vertebrae with outward extension into the adjacent vertebral bodies. Multiple, white, variably-sized, discolored foci were encountered in the myocardium of both ventricles, along the capsular and cut surfaces of both kidneys, and in the muscles and subcutaneous fat overlying the right side of the lumbosacral area (Figure 2). Histologically, stained sections of heart, kidney, eye, subcutaneous fat, skeletal muscle, colon, thyroid, and parathyroid gland revealed multifocal areas of necrosis and minimal granulomatous inflammation associated with numerous intralesional algae compatible with Prototheca spp. Those organisms were round to oval, measured 8–22 µm in diameter, had a clear 2–4 µm thick wall, and contained either central granular amphophilic material or multiple (2–5) wedge-shaped endospores. Focal osteomalacia with fibrosis and intralesional algae was identified within the spinal column at the first and second lumbar vertebrae (Figures 3, 4) with outward extension of exudate into the dura mater surrounding the adjacent segment of spinal cord. Examined sections of cerebrum, cerebellum, and medulla were histologically unremarkable. Cultures of samples taken from heart, kidney, muscle, and subcutaneous fat all yielded heavy growth of yeast-like organisms later identified as Prototheca spp. based on their morphologic characteristics and the results of two separate yeast identification systems^l. Unfortunately, neither of those tests was reliably able to determine the species of the organism.

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Discussion

The clinical features of protothecosis in dogs are unique in many respects compared with those observed in other species. Although usually restricted to the skin of cats, the mammary glands and corresponding lymph nodes of cows, and the skin and bursae of people, infections in dogs tend to be insidious and relentlessly progressive with widespread dissemination common.^2–6,9 At the time this article was written, 52 cases of canine protothecosis were documented in the veterinary literature. Of those, 28 (53.8%), 14 (26.9%), and 6 (11.5%) presented with a clinical history of diarrhea, ocular abnormalities, or neurologic deficits, respectively.^8–13 Within that same population of dogs, another 9 (17.3%), 17 (32.7%), and 17 (32.7%) had each of the aforementioned problems, respectively, by the time of either diagnosis or death.^8–14 Adhering somewhat to that general trend, the dog in this report presented with an acute onset of vision loss that quickly progressed to include neurologic deficits. Interestingly, despite histological evidence of colonic involvement, signs referable to the lower gastrointestinal tract were neither reported nor observed until just days before the dog's euthanasia. That was surprising given the extensive involvement of other organ systems documented and the general consensus that most canine protothecal infections arise and disseminate from the colon.^1,3,4 A detailed review of the dog's history failed to identify any significant episodes of diarrhea or other pertinent gastrointestinal signs within 3 yr of the dog’s presentation to the authors’ hospital. However, as the dog was allowed to roam freely outdoors, it is possible that episodes of diarrhea went unnoticed by the owners.

In addition to its ocular and neurologic abnormalities, the dog described in this report was also found to exhibit signs of spinal hyperpathia and was subsequently diagnosed with vertebral spondylitis at the first and second lumbar vertebrae. Although initially thought to represent a lesion associated with a rare case of disseminated candidiasis, postmortem histological analysis of the site confirmed infection with Prototheca spp. Apart from the more common sites of the colon, eyes, and brain, protothecal dissemination has also been documented to affect the kidneys, heart, skeletal muscle, skin, spinal cord, stomach, liver, spleen, pancreas, ileum, cecum, lungs, diaphragm, lymph nodes, omentum, ears, tongue, thyroid gland, aorta, and bone of dogs.^{3,4,7,9,11,15–21} Until this current publication, disseminated protothecosis involving skeletal lesions in dogs was only reported in two cases.⁹ In both of those cases proliferative periosteal lesions confined to the long bones of the appendicular skeleton were documented in two young boxers from Australia with chronic histories of diarrhea. The present case is the third case to document protothecal dissemination to bone, and it is the first to report involvement of a vertebral endplate. Although there exists a paucity of cases documenting bone involvement, it is possible that skeletal lesions actually occur with greater frequency and are merely overshadowed by more prominent clinical signs of neurologic dysfunction. Such a contingency could very easily have been the case in this dog if abdominal radiographs were not performed. Indeed, in reviewing the cases of canine protothecosis in which signs of lameness or paresis were described, radiographic studies were either omitted, not performed, or incomplete in almost all situations.^{7,9,10,17,20,22–25}

The pathogenesis of canine protothecosis remains poorly understood, but despite the causative organism's ubiquity within the environment, the sporadic occurrence of the infection suggests some form of immune dysfunction is likely involved.^1,4,7,8,26 Among the mechanisms postulated to play a role, defects in cell-mediated immunity have garnered the most attention.^2,4–8 Yet, whether such defects are the cause or effect of infection remains to be clearly elucidated. Within the veterinary literature, evidence supporting a mechanism by which an inherited or acquired defect in cell-mediated immunity contributes to the development of disease is furnished by an apparent predisposition to infection among collies and boxers, observations of suppressed T lymphocyte and neutrophil function in a collie with disseminated protothecosis, and an inverse relationship between the number of protothecal organisms and the surrounding cellular infiltrate in a dog with cutaneous protothecosis.^7,9,26 Although specific immunologic studies were not conducted in this case report, further support for such defects can be inferred from several observations. First, as has been reported in other cases, a significant cellular response to the infecting organisms was not recognized within any of the histological samples examined.^7,8,18,24,26 Second, the inflammatory responses that were recognized were universally characterized by minimal numbers of plasma cells, lymphocytes, and macrophages, and a general lack of neutrophils. Lastly, although firm conclusions regarding any association between the dog's heat cycles, chronic pyoderma, or recurrent antibiotic and corticosteroid use and immune dysfunction could not be made, the role those potential predisposing factors may have played in contributing to an immunocompromised state cannot be ignored.

Additional support for underlying dysfunctional immunity can also be garnered from the possibility of a concurrent infection with the opportunistic fungus Candida. Implicated primarily in disease states of immunocompromised animals and people, that possible infection was one of the most intriguing aspects of this case and yet could conceivably have been the result of misidentification.^27,28 Instead of representing a species of Candida as reported, the positive blood and urine cultures obtained may have represented Prototheca. Such a contingency would certainly follow with what would seem more plausible in a case of disseminated protothecosis and could be supported by the absence of Candida within any of the affected tissues examined. Although that failure could have been caused by the effects of voriconazole, a medication shown to possess excellent activity against the various species of Candida, the time span between the initial administration of the medication and the collection of histologic samples would seem inadequate to completely eradicate all evidence of the organism in tissue.²⁹

The limitations inherent within yeast identification systems could also potentially account for the possibility of misidentification of Prototheca as Candida. One such limitation involves the possibility of discordant results with differing incubation periods. The laboratory that reportedly isolated Candida zeylanoides from blood and urine used an identification kit that can provide results in as little as 4 hr (incubation time). Although such kits work well for many species of Candida, one study determined that misidentifications were more likely for less commonly isolated yeasts and yeast-like organisms compared with systems employing 24 hr incubation periods.³⁰ Moreover, in another study, the manufacturer of an automated identification system reported that Prototheca zopfii could not be accurately identified before a 48 hr incubation was completed, thus providing further evidence that longer incubation periods may be necessary to correctly identify Prototheca from samples assayed.³¹ Unlike the first laboratory, the laboratory that isolated Prototheca from the necropsy-derived samples did so utilizing both a 4 and 24 hr kit but opted to use an optional longer incubation period (24–72 hr) thereby ensuring adequate growth and accuracy of identification.

Additional limitations often associated with the use of various yeast identification systems include a need for both relatively high inoculum densities and strict adherence to kit instructions. Yeasts and yeast-like organisms tend to be much slower growing than bacteria; therefore, if any single inoculum is too light, inadequate growth will likely ensue, thereby resulting in discordant substrate reactions. If the laboratory that identified Candida in this case used a light inoculum and inadequate growth resulted, one could expect false-negative reactions to occur. In turn, those factors could increase the odds of inaccurate identification. Additionally, if a laboratory technician in that same laboratory underestimated the importance of following kit manufacturer instructions, inaccurate results may have been provided. Indeed, in one of the aforementioned studies only 48 of 97 yeast isolates were correctly identified when the inoculum was prepared directly from the primary isolation media rather than from the media specified by the kit manufacturer.³⁰ Because the organism isolated from blood and urine was not available to be compared with the organism isolated from tissues and identified as Prototheca, it is impossible to know the exact reasons for the discrepant identification results. Nonetheless, it is probable that Prototheca was simply misidentified as Candida zeylanoides from the initial blood and urine samples obtained in this case report.

Despite a rather comprehensive diagnostic evaluation, successful antemortem diagnosis proved elusive in this case. This, undoubtedly, was due in large part to the isolation of an organism identified as Candida from blood and urine cultures that lead to a preliminary diagnosis of disseminated candidiasis. The general lack of signs referable to the gastrointestinal tract also proved misleading, decreasing the authors’ index of suspicion for protothecosis. In most cases, an accurate diagnosis of protothecosis can usually be made by identifying the organism within either cytological or histological samples of affected tissues.³ Although fungal-like organisms were recognized within the subretinal and vitreal aspirates the authors obtained, unfortunately those organisms proved difficult to definitively identify and were ultimately presumed to represent Candida. This difficulty was likely due to the high degree of degenerative changes present within each of the aspirates examined, complicating the ability to distinguish between pathologic organisms and surrounding inflammatory cells.

Apart from the spinal lesion identified radiographically, the purported isolation of Candida from blood and urine cultures, and the results of our CSF analysis, few of the diagnostic findings in this case deviated from what has been reported in other cases of canine protothecosis. Indeed, in accordance with previously published cases, the results of most of the laboratory diagnostics were either normal or nonspecific.^{7–10,14–25,32} Moreover, on histologic samples of infected tissues, Prototheca were usually readily identifiable and often associated with only minimal cellular responses. Of notable exception to that general trend was the lack of either identifiable organisms or lesions within any sections of brain tissue evaluated despite obvious signs of central neurologic dysfunction. Considering the multifocal distribution of organisms and lesions manifested in other affected tissues, that absence was likely the result of the chance sampling of unaffected brain tissue.

Another somewhat deviant result among all the diagnostics performed in this case was the cytological findings of the CSF obtained. Despite the frequency with which neurologic signs have been reported in cases of canine protothecosis, few dogs exhibiting such abnormalities have actually undergone a comprehensive neurodiagnostic evaluation. As such, information regarding expected findings of advanced imaging or CSF analysis remains somewhat lacking. Of the three cases of canine protothecosis in which abnormal CSF findings have been described within the veterinary literature, all reported an inflammatory response characterized by an eosinophilic pleocytosis.^9,24,25 In contrast to those findings, a mixed pleocytosis composed predominantly of lymphocytes and mononuclear cells was recognized in this case. Although commonly associated with mycotic or protozoal encephalitides, the current study’s findings may have differed from those of previous reports because of the early corticosteroid use used by the dog's referring veterinarian.³³ That theory could be supported by one of the aforementioned reports in which a follow-up CSF analysis revealed resolution of an eosinophilic pleocytosis after treatment with systemic steroids had been implemented.²⁴ As the CSF sample was collected after euthanasia, postmortem changes could not be excluded as contributing factors to the discrepancy between the study authors’ findings and those of previous cases.

The treatment of disseminated protothecosis in dogs remains challenging and although various regimes have been attempted, optimal therapeutic guidelines remain poorly established.^1,4,9 To date, treatment with amphotericin B, fluconazole, itraconazole, ketoconazole, clotrimazole, aminoglycosides, and tetracycline have all proven unsuccessful in the dog.^1,4 Among those medications, however, both amphotericin B and itraconazole have exhibited some ability to slow the progression of the disease.^9,15,34 Driven out of concern for disseminated candidiasis, voriconazole was selected in this case. A synthetic derivative of fluconazole, this newer triazole was specifically chosen because of its widespread distribution, superior ability to penetrate nervous tissues, and efficacy against increasingly prevalent nonalbicans Candida infections with resistance to traditional antifungal medications.²⁹ Although unintentional, this is the first report to document the use of voriconazole in a dog with disseminated protothecosis. Unfortunately, like so many other antimicrobials used in the treatment of this devastating disease, voriconazole failed to elicit any significant positive clinical outcome. That failure aside, the delay with which that medication was started, coupled with the extent of systemic involvement manifested in this dog, make drawing definitive conclusions regarding the efficacy of voriconazole in the treatment of canine protothecosis difficult.

Conclusion

Protothecosis remains an uncommon but devastating disease in dogs. Although most often associated with signs referable to the gastrointestinal tract, this report underscores the importance of not excluding protothecosis as a differential diagnosis when such symptoms are absent. Moreover, this report also validates the need to consider protothecosis in any dog presenting with signs of diskospondylitis. Given the immunocompromised state believed to be inherent in many protothecal infections, the possibility of a concurrent infection with another infectious disease cannot be ignored. Lastly, although firm conclusions regarding the therapeutic efficacy of voriconazole in all cases of protothecosis cannot be made from a single case, it appears this newer triazole is of little utility in treating late-stage disseminated infections.