Questioning the Use of Zika Virus Injection in Dogs with Advanced-Stage Brain Tumors

Main Text

To the editor:

In the article by Kaid et al.,¹ intrathecal injection of Brazilian Zika virus (ZIKV^BR) was reported as a potential oncolytic viral therapy for dogs with advanced CNS tumors. While harnessing a neurotropic virus for lysis of otherwise difficult-to-treat CNS tumors is an attractive proposition, the data presented are convincing of neither safety nor efficacy of such therapy.

Paragraph 1 of Materials and Methods states that “Dogs were considered eligible for the oncolytic therapy if they were diagnosed by a veterinary neurologist with advanced CNS primary tumor with neural origin, excluding meningioma and other non-neural tumors.” However, two of the three cases were diagnosed with meningioma; presumptively based on MRI findings (case 1) and subsequently confirmed by postmortem histopathology (cases 1 and 3 of Table 1). Hence, these cases should have been excluded once the identity of the neoplasms was confirmed.

The MRI interpretation in case 1 was of a “possible multicentric tumor in the brain white matter of the left frontoparietal lobe region, and the second a possible meningioma in the right ventrolateral side of the brainstem.” The diagnosis of meningioma was histologically confirmed; however, concurrent presence of a multicentric tumor in the frontoparietal lobe in case 1 is unsubstantiated. Supplemental figures labeled “Pirata-MRI 1- Day 0 post-contrast T2 weighted transversal plane” and “Pirata – MRI 2- Day 7 post-contrast T2 weighted sagittal plane” are not T2-weighted sequences based on the hypointense signal from cerebrospinal fluid (CSF) in the lateral ventricles.² The images might reflect a T2-FLAIR sequence; however, the signal of the brain gray and white matter is similar as in the T1-weighted images. Therefore, it is likely that both sequences are T1W rather than T2W images. All MR images lack left/right markers for orientation. Therefore, based on MR images in Data S1, the lesions in the frontal lobe are equivocal, and volume averaging cannot be excluded since a transverse T2W sequence was not provided. To confidently identify a neoplastic, inflammatory, or neurodegenerative lesion by MRI, multiple orthogonal sequences, both with T1 and T2 weighting, are required.^3,4

The authors suggest that since the putative frontal lobe multicentric neoplasm identified by MR imaging in case 1 was not apparent postmortem, it was more likely a neoplasm rather than a neurodegenerative condition that disappeared in the 2 weeks following intrathecal ZIKV injection. There is a lack of imaging and histopathological evidence for this supposition. The authors’ differential diagnoses for multifocal intra-axial lesions in the brain of dogs does not consider encephalitis, either immune-mediated or infectious.⁵ The standard of care for a dog with the described neurological and imaging abnormalities includes anti-inflammatory glucocorticoid therapy, which would be expected to reduce encephalitis and swelling.^6,7 It is unclear whether case 1 received glucocorticoid therapy, but such therapy would alter interpretation of progression or regression of clinical and imaging abnormalities.

Case 1 survived 18 days after intrathecal ZIKV infection, and sudden deterioration was attributed to pneumonia and pre-existing cardiac disease. No evidence of pneumonia was reported from postmortem evaluation, and cardiomyopathy sufficiently severe to cause sudden de-compensation should have been detectable in vivo prior to terminal heart failure. The dog was also reported to be immune compromised but had only mild lymphopenia, which is a common finding in hospitalized ill dogs unrelated to immune compromise. Furthermore, the dog had hyperglobulinemia, which suggests immune stimulation rather than compromise. Abnormalities of the lymphoid system were not reported postmortem. Furthermore, although this dog had a meningioma similar to case 3, only this case was censored from the survival interpretation.

Case 2 was castrated 60 days after the first ZIKV injection due to putative right testicular inflammation. A Leydig cell tumor and seminoma were diagnosed in the right and left testicle, respectively, concurrent with ZIKV antigen detection by immunohistochemistry in both testes. While there appear to be no prior data concerning ZIKV infection of testicular cells or testicular neoplasms in dogs, it appears unlikely that the typically incidentally diagnosed and slow-growing Leydig cell tumor in case 2 caused acute testicular inflammation and necrosis. Rather, ZIKV-induced inflammation should be considered as a potential etiology, even if the testis with a seminoma did not show inflammation and necrosis. Case 1 was reported to have very high ZIKV RNA in the testes (Figure 1B).

Case 2 had an apparent ∼14 cm³ tumor region in the right intra-axial frontal lobe on MRI with variable contrast enhancement. The first injection of ZIKV was reported to result in “tumor pseudo-progression,” while a 10-fold higher second ZIKV injection on day 21 was considered to lead to “substantial tumor remission.” Postmortem, the tumor was diagnosed as a glioma. Canine gliomas are morphologically and molecularly heterogeneous, and robust survival data associated with subtypes is lacking. However, large areas of necrosis are a common and defining feature of high-grade gliomas.^8,9 Furthermore, supratentorial location of tumors, as in this case, was associated with longer survival than infratentorial location.¹⁰ Therefore, findings of “extensive necrosis, cellular debris, and immune-cellular infiltration” on histopathology (Figure S3), including grossly apparent cavitation (Figure 2D), are typical of high glade gliomas in dogs and less likely due to ZIKV anti-tumoral effects.^8,9

Case 3 “had been misdiagnosed as carrying a highly suggestive glioma in the left frontal lobe by MRI.” It appears that, at postmortem, this dog was determined to have an intracranial meningioma rather than a glioma. Therefore, it should have been excluded from the study. This dog was leukopenic and should have been considered as potentially immune compromised.

The authors are to be commended on their success with identifying β3 tubulin staining in adipose-derived stem cells and the canine glioblastoma cell line (Figures S4D and S4E). Expression of this microtubule has generally been considered specific for neuronal and testicular cells. Hence, these findings extend the range of cell types to be considered positive for β3 tubulin.

Based on Figures 1C, 1E, 2E–2H, and S1, neurons are considered uninfected by ZIKV, but it is unclear which cells in the CNS are infected by ZIKV. What do the authors consider as the identity of the immunohistochemistry (IHC)-labeled cells in Figures S1D and S1F? And what might be an explanation for the absence of DAPI nuclear staining in neurons in Figures S1G–S1I?

Some laboratory parameters, such as a serum glucose of 0 mg/dL in cases 1 and 3 (Table S1), are incompatible with life, and therefore erroneous. Other parameters are reported or interpreted incorrectly, such as a putative leukopenia in case 1 indicating immunodeficiency and hematocrit > 25% indicating “adequate organ function.”

This report promotes the potential value of intrathecal injection of ZIKV as an oncolytic agent for brain tumor therapy in dogs. However, since only one case had a tumor of neural origin, patients were assessed inconsistently, and therapies other than ZIKV were not or only partially provided or reported, the authors’ conclusions are premature. The patients were not “normal pet dogs,” but rather they were debilitated geriatric patients with one or several neoplasms and substantial co-morbidities who were then subjected to multiple general anesthetics for imaging, intrathecal injections, CSF collection, and surgery. End-of-life considerations should be considered in such situations.¹¹ There was no indication of consistent assessment for fever, systemic inflammation, or potentially adverse effects from more than three general anesthetics over a course of 10 days—in particular, in dogs with advanced cardiac disease (case 1). There is also no or only limited indication of concurrent standard therapy for neurological disease, such as hyperosmotic fluids, glucocorticoid, and anti-seizure medications. Such therapy could contribute to or be responsible for transient clinical improvement and amelioration of MRI changes by reducing peritumoral edema.⁷ There is no comprehensive description of postmortem lesions.

The authors suggest that survival of 153 and 80 days in cases 2 and 3, respectively (case 1 survived only 18 days), “represents a significant increase of survival rate since the literature reports a survival of 24–33 days for dogs diagnosed with brain tumor.” This statement is in reference to 38 dogs with brain tumors that received symptomatic treatment only, which was published in 1991.¹¹ Treatment of intracranial tumors in dogs has advanced greatly in the past three decades. Recently, 128 dogs with brain tumors treated solely for symptom control with glucocorticoid and anti-seizure medications were reported to have a median survival of 65 days (range 6 to 359 days), while dogs with brain tumors (including gliomas and meningiomas) receiving radiotherapy survived a mean of 18.1 months (95% confidence interval of 12.3–26.6 months).^12,13 Therefore, the suggestion that outcomes in two dogs that received intrathecal ZIKV “represent a significant increase in survival rate” is not substantiated.

To support a claim of efficacy of a novel therapy in a “small clinical trial,” the outcome in subjects receiving standard care plus a treatment under investigation should be compared to the outcome in a control group receiving the standard of care plus a “sham” treatment. The authors of this study indicate it would be ethically challenging to recruit a control group of untreated dogs with brain tumors. We agree that that this would inappropriate. Dogs with seizures and other illness should receive appropriate therapy as described above, whether owners consent to additional experimental treatment or not. Findings from this study do not provide scientifically sound or clinically meaningful data in support of safety or efficacy of intrathecal ZIKV injection, and the study was conducted in discordance with recommendations for studies in companion animals with cancer.¹⁴

Finally, as veterinarians, we are well aware of the challenges in regard to testing and validating new therapies for difficult-to-treat conditions in patients. However, studies purporting novel therapies while lacking consistent assessment and provision of standard therapy have potential for inappropriate interpretation by non-veterinarians. Thus, while the use of ZIKV as an oncolytic agent might have potential for humans and dogs with brain cancer, the data presented in this report provide insufficient evidence.