**1. Introduction**

Harmful algal blooms (HABs) resulting in animal intoxications are a worldwide occurrence, with reports of mortality becoming more prevalent [1,2]. Shoreline cyanobacteria blooms are one

type of HAB that can lead to exposure of those living near lakes and streams, such as domestic dogs. Dogs represent a sentinel species due to their shared environment with humans [3], supporting the framework for a One Health approach when investigating cyanotoxins. Cyanobacteria poisoning of canines has been well documented, but comprehensive reporting on toxin levels detected in specimens is sparse [2,4–7]. Reports of canine intoxication by cyanobacterial neurotoxins are more prevalent, but intoxication to hepatotoxins such as microcystin (MC) have increased in frequency [2]. Over a four-year period (2007–2011), Departments of Health and/or Environment from 13 states reported 43 dogs suspected of poisoning by MCs with a moderate to high probability, based on clinical and diagnostic pathology [2]. Awareness of these events is spreading; however, MC intoxication events likely go under reported. This could be due to a multitude of contributing factors, such as; insufficient exposure history, lack of supportive environmental data, lack of standard HAB protocols or policies leading to inadequate sample acquisition and/or handling, improper analytical test selection, or misdiagnosis due to commonality of symptoms to other hepatotoxins. Furthermore, monetary restrictions may hinder testing beyond preliminary veterinary intervention. Therefore, providing veterinarians and analytical laboratories information on proper specimen collection protocols is of the utmost importance. Protocol dissemination will help to minimize costs, provide clinically relevant information, and compile data to inform the community of local environmental threats.

Typical MC canine toxicosis cases are the result of cyanobacteria ingestion. Post ingestion, it has been shown that MCs make their way through the gastrointestinal tract and into the liver, presumably through first pass effects through the bile-acid transport system [8–10]. MCs do not passively enter cells, but require active transport [11], mediated by organic anion transporters (OATs) [12]. These transporters are not only significantly expressed in the liver (Oatp1b2, OATP1A2, OATP1B1, OATP1B3) [12], but are also expressed in the brain and kidney (OATP1A2) [12,13]. The OAT facilitated uptake of MCs is one of the key steps in the pathogenesis of the reported hepatocellular damage and may account for the neurological effects observed in animals and humans following exposure [14,15]. The predominant OATs in the liver and kidney of canines (Oatp1b4 > Oatp2b1 > Oatp1a2) appear to exhibit similar substrate specificity to that of the human OATP1B3 [16]. Human OATP1B1 is abundant in lobular hepatocytes, while OATP1B3 is predominantly expressed in hepatocytes near the central vein [17], indicating interspecies differences in OAT location may play a role in clinical presentation of toxicosis. Once in the cytoplasm, MCs can affect a variety of cellular pathways, including regulation of DNA repair, regulation of protein activity, cell signaling, cell cycle, gene expression, apoptosis, and metabolism of endogenous or cytotoxic compounds [18]. The most studied of these pathways is the inhibition of the essential members of the protein phosphatase (PP) family. The reversible phosphorylation of proteins is an integral part of metabolism, which MCs inhibit by binding to serine/threonine PP1, PP2A [19] and PP3 [20]. PP inhibition can result in hyperphosphorylation [21], increases in reactive oxygen species (ROS) [22] and/or inflammation. These effects result in the disruption of cytoskeletal components, rearrangement of actin filaments within hepatocytes, and ultimately cellular death, which elucidates the observed morphological changes post-mortem [23].

The main pathway of hepatic elimination of MCs is Phase II biotransformation through conjugation with glutathione (via glutathione-S-transferase or non-enzymatically) [21,24–26] and through an elimination conjugation reaction with cysteine [24]. In mammals, MCs are primarily eliminated by both biliary and renal routes, with conjugated forms excreting mainly through the kidneys [27–29]. Since MC conjugates retain some of their toxic potential [30], metabolites may lead to continued insult to vital organ systems. The extent of intoxication and ability to recover from MC exposure is dependent on dose and the animal's capability to metabolize MCs. Since the antioxidant glutathione is integral to the detoxification and elimination of MCs, depletion after a high dose or in the presence of concomitant contaminants has been observed [31]. The loss of active glutathione coupled with continued hyperphosphorylation and resultant ROS formation likely contributes to the necrosis and apoptosis of hepatocytes, as well as the breakdown of the hepatocyte cytoskeleton. As the primary site of MC detoxification, the hepatic parenchyma exhibits the most striking damage to intoxication;

however, the renal nephron can also be negatively affected [5,32,33]. Although the cause of renal parenchymal damage has not been identified, tubular ischemia has been proposed as one mechanism [5]. Other probable mechanisms include hepatic shock and direct toxic action to the renal tubules of conjugated and free MCs.

The proper analytical approach is key to confirmation of MC exposure. Advances in MC research have elucidated numerous structurally related congeners, which have increased from 60+ known in the 1990's [34] to over 250 described to date [35]. Variations along the structure occur mainly in two amino acid positions (X2, Y4; Figure S1), but modifications, such as desmethylation, may happen along other parts of the structure. This structural variation coupled with protein binding and conjugate formation present significant analytical challenges. Therefore, widely available commercial enzyme-linked immunosorbent assays (ELISAs) are frequently used to test MCs due to their broad specificity to the various congeners (and potentially conjugates). However, MC ELISAs have a narrow range of applicability to complex matrix testing (e.g., urine, tissue, blood) from different animal species, with false positive results reported when analyzing mammalian livers [36]. While their availability and ease of use make them convenient, they also require an alternate method of confirmation, such as liquid chromatography tandem-mass spectrometry (LC-MS/MS). The specificity achieved when targeting MCs via LC-MS/MS provides quantitative accuracy, but also results in the under-reporting of total MCs [37,38]. This is due to a lack of commercially available reference materials for method calibration coupled with the extensive variability in MC forms. Other techniques utilized to address this include non-targeted high- and low-resolution LC-MS, but this requires an intimate knowledge of MC chemistry. An alternate approach involves the oxidative cleavage of the unique Adda (3-amino-9-methoxy-2,6,8-trimethyl-10-phenyl-4,6-decadienoic acid) side chain and subsequent quantitative analysis of MMPB (2-methyl-3-methoxy-4-phenylbutyric acid). The MMPB technique provides a relatively straightforward protocol accounting for total Adda containing MCs or nodularins [38–40]. This test allows for the quantification of free MCs and those modified during metabolism as long as there is conservation of the Adda side chain.

At present, the diagnosis of cyanobacteria poisoning requires a thorough history of the exposed patient in relation to the contaminated source. A two-tier analysis should include identification of the dominant cyanobacteria genera present and toxin analyses. In the absence of an algal grab sample, analyses can be conducted on the vomitus/stomach contents of a recently exposed individual, which is representative of unmetabolized cyanotoxins. However, due to the low pH of gastric contents, degradation of the organisms may impede identification of cyanobacteria genera. Thus, broad screening of multiple cyanobacterial toxins or targeted analysis for specific toxins based on clinicopathological data may be required. Once toxins are confirmed in the source, additional targeted analyses should be conducted on other specimens (e.g., liver, kidneys, feces, and urine) to confirm exposure and metabolism. Data achieved from the source of exposure, coupled with clinical/pathological observations and analytical data have been utilized to confirm MC intoxication in previous studies [4–6]. This process is time consuming, cost prohibitive, requires significant knowledge of cyanobacteria and the toxins they produce, and many specimens are not ideal for antemortem testing. Therefore, a standard protocol for diagnosing MC toxicosis should be developed for non-invasive specimen collection and sensitive accurate testing.

The present study illustrates the most comprehensive report on the pertinent clinicopathological data, pathological characteristics, supportive care, and novel diagnostic testing performed during and after an exposure event involving dogs. Results from this investigation provide support of viable antemortem testing methods for detection of MCs in canines during and after suspected exposure to microcystin producing cyanobacteria.

### **2. Results**

### *2.1. Presentation, Clinical Data and Treatment*

Between 26 August to 8 September 2018, six dogs were admitted for medical care at Pet Emergency of Martin County, Florida, USA. Information pertaining to the exposed animals and negative controls used in this study are presented in Table 1. The patient history for the six hospitalized cases included access to the Indian River and potential ingestion of decaying fish, organic debris, or water from the waterway. The onset of clinical signs varied from 2–48 h post exposure, with the most common signs being vomiting and depression. Weakness, collapse, tachycardia, petechia/ecchymosis and melena were also noted in a subset of patients. Clinicopathological findings included but were not limited to: elevated ALT (Alanine aminotransferase), thrombocytopenia, prolonged partial thromboplastin time (PTT) and prothrombin time (PT), peritoneal and/or pleural effusion, hypoglycemia and hyperbilirubinemia. For a full list of abnormalities and values, refer to Table 2.

**Table 1.** Subjects examined in this study shown with weights, age, sex, date of exposure and status. UE = Unexposed Individual. N = neutered, S = spayed.


**Table 2.** Clinicopathological abnormalities noted during hospitalization of six dogs exposed to the St. Lucie River HAB event. ID = Identification, APTT/PT = activated partial thromboplastin time/prothrombin time, ALT = Alanine aminotransferase, >DL = greater than detection limit.


After presentation to the emergency clinic, a complete blood count, blood chemistry, and clotting profiles were performed. Decontamination through bathing was initiated in a subset of patients prior to arrival. One patient presented with productive emesis and vomitus was saved for cyanotoxin evaluation. The Animal Poison Control Center was contacted for advice on the cases but ultimately treatment was tailored for each dog by the attending veterinarian with the emphasis on acute liver injury. Therapy included intravenous fluids (with dextrose supplementation as indicated), gastroprotectants, antibiotics, antiemetics, analgesics, fresh frozen plasma (FFP), cholestyramine, vitamin K, N-acetylcysteine and various oral liver protectants. For the dogs that required intensive overnight care, continuous monitoring of electrocardiogram and blood pressure were performed. Furthermore, other essential parameters were monitored at varying intervals such as blood glucose, electrolytes, activated partial thromboplastin time/prothrombin time (APPT/PT), complete blood count (CBC), and blood chemistry. Patient hospitalization ranged from one day to nine days.

#### *2.2. Pathology*

One of the six dogs succumbed to fulminant liver failure, coagulopathy, and shock. A full postmortem examination was performed (Figure S2). On preliminary gross examination, the dog was in good body condition with a body condition score of 5/9 and mild post-mortem autolysis [41]. The skin in areas without hair appeared slightly yellow with multifocal areas of ecchymosis and petechiation. Diffuse icterus of the mucous membranes and subcutaneous adipose tissues was noted. Abundant dark red to black fluid drained from the nares and oral cavity upon manipulation of the head. Entry into the abdominal cavity showed up to 1L of serous red tinged fluid. The length of the intestines was dark pink to red with red streaking down the serosa, which had a granular appearance. Abundant edema and coagulated blood expanded the mesentery and omentum adjacent to the spleen as well as around the pancreas. Inspection of the esophagus, stomach, and intestines revealed abundant dark red to black fluid that filled the entire gastrointestinal tract. The gastric wall at the pylorus was diffusely expanded by submucosal hemorrhage and edema. The liver was diffusely dark red with sharp margins, a reticular pattern, and had a normal consistency. The gall bladder was filled with dark green bile and the wall of the bladder was thickened by edema. Abundant bright yellow, granular thick fluid was present in the urinary bladder. However, the kidneys and ureters were intact. The spleen was diffusely pale red and had multiple <3 mm fibrotic nodules on the serosal surface. The right lung lobes appeared pink except for a 3–4 cm red focus in the cranial lobe and mild dark red mottling in the middle lobe. The left lobes were diffusely red, wet, and oozed abundant red fluid upon transection. The adrenal glands had bilateral hyperplasia of the cortical layers. Multifocal petechia and ecchymosis were observed in the wall of the great vessels of the heart and in the endocardium of the left ventricle. There was mild multifocal, nodular thickening of the mitral valve leaflets.

Compared to normal canine microscopic anatomy (Figure 1A,C,E), tissues of the deceased dog showed several significant microscopic changes including; acute, severe, massive hepatocellular necrosis (Figure 1B), acute, moderate, multifocal, tubular necrosis with granular casts and intracellular iron (Figure 1D), acute, severe, multifocal to coalescing, hemorrhage in the thymus, gastrointestinal tract, mesentery, omentum, lymph nodes, pancreas, lungs, pulmonary artery, endocardium (Figure 1F), acute, diffuse, gall bladder edema, and acute, diffuse, splenic contraction with multifocal siderofibrotic plaques. Mild mitral valve endocardiosis was also noted in this patient as an incidental finding.

Ancillary testing included aerobic culture of the liver and leptospirosis PCR. Results of the culture revealed bacterial organisms *Enterococcus faecalis* and *Erysipelothrix rhusiopathiae.* Leptospirosis PCR was negative.

**Figure 1.** Photomicrographs (canine) of hematoxylin and eosin stained (H&E) normal liver, renal cortex, and thymus (**A**,**C**,**E**) as compared to the MC exposed dog (C-SP) (**B**,**D**,**F**). (**B**) Severely disrupted hepatic cords characterized by massive hepatocellular necrosis and hemorrhage (asterisk). Low numbers of hepatocytes adjacent to a central vein are spared. (**D**) Renal cortex with a locally extensive area of acute tubular necrosis. Note accumulation of brown granular pigment within tubular epithelial cytoplasm or sloughed cellular debris within the tubular lumina (arrow). (**F**) Mediastinal adipose and thymus expanded by hemorrhage, fibrin and edema (asterisks). Scale bars are 100, 50, and 500 μm for liver, renal cortex and thymus, respectively.
