Methods

In the current study, urine was analyzed from client-owned dogs with DCM that had been enrolled from September 2018 through March 2020 in a prospective 9-month study⁷ on diet-associated DCM and healthy controls from that study, although not all dogs from the original study had urine available from the time of diagnosis. Dogs from the original study with subclinical cardiac abnormalities that did not meet the criteria for DCM were not included in the current analysis.⁷ The original study was approved by the Cummings School of Veterinary Medicine Clinical Studies Review Committee. Owners had signed an informed consent form and completed a diet history form at the time of enrollment. The original study's definition of DCM consisted of M-mode fractional shortening ≤ 25%, normalized left ventricular internal diameter in diastole ≥ 1.8, and normalized left ventricular internal diameter in systole ≥ 1.2 (or breed-specific criteria for Doberman Pinschers or Boxers).^5–7,18,19 To be eligible for the original study, dogs had to be eating a commercial extruded (kibble) diet as their main source of calories for at least 6 months. In the original study, diets were categorized as “nontraditional” if they were grain-free or included pulses or potatoes/sweet potatoes in the top 10 ingredients and “traditional” if they were grain inclusive and had no pulses or potatoes/sweet potatoes in the top 10 ingredients.⁷ However, as information on diet-associated DCM has expanded, the definition for diet groups in the current analysis was refined to HP and LP. High-pulse diets were defined as those with ≥ 1 pulse in the first 10 ingredients on the ingredient list or ≥ 2 pulses anywhere on the ingredient list and LP as those with no pulses in the first 25 ingredients. One healthy dog from the original study whose diet had been categorized as nontraditional was excluded from the current analysis because the diet did not contain any pulses (it was high in potatoes and sweet potatoes). Ingredients were determined based on the ingredient list of the diet providing the majority of calories to each dog. Diet pulse scores were calculated for each dog's diet as previously described as an estimate of the amount of pulses in the dogs' main diets.^7,20

As part of the original study, voided (free catch) urine from the time of diagnosis was collected and stored at −80 °C until analysis as a single batch. Urine di-22:6-BMP was measured by LC-MS-MS (in ng/mL; Nextcea) and was normalized to urine creatinine (in mg/mL) for final results, which were expressed as ng/mg creatinine.

Results

Fifty-three dogs were included in the analysis: 29 with DCM (DCM-HP, n = 25; DCM-LP, n = 4) and 24 healthy controls (HP diets [control-HP], n = 10; LP diets [control-LP], n = 14). There were no significant differences among the groups for age, sex, breed, body weight, or body condition score, although both groups of dogs with DCM were more likely to have cachexia (ie, muscle loss) than healthy controls (none of which had muscle loss; P < .001; Table 1).

Normalized urine di-22:6-BMP concentrations for the 4 groups (ie, DCM-HP, DCM-LP, control-HP, and control-LP) are shown in Figure 2. Mixed-effects models adjusted for age and sex showed that HP diet was significantly associated with higher normalized urine di-22:6-BMP concentrations (P = .01). A 1-way ANOVA showed that there was a significant difference among the 4 groups (P = .01), with Tukey post hoc analysis showing that the DCM-HP group had significantly higher normalized urine di-22:6-BMP concentrations compared to the control-LP group (P = .02). Other pairwise comparisons were not significantly different. For all 53 dogs, normalized urine di-22:6-BMP concentrations were significantly positively correlated with pulse score (r = .52; P < .001). Results were similar whether the 1 healthy dog from the original study whose diet contained potatoes but no pulses was excluded or included in the analysis.

View Full Size

In this cross-sectional analysis, voided urine was collected from client-owned dogs with dilated cardiomyopathy (DCM) from September 2018 through March 2020 and from healthy controls. Box-and-whisker plots representing urine concentrations of di-docosahexaenoyl (22:6)-bis(monoacylglycerol)phosphate (di-22:6-BMP) normalized to urine creatinine for 29 dogs with DCM eating high-pulse (HP) diets (n = 25) or low-pulse (LP) diets (n = 4) and 24 healthy controls eating HP diets (control-HP; n = 10) or LP diets (control-LP; n = 14). Mixed-effects models adjusted for age and sex showed that HP diet was significantly associated with higher normalized urine di-22:6-BMP concentrations (P = .01). A 1-way ANOVA showed that there was a significant difference among the 4 groups (P = .01), with Tukey post hoc analysis showing that the DCM-HP group had significantly higher normalized urine di-22:6-BMP concentrations compared to the control-LP group (P = .02). Other pairwise comparisons were not significantly different. Each box represents the central 50% of the values. The horizontal line in each box represents the median value, and whiskers indicate the range (± 1.5X IQR). Asterisks indicate outliers, and circles represent extreme (far outside) values.

Citation: American Journal of Veterinary Research 86, 1; 10.2460/ajvr.24.07.0211

• Download Figure
• Download figure as PowerPoint slide

Discussion

In this study, HP diets were significantly associated with higher normalized urine di-22:6-BMP concentrations. While this is only a first step in studying this possible connection, these findings support primary or secondary phospholipidosis as a potential mechanism for diet-associated DCM. Phospholipidosis can be associated with certain drugs (especially cationic amphiphilic drugs) and can also be seen in some lysosomal storage disorders.^11,13,15,21 Drug-induced phospholipidosis can be reversible,^13,15,21 which is compatible with findings from dogs with diet-associated DCM in which improvement in cardiac size and function can be seen, especially if the disease is caught early.^4–9 We were unable to find any reports in the literature of an association between dietary ingredients/contaminants and phospholipidosis.

The exact mechanism for the accumulation of phospholipids in late endosomes and lysosomes in phospholipidosis is unknown, but proposed mechanisms include drug-induced increases in lysosomal pH, lysosomal membrane conjugates, inhibition of phospholipases, inhibition of lysosomal export processes, and overwhelming of normal lysosomal activity by increased rates of membrane/organelle degradation.^13,21 Urinary concentrations of di-22:6-BMP do not necessarily implicate primary cardiac disease and can reflect changes in other organs; thus, higher di-22:6-BMP concentrations in the urine could derive from organs other than the heart. Therefore, systematic ultrastructural evaluation of other tissues is important and could help to advance understanding of this disease.

Lamellar bodies are not specific to drug-induced phospholipidosis and can also be seen in hearts and other tissues of people with Fabry disease or Niemann-Pick disease.^11,15,22 Lamellar bodies can also sometimes be seen in diseased hearts, especially in areas of advanced degeneration.^23–26 Therefore, lamellar bodies could also be the result of other lysosomal abnormalities associated with diet. Current findings that higher di-22:6-BMP concentrations were associated more with HP diets than with DCM suggest that the presence of lamellar bodies in these dogs is less likely due to the underlying heart disease and more likely related to diet. However, additional research is needed before phospholipidosis can be determined to be the specific mechanism for diet-associated DCM.

Susceptibility to phospholipidosis and the organs affected vary among species.¹⁶ In preclinical drug approval studies in dogs, the most commonly affected target organs of phospholipidosis are reported to be lymphoid tissue > liver > biliary system > gastrointestinal tract.¹⁶ Dogs' hearts were also reported to be affected in preclinical drug studies¹⁶ but less commonly. In humans with drug-induced phospholipidosis, morphological changes at the light microscopic level typically trigger ultrastructural evaluation of the myocardium. Although we have found minimal changes in the myocardium of most dogs with diet-associated DCM, some dogs have had regional vacuolar degeneration on light microscopy. More extensive and systematic characterization of the myocardium with light microscopy in dogs with diet-associated DCM is warranted.

In addition to the association between HP diets and normalized urine concentrations of di-22:6-BMP, there was a positive correlation between urine di-22:6-BMP concentrations and diet pulse scores. It is important to note that the diet pulse score is not an exact quantitative measure of pulse concentrations in diet and has not been validated but was developed to provide a rough estimate of relative levels of these dietary ingredients. Numerous other factors besides dietary ingredient levels could also influence di-22:6-BMP concentrations, such as the duration dogs had been eating the diet, food intake, and the stage of heart disease.

This study has additional important limitations. The sample size is small, especially for the dogs with DCM eating LP diets. In the original study,⁷ only 9 of 60 (15%) dogs diagnosed with DCM were eating LP diets, and only 4 of those dogs had urine available at the time of diagnosis for analysis of di-22:6-BMP concentrations. One dog eating a “nontraditional diet” in the original study was excluded from the current analysis because that dog's diet contained potatoes but no pulses, so it did not meet the current study's diet group definition. However, including or excluding that dog in the analysis did not change the findings. Definitions for diets associated with this disease have been refined over the last several years as additional data have accumulated. The definition used in the current study focused on the presence of pulses. This definition still might need to be refined because, until the exact cause is known, it is impossible to target specific compounds or ingredient levels. In rodent studies¹⁵ of drug-induced phospholipidosis, urine di-22:6-BMP concentrations decrease after discontinuation of the drug. It is unknown whether this will be true for dogs with DCM eating HP diets after the diet is changed, but these additional analyses are being planned. Although di-22:6-BMP concentrations have previously been measured in dogs,¹⁵ its use as a biomarker for phospholipidosis in dogs has not been published. Therefore, complementary techniques are needed, such as EM, histochemical stains, and immunohistochemistry (eg, lysosome-associated membrane protein-2). Finally, higher urine concentrations of di-22:6-BMP do not prove that phospholipidosis is the mechanism involved in diet-associated DCM and could be a secondary effect of cardiomyocyte injury. Therefore, additional research on lysosomes, autophagy, and other complex cellular functions in diet-associated DCM is needed before phospholipidosis can be considered a mechanism for diet-associated DCM.

Despite these limitations, these results support the possible presence of primary or secondary phospholipidosis in dogs with diet-associated DCM and provide a plausible mechanism that fits with EM findings in these dogs. Therefore, this provides a pathway for additional research studies to better understand, prevent, and treat this disease.

Disclosures

In the last 3 years, Dr. Freeman has received research or residency funding from, given sponsored lectures for, or provided professional services for Aratana Therapeutics, Elanco, Guiding Stars Licensing Co LLC, Morris Animal Foundation, Nestlé Purina PetCare, and P&G Petcare (now Mars). In the last 3 years, Dr. Rush has received research, residency funding, or travel support from; given sponsored lectures for; attended sponsored seminars; or provided professional services for Aratana Therapeutics, Boehringer Ingelheim, CEVA Animal Health, Elanco, Increvet, MedVet, Morris Animal Foundation, Nestlé Purina PetCare and Veterinary Professional Development.

No AI-assisted technologies were used in the generation of this manuscript.

Funding

Funded by the Barkley Fund, Morris Animal Foundation, and Nestlé Purina PetCare.

ORCID

L. M. Freeman https://orcid.org/0000-0002-0569-9557

J. E. Rush https://orcid.org/0000-0002-8277-8996

E. G. Martinez-Romero https://orcid.org/0000-0001-5412-340X