The endocannabinoid system is a widely expressed entity that has many roles in the physiologic events that are required to maintain and restore cell and tissue homeostasis. This signaling system has several important roles in the pathways involved in tumor growth and progression which has been previously documented in DLBCL and CLL in humans (14, 15). Here we demonstrated for the first time significantly higher levels of OEA, AEA, and PEA in dogs with lymphoma compared to healthy controls. The involvement of the endocannabinoid system in tumor progression is still under investigation and this relationship can be challenging to elucidate as endocannabinoids can have both pro- and anti-cancer effects. However, endocannabinoids have been postulated to potentiate cancer cell proliferation and survival through a variety of mechanisms. Zhang et al. demonstrated that AEA has pro-angiogenic properties through potentiation of fibroblast growth factor 2, which stimulates endothelial cell proliferation (25). Additionally, it has been demonstrated that exogenous endocannabinoids applied at nanomolar concentrations can induce cancer cell proliferation in glioblastoma and lung carcinoma cell lines through activation of epidermal growth factor receptor (EGFR), a receptor tyrosine kinase commonly upregulated in cancer that can activate downstream pro-oncogenic signaling pathways resulting in cancer cell proliferation (26, 27).

Our data found that in opposition with the other measured endocannabinoids, total AG is higher in healthy control dogs compared to dogs with lymphoma. This contrasts with the study by Zhang et al. that found significantly elevated levels of 2-AG in humans with DLBCL (14). In aqueous environments 2-AG spontaneously isomerizes to 1-AG via acyl migration and can be challenging to measure independently (19, 23). Most methods assume that 2-AG and 1-AG both originate from biological 2-AG (19); therefore, in our study the peaks were measured together and summed to total AG. However, a recent study found that 1-AG is a biologically active compound with its own effects on CB1. Therefore, it is possible that measuring total AG instead of 2-AG led to the opposing results from Zhang et al. (14). Alternatively, this could imply that downregulation of the 2-AG signaling system is involved in the pathogenesis of canine multicentric lymphoma. Elevated levels of MAGL, which is the enzyme responsible for degradation of 2-AG, has been shown to be highly expressed in a variety of aggressive human cancer cells and primary tumors (28). Additionally, activation of CB1 by 2-AG has been demonstrated to induce cell apoptosis through activation of ERK 1/2 expression and increased ceramide synthesis (29). Therefore it is possible downregulation of 2-AG is contributing to the progression of canine multicentric lymphoma and further investigation is warranted to evaluate MAGL levels and CB1 activation.

We found a significant relationship between endocannabinoid levels and age, with decreasing endocannabinoid levels in older normal dogs and older dogs with T cell lymphoma. Interestingly, the reverse was true in dogs with B cell lymphoma, with increased endocannabinoids in older dogs with B cell lymphoma. It has been shown in both murine and human experimental and preclinical models that the endocannabinoid system varies with age. Although there has been some conflicting reports, studies have overall shown a reduction in CB1 stimulated activity and reduction in 2-AG with increasing age (30). It is thought that these changes are particularly linked to learning disturbances and sleep disorders (31, 32). The changes in endocannabinoid levels in patients with neoplasia have not been specifically evaluated in humans. In a study of patients with DLBCL, participants were age matched with their controls, thus the effect of age was not evaluated (14). Masoodi et al. found that there was no significant relationship between OEA levels and age in their control population or the CLL population; however, it is important to note the majority of their CLL patients were older with 97% being above 50 years and 80% above 60 years old (15). Our results contrast a study on canine chronic enteropathy by Febo et al. in which there was no significant association between endocannabinoids and age in healthy dogs (33). This may have been influenced by the age of the control group, the median age of our control group was 84 months as compared to 60 months in their population, however further research evaluating the impact of age is needed.

There were no correlations between endocannabinoid levels and lymphoma immunophenotype. In our population, the dogs with T cell lymphoma had a shorter PFI (79.5 days) compared to the dogs with B cell lymphoma (237 days) which is consistent with prior reports of shorter PFI in multicentric T cell lymphoma (1, 34). The lack of difference in the measured endocannabinoids suggests that they are not a factor responsible for the more aggressive biologic behavior of T cell lymphoma.

Elevated PEA levels were significantly associated with shortened PFI. PEA has not been previously evaluated in the context of human or canine lymphoma. PEA is an endogenous compound that has been primarily investigated for its anti-inflammatory properties, specifically its impact on reducing degranulation of mast cells in response to substance P (35). Mast cells have various roles in neoplasia and their accumulation can be either beneficial or detrimental based on the tumor microenvironment and the stage of disease. A protective role for mast cells has been demonstrated in early stages of human intestinal carcinoma where mast cells promote apoptosis (36). A similar finding was reported in canine mammary gland tumors where the density of stromal mast cell tumors predicted disease outcome, with higher metastatic rates in tumors with lower numbers of stromal mast cells (37). The literature with respect to lymphoma is conflicting but Hedstrom et al. found that mast cell infiltration is a favorable prognostic factor in people with DLBCL; patients with increased numbers of nodal mast cells had significantly better event-free survival (38). Studies in canine nodal lymphoma found significantly higher mast cell counts compared to control lymph nodes (39); however, they did not evaluate the prognostic implications. Overall, intratumoral mast cells may have a protective role in canine lymphoma that is being suppressed by elevated PEA. Alternatively, PEA has been shown to have both direct and indirect effects on a variety of different receptor targets (40) including CB2 and PPAR-α. Therefore, it is possible that PEA has other implications in canine lymphoma not related to mast cells and further investigation is needed.

OEA was found to be elevated in dogs with lymphoma compared to healthy controls. OEA is a naturally occurring ethanolamide lipid that acts as a satiety factor. OEA has been demonstrated to stimulate lipolysis in adipose tissue, bone marrow, and lymph nodes through its activation of PPAR-α (41) leading to increased levels of circulating free fatty acids. In humans, abnormalities in lipid metabolism have been previously reported in leukemia and lymphoma with increased levels of triglycerides and free fatty acids being demonstrated at diagnosis especially in patients with a high disease burden or bone marrow involvement (42, 43). Masoodi et al. found high levels of free fatty acids accompanied the elevated OEA levels in patients with CLL which correlated with OEA causing adipocyte lipolysis. They also found that adipocyte-derived lipid factors protected the CLL cells from chemotherapeutic drugs in vitro with increased cell survival when the CLL cells were cultured with adipocyte-conditioned media and vincristine as opposed to vincristine alone (15). Although it did not reach significance, OEA was trending toward being negatively associated with PFI. Therefore, it is possible that OEA may represent a non-invasive measurement of prognosis and a potential therapeutic target; however, further research is needed with increased number of dogs.

There was no significant association between the endocannabinoids measured and the OST in this population of dogs. Previous studies in humans have not evaluated the relationship between endocannabinoids and OST. OST can be challenging to evaluate in veterinary retrospective studies due to lack of standardization of the type and number of rescue protocols pursued. These differences may have influenced the OST in our population.

The limitations of this study are predominantly due to its retrospective nature. First, the majority of serum samples were obtained between 8:00 a.m. and 12:00 p.m.; however, some samples were obtained outside of this time frame. Although circadian rhythm had been shown to affect endocannabinoid levels in people and mice (44, 45), this has not been demonstrated specifically in dogs; however, in future studies this aspect should be considered and controlled. Additionally, the duration of storage of serum prior to the measurement of endocannabinoids was variable, raising the potential for inconsistency due to endocannabinoid breakdown. To date the stability in canine serum has not been validated, however a human study by Bobrich et al. found that AEA and 2-AG in serum were stable at 60 days when stored at −80°C (46). Thirdly, the control group was small and was not intentionally age or weight matched to the lymphoma population. For the lymphoma patients, full staging was not performed for all patients and substage was assigned retrospectively based on information in the medical record. The case management was not standardized, leading to some variation in treatment decision making and case follow up. Lastly, the endocannabinoid levels were only measured once at the time of diagnosis. Further studies with repeated endocannabinoid measurements throughout the CHOP protocol are necessary to evaluate the changes with respect to response to treatment and disease progression.

In conclusion, we found that OEA, AEA, and PEA levels are significantly higher in dogs with lymphoma compared to healthy controls at the time of diagnosis. Furthermore, elevated PEA at the time of CHOP chemotherapy initiation is associated with a shorter PFI. Further research into both the endocannabinoid receptors and the enzymes involved in their synthesis and degradation are needed to better elucidate the underlying mechanisms of this system and its effects on canine lymphoma. Additionally, prospective studies evaluating dynamic changes of endocannabinoids throughout the CHOP protocol and evaluation of potential therapeutic targets are warranted.