Glucose Levels: Hyperglycemia and glucosuria are hallmarks for CDM and a requirement for the diagnosis and study entry. The CDM study dogs included nine new onset (diagnosed within –1 to 6 months) and three long-term (diagnosed greater than ≥18 months) diabetic dogs. The age, breed, and disease duration of participants is shown in Table 1. All diabetic dogs were greater than 5 years of age, which is typical of CDM [2]. Regardless of sex or breed, all dogs demonstrated significantly (p < 0.0001) elevated blood glucose compared to HC (Fig. 1A). The CDM dogs had a glucose range from 176 to 616 mg/dl, with a mean 394.0 ± 42.90 mg/dl (Fig. 1A). In human subjects, diabetes is diagnosed with a fasting blood glucose level of > 200 mg/dl [16]. In a laboratory setting, diabetes prone mice are diagnosed when blood glucose levels reach and maintain 250 mg/dl [17]. Glucose was detected in the urine of all the diabetic dogs, demonstrating glucosuria, which occurs when the blood glucose concentration exceeds the renal threshold for kidney regulation; approximately 180 mg/dL in canines.
Diabetic Ketoacidosis: Unregulated diabetes can progress to diabetic ketoacidosis (DKA) in both canine [18] and human subjects [19]. DKA clinical conditions include blood glucose greater than 250 mg/dl, the presence of ketone bodies by urinalysis, and serum bicarbonate levels less than 18 mmol/L, suggesting acidic conditions [19]. Of the 12 dogs in the study, four presented with likely DKA (Table 2). They had the highest blood glucose levels, each had detectable ketone bodies in urine, each had a urine pH under physiologic 7.0 indicating acidity, and each had bicarbonate levels under 18 (Table 2). Three other dogs were at risk, with a urine pH under 7 and urine glucose ≥ 2.0. Because CDM often goes undetected for long periods of time, DKA development can be common.
Hypercholeterolemia: During inflammation the increasing demand for adaptive and innate immune cells causes increased metabolic demands. A common mechanism to satisfy metabolic demands is increased cholesterol production and metabolism [20]. Cholesterol metabolism was shown to play a key role in maintenance of innate and adaptive immune cells [20]. Nine of the twelve CDM dogs demonstrated moderate to severe hypercholesterolemia (Fig. 1B). Two of the dogs had normal cholesterol levels and the other was borderline high (Fig. 1B). There was no direct correlation between serum glucose levels and cholesterol levels however (Observations).
Alkaline Phosphatase: Alkaline phosphatase (ALP) is a zinc modulated enzyme involved in protein metabolism. Increases in ALP often are associated with cholestatic disease, cancers, liver problems, bone disorders, or reactive hepatopathies. In human diabetic subjects increases in ALP are reported [21].In fact, high ALP is associated with increased risk for cardiovascular disease and mortality [21], and for changes in bone density [22]. In CDM dogs, ALP was significantly (p < 0.0001) increased compared to HC dogs (Fig. 2A). All CDM were substantially outside the normal range (Fig. 2A).
Fructosamine: A measure of disease control in human diabetes is glycated hemoglobin A1c (Hb A1c). In veterinary patients fructosamine, glycosylated albumin, is examined [23, 24]. Fructosamine turns over more rapidly than Hb A1c, every 14 days compared to 35 days [25]. Fructosamine levels in all diabetic dogs were substantially above the canine normal range (Fig. 2B). This data is consistent with unregulated disease.
C-peptide: A measure of beta cell function is detection of C-peptide. C-peptide is generated when pro-insulin is processed in the islet beta cell to mature insulin. The process creates the mature insulin A- B chain molecule and releases C-peptide in serum. When beta cell function is fully dysregulated, C-peptide is no longer detected. The CDM dogs each had detectable C- peptide, but at levels below the established normal range (Fig. 2C). This suggests some active beta cells are present in these dogs, but the activity is not sufficient to regulate the blood glucose concentration.
Lymphocyte Dysregulation: Autoimmune disease necessarily involves immunocyte dysfunction, but in CDM the question arises as to whether the disease etiology is autoimmune. Focusing on complete blood counts from patients, we examined total lymphocyte counts, and innate cells including monocytes and neutrophils. In diabetic subjects, lymphocyte numbers were significantly (p < 0.001) lower than in HC dogs (Fig. 3A). This observation includes both CD4 and CD8 T cells and B cells as cell type differentiation was not performed; clinical laboratories do not perform this differentiation on canine blood. A decrease in lymphocytes likely relates to changes in metabolism, to be discussed further.
Innate immune dysregulation: Monocyte numbers representing innate immune cells in CDM dogs were increased significantly (p = 0.0166) compared to HC dogs (Fig. 3B). Neutrophils, another innate immune cell, produce high levels of NADPH oxidase (NOX-2) which is implicated in autoimmunity and hydrogen peroxide generated from neutrophils suppresses lymphocyte activation and cell numbers [26]. While the overall number of neutrophils in CDM was increased there was not significant difference between CDM and HC neutrophil counts (Fig. 3C). However, the two CDM dogs that had substantially high neutrophil numbers were new onset diabetics.
Platelet dysregulation: Platelets play a clear role in hemostasis and thrombosis, but growing evidence suggest a contributory role to inflammation [27, 28]. Platelets are derived from megakaryocytes and inflammatory cytokines induce rupture of megakaryocytes to increase platelet numbers [29]. In both new onset and longer-term CDM, platelet numbers were significantly (p = 0.0007) elevated compared to HC (Fig. 3D). A feature of platelets that reflects both age of the platelet and platelet activity, i.e. more activated, is platelet size registered by mean platelet volume (MPV) [29]. Mean platelet volume in 7 of the 12 CDM dogs was above normal range (Fig. 3E).
Systemic/Chronic Inflammation: Inflammation involves interactions between adaptive immune cells, and innate cells. Attempting to develop inflammatory biomarkers, an index of neutrophil to lymphocyte ratio (NLR) has been created [30, 31]. The ratio accounts for dynamic changes in neutrophils and lymphocytes during disease development. NLR has been described in chronic diseases including coronary heart disease, stroke, diabetes, and cancer of solid organs [31]. In one study NLR increases were significantly associated with early neurological deterioration differentiating diabetic and non-diabetic subjects [30]. We compared NLR between CDM and HC dogs (Fig. 4A) with significantly (p < 0.0001) striking differences. All CDM dogs had substantially greater NLR. Another approach focusing on innate to adaptive immune cell differences compared the monocyte to lymphocyte ratio (MLR). MLR is more generally associated with cancers, but has been examined in autoimmune disorders including cardiovascular issues [32]. In CDM the MLR was significantly (p = 0.013) greater than in HC dogs (Fig 4B). Given the noted differences in platelets between CDM and HC we determined platelet to lymphocyte ratio (PLR). In CDM the PLR was significantly (p < 0.0001) higher than the value in HC dogs (Fig. 4C).
Further study of dynamic changes in immune cells as a measure of chronic inflammation led to the creation of the systemic-immune-inflammation index (SII), which is derived by multiplying the number of neutrophils and platelets then dividing by the number of lymphocytes [33, 34]. This index is used for prognosis of solid tumors but also to predict autoimmune outcomes, predominantly heart disease and rheumatoid arthritis [34]. In CDM the SII was significantly (p < 0.0001) greater than that in HC dogs (Fig. 5A). We considered that another analysis could be achieved by substituting monocytes for neutrophils in the index. A chronic inflammation index (CII) was derived by multiplying absolute number of monocytes and platelets then dividing by absolute number of lymphocytes (Fig. 5B). The difference between CDM and HC dogs was significant (p = 0.0008) and demonstrated a narrower range than the SII.
Th40 cells indicate chronic inflammation: In previous work we showed that a T cell subset described as Th40; CD3+CD4+ helper T cells that express the CD40 receptor [13], is increased in number in murine autoimmune diabetes [17, 35, 36] and in human autoimmune (type 1) diabetes [15, 37]. In the murine T1D model, Th40 cells proved to be pathogenic in adoptive transfer studies [17, 35]. In human DM Th40 cells were present in healthy controls and in type 2 (non-autoimmune) diabetes but at significantly lower numbers compared to T1DM [15]. We compared Th40 cell numbers in CDM to HC dogs. Three of the HC dogs were seen for osteoarthritis, a chronic but not autoimmune condition, otherwise they were seen for a healthy dog check-up, vaccines, or elective surgery (spay and neuter). As reported for laboratory mice and human diabetic subjects, CDM dogs had significantly (p < 0.0001) increased Th40 cell numbers (Fig. 6A). PBMC were analyzed for CD3 and then CD4+CD40+ within the CD3 gated subset (Fig. 6A for controls and 6B for CDM). In the forward versus side scatter plots, CDM demonstrated a high side scatter, lower forward scatter population (Fig. 6B) compared to HC dogs (Fig. 6A). In CDM the CD3 expression was lower than in HC. HC were further examined by age for Th40 cell differences, but no differences were noted (Suppl Fig 1A). There were no differences in Th40 cell numbers when differentiated by sex (Suppl Fig. 1B).