Effects of a therapeutic weight loss diet on weight loss and metabolic health in overweight and obese dogs

Abstract

Obesity has become a major health issue in dogs. Obesity in dogs increases the risk of many chronic diseases and chronic low-grade inflammation. The objective of this study was to determine the effect of a therapeutic weight loss (TWL) diet on weight loss and metabolic health in overweight and obese dogs. Thirty overweight and obese dogs were randomized into two groups with 15 dogs per group based on key baseline (BSL) parameters and allotted to either a control or TWL diet for 6 mo. At the start of the study, the control group had six females and nine males with mean age of 9.12 ± 0.48 (mean ± SEM) yr; there were seven females and eight males with mean age of 9.73 ± 0.63 yr in the TWL group. The control group and the TWL group had comparable body weight (34.78 ± 0.76 and 34.63 ± 0.86 kg, respectively), % body fat (BF; 39.77 ± 1.18 and 39.89 ± 0.93, respectively), and body condition score (BCS; 7.80 ± 0.14 and 7.67 ± 0.16 on a 9-point BCS scale, respectively). The control (CTRL) diet was formulated based on the macronutrient ratio of a commercial metabolic diet, and the TWL diet was enriched with dietary protein, fish oil, and soy germ meal. Both diets were fortified with essential nutrients to account for caloric restriction during weight loss. Dogs were fed with 25% less than BSL maintenance energy requirement (MER) for the first 4 mo and if they did not reach a BCS of 5, they were fed 40% less than BSL MER for the last 2 mo. Body composition was determined by dual-energy x-ray absorptiometry. Postprandial glucose profiles were determined by continuous glucose monitoring devices. Serum samples were collected for analyses of blood parameters, hormones, and cytokines. All data were analyzed using SAS 9.3, with significance being P < 0.05. At the end of the study, the control group and the TWL group had comparable weight loss (−5.77 ± 0.31 and −6.14 ± 0.32 kg, respectively; P = 0.4080). But the TWL group lost significantly (P = 0.034) more BF (−13.27 ± 1.28%) than the control group (−9.90 ± 1.23%). In addition, the TWL diet completely prevented loss of lean body mass (LBM) in dogs compared with BSL. Dogs fed with the TWL diet had significantly lower fasting serum cholesterol, triglycerides, insulin, leptin, mean postprandial interstitial glucose, and pro-inflammatory cytokines compared with dogs fed with the CTRL diet. In summary, the TWL diet prevented loss of LBM, promoted weight loss and metabolic health, and reduced pro-inflammatory cytokines and chemokines in overweight and obese dogs during weight loss.

A therapeutic weight loss diet prevented loss of lean body mass, promoted body fat loss and metabolic health, and reduced pro-inflammatory cytokines in overweight and obese dogs.

Introduction

Obesity has become a major health issue in dogs (German 2006, 2010; Laflamme, 2012). About 38.8% to 65% of dogs in developed countries were reported as overweight or obese (Courcier et al., 2010; German et al., 2018). Known risk factors for overweight and obesity in dogs include middle age, spaying/neutering, genetics (breed), indoor dwelling, owner age, hours of weekly exercise, and frequency of snacks and treats (German, 2006; Courcier et al., 2010).

Excess weight can reduce longevity and adversely affect quality of life in dogs (Kealy et al., 2002). Obesity is associated with chronic low-grade inflammation due to the increased levels of pro-inflammatory hormones, cytokines, and chemokines released by excess adipose tissue (Wakshlag et al., 2011). Although the impact of this is not completely understood (Wakshlag et al., 2011), the chronic inflammatory state as well as obesity- associated oxidative stress may contribute to the elevated risk of many chronic diseases including osteoarthritis, dermatopathy, gastrointestinal disease, neoplasia, oral disease, and urinary diseases (Pérez et al., 1998; Marshall et al., 2009; Laflamme, 2012). In addition, increased pro-inflammatory leptin, resistin, and tumor necrosis factor alpha (TNF-α) reduce insulin sensitivity throughout the body and thus likely contribute to the insulin resistance noted in overweight and obese dogs (Blanchard et al., 2004; (German et al., 2009); Radin et al., 2009; Laflamme, 2012).

Weight loss strategies typically involve reducing caloric intake while increasing energy expenditure with increased exercise. One major side effect of weight loss is loss of lean body mass (LBM) in addition to body fat (BF). A high-protein, extremely low-starch diet (47.5% crude protein and 5.3% starch, dry matter basis) was able to limit loss of LBM to about 20% of weight loss in obese dogs (Diez et al., 2002). Bierer and Bui (2004) reported that under mild caloric restriction (85% of baseline maintenance energy requirement [BSL MER]), a high-protein, low-carbohydrate diet (52% protein, 22% carbohydrate, 9% fat) led to a higher percentage of weight and fat mass loss in overweight and obese dogs compared with a high-carbohydrate diet (28% protein, 43% carbohydrate, 11% fat). Furthermore, dogs fed with the high-protein, low-carbohydrate diet showed decreased serum triglyceride levels at the end of the 12-wk study period, while serum triglycerides increased in dogs fed with the high-carbohydrate diet (Bierer and Bui, 2004). In people, high-protein diets have been shown to promote weight and BF loss, maintain LBM, and reduce blood triglycerides (Westerterp-Plantenga, 2003). Research in obese cats reported that high-protein diets promoted more loss of BF and less loss of LBM during weight loss than diets containing more moderate levels of protein (Laflamme and Hannah, 2005; Hoenig et al., 2007; des Courtis et al., 2015).

When compared to a low-fat diet in obese people, a low-carbohydrate diet resulted in significantly more weight loss and greater reduction in triglyceride level and fasting blood glucose (Samaha et al., 2003). In addition, the low-carbohydrate diet improved insulin sensitivity while the low-fat diet failed to improve it (Samaha et al., 2003). These results suggest that feeding overweight or obese dogs a weight loss diet with a high-protein, low-starch content may promote greater loss of BF and body weight (BW) while improving metabolic health.

Studies have shown that feeding spayed and neutered dogs a diet supplemented with soy isoflavones can help reduce weight gain (Pan 2006, 2007, 2012) by increasing energy expenditure (Pan, 2012). In addition, a diet containing isoflavones was able to reduce plasma isoprostanes (a biomarker of oxidative stress) and had a trend to reduce loss of LBM in overweight dogs during weight loss (Pan et al., 2008). Omega-3 (n-3) polyunsaturated fatty acids (PUFA) including eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) have anti-inflammatory properties (Oppedisano et al., 2020). Research demonstrated improved mobility in dogs with osteoarthritis fed with a therapeutic diet high in EPA and DHA from fish oil (Moreau et al., 2013).

Although many weight loss diets are available on the market, most have been developed simply to promote weight loss in obese dogs and fail to prevent the loss of LBM and address the above-mentioned comorbidities associated with chronic obesity. The main objective of this study was to develop a weight loss diet that promotes healthy weight loss (promotes BF loss and maintains LBM) and metabolic health while reducing pro-inflammatory cytokines and chemokines (indicators of chronic inflammation) in overweight and obese dogs. Our hypothesis was that a diet high in protein and omega-3 PUFA, rich in isoflavones, and low in starch could help us achieve the main objective.

Materials and Methods

Animals and BSL measurements

Thirty Labrador Retriever dogs with a body condition score (BCS) of 7 or higher on a 9-point BCS scale were recruited in this weight loss study. The sample size of 15 dogs per group was based on our previous study showing a significant difference in fat loss between treatment groups (Pan et al., 2008). At the start of the study, the control group had six females and nine males with mean age of 9.12 ± 0.48 (mean ± SEM) yr; there were seven females and eight males with mean age of 9.73 ± 0.63 yr in the TWL group. The control group and the TWL group had comparable BW (34.78 ± 0.76 and 34.63 ± 0.86 kg, respectively), % BF (39.77 ± 1.18 and 39.89 ± 0.93, respectively), and BCS (7.80 ± 0.14 and 7.67 ± 0.16, respectively). The study protocol was approved by the Nestlé Purina Institutional Animal Care and Use Committee. Standard husbandry and management practices were used throughout the study. Dogs were housed in a climate-controlled building with indoor/outdoor access. Indoor areas had both natural (windows) and artificial lighting, air conditioning/heating, and fans for air circulation. The kennel design provided an open visual field to allow each dog to see all other dogs in the kennel as well as any human caretakers or researchers coming in and out of the kennel. The inside kennel size was 50” × 108” (~37 square feet), and the outside kennel size was 52” × 180” (~65 square feet). Dogs were pair-housed with access to two individual indoor runs and a shared outdoor run at all times except during feeding, when dogs were separated into individual indoor runs to allow for accurate tracking of consumption. Each pair of dogs had access to elevated beds, environmental enrichment (toys), and an automated drinking system providing ad libitum access to fresh drinking water. Indoor/outdoor runs were cleaned once daily. Feeding bowls were cleaned every day. Dogs had daily contact with the care staff and their kennelmate. Every day, dogs were taken out for walks, play yard sessions, and/or socialization events together with other dogs from their established play groups.

The BSL MER was determined as the amount of the control (CTRL) diet that stabilized the dog’s BW (defined as less than 5% change over 3 wk). The amount of food was adjusted by decreasing or increasing 5% if a dog gained or lost weight by more than 5% after 3 wk of feeding. The adjustment was repeated until the dog’s BW was stabilized.

After the BSL MER was determined, BSL body composition was determined by dual-energy x-ray absorptiometry (DEXA; Lunar Prodigy model 8743, GE, Madison, WI). BSL postprandial glucose profiles were determined by Abbott FreeStyle Libre Pro continuous glucose monitoring devices (Abbott Diabetes Care Inc., Alameda, CA). Serum samples were collected for BSL analyses of metabolism, hormones, and cytokines.

Diets

A CTRL diet was formulated based on the macronutrient ratio of a commercial metabolic diet, and a therapeutic weight loss (TWL) diet was formulated to be enriched with dietary protein, fish oil, and soy germ meal. Both diets were fortified with essential nutrients to account for caloric restriction during weight loss. The nutrient profiles of both diets are summarized in Table 1. Both diets were manufactured by Nestlé Purina Petcare Company. Diet samples were submitted to Nestlé Purina Analytical Laboratories (Nestlé Purina Petcare, St. Louis, MO) for analysis of ash (AOAC, 2008; method 942.05), crude fat (AOAC, 2005; method 922.06), crude fiber (AOAC, 2010; method 962.09), total dietary fiber ((NO LABEL, 1995); method 991.43), crude protein (AOAC, 2005; method 990.03), moisture (AOAC, 2005; method 930.15), and fatty acid profile (n-3 PUFA, n-6 PUFA; AOAC, 2008; method 996.06). Isoflavone levels were measured by the analytical method reported by Klump et al. (2001).

Table 1.

The nutrient profiles of the diets¹

	CTRL diet	TWL diet
Moisture, %	8.09	8.07
Protein, %	26.47	48.70
Carbohydrate, %	34.12	22.19
Starch, %	31.60	15.65
Fat, %	14.73	10.1
Crude fiber, %	11.40	5.00
Total dietary fiber, %	19.70	12.93
Ash, %	5.19	5.94
n-3 PUFA, %	0.08902	1.2166
n-6 PUFA, %	1.62164	1.53587
Total isoflavone, aglucon units mg/kg	89.33	579.67
Calculated ME, kcal/kg	3181	3200

¹Analytical values except calculated ME, metabolizable energy.

Randomization

After the BSL MER for each dog was determined, the dogs were randomized into two groups (Control and TWL) with 15 dogs per group to ensure that there were no significant differences in MER, percentage of BF, BCS, and BW (see Table 2 for details). One dog from the test group was removed from the study due to an unrelated health issue.

Table 2.

BSL parameters of the dogs

	CTRL diet	TWL diet
Body weight, kg	34.78 ± 0.76	34.63 ± 0.86
LBM, kg	20.29 ± 0.65	20.09 ± 0.50
Body fat, kg	13.35 ± 0.42	13.38 ± 0.52
Body fat, %	39.77 ± 1.18	39.89 ± 0.93
BCS	7.80 ± 0.14	7.67 ± 0.16
MER, g of CTRL diet	472 ± 16.83	472 ± 18.59

Feeding protocol

Dogs were fed with 75% of their BSL MER for up to 4 mo and were removed from the study when their BCS reached 5. Dogs that had not reached a BCS of 5 by the end of the first 4 mo were energy-restricted to 60% of their BSL MER for an additional 2 mo of feeding to further promote weight loss or until their BCS reached 5.

Measurements and sample collections

During the 6 mo weight loss study, food intake was recorded daily, BW was recorded weekly, and BCS was recorded monthly. Body composition was determined with DEXA at BSL and every 2 mo (months 2, 4, and 6) or when a dog’s BCS reached 5. Complete blood count (CBC) and blood chemistry were assayed at BSL and every 2 mo (months 2, 4, and 6) or when a dog’s BCS reached 5. Serum separator tubes (SST; 456073P, Greiner Bio-One, Monroe, NC) were used to collect serum samples at BSL and every 2 mo (months 2, 4, and 6) or when a dog’s BCS reached 5. Serum cytokines, leptin, insulin, and adiponectin were measured. Cytokines were measured on the Luminex 200 instrument (Luminex Corp., Austin, TX) using a commercial multiplex assay (Millipore-Sigma, Burlington, MA). This canine multiplex panel contained the following analytes: granulocyte-macrophage colony-stimulating factor (GM-CSF), interferon gamma (IFN-γ), interleukin (IL)-2 (IL-2), IL-6, IL-7, IL-8, IL-15, IFN-γ-inducible protein of 10kDa (IP-10), keratinocyte chemotactic-like (KC-like), IL-10, IL-18, monocyte chemoattractant protein 1 (MCP-1), and tumor necrosis factor alpha (TNF-α) as reported previously (Bastien et al., 2015). Serum samples were submitted to Michigan State University (East Lansing, MI) for insulin assay and Metabolon for metabolomic analysis (Morrisville, NC). Commercial ELISA kits were used to measure adiponectin (KA0017, Abnova, Taipei, Taiwan) and leptin (EZCL-31K, Millipore-Sigma). ELISA plates were read at 450 nm on the Synergy H1 plate reader with the Gen 5 software (BioTek, Winooski, VT).

Postprandial interstitial glucose profiles

On days 3 and 60, 6 h postprandial glucose profiles for all dogs were determined with the continuous glucose monitoring sensors and readers. Sensors require 1 h start-up period following application before interstitial glucose was measured. The sensors measured interstitial glucose every 15 min and were scanned before feeding and rescanned at 6 h after the feeding to determine 6 h postprandial glucose profile. Vet wrap, collars, or ThunderShirts (ThunderWorks, Durham, NC) were used to ensure that sensors remained in place. Readers could detect sensors even if the sensor was covered.

Statistical analysis

SAS 9.3 (Copyright (c) 2002-2010 by SAS Institute Inc., Cary, NC) was used to conduct all statistical analysis.

The following repeated measures analysis of variance (ANOVA) model was used in the analysis of data including DEXA, blood chemistry, and leptin.

$${\displaystyle \begin{array}{l}{\displaystyle Y_{ijk}=\mu +{\alpha }_i+{\beta }_{j(i)}+{\tau }_k+{(\alpha \tau )}_{ik}+{\in }_{ijk}}\end{array}}$$

where

• µ = overall mean

• α _i  = effect of treatment i

• β _j(i) = random effect of dog j receiving treatment i

• τ _k  = effect of time k

• (ατ)_ik = treatment by time interaction

• ∈ _ijk  = experimental error

Data were analyzed with paired t-tests to compare change from BSL to end of study within each group; two sample t-tests were then used to compare those changes in IL-6, TNF-α, IL-18, and insulin from BSL to the end of the study between the two groups. Previous studies showed that high-protein diets (Hannah and Laflamme, 1998; German et al., 2010) promoted more fat loss than a standard low-fat weight loss diet. Since the TWL diet contained dietary protein levels higher than those of previous studies, we hypothesized that our high-protein TWL diet would lead to greater fat loss than the CTRL. Therefore, one-tail T test was performed on the data of BF loss in this study after the analysis of the repeated measures ANOVA model. Postprandial interstitial glucose levels were recorded every 15 min continuously. The mean postprandial glucose values on days 3 and 60 were calculated as the average of the recorded values within 5.5 and 6 h after the dogs consumed the meals, respectively. One dog in the TWL diet group was removed from the study due to an unrelated health condition after 4 mo. Data from 15 dogs in the control group and 14 in the TWL group were used in the statistical analysis.

Results

Effects of diets on weight loss, body composition, and BCS in dogs

Weight loss for both groups remained less than 1% per week for the duration of the study. Dogs in both groups had comparable rates of weight loss during the first 4 mo of 75% MER feeding and the last 2 mo of 60% MER feeding (0.59 ± 0.06%/wk for control vs. 0.65 ± 0.06%/wk for TWL during 75% MER feeding; 0.79 ± 0.08%/wk for control vs. 0.85 ± 0.08%/week for TWL during 60% MER feeding). Both diets promoted BW loss and BF loss (P < 0.05) compared with BSL (Figures 1 and 2). Repeated measures ANOVA showed significant effects of the diets on BF (P = 0.0408), % BF (P = 0.0265) and lean body mass (P = 0.0025) and did not detect significant interaction between diet and time for those three parameters. The ANOVA analysis did not detect significant effects of the diets on BW (P = 0.5560). While dogs fed with the CTRL diet lost a significant amount (P < 0.05) of LBM compared with BSL, dogs receiving the TWL diet did not (Figure 3). Dogs in both groups lost comparable amounts of BW at 2, 4, and 6 mo (Figure 1), but control dogs lost more (P < 0.05) LBM (Figure 3) and less (P < 0.05) BF (Figures 1 and 2) than dogs fed with the TWL diet at 2, 4, and 6 mo. Both diets lowered (P < 0.05) BCS at the end of the study compared with BSL, but dogs fed with the TWL diet had lower BCS than dogs fed with the CTRL diet (5.43 ± 0.17 vs. 5.93 ± 0.12, respectively; P = 0.0215).

Figure 1.

Change in BW and BF, kg, from BSL in dogs fed with CTRL (blue bars) or TWL(pink bars) diet. Data represent the mean ± SEM. Dogs in both groups lost significant amounts of BW and BF at months 2, 4, and 6 compared to BSL (P < 0.05). The amount of BW lost at each timepoint was not significantly different between dogs in CTRL group (solid blue bars) and TWL group (solid pink bars). However, dogs fed with the TWL diet (hashed pink bars) lost significantly more BF at each timepoint compared to dogs fed with the CTRL diet (hashed blue bars).

Figure 2.

Change in BF, %, from BSL in dogs fed with CTRL (blue bars) or TWL (pink bars) diet. Data represent the mean ± SEM. Dogs in both groups lost a significant percentage of BF at months 2, 4, and 6 compared to BSL (P < 0.05). However, dogs fed with the TWL diet lost a significantly greater percentage of BF at each timepoint compared to dogs fed with the CTRL diet.

Figure 3.

Change in LBM, kg, from BSL in dogs fed with CTRL (blue bars) or TWL (pink bars) diet. Data represent the mean ± SEM. Dogs fed with the TWL diet maintained LBM, while dogs fed with the CTRL diet lost a significant amount of LBM at months 2, 4, and 6 compared to BSL (P < 0.05). Thus, there was a significant difference between the dogs fed with the TWL diet and those fed with the CTRL diet at all timepoints.

Effects of diets on CBC, blood chemistry, and blood lipid profiles in dogs

All parameters of CBC and blood chemistry were within the normal ranges during the study (data not shown). Repeated measures ANOVA showed significant effects of diet (P < 0.0001), time (P < 0.0001), and interaction between diet and time (P < 0.0001) for cholesterol. In addition, repeated measures ANOVA showed significant effects of time (P < 0.0001), and an interaction between diet and time (P = 0.0369) for triglycerides. Dogs in both groups had elevated fasting serum cholesterol at BSL (Figure 4). Both diets reduced (P < 0.05) fasting serum cholesterol and triglycerides compared with BSL, with the exception of triglycerides in the control group at month 2 (Figures 4 and 5). More interestingly, the TWL diet was more effective (P < 0.05) in reducing fasting serum cholesterol and triglycerides compared to the CTRL except for triglycerides at month 4 (Figures 4 and 5).

Figure 4.

Effects of the diets on fasting serum cholesterol, mg/dL, in dogs fed with CTRL (blue bars) or TWL (pink bars) diet. Data represent the mean ± SEM. Fasting serum cholesterol was significantly reduced in both groups of dogs at months 2, 4, and 6 compared to BSL (P < 0.05). However, the reduction in serum cholesterol was significantly greater for dogs fed with the TWL diet at each timepoint compared to dogs fed with the CTRL diet.

Figure 5.

Effects of the diets on fasting serum triglycerides, mg/dL, in dogs fed with CTRL (blue bars) or TWL (pink bars) diet. Data represent the mean ± SEM. Fasting serum triglycerides were significantly reduced in dogs fed with the TWL diet at month 2 and in both groups at months 4 and 6 compared to BSL (P < 0.05). The reduction in serum triglycerides was significantly greater for dogs fed with the TWL diet at months 2 and 6 compared to dogs fed with the CTRL diet.

Effects of diets on blood hormone profiles in dogs

Dogs fed with the TWL diet had a significant decrease (P < 0.05) in fasting insulin compared to dogs fed with the CTRL diet at the end of the weight loss study (−3.04 ± 1.07 uIU/mL and 0.76 ± 0.77 uIU/mL, respectively). Repeated measures ANOVA showed significant effects of diet (P = 0.0160), and time (P < 0.0001), but no significant interaction between diet and time (P = 0.1079) for leptin. Both diets resulted in a decrease (P < 0.05) in fasting serum leptin, but the TWL diet led to greater changes from BSL in fasting serum leptin compared with those of the CTRL diet at 2 and 6 mo of the study (Figure 6). Neither diet affected fasting serum adiponectin (data not shown).

Figure 6.

Change in serum leptin, %, from BSL in dogs fed with CTRL (blue bars) or TWL (pink bars) diet. Data represent the mean ± SEM. Serum leptin was significantly reduced in both groups of dogs on months 2, 4, and 6 compared to BSL (P < 0.05). Dogs fed with the TWL diet had a significantly greater reduction on months 2 and 6 compared to dogs fed with the CTRL diet.

Effects of diets on blood cytokine and chemokine profiles in dogs

No significant differences were observed between the CTRL group and TWL group in BSL cytokines and chemokines (data not shown). The TWL diet resulted in greater (P < 0.05) percentage changes from BSL in IL-6, IL-18, and TNF-α compared with the CTRL diet at the end of the study (Figure 7). Dogs fed with the TWL diet had significantly reduced (P < 0.05) pro-inflammatory cytokines (IL-2, IL-6, IL-7, IL-15, IL-18, INF-γ, and TNF-α), anti-inflammatory cytokine (IL-10), and pro-inflammatory chemokines (IP-10 and MCP-1) at the end of the study compared with BSL levels (Figure 8). The CTRL diet significantly reduced (P < 0.05) only one pro-inflammatory chemokine (IP-10) at the end of the study compared with BSL levels in the dogs (Figure 8).

Figure 7.

Change in serum IL-6, IL-18, and TNF-α, %, at the end of the study period (month 6) from BSL in dogs fed with control (CTRL; blue bars) or TWL (pink bars) diet. Data represent the mean ± SEM. At month 6, serum IL-6, IL-18, and TNF-α were significantly reduced in dogs fed with the TWL diet compared to BSL (P < 0.05) and compared to dogs fed with the CTRL diet.

Figure 8.

Changes in serum cytokines and chemokines, pg/mL, at the end of the study period (month 6) from BSL in dogs fed with CTRL (blue bars) or TWL (pink bars) diet. Data represent the mean ± SEM. At the end of the study, all serum cytokines and chemokines measured were significantly reduced in dogs fed with the TWL diet compared to BSL; only serum IP-10 was significantly reduced in dogs fed with the CTRL diet compared to BSL (*P < 0.05).

Effects of diets on postprandial interstitial glucose profiles in dogs

No significant differences were observed between the CTRL group and TWL group in BSL fasting interstitial glucose (88.46 ± 4.91 mg/dL and 93.50 ± 5.60 mg/dL, respectively; P = 0.5062). Dogs fed the TWL diet had lower (P = 0.0027) mean postprandial interstitial glucose within 5.5 h after a meal compared with the dogs fed the CTRL diet on day 3 of weight loss feeding (Figure 9). After 2 mo of weight loss feeding, dogs fed with the TWL diet had lower (P < 0.0001) mean postprandial interstitial glucose than dogs fed with the CTRL diet within 6 h of food consumption (Figure 9).

Figure 9.

Mean postprandial interstitial glucose at days 3 and 60 in dogs fed with CTRL (blue line) or (pink line) diet. Data represent the mean ± SEM. Dogs fed with the TWL diet had significantly lower mean postprandial glucose values at both timepoints compared to dogs fed with the CTRL diet.

Effects of diets on biomarkers of skeletal muscle physiology in dogs

Because dogs in both groups were fed the CTRL diet to determine individual MER, both 1-methylhistidine and 3-methylhistidine levels were comparable at BSL (Figures 10 and 11). Dogs fed with the CTRL diet had higher (P < 0.05) 1-methylhistidine than their BSL value and dogs fed with the TWL diet at the end of the study (Figure 10). There was no significant difference in 3-methylhistidine between BSL and the end of the study in dogs fed with the CTRL diet (Figure 11). Concentrations of both 1-methylhistidine and 3-methylhistidine at the end of the study in dogs fed with the TWL diet were lower (P < 0.05) than those at their BSL and those at the end of the study in dogs fed with the CTRL diet (Figures 10 and 11). While other metabolomics analysis was completed, those results will be presented in a separate manuscript.

Figure 10.

Changes in serum 1-methylhistidine in dogs fed with CTRL (blue) or TWL (pink) diet. Serum samples collected at BSL and end of the study were subject to metabolomic analysis. Plus signs represent the mean. ^*Significantly different (P < 0.05) from BSL. ^§Significantly different (P < 0.05) from BSL and from CTRL dogs. N = 15 for CTRL and N = 14 for the TWL diet group.

Figure 11.

Changes in serum 3-methylhistidine in dogs fed with CTRL (blue) or TWL (pink) diet. Serum samples collected at BSL and end of the study were subject to metabolomic analysis. Plus signs represent the mean; and open circles represent extreme data points. *Significantly different (P < 0.05) from BSL and from CTRL dogs. N = 15 for CTRL and N = 14 for the TWL diet groups.

Discussion

While dogs in both diet groups lost a comparable amount of BW, the TWL diet promoted more BF loss compared to the CTRL diet. This observation is consistent with previous reports showing that high-protein diets lead to greater fat loss in dogs (Hannah and Laflamme, 1998; German et al., 2010). More interestingly, the TWL diet prevented significant loss of LBM compared with BSL, while dogs fed with the CTRL diet had significant loss of LBM compared with BSL. The effect of the TWL diet on preserving LBM was quite unexpected, as Diez et al. (2002) reported that obese dogs fed with a high-protein (47.5%), extremely low-starch (5.3%) diet with a protein to starch ratio of 9:1 had 20% of weight loss coming from LBM. The protein level of the TWL diet (48.7%) was similar to that of the Diez et al. test diet, but our test diet contained more starch (15.65%). A weight loss diet enriched with isoflavones from soybean germ meal had a trend to reduce loss of LBM in overweight dogs during weight loss (Pan et al., 2008). Therefore, it is possible that nutrients in the TWL diet, such as high protein, n-3 PUFA, and isoflavones, may work synergistically to preserve LBM during weight loss in dogs. Another possible explanation for the TWL’s effects on LBM was its slower weight loss rate (0.85%/wk) compared with that of the Diez study (1.95%/wk).

3-methylhistidine is a marker of muscle protein breakdown (Young et al., 1973; Bilmazes et al., 1978), and 1-methylhistidine is a component of anserine (beta-alanyl-1-methyl-histidine), which is mainly present in skeletal muscle in animals except people (Sjölin et al., 1987). Therefore, 1-methylhistidine level has been used as an index of meat consumption in people because human skeletal muscle does not contain anserine (Sjölin et al., 1987). Since dogs’ skeletal muscle contains anserine (Harris et al., 1990), blood 1-methylhistine in dogs comes from both dietary intake of animal protein and their own skeletal muscle, which may explain why the 1-methylhistine in the blood of CTRL dogs increased over BSL even though their dietary intake of protein was reduced during the study (due to caloric restriction). In dogs, 3-methylhistidine can also come from dietary animal protein and muscle protein breakdown. Again, even though intake of 3-methylhistidine from protein was reduced during the study in CTRL dogs due to caloric restriction, no significant difference was observed between the BSL and end of the study. This data indicates that dogs fed with the CTRL diet had higher skeletal muscle protein breakdown compared to BSL, which led to release of more 3-methylhistine and 1-methylhistine from muscle into the blood in the CTRL dogs. These data are consistent with significant loss of LBM in the CTRL dogs. Interestingly, both biomarkers were lower (P < 0.05) in the TWL diet dogs at the end of the study compared with BSL. The TWL diet, while an overall higher protein diet, contained a lower inclusion rate of animal protein compared to CTRL (13.16% vs. 20.75%, respectively). The fact that the dogs fed with the TWL diet did not lose a significant amount of LBM supports the hypothesis that the reduction in these biomarkers at the end of feeding is due to both reduced dietary intake and decreased skeletal muscle breakdown compared to BSL.

All CBC and blood chemistry parameters were within the normal range during the study, which confirmed that the dogs were healthy and consumed adequate amounts of essential nutrients during the study. Both diets reduced (P < 0.05) fasting serum cholesterol in dogs after weight loss, which is consistent with the report by Tvarijonaviciute et al. (2012), showing that cholesterol was significantly reduced in obese dogs after weight loss. However, the TWL diet reduced fasting cholesterol more (P < 0.05) than the CTRL diet. Both diets also reduced (P < 0.05) fasting serum triglycerides compared with BSL at all timepoints except the CTRL diet after 2 mo of weight loss. In addition, the TWL diet reduced fasting serum triglycerides more than the CTRL diet except at month 4. On the contrary, Tvarijonaviciute et al. (2012) reported a nonsignificant reduction in triglycerides in obese dogs after weight loss.

While the CTRL diet did not affect fasting serum insulin during weight loss, the TWL diet reduced (P < 0.05) fasting serum insulin at the end of the study. Tvarijonaviciute et al. (2012) also reported a significant reduction in insulin in obese dogs after weight loss. Blanchard et al. (2004) reported a significant reduction in insulin sensitivity in dogs when they are obese compared to an ideal body condition. Both diets reduced (P < 0.05) fasting serum leptin compared with BSL, and the TWL diet resulted in a greater decrease (P < 0.05) in fasting serum leptin compared to CTRL diet except after 4 mo of weight loss. Leptin is strongly correlated with total BF and percentage BF (Kempf et al., 2006). The TWL diet promoted more fat loss than the CTRL diet, which may explain the greater reduction (P < 0.05) in fasting leptin in dogs fed with the TWL diet compared with the CTRL diet.

The TWL diet resulted in greater (P < 0.05) percentage changes from BSL in IL-6, IL-18 and TNF-α compared with the CTRL at the end of the study. The dogs fed with the TWL diet had reduced (P < 0.05) pro-inflammatory cytokines (IL-2, IL-6, IL-7, IL-15, IL-18, INF-γ, and TNF-α), anti-inflammatory cytokine (IL-10), and pro-inflammatory chemokines (IP-10 and MCP-1) at the end of the study compared with BSL levels. Bastien et al. (2015) also reported significant decreases in several pro-inflammatory cytokines or chemokines including IL-2, IL-7, IL-18, and MCP-1 in overweight dogs after weight loss. On the other hand, the CTRL diet reduced (P < 0.05) only one pro-inflammatory chemokine (IP-10) at the end of the study compared with BSL levels.

Dogs fed the TWL diet had lower (P < 0.05) mean postprandial interstitial glucose within 5.5 h after a meal compared with the dogs fed with the CTRL diet on day 3 of weight loss feeding. These data indicate that the glucose-lowering effect of the TWL diet is independent of weight loss since the dogs did not have significant weight loss on day 3 of feeding. After 2 mo of weight loss, dogs fed with the TWL diet had lower (P < 0.05) mean postprandial interstitial glucose than the dogs fed with the CTRL diet within 6 h after a meal. The TWL diet contained almost 12% less carbohydrate and 22% more protein than the CTRL diet. Starch in the CTRL diet was nearly twice the amount of the TWL diet. Postprandial glucose response in dogs is most impacted by the amount of starch in a diet (Nguyen et al., 1994; André et al., 2017). Dietary protein has a limited impact on postprandial glycemic response in both humans and dogs, potentially due to the effect of insulin on hepatic gluconeogenesis (Elliott et al., 2012). The impact of reduced starch on postprandial glucose has been noted in both healthy and obese dogs. This data indicates that the glucose-lowering effect of the TWLdiet in this study remains after significant weight loss in dogs.

Because the dogs were housed in a controlled facility, compliance in following the study protocol was ensured. The observed benefits in the TWL diet may apply to obese humans during weight loss. However, one limitation is that the TWL diet contained multiple nutrients including high protein, low starch, high omega-3 PUFA, and isoflavones, so it is not possible to conclude which of the nutrients in the TWL diet contributes to the LBM-sparing benefits.

In summary, dogs fed with both diets lost (P < 0.05) BW and BF compared with BSL. While both groups lost a comparable amount of BW, dogs fed with the TWL diet lost more (P < 0.05) BF and percentage BF than the dogs fed with the CTRL diet. In addition, dogs fed with the CTRL diet lost (P < 0.05) LBM, while the TWL diet prevented significant loss of LBM during weight loss. Metabolomic analysis suggested that dogs fed with the TWL diet had less muscle protein breakdown at the end of the study compared with BSL. The TWL diet promoted better metabolic health by reducing (P < 0.05) fasting blood cholesterol, triglycerides, insulin, and leptin and by maintaining lower (P < 0.05) postprandial interstitial glucose compared to the CTRL diet. The TWL diet lowered (P < 0.05) many pro-inflammatory cytokines and chemokines in dogs compared with BSL, while the CTRL diet lowered (P < 0.05) only one pro-inflammatory chemokine. These data confirm that the TWL diet was able to promote healthy weight loss and metabolic health and reduce chronic inflammation in overweight and obese dogs.

Glossary

Abbreviations:

BCS

body condition score

BF

body fat

BSL

baseline

BW

body weight

CBC

complete blood count

CTRL

control

DEXA

dual-energy x-ray absorptiometry

DHA

docosahexaenoic acid

ELISA

enzyme-linked immunosorbent assay

EPA

eicosapentaenoic acid

GM-CSF

granulocyte-macrophage colony-stimulating factor

IFN-γ

interferon gamma

IL

interleukin

IP-10

INF-γ-inducible protein of 10kDa

KC-like

keratinocyte chemotactic like

LBM

lean body mass;

MCP-1

monocyte chemoattractant protein 1

MER

maintenance energy requirement

n-3

omega-3

PUFA

polyunsaturated fatty acids

SST

serum separator tubes

TNF-α

tumor necrosis factor alpha

TWL

therapeutic weight loss

Author Contributions

Yuanlong Pan (Study design, Study protocol setup, Supervision of the study, Result interpretation, Writing of the manuscript), Julie K. Spears ( Study design, Study protocol setup, Study execution, Result interpretation, Writing of the manuscript), Hui Xu (Dietary formulation, Result interpretation), Sandeep Bhatnagar (Diet production, Sample submission). All authors read and approved the final manuscript.

Conflict of interest statement. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.