Lowering fasting blood glucose with non‐dialyzable material of cranberry extract is dependent on host genetic background, sex and diet

Abstract Background Type 2 diabetes (T2D) is a polygenic metabolic disease, characterized by high fasting blood glucose (FBG). The ability of cranberry (CRN) fruit to regulate glycemia in T2D patients is well known. Here, a cohort of 13 lines of the genetically diverse Collaborative Cross (CC) mouse model was assessed for the effect of non‐dialyzable material (NDM) of cranberry extract in lowering fasting blood glucose. Methods Eight‐week‐old mice were maintained on either a standard chow diet (control group) or a high‐fat diet (HFD) for 12 weeks, followed by injections of intraperitoneal (IP) NDM (50 mg/kg) per mouse, three times a week for the next 6 weeks. Absolute FBG (mg/dl) was measured bi‐weekly and percentage changes in FBG (%FBG) between weeks 0 and 12 were calculated. Results Statistical analysis showed a significant decrease in FBG between weeks 0 and 12 in male and female mice maintained on CHD. However, a non‐significant increase in FBG values was observed in male and female mice maintained on HFD during the same period. Following administration of NDM during the following 6 weeks, the results show a variation in significant levels of FBG lowering between lines, male and female mice and under the different diets. Conclusion The results suggest that the efficacy of NDM treatment in lowering FGB depends on host genetic background (pharmacogenetics), sex of the mouse (pharmacosex), and diet (pharmacodiet). All these results support the need for follow‐up research to better understand and implement a personalized medicine approach/utilization of NDM for reducing FBG.


| INTRODUC TI ON
Many studies aimed at developing glucose-lowering drugs in patients with type 2 diabetes (T2D) have been published during the last three decades. [1][2][3][4] However, most pharmaceutical glucose-lowering drugs approved for clinical use do not lead to satisfactory treatment. The safety of such drugs is a major drawback. In particular, the adverse effects of such drugs on cardiovascular complications have been of major concern. 5 The Food and Drug Administration (FDA) has issued a request to pharma companies for more stringent measures in assessing the safety of glucose-lowering drugs. Consequently, research has focused on alternative glucose-lowering drugs with minimal or no side effects obtained from plant extracts. Vieira et al. 2 have reviewed the most recent studies on novel natural sugar lowering compounds and plant extracts in the treatment of T2D. Most of these studies employed animal models (rats or mice) to evaluate the glucose-lowering activity of the test plant extracts. Significantly, cranberry extracts were shown to reduce insulin resistance, as well as other markers of diet-induced metabolic syndrome, 6,7  Full details of CC lines and their power of mapping quantitative trait loci (QTL) with host susceptibility to complex traits are presented in various publications. [8][9][10][11] Briefly, the CC lines were generated using a breeding scheme in which the genomes of the eight founders were combined in three out-breeding generations, followed by inbreeding generations through sibling mating. [8][9][10][11][12] The CC is a large panel of  [8][9][10][11][12] They allow a mapping resolution down to less than one mega base pair (Mb) by phenotyping only a relatively modest number -<100 -of CC lines. 8,9,12 The choice of CC mouse lines also relies on many studies showing that this genotyped mouse population allows determination of the heritability of a test trait employing a relatively low number of mice lines. 11 Based on these findings, in the current study, CC mice lines on the HFD were treated with cranberry extract termed NDM. [13][14][15][16] This unique model is a platform for research of complex traits that overcomes the limitation of existing animal models. 9 The CC population is a powerful tool for genetic dissection of the different phenotypes. This large, multi-parental, recombinant inbred strains panel exhibits recombinant chromosomal segments of approximately 2-5 Mb in size. The eight founder strains capture a much greater level of genetic diversity than existing RIL panels or other extant mouse genetic resource populations. [17][18][19] The high molecular genomic DNA of the CC lines was genotyped in three ways: once using an array of 620 000 single nucleotide polymorphism (SNP) markers of mouse diversity, 8 once with a mouse universal genotyping array (MUGA) using 7500 markers, and finally with a mega mouse universal genotyping array (MegaMUGA) of 77 800 markers. The updated genotype status of the population has been presented, 17 showing the diversity of the phenotypic response of the CC lines to complex disease and their power to map quantitative trait loci (QTL) associated with a particular trait to a small genomic interval less than 0.5 cM, which consists of few genes. [20][21][22][23][24][25][26][27][28] Recently, we have shown that the host genetics background shapes the pharmacogenetics effect of NDM on reducing body weight using the CC mouse population. 29 Here, we present results showing that the NDM effect in lowering fasting blood glucose (FBG) levels in CC mice is significantly dependent on the genetic background of the mice (pharmacogenetics), the sex of the mice (pharmacosex) and diet (pharmacodiet). The results presented in this study support the need for follow-up research to better understand and implement a personalized medicine approach/utilization of NDM for reducing FBG levels.

| Ethical statement
All experimental mice and protocols were reviewed and approved by the Institutional Animal Care and Use Committee of Tel-Aviv University (TAU) (IACUC Approval number-01-17-056).

| Mouse experiments
Overall, 280 mice (137 males and 143 females) from 13 different CC lines were analyzed, as shown in Table 1. TAU and CC line designations, as well the number of contributors from the eight founders, are detailed in Table 1. We used a longitudinal experiment design, where we employed a large number of CC lines (13 lines) and, except for two female lines, at least 4 mice per sex per line, to provide sufficient statistical power for calculating the mean value of trait per sex and line. At the age of 3 weeks old, mice were weaned to separate cages based on sex and CC line, with free access to standard rodent chow diet (CHD) and water.
The experiment started when the mice were 8 weeks old (week zero (w0)). Throughout the following 18 weeks, the mice were maintained either on CHD or on a high-fat diet (HFD), considered to be a Western diet, or TD. 8137 (Teklad Global, Harlan Inc.), consisting of 42.0% (Kcal) fat, 15.3% protein, and 42.7% carbohydrates. At the end of the experiment, the mice were 26 weeks old.
At week 12, cranberry extract was administered intraperitoneally (see below) and measurements of FBG were taken at weeks 0, 12, and 18.

| Measuring fasting blood glucose levels (FBG)
The FBG test is used to detect disturbances in glucose metabolism that can be linked to diabetes or metabolic syndrome. The absolute FBG (mg/dl) was assessed at the start of the experiment (w0), 12 weeks into the experiment, and 6 weeks after NDM treatment (week 18 (w18)).
Mice were fasted for 6 h (6:00 AM to 12:00 AM) with free access to water. After 6 h fasting, blood glucose levels (mg/dl) were measured using a U-RIGHT glucometer TD-4267 (TaiDoc Technology Corporation 3 F, 5 F) as described elsewhere. 32 Subsequently, the percentage change in FBG (%FBG) was calculated as the difference (delta) between FBG values at w0 and w12 divided by FBG value at w0.

| Cranberry extract (CRE)
Non-dialyzable material (NDM) of cranberry extract was obtained from a cranberry juice concentrate (50 Brix, Ocean Spray) as described previously. 13 Briefly, the concentrated juice was dialyzed extensively in dialysis tubes of pore size 12-14 kDa (kiloDalton) against distilled water (>10× the volume of concentrated juice) with six changes over 6 days, and the non-dialyzable material was collected, lyophilized, and chilled until use (about 6 g NDM from one-liter concentrated juice). NDM was found to contain A-type proanthocyanidin oligomers (PACs) of 3-6 degrees of polymerization composed of catechin units, with some gallocatechin and anthocyanin units also present, as well as quercetin derivatives. 14 It also contains putative xyloglucans, which together with the mixed polyphenols exhibit anti-bacterial biofilm activity. NDM extract was dissolved in distilled water to provide a dose of 50 mg/kg mouse body weight, and 200 μl was injected intraperitoneally (IP) 3 times a week for 6 weeks (from w12 until w18).

| Statistical analysis
Statistical analysis was performed using the software package IBM SPSS Statistics version 24. One-way ANOVA was carried out for testing the significance level of trait variation among CC lines. A t test was also performed to assess the changes and significance before and after diets (between weeks 0 and 12), as well after NDM treatment between weeks 12 and 18.

| Broad sense heritability and genetic coefficient of variation
The broad-sense heritability and genetic coefficient of variation were estimated as described in detail elsewhere. 11 Briefly, heritability (H2) refers to the proportion of variation between individuals in a population that can be influenced by genetic factors. Here, we used the one-way ANOVA output of the phenotypic traits to calculate the heritability (including epistatic, but not dominance effects), using the following equation: where, H 2 = heritability, V g = genetic variance, and V e = environment variance. To evaluate the genetic dispersion of the monitored phenotypes, the genetic coefficient of variation (CV g ) was estimated using

| NDM affects male and female mice differently
The data were further scrutinized by individual lines of mice on HFD for both female and male mice ( Figure 2B,D, respectively). It is clear that the pattern of change in FBG values of the male lines is markedly different from the counterpart female lines, consistent with the differences in the average FBG values between the two sexes that were maintained on this diet.

| NDM effects vary on CC lines with different diets
In contrast to the CC male lines, their female counterparts responded differently to the two diets and to the NDM treatment, as shown in Tables 3 and 4 for CHD and HFD, respectively. The mean FBG levels of the HFD fed and NDM treated females showed a nonsignificant increase from a mean value of 130.29 ± 2.78 mg/dl at w12 to 133.58 mg/dl at w18, whereas during the same period of time in the CHD fed and NDM treated mice the mean value of the male FBG dropped significantly ( Table 3). These remarkable sex-associated differences between males and females of the 13 lines can be ap-  Statistical analysis of the overall %FBG -glycemic -levels of CHD-fed male mice at w12 showed a non-significant decrease compared to FBG levels at w0, and none of the mice exceeded 200 mg/ dl of blood glucose (Table 5). During the following 6 weeks on the same diet and NDM treatment, the overall male population showed a non-significant glycemic drop, while only three CC lines (IL5000, IL5003, IL5004) showed a significant drop in mean FBG levels from 134.31 ± 4.6 mg/dl at w12 to a mean FBG of 120.25 ± 3.64 mg/dl at w18 ( Table 5) .
The HFD in male mice caused a significant increase in FBG at w12 in three lines (IL5000 IL5023, and IL6009) (

| Broad sense heritability and genetic coefficient of variation
The H2 and CV g of the NDM effects on absolute and %FBG changes for females and males of the 13 mice lines maintained on CHD (control group) and HFD are shown in Table 7. H2 values were estimated for the FBG changes in experimental mice maintained on HFD before and after NDM treatment, and were found to be 0.60 and 0.50, respectively, for female mice and 0.64 and 0.73, respectively, for male mice. The estimated H2 for the FBG changes in control mice maintained on CHD before and after NDM treatment were found to be 0.39 and 0.80, respectively, for female mice and 0.63 and 0.43, respectively, for male mice (Table 7A) (Table 7A).

F I G U R E 2
Percentage changes (Delta) in fasting blood glucose (FBG) between weeks 12 and 0 of the female and male mice of 13 different collaborative cross (CC) lines maintained on either a standard chow diet (CHD) or a high-fat diet (HFD) for 12 weeks. Panels A and B show the profiles of the female mice of 13 different CC lines maintained on CHD and HFD, respectively, for 12 weeks. Panels C and D show the profiles of the male mice of 13 CC lines maintained on CHD and HFD, respectively, for 12 weeks. In all the panels A-D, the x-axis shows the thirteen CC lines, while the y-axis represents the percentage (%) FBG differences between weeks 12 and 0 during the experiment.

TA B L E 3
Effect of NDM on fasting blood glucose (FBG) levels in CC female mouse lines maintained on standard chow diet (CHD) for 12 weeks, and subsequently treated with NDM for six more weeks while maintained on the same diet. Summary of FBG (mg/dL) of female mice of 13 different of CC lines at age 8 weeks (w0 of the beginning of the experiment), at 12 weeks after maintaining on CHD (w12 + SE), changes in FBG (ΔFBG) between week 12 and week zero ((ΔFBG) (w0-12 + SE)) and %FBG changes between this period including standard error (% change ± SE) and indicating the significance status ( S ), FBG after 6 weeks treatment with NDM FBG (w18 ± SE), differences in %FBG changes (ΔFBG(w12-18 + SE)) between weeks 18 and 12, percentage of FBG changes between this period including standard error (% change ± SE) and indicating significant (S) or non-significant (NS) status

CHD group FBG ± SE
Week 0 to 12 In female mice, the CV g values of the %FBG 0-12 were −1. The calculated heritability for absolute FBG values was found to be high, which indicates and confirms that these assessed traits are strongly influenced by host genetic background. Furthermore, the calculated CV g shows that the genetic diversity is high in these tested 13 CC lines and is a promising resource for identifying genetic factors underlying these tested traits.

| DISCUSS ION
Previous studies employing HFD-fed C57BI/6J male mice showed that cranberry extract treatments caused a beneficial effect on several metabolic health markers such as reversal of insulin resistance and glucose tolerance. 7,35,36 Nevertheless, cranberry treatment of the C57BI/6J male mice did not lower fasting glycemia. 36 In contrast, in the present study cranberry significantly lowered

TA B L E 5
Effect of NDM on fasting blood glucose (FBG) levels in CC male mouse lines maintained on standard chow diet (CHD) for 12 weeks, and subsequently treated with NDM for six more weeks while maintained on the same diet. Summary of FBG (mg/dl) of male mice of 13 different of CC lines at age 8 weeks (w0 of the beginning of the experiment), at 12 weeks after maintaining on CHD (w12 + SE), changes in FBG (ΔFBG) between weeks 12 and week zero ((ΔFBG) (w0-12 + SE)) and %FBG changes between this period including standard error (% change ± SE) and indicating the significant status (S) , FBG after 6 weeks treatment with NDM FBG (w18 ± SE), differences in %FBG changes (ΔFBG (w12-18 + SE)) between weeks 18 and 12, percentage of FBG changes between this period including standard error (% change ± SE) and indicating the significant (S) or non-significant (NS) status

CHD group FBG ± SE
Week 0 to 12  The eight founders of the CC mouse population contain 42 million genetic variations (single nucleotide polymorphisms (SNPs)), which is higher than the human population (20 million SNPs). 16,37 Therefore, there is a strong case for the CC mouse population to be used for human studies to capture genes which may be translated to humans. The Collaborative Cross (CC) mouse model is a new and high genetically diverse reference population. 16,38 This unique model is a platform for research of complex traits that overcomes the limitation of existing animal models. 8,9,16 The CC population In the current study, we have TA B L E 7 Broad sense heritability (H2) and genetic coefficient of variance (CVg) at the initial and pretreatment as well post-treatment of NDM of the absolute (A) and percentage (B) FBG in female and male mice of 13 different CC lines at the initial time point of the experiment (FBG0), week 12 (pretreatment (FBG12)) and at week 18 weeks (post-treatment (FBG18)) maintained on CHD or HFD and after treatment with NDM between weeks 12 and 18 In our study, we injected the NDM into the peritoneal cavity of the animals. This is a common technique in laboratory rodents but is rarely used in larger mammals and humans. Intraperitoneal (IP)

CRE FBG wk18 ± SE
injection is used for small species for which intravenous access is challenging and it can be used to administer large volumes of fluid safely or as a repository site for surgical implantation of a preloaded osmotic minipump. Although IP delivery is considered a parenteral route of administration, the pharmacokinetics of substances administered intraperitoneally are more similar to those seen after oral administration, because the primary route of absorption is into the mesenteric vessels, which drain into the portal vein and pass through the liver. 40 Therefore, substances administered intraperitoneally may undergo hepatic metabolism before reaching the systemic circulation.
In addition, a small amount of an intraperitoneally injected substance may pass directly across the diaphragm through small lacunae and into the thoracic lymph. 41,42 The route of administration is probably important, especially the oral route employed by Anhe et al. 7 is likely that the oligomeric cathechin is absorbed by the epithelial barrier and degraded to biologically active phenolic compounds by internal organs such as the liver. In fact, it has been argued that although the metabolism of phenolic compounds in vivo is complex and poorly understood, 51 they are degraded by the liver and other organs. We hypothesize that a similar fate of adsorption and degradation of phenolic oligomers in the intraperitoneal cavity may occur.
Certainly these molecules can be adsorbed by the epithelium barrier to enter the gut and then reabsorbed to reach the internal organs, as occurs with the ingestion of such compounds. It follows that the extent of adoption and degradation of the polyphenolic molecules is dependent on a particular genetic background, which explains the variability of the effect of NDM in the CC lines.
In conclusion, our study shows an effect of cranberry extract in lowering fasting blood glucose levels and that this effect is dependent on the host's sex (pharmacosex), genetic background (pharmacogenetics), and diet type, i.e. CHD or HFD (pharmacodiet). Because the DNA sequence of both the mouse founders and the CC lines are known, there is no doubt that the genetic loci or genes involved in such an effect may be determined once more CC lines of mice are employed. This will allow identification and use of mouse lines showing the maximum response to NDM to eventually determining the appropriate dosage and route of administration of the extract in humans. All these results support the need for further research to better understand and implement a personalized medicine approach/utilization of NDM for reducing FBG levels.
Finally, the results of our study have shown that amount of NDM we used induces a significant effect, and this is believed to be due to the high purity and concentration of the active component in our NDM extract.

AUTH O R CO NTR I B UTI O N S
Fatima Amer was involved in project design, execution, data analysis, and MS preparation, Rana Tarabeih was involved in project execution, and Itzhak Ofek was involved in project design, data analysis, and MS preparation. Fuad A. Iraqi was involved in the project design, data analysis, and MS preparation and approving its final version.

ACK N OWLED G M ENTS
The authors declare no competing financial interests or other associations that may pose a conflict of interest (e.g. pharmaceutical stock ownership, consultancy). This report was supported by a core fund from Tel-Aviv University. The authors thank Iqbal M. Lone for his comments on the manuscript.

CO N FLI C T O F I NTE R E S T
None.