Oral GLP-1 Analogue Ameliorates Obesity-Induced Diabetes In Db/Db Mouse

Type 2 diabetes is currently experiencing an outbreak worldwide. GLP-1 effectively lowers blood glucose level as an emerging target for the treatment of type 2 diabetes. However, the application of GLP-1 is limited by short half-life and too expensive cost in clinic. In this study we employed the food-grade probiotics as delivery system to express human GLP-1 and its analogue. Recombinant lactococcus lactis could express GLP-1 and analogue in vitro and modied GLP-1 analogue was more resistant to DPP-4 degradation. Oral administration of GLP-1 analogue could reduce the fat mass. More importantly, GLP-1 analogue improved hyperglycemia and insulin resistance in Db/Db mouse although the insulin secretion is not observed in vitro. Our study demonstrates that lactococcus lactis genetically modied with single amino acid mutation could prolong half-life of GLP-1 and increase insulin sensitivity in Db/Db mouse model as an oral drug delivery system driving the development and innovation of drug therapy for type 2 diabetes. (Sigma) was oral administered and blood glucose level was measured with Accu-Chek Performa test strips (Roche Diagnostics Corp) at each indicated time point. If the blood glucose value is over the limit of detection, blood would be diluted immediately into saline and repeated this test. The individual glucose concentration would be calculated and recorded by dilution proportion itself. The situation of insulin resistance would be reected by HOMA-IR (Formula by Fasting glucose (mM) × Fasting insulin (µU/mL)/22.5. The serum insulin was measured by Rat/Mouse Insulin ELISA (Merck Millipore, EZRMI-13K) as delivery system to express human GLP-1 and its analogue. Recombinant lactococcus lactis could express GLP-1 and analogue in vitro and GLP-1 Mutant was more resistant to DPP-4 degradation. More importantly, GLP-1 analogue reduced the fat mass, improved hyperglycemia and insulin resistance in Db/Db mouse. Our study demonstrates that lactococcus lactis genetically modied with single amino acid mutation could prolong half-life of GLP-1 and increase insulin sensitivity in Db/Db mouse model as an oral drug delivery system driving the development and innovation of drug therapy for type 2 diabetes. In summary, feasibility design of GLP-1 mutation has a potential prospective in clinical application furtherly more improvement needs to be performed in the future.


Introduction
Diabetes is a metabolic disease characterized by chronic hyperglycemia caused by multiple causes. According to the global report of World Health Organization (2016), more than 420 million adults were living with diabetes in 2014. Not only in high-income countries, but diabetes spread very quickly in lowand middle-income countries. Due to the high calorie diets and increasingly sedentary lifestyles prevalent in modern society, fat accumulation has become a major risk factor for type 2 diabetes [1]. Unlike the type 1, type 2 diabetes has been certain that the disease process could be halted with balanced diet to correct the poor fat metabolism [2]. Insulin resistance and islet dysfunction are the major pathophysiologic features of type 2 diabetes. The necessary therapeutic interventions are also helpful in accelerating the course of the diabetes, for example, bariatric surgery [3,4], insulin administration [5,6], sulfonamides [7,8] or metformin treatment [8]. However, there are still problems with the aforementioned treatments. For those patients whose condition is not improved by oral medication, bariatric surgery can only be used in a small number of people [9][10][11], and interventional administration of insulin injection also brings endless torment. Reverse-engineer would replace the surgery to nd less invasive and more equivalent therapeutics. Moreover, some Prudent advice have been offered that combining either pre-or probiotic approaches with gut-peptide-based therapies may ultimately achieve e cacy similar to bariatric surgery [12].
GLP-1 is a very short gut-peptide from L cells which could induce glucose-dependent insulin secretion, reduce glucagon secretion, delay gastric emptying, increase β-cell survival [13][14][15][16] and decrease blood glucose and food intake [17]. Thus as a potential drug therapy, the development of GLP-1 preparations has been highly sought after. Considering of lower half-life of wildtype GLP-1 for less than 2 min [17], GLP-1 analogues is been investigated to remedy their de ciency, such as restriction enzyme mutation, unnatural amino acid substitution, covalent complex and Exendin family from Glia monster [18]. Another problem, however, is that GLP-1 analogues are too expensive to achieve widespread use as rst-line therapy compare with oral anti-diabetes drugs.
As a food grade engineering bacteria, Lactococcus lactis is emerging as a potential candidate for foreign protein delivery [19][20][21]. Genetically modi ed lactococcus lactis has been utilized in many disease model including in ammatory bowel disease [22,23], anti-tumour [24], anti-obesity [25], anti-oxidation [26], indicating that industrial bacteria could offer a prospective living biotherapeutics for the treatment of human disease.
Nevertheless, a new amino acid mutation of GLP-1 expressed in lactococcus lactis hasn't been investigated to con rm the anti-diabetes effect in Db/Db mouse. In this study, we combined lactococcus lactis and functional GLP-1 analogue to construct an anti-hyperglycemia probiotic expression system and the process of protein production is transferred to gut. GLP-1 Mutant could prolong half-life in vitro and have been veri ed the hypoglycemic effect in Db/Db mouse. This design solves the problems of short half-life and high cost of injectable GLP-1. In summary, feasibility design of GLP-1 mutation has a potential prospective in clinical application furtherly more improvement needs to be performed in the future.

Materials and animals
Lactococcus lactis cremoris NZ3900 (Cat# VS-ELS03900-01) and plasmids pNZ8149 (Cat# VS-ELV00300-01) were purchased from MoBiTec GmbH, Germany. The standard substance of GLP-1 and GLP-1 Mutant used in this study was arti cial synthesis by Genscript Biotech Corporation. All reagents are from Sigma unless otherwise speci ed.
Genetic obesity model C57BLKS/J-Lepr-/Lepr-(Db/Db) male (7-week-old) was purchased from Model Animal Research Center of Nanjing University. After a week of adaptive breeding, mice were divided into three groups according to body weight and oral administration daily for 4 weeks with 1.0×10 8 CFU frozendried powder of vehicle, GLP-1 and GLP-1 Mutant probiotics, respectively. During this phase, body weight and blood glucose would be recorded to assess the function of GLP-1 analogue.
Animals were housed under a 12/12 hr light/dark cycle with free access to food and water. Before sacri ce, animals were given an anesthetic and cervical dislocation rapidly to relieve pain. All animal studies followed the National Institutes of Health guide for the care and use of Laboratory animals (NIH Publications No. 8023, revised 1978) and were approved by the Institutional Animal Care and Use Committee of Institute of Zoology, Chinese Academy of Sciences.

Plasmid cloning and transformation
The human GLP-1 (7-37aa) followed by USP45-LEISS, 6×his-tag and enterokinase restriction site (EK) was sub-cloned into pNZ8149. Recombinant vector above was transferred into home-made NZ3900 competent cell through Gene PulserXcell™ electroporation system (BioRad, California, USA). That goes for preparation of GLP-1 analogue and these two freeze-dried powders were registered under GLP-1 and GLP-1 Mutant for short, respectively. The pure culture of GLP-1 and GLP-1 Mutant was transferred into 10 ml fresh M17 broth for stationary culture until OD600 around 0.4-0.6, 25ng/ml nisin was added to induce the target protein expression. Given the very small peptide chain of GLP-1 and its mutant, the supernatant of the pure culture was enriched by his-tag a nity chromatography and detected by tandem high performance liquid chromatography and mass spectrometry (HPLC-MS, Orbitrap Elite, Thermo sher) after enrichment.

Stability detection of analogues
In the 96 well plate, 1.5ng DPP-4 was mixed with 10 ug GLP-1 or GLP-1 Mutant under sodium phosphate buffer (0.1 M, pH 7.4), respectively. After incubation for 60 min in 37℃ incubator, tri uoroacetic acid (TFA, 0.1%) was added in well to terminate reaction. Then sample mixture was puri ed by Oasis HLB extraction cartridges (Waters, WAT094225). Following the elution of methanol and vacuum drying, puri ed protein was re-dissolve with 0.1% methanoic acid. At the end, samples through 0.45um centrifugal lters (Millipore, UFC30HV00) could be tested by mass spectrometry.

Body composition analysis
The total fat and lean masses of mice after a 4-week-treatment with either vehicle or GLP-1 analogue were assessed with the Small Animal Body Composition Analysis and Imaging System (MesoQMR23060H-I, Nuimag Corp., Suzhou, China).

Glucose homeostasis analysis
Mice would be fasted for 12 hours (21:00-9:00) with free access to water. Then refeed glucose was measured after cafeteria feeding for 2 hr and random glucose could be detected at any time of maintenance feeding.
For oral glucose tolerance test, 0.75 g/kg D-glucose (Sigma) was oral administered and blood glucose level was measured with Accu-Chek Performa test strips (Roche Diagnostics Corp) at each indicated time point. If the blood glucose value is over the limit of detection, then blood would be diluted immediately

Isolation of islets
Common bile duct perfusion method has been referenced for islet separation. Detailed, the liver-pancreatic junction was found to reveal the position of common bile duct after cervical vertebrae dislocation. Clamping the junction of common bile duct and duodenum with tweezers, digestive solution with 1 mg/ml collagenase P was injected into common bile duct with 31G needle and complete in ltration of the pancreas. Then the whole pancreas was rapidly stripped and digested in vitro at 37℃ for 15 min. At the end, the islets were washed twice and manual picked under stereoscope for quantitative statistics.

Statistical Analysis
All data are presented as means ± SEM. Data from multiple groups were analyzed by one-way ANOVA. Data from two groups were analyzed by unpaired Student's t-test. P < 0.05 was considered to indicate a statistically signi cant difference.

Results
The design and construction of GLP-1 analogue GLP-l analogues, such as exendin-3, have almost exactly the same biological function and resistant to DPP-IV degradation with a longer half-life. Thus we referred to the structure of exendin-3 and eliminated the DPP-4 recognition site by transferring alanine to serine (Fig. 1A). According to the construction of oral protein expression system, GLP-1 and GLP-1 Mutant were sub-cloned into pNZ8149 following the signal peptide Usp45, enhancer sequence LEISS, his tag and enterokinase recognition site under nisin induced promoter (Fig. 1B). We successfully harvested the recombinant lactococcus lactis with GLP-1 or GLP-1 Mutant (Fig. 1C).
Due to very short peptide chain of GLP-1 and its analogue and the difference of individual amino acid, the western blotting is not adoptable for GLP-1 analogue. Here we made a synthesis of GLP-1 and analogue in vitro as standard substance. After comparison, GLP-1 and analogue could be detected in culture supernatant consistent with ngerprint of standards ( Fig. 2A). In addition, the sequence of peptide matched design that recombinant lactococcus lactis could express the GLP-1 and analogue in vitro (Fig. 2B).

Stability study of GLP-1 analogue and insulin secretion in vitro
In order to assess the stability of GLP-1 analogue, we employed HPLC-MS to detect the degradation of GLP-1 and its analogue in vitro with adding of DPP-4. After 1 hr treatment, GLP-1 content quickly dropped to 60% due to carrying enzyme cleavage sites of DPP-4. On the other hand, GLP-1 analogue only decreased by 15% (Fig. 3A). Therefore, this mutant of GLP-1 extended half-life of wildtype GLP-1 in some degree.
On the functional veri cation, we used INS-1 cell to check the insulin secretion under low and high glucose stimulation of physiological conditions. Consistently with GLP-1, GLP-1 Mutant also induced insulin secretion whatever 2.8 mM or 16.7 mM of glucose (Fig. 3B).
GLP-1 analogue reduce the fat mass without changing body weight Db/Db mouse is a very good model of obesity and diabetes which is suitable for verifying the function of GLP-1 analogues. Here we employed Db/Db mouse to verify the function of GLP-1 analogue (Fig. 4A). During four weeks of continue oral administration, there was no signi cant change in body weight between the treatment group and the control group (Fig. 4B). However, it was found that only fat mass of GLP-1 Mutant was decreased by 6.6% signi cantly, but the lean mass remained unchanged (Fig. 4C, D). Tissue weight is also shown that GLP-1 Mutant signi cantly reduced the weight of adipose tissue (Fig. 4E).

GLP-1 analogue improved the glucose homeostasis in Db/Db mice
The function of GLP-1 is to promote glucose dependent insulin secretion and reduce blood glucose. Firstly we tested random blood glucose and found that the blood glucose of GLP-1 Mutant gradually down-regulated after 4 weeks treatment, and GLP-1 group has a signi cant hypoglycemic effect in the early stage of treatment (Fig. 5A). Then we investigated the change of blood glucose before and after cafeteria feeding. Whatever GLP-1 or GLP-1 Mutant didn't in uent the 12 hr-fasting glucose, but GLP-1and GLP-1 Mutant indeed mitigated the surge in postprandial blood glucose after 2 hour-cafeteria feeding (Fig. 5B). At posttreatment of 2 week, glucose tolerance test was also con rmed that GLP-1 and GLP-1 Mutant could signi cantly reduce the fasting glucose and enhance the capability of tolerance for oral glucose comparing with control group (Fig. 5C).

GLP-1 analogue improved the insulin resistance without changing insulin secretion in Db/Db mice
Given the GSIS function in INS-1 cell in vitro, we tried to con rm that insulin secretion in Db/Db mouse. The recombinant lactococcus lactis didn't change the intestinal length in three groups (Fig. 6A). The GLP-1 and GLP-1 analogue could up-regulate the islet number which means insulin secretion might be enhanced (Fig. 6B). Combined with the results of insulin secretion experiment in vitro, we hypothesized that GLP-1 also increases insulin production in vivo and detected the insulin level in tail tip blood glucose of Db/Db mouse. Unexpectedly, we didn't observed insulin concentration increasing in peripheral blood due to abnormal uctuations (Fig. 6C). But more powerful than GLP-1, GLP-1 Mutant indeed improved the insulin resistance in Db/Db mouse depending on the careful calculation of glucose and insulin concentration (Fig. 6D).

Discussion
The number of patients with type 2 diabetes increases year by year worldwide. Although diabetes does not directly cause death, the pain brought by long-term medication seriously reduces the quality of life of patients because of inconvenient operation and poor patient compliance, especially injection of insulin and other protein drug [27]. GLP-1 and its analogues are one of the most popular drugs for type 2 diabetes. Long acting GLP-1 analogues have been developed to reduce the number of injections needed recently like amino acid sequence modi cation, chemical modi cation, fusion protein polymers and special material package, and mode of administration for peptide drugs and higher cost from complicated preparation process still not be changed. Therefore an appropriate oral drug delivery system has long been the ultimate goal of scientists. Several studies have attempted to use lactic acid bacteria for mucosal drug delivery [28]. It has been widely approval that food-grade lactococcus lactis is safe enough to be mucosal delivery therapeutic drugs [29][30][31]. Herein we combined lactococcus lactis with functional GLP-1 analogue to construct an anti-hyperglycemia probiotic. For the sake of prolongation of GLP-1 half-life, we referred to the structure of exendin-3 and mutated the DPP-4 enzyme restriction site to serine. By mass spectrometry, it was found that GLP-1 Mutant was indeed resistant to enzymatic hydrolysis and signi cantly prolonged the action time of GLP-1.
Because of the anti-obesity and anti-diabetes effect of GLP-1, Db/Db mouse are employed to verify the function of GLP-1 analogues. Db/Db mouse is a common obesity mouse model which exhibit type 2 diabetes, dyslipidemia and liver damage during the course of growth and development [32][33][34]. We found that the GLP-1 Mutant signi cantly remitted adiposity of Db/Db mouse after 4 week treatment without changing body weight. Unlike previous understandings of GLP-1 classical function, GLP-1 analogues could induce the body weight loss obviously [35]. But oral semaglutide is mostly attributable to body fat mass loss while causing weight loss [36,37]. Consistently, recombinant lactococcus lactis with natural GLP-1 still didn't affect the body weight and fat mass of chow-diet and high fat diet feed mouse [38]. Taken together, fat mass loss of GLP-1 Mutant is same as commercial GLP-1 analogues and prolonged, sustained and stable GLP-1 stimulation might lead to obvious changes in body weight.
The insulin secretion is another identify object for GLP-1 function. In our study, insulin secretion is modestly up-regulated by 11% under high glucose stimulation in vitro. This up-regulation is not enough to elevate the serum insulin production in Db/Db mouse. Thus serum insulin was not detection by any difference between three groups as expected. Interestingly, Arora and his colleagues didn't nd changes in insulin secretion in chow-diet mouse [38], and the uctuation of insulin levels was also violently in ZDF rats totally different from the cell experiments [39,40]. On the other hand, the dosage and method of oral delivery is different from previous studies (1.0×10 8 CFU frozen-dried powder in our method v.s. ≥1.0×10 9 fresh culture suspension ), and this might be another factor contributing to insigni cant increasing of insulin release. Next we will attempt to use higher dosage of GLP-1 Mutant to check the function of glucosedependent insulin secretion.
For safety reasons, antibiotic resistance and labeled proteins are not suitable for clinical research. Recombinant lactococcus lactis doesn't carry any antibiotics resistance gene and doesn't seriously affect ora balance in theory. His tag was linked before GLP-1 fusion protein for convenient puri cation and detection in this study. There is a possibility that the tagged proteins could induce antibodies production in animals and trigger an in ammatory response.
Alternatively, the enterokinase recognition site was loaded between his tag and GLP-1 sequence to crack the mature GLP-1 analogue in the environment of digestive juices. Thus unlabeled fusion protein would be considered to facilitate clinical transformation. Safety evaluation would be executed to verify the possible side effects of oral GLP-1 analogue. In addition, pharmacodynamic studies are not su cient to expand GLP-1 Mutant analogue. The in uence of gut microbiota on type 2 diabetes should not be ignored completely. Abnormalities in the composition of the gut microbiota might contribute to the development of type 2 diabetes [41]. Chinese herbal formula may ameliorate type 2 diabetes via enriching bene cial bacteria with metformin [42]. Thus potential therapeutic strategies might relief hyperglycemia from both molecular and microbial dimensions. Faece metagenomic analysis should be carried out to reveal the exact molecular mechanism.
Unlike expensive eukaryotic cell culture situation and puri cation process, lactic acid bacteria production is relatively affordable. The feature of living biotherapeutics lies on the production process postposition that designed process ow are performed in gut. The advantage of this process is that it reduces production costs and side effect [43]. In partially, the problem of excessive economic burden of biological medicine has been solved and living biotherapeutics with versatility would provide us with more and more options [44].
In our study, we employed the food-grade probiotics as delivery system to express human GLP-1 and its analogue. Recombinant lactococcus lactis could express GLP-1 and analogue in vitro and GLP-1 Mutant was more resistant to DPP-4 degradation. More importantly, GLP-1 analogue reduced the fat mass, improved hyperglycemia and insulin resistance in Db/Db mouse. Our study demonstrates that lactococcus lactis genetically modi ed with single amino acid mutation could prolong half-life of GLP-1 and increase insulin sensitivity in Db/Db mouse model as an oral drug delivery system driving the development and innovation of drug therapy for type 2 diabetes. In summary, feasibility design of GLP-1 mutation has a potential prospective in clinical application furtherly more improvement needs to be performed in the future.

Declarations Ethical Statement
All applicable international, national, and/or institutional guidelines for the care and use of animals were followed.

Availability of data and materials
The data used to support the ndings of this study are available from the corresponding author upon request.

Con icts of Interest
All authors declare that they have no con ict of interest.   GLP-1, n=7 and GLP-1 Mutant, n=8). Multiple comparisons which were signi cant were represented with alphabetic notation, abc recorded as p value < 0.05. were signi cant were represented with alphabetic notation, abc recorded as p value < 0.05.