Simple Continuous Glucose Monitoring in Freely Moving Mice

Diabetes mellitus (DM) is a devastating disease characterized by high blood glucose levels. Type 1 DM can be a result of an autoimmune attack on the insulin-producing beta cells in the pancreas. Type 2 DM and gestational DM, on the other hand, are characterized by a failure of the beta cells to secrete sufficient insulin in response to a rise in glucose levels¹. The mouse is a common model organism used to study DM since it has similar physiology, and its normal glucose levels are close to those of humans. Furthermore, specific mouse strains may develop DM due to mutations in key signaling pathways or following exposure to specific diets, which enables disease modeling^2,3,4.

Blood glucose is commonly measured in mice using glucometers designed for patients by extracting a small drop of blood (1-2 µL) from the tip of the tail of the mouse. This method causes stress and requires handling of the mouse, which affects the glucose levels and prohibits the measurement of blood glucose levels in freely behaving mice or when the researcher is not close by⁵. Bleeding the mice may cause stress to nearby mice, particularly to mice of the same cage whose glycemia has not been measured yet, thus affecting the results. Mice respond differently depending on the handler, and the person measuring glucose may affect the glucose levels of the mice. These pitfalls call for careful experimental design and underlie some inconsistencies between experiments.

It is possible to measure glucose in freely moving mice with no bleeding by implanting glucose sensors into the aortic arch of the mice using state-of-the-art telemetry⁶. The resulting measurements are very good and can be sustained over a long period, but it is challenging to implant these sensors, and the telemetry system is expensive, leading to a moderate adoption of this methodology and no adoption in non-specialized labs. Subcutaneous or other glucose sensors that are tailored to the dimensions of the mice and their physiology have been developed in recent years, but these again require highly skilled experts and are in some cases costly^6,7,8,9,10.

Commercial continuous glucose monitors (CGMs) that were originally developed to monitor the glucose levels of DM patients offer another option to measure glucose in freely moving mice, with lower cost and technical expertise requirements than implanted probes. Such probes have been used in basic research by a few labs^{5,11,12,13,14,15} including our colleagues who used this protocol¹⁶. These devices typically include a sensor, a mounting device, a receiver, and a software application. The sensor has a cannula guiding the enzymatic glucosensor, adhesive tape, an energy source, short-term memory, and a wireless communication module that stores and sends the data to the receiver. The receiver can show the current glucose levels and sends the data to a server; this receiver can be a cellphone. The software application provides data for the patient and the medical care team on the glycaemia of the patient. In patients, the sensor is attached easily using the mounting device. The cannula is inserted subcutaneously by pressing the mounting device against the skin, and the sensor stays in place with the help of adhesive tape.

This is a detailed protocol for adapting a commercial CGM device to measure glucose levels in mice. This protocol describes how to surgically insert the glucose sensor and attach it to the mouse. Scripts for basic data analysis and data visualization are provided. The potential pitfalls, troubleshooting, and examples of standard results are provided. The protocol below is specific for a certain CGM but can be easily adapted to other types of commercial CGMs as they become available.

The experiments were approved by the Hebrew University's Institutional Animal Care and Use Committee (IACUC).

NOTE: All tools must be sterilized, and handling of the cannula must be performed using a sterile technique. The protocol below is fine-tuned to a specific CGM. The protocol can be adapted to other CGMs.

1. Analgesic administration before the procedure

1. Administer 5% dextrose and 0.45% saline with meloxicam at 5 mg/kg body weight subcutaneously.

2. Anesthesia administration

1. Place the mouse in the induction chamber, closing the lid tightly. Set the induction of anesthesia in the induction chamber at 3% isoflurane at a flow rate of 500 mL/min.
2. Once the mouse is unresponsive, remove the mouse from the chamber, and fit the nose cone to the mouse. Confirm the anesthesia level with an interdigital pinch. Set the concentration to 1%-1.5% isoflurane and the flow rate to 100 mL/min in a mouse weighing 30 g.
3. Apply ophthalmic ointment to the eyes to prevent dryness during anesthesia.

3. Sensor preparation

1. Mount the sensor onto the sensor mounting device to expose the sensor's tape and cannula side (Figure 1A). Be cautious since the needle is inserted into the cannula and exposed.
2. Suture two 5-0 taper point sutures to the tape on both sides of the cannula (Figure 1A).

4. Hair removal and disinfection

1. Shave an area of approximately 4 cm x 4 cm on the midline of the back of the mouse.
2. Administer a depilatory cream to the shaved area to ensure complete hair removal.
3. Wipe the skin, and disinfect it using an antiseptic solution containing 2% chlorhexidine gluconate and 70% isopropyl alcohol.

5. Dorsal skin preparation

1. Make a 2 mm incision in the center of the shaved area above the spine using sharp scissors (Figure 1B).
2. Briefly, insert small forceps with a blunt edge under the skin to form a small subcutaneous pocket so that the cannula can be inserted easily into the subcutaneous pocket (Figure 1B).
3. Pass a suture from step 3.2 through the skin on each side of the incision (Figure 1C).

6. Sensor insertion

1. Remove the sensor completely from the sensor mounting device (the cannula is void from the needle), and hold the sensor with forceps to keep the surrounding tape from sticking to itself.
2. Carefully insert the cannula into the subcutaneous pocket.
3. Pull the sutures on each side, and tighten and tie them to attach the sensor firmly in place, thus preventing the cannula from slipping out from the subcutaneous pocket once the adhesive tape loosens over time.

7. Sensor attachment and suturing

1. Attach the sensor to the back firmly by tying the inner sutures and using the adhesive tape surrounding the sensor.
2. Make eight discontinuous sutures around the sensor, attaching the border of the sensor's tape to the skin (Figure 1D).

8. Activation of the reader

1. Once the sensor has been inserted, activate the reader by turning on the reader, pressing Start New Sensor, and swiping the sensor according to the manufacturer's instructions.
2. The first reading can only be taken a few minutes after installing the CGM. In the case of this CGM, the first reading can be taken after 60 min.

9. Reading results

1. Place the reader close to the mouse (there is no need to touch it). All data stored in the sensor are transmitted to the reader.
   ​NOTE: Different CGM devices may differ in the period of historic data saving capacity. In the case of this CGM, a maximum of 8 h can be stored between two readings.

10. Removing the sensor

1. Anesthetize the mouse (see section 2).
2. Cut the sutures connecting the sensor to the back of the mouse using sharp scissors.
3. Remove and cut the sutures at the incision by gently removing the sensor.
4. If necessary, use a single suture to close the incision in the back of the mouse.

11. Data analysis

1. Data download: Download the data as per the instructions provided by the CGM manufacturer.
   ​NOTE: Each CGM has a different format, which may or may not be easily accessible to the user. This is an important consideration in choosing the CGM.
2. For analysis with the software provided, format the data according to the instructions in the readme file on Github (https://github.com/mika-littor/CGM-in-Mice-Analysis.git).

Figure 1: Attachment of the sensor to the mouse. (A) Two sutures marked by red arrows are passed through the sensor tape on both sides of the cannula on the bottom side of the CGM sensor, marked by a white arrow. (B) A small 2 mm incision is made in the center of the shaved area along the spine using sharp scissors. Small forceps with a blunt edge are briefly inserted under the skin to form a small subcutaneous pocket so that the cannula can be inserted subcutaneously. (C) The same sutures from A are passed subcutaneously on each side of the incision. The red arrows mark the sutures attached to the sensor as in A, the blue arrows mark the location the sutures passed through to the skin in the back of the mouse, and the black arrow shows the incision. (D) After the cannula is inserted, the inner sutures are tightened and tied close to the incision to secure the CGM. The sensor is then sutured to the skin. Please click here to view a larger version of this figure.

Surgical outcome
Results from eight HSD:ICR mice (aged 8 weeks) fed a high-fat high-sucrose diet (HFHS) for 18 weeks and five lean HSD:ICR mice (aged 12 weeks) are shown. The device we used stores data for up to 8 h. Access to the local animal facility was restricted to 07:00-19:00, thus prohibiting data collection during the late P.M. hours, when the mice are active. The mice were, therefore, placed in a room with reverse lighting for 7 days before the surgical procedure, with dark hours between 8:30 and 20:30. This is not necessary for all devices or animal facilities, and we recommend using devices that can store information for over 12 h.

There was no mortality following the surgery. The surgery led to an approximately 10% weight loss during the time of the experiment (Figure 2A). Therefore, the measurements taken in the first few days after surgery, during weight loss, should be interpreted with care. Weight loss was not due to the inability of the mice to reach for food and water. The comparison of CGM measurements and tail-tip blood glucose measurements showed good agreement in fasting and non-fasting states (Figure 2B). The CGM was active for 11 days on average (Figure 2C). The maximum number of days for this type of device is 14 days. When the device became inactive earlier, it was not due to the CGM falling off.

Figure 2: General outcomes of applying the CGM. (A) Mean weight reduction during the time the CGM was active. n = 8 mice. (B) The mean difference between the glucose readings by the hand-held glucometer and the CGM device. The difference was not significantly different from 0 mg/dL. n = 10 readings in six mice. (C) The mean time the CGM was active for in n = 8 mice. Error bars represent the standard error of the mean. Please click here to view a larger version of this figure.

Raw output
The glucose levels of a single day are shown on a graph produced by the CGM software (Figure 3A). The data from a few days can be viewed using the code provided (Figure 3B). We show data from 3 days for clarity.

Figure 3: Data analysis. (A) Commercial output. Data were not collected between 18:00 and midnight. The shaded area shows normoglycaemia values in patients, which are between 70-100 mg/dL. (B) Raw output data of 3 days from a single mouse using the code provided. Note the difference in scale in the y-axis between A and B. The axis parameters, and all the other parameters, can be modulated in the code. Data for 3 days are shown for clarity. Please click here to view a larger version of this figure.

Analysis
Once the data are extracted, analysis can be done using the code provided or any other custom-built software. Below is the mean (Figure 4A,C) and median (Figure 4B,D) glucose levels at every time point for a single mouse. A sliding window can be used to smoothen the plot. Only two mice are shown for clarity.

Figure 4: Data analysis output using the provided scripts. (A) Mean and (B) median glucose levels at each time point in a specific mouse. The shaded area denotes the standard deviation in glucose levels. (C) Mean and (D) median glucose levels from two mice. Only two mice are shown for clarity. A dashed line denotes the transition from light (20:30-08:30) to dark (8:30-20:30). A sliding window of 20 min was used to smooth the curve. The size of the window, and all the parameters, can be modified in the code. Please click here to view a larger version of this figure.

This protocol offers a simple, inexpensive method to monitor glucose levels in mice that does not require challenging microsurgery and does not involve bleeding or handling the mice. The method is easy to implement in every facility and does not cause mortality, pain, or excessive discomfort to the mice. The most critical step in the protocol is inserting the cannula of the glucose sensor under the skin of the mouse. The addition of a few sutures allows the cannula to stay in place for a longer time. The sensors are small and can become blocked or misplaced as the mouse moves. The sensor must be secured by the inner sutures and attached to the back of the mouse by a few sutures. The protocol can be adapted to many commercial CGM systems.

The mice lost weight after the surgery, which is expected following surgery and can be associated with the stress of surgery and anesthesia, the burden of carrying the device, and the single-housing imposed by the specific protocol used in this study. As CGM devices become progressively lighter and more accurate, these effects are expected to become smaller. We observed that mice carrying the CGM moved freely in the cage and reached food pellets that required climbing. The method is limited by the ability of each CGM to function over time. Researchers should monitor for weight loss and consider the advantages and disadvantages of this method compared to standard alternatives that require frequent handling and bleeding of the mice, which assume that bleeding one mouse has little effect on the glycemia of its neighbors and involve measuring glucose levels in a stressful condition during daytime.

The protocol described above is fairly simple, fast, and scalable. It does not require a special setting in the animal facility or expensive equipment, can be used in parallel with other procedures that are part of the experiment, and can be used on any genetic background or with any nutrition. Data analysis scripts are provided to facilitate analysis by research groups that are less experienced in data analysis. Hopefully, this detailed protocol and data analysis scripts will enable other labs to measure glucose in their experiments in metabolic and other research fields.