The hyperinsulinemic-euglycemic clamp, or insulin clamp, is the gold standard for assessing insulin action in vivo. A method for performing insulin clamps in mice is described. This includes a method for arterial catheterization that permits experiments to be performed in conscious, unrestrained mice with minimal stress.
Type 2 diabetes is characterized by a defect in insulin action. The hyperinsulinemic-euglycemic clamp, or insulin clamp, is widely considered the “gold standard” method for assessing insulin action in vivo. During an insulin clamp, hyperinsulinemia is achieved by a constant insulin infusion. Euglycemia is maintained via a concomitant glucose infusion at a variable rate. This variable glucose infusion rate (GIR) is determined by measuring blood glucose at brief intervals throughout the experiment and adjusting the GIR accordingly. The GIR is indicative of whole-body insulin action, as mice with enhanced insulin action require a greater GIR. The insulin clamp can incorporate administration of isotopic 2[14C]deoxyglucose to assess tissue-specific glucose uptake and [3-3H]glucose to assess the ability of insulin to suppress the rate of endogenous glucose appearance (endoRa), a marker of hepatic glucose production, and to stimulate the rate of whole-body glucose disappearance (Rd).
The miniaturization of the insulin clamp for use in genetic mouse models of metabolic disease has led to significant advances in diabetes research. Methods for performing insulin clamps vary between laboratories. It is important to note that the manner in which an insulin clamp is performed can significantly affect the results obtained. We have published a comprehensive assessment of different approaches to performing insulin clamps in conscious mice1 as well as an evaluation of the metabolic response of four commonly used inbred mouse strains using various clamp techniques2. Here we present a protocol for performing insulin clamps on conscious, unrestrained mice developed by the Vanderbilt Mouse Metabolic Phenotyping Center (MMPC; URL: www.mc.vanderbilt.edu/mmpc). This includes a description of the method for implanting catheters used during the insulin clamp. The protocol employed by the Vanderbilt MMPC utilizes a unique two-catheter system3. One catheter is inserted into the jugular vein for infusions. A second catheter is inserted into the carotid artery, which allows for blood sampling without the need to restrain or handle the mouse. This technique provides a significant advantage to the most common method for obtaining blood samples during insulin clamps which is to sample from the severed tip of the tail. Unlike this latter method, sampling from an arterial catheter is not stressful to the mouse1. We also describe methods for using isotopic tracer infusions to assess tissue-specific insulin action. We also provide guidelines for the appropriate presentation of results obtained from insulin clamps.
1. Preparation of catheters and Mouse Antenna for Sampling Access (MASAtm)
2. Surgical catheterization
3. Hyperinsulinemic-euglycemic clamp
Note: The volume of blood sampled depends on the analysis being performed. For example, analysis of [3-3H]glucose concentration requires 10 μl of plasma, so 50 μl of blood are drawn. This yields 20-30 μl of plasma, which is sufficient for the analysis plus additional plasma if needed. Measurements of hormones and other metabolites (e.g. insulin, free fatty acids) require sampling of additional blood.
4. Representative Results
An example of results obtained from an insulin clamp experiment is shown in Figure 4. This example shows the ability of a high fat diet to precipitate insulin resistance in mice. All presentations of insulin clamp results must include the following to be interpretable: a time course of blood glucose levels, a time course of the GIR and plasma insulin levels (baseline and clamp). As shown here, fasting glucose (Figure 4A) and insulin (Figure 4C) levels are higher in mice fed a high fat diet, indicative of insulin resistance. Presenting a time course of glucose levels throughout the clamp study (Figure 4A) allows the reader to assess how well euglycemia was maintained, which is indicative of the quality of the clamp. Similarly, a time course of the GIR (Figure 4B) allows the reader to determine how quickly a steady-state was achieved. Showing these data as time courses is significantly more informative than the conventional practice in the mouse insulin clamp literature of presenting a 2-hour experiment as a single datum point representing average values from an undefined “clamp” period (4-13). In the present example, glucose levels were equal between the control and high fat-fed groups, but the GIR was significantly lower in the high fat-fed group (Figure 4B). This is indicative of an impairment in whole-body insulin action. Clamp insulin levels were also higher in the high fat-fed group (Figure 4C), further supporting the presence of an insulin resistant phenotype in these mice. The use of isotopic tracer infusions allows for the assessment of insulin action in specific tissues. [3-3H]glucose is used to estimate the rate of endogenous glucose appearance (endoRa), which is an index of hepatic glucose production (HGP) and the rate of whole-body glucose disappearance (Rd). Whereas insulin completely suppresses HGP in control mice, this is impaired in mice fed a high fat diet (Figure 4D). Similarly, the ability of insulin to stimulate Rd in control mice is compromised in mice fed a high fat diet (Figure 4E). 2[14C]deoxyglucose is used to assess the glucose metabolic index (Rg), a measure of tissue-specific glucose uptake. As seen in this example, insulin-stimulated glucose uptake into skeletal muscle is impaired in mice fed a high fat diet (Figure 4F).
Figure 1: Preparation of arterial (A) and (B) venous catheters and (C) MASAtm. Arterial catheters are prepared by inserting a 1.3 cm piece of PE-10 about 3 mm into a 6 cm piece of 0.012″ID silastic. The PE-10 tip is beveled such that the length from the bevel to the silastic is 0.9 cm. Venous catheters are made by sliding a small piece of 0.020″ID silastic 1.1 cm from the beveled end of a 6 cm piece of 0.012″ID silastic. The 0.020″ID silastic piece acts as a restraining bead to secure the catheter to the jugular vein. For assembly of the MASAtm, each of two 1.3 cm 25-gauge connectors is inserted into each of two 3 cm pieces of PE-20. These are held together by a small piece of 0.040″ID silastic. The connectors are bent to a 120° angle and separated at a 45° angle. The entire assembly is immersed in medical silicone adhesive.
Figure 2: Catheterized mouse. Catheters are surgically implanted in the left carotid artery and the right jugular vein. The free ends of the catheters are externalized behind the head and connected to a MASAtm. The MASAtm is inserted subcutaneously between the shoulder blades. This allows for vascular access during insulin clamp experiments without the need to restrain, handle or anesthetize the mouse.
Figure 3: Depiction of the setup and time line for an insulin clamp experiment. The mouse is tethered to a dual-channel swivel that acts as a hub for infusion and sampling syringes. Typical setups for experiments not using tracer infusions (A) and using both [3-3H]glucose and 2[14C]deoxyglucose (B) are shown. A time line of procedures for setting up and performing the insulin clamp (C) is also shown. During the clamp, blood samples () are taken every 10 min to measure blood glucose. The GIR is adjusted accordingly to maintain euglycemia. Samples for baseline blood glucose, plasma insulin, and plasma [3-3H]glucose are taken at t = -15 and -5 min. Samples for clamp plasma [3-3H]glucose are taken at t = 80, 90, 100, 110 and 120 and for clamp insulin at t = 100 and 120 min. 2[14C]deoxyglucose is administered after the sample at t =120 min and blood is collected at t = 2, 15, 25 and 35 min after. Tissues are taken after the t = 35 min sample.
Figure 4: Results from an insulin clamp experiment comparing mice on a control diet (Chow) to mice on a high fat diet (HFD). Time course of arterial glucose (A) and GIR (B), baseline and clamp insulin (C), EndoRa (D), and Rd (E) and skeletal muscle (gastrocnemius and vastus lateralis) Rg (F) are shown. All results indicate the effect of high fat feeding to induce insulin resistance.
The hyperinsulinemic-euglycemic clamp, or insulin clamp, is widely considered the “gold standard” method for assessing insulin action in vivo. This technique has been applied to several species including humans, dogs, rats and mice. Given the growing number of transgenic mouse models for metabolic disorders, the miniaturization of the technique for use in the mouse has provided significant advances to metabolic research.
While the concepts behind the insulin clamp are straightforward, in practice there are different approaches for performing insulin clamp experiments. This is not a trivial point, since the manner in which the experiment is performed affects results obtained1. Here we present the protocol used by the Vanderbilt MMPC. The key distinction between our protocol and that of others is that we use an arterial catheter for obtaining blood samples. This is in contrast to the more widely used approach of obtaining blood samples by severing the tip of the tail4, 7, 11, 12, 14-17. The advantage of sampling from an arterial catheter is that the experiment is conducted in a conscious and unrestrained mouse. Sampling from the tail often requires restraint and increases indices of stress when large blood samples are acquired1. Stress hormones stimulate endogenous glucose production and impair glucose disposal18, 19, potentially giving the appearance of an insulin resistant phenotype. Sampling from the severed tail may require special Institutional Animal Care and Use Committee approval because of its stressful nature. The arterial catheterization procedure was developed to avoid the stress to the mouse of severing the tail.
A key aspect of performing insulin clamps is the ability to maintain euglycemia. There are no algorithms that can correctly predict how the GIR should be adjusted based on blood glucose readings. Like the surgery, personnel conducting insulin clamp experiments will become proficient in maintaining reasonable euglycemia only through experience. It is important to note that because of their higher metabolic rate the data obtained from mouse studies will be inherently noisy. This makes the complete presentation of data, including time courses of glucose and GIR and absolute values for plasma insulin, endoRa, Rd and Rg crucial for the ability of any reader to interpret results. The high glucose flux rates in the mouse (approximately 5 times higher than the rate in humans) warrant a high frequency of glucose sampling. While the blood volume of the mouse is limited, a minimum sampling frequency of once every 10 minutes is necessary to be certain that an adequate clamp has been achieved.
As shown in Figure 4, clamp insulin levels can be different between groups. Factors such as diet interventions, transgenic manipulations or differences in background strains can affect fasting insulin levels, which can subsequently affect clamp insulin levels. Interpreting results when clamp insulin levels are different can be problematic. This can be experimentally addressed by performing pilot experiments to select insulin infusion rates that achieve equivalent clamp insulin levels between groups. Alternatively, somatostatin can be used to inhibit pancreatic hormone secretion, and insulin and glucagon can be replaced at experimentally controlled rates. This latter approach is more commonly done in insulin clamps on rats than mice. If these experimental approaches are not taken, the steady-state GIR can be normalized to the clamp insulin level, or an insulin sensitivity index (SI) can be derived from clamp data as SI=GIR/(G•ΔI), where G is the steady-state glucose concentration and ΔI is the difference between fasting and clamp insulin concentrations. One assumption with either approach is that the clamp insulin level achieved is within the range where insulin sensitivity is linearly related to the insulin level according to the group being studied. This latter assumption may not apply when comparing insulin resistant and insulin sensitive groups. Ideally, an insulin dose response curve should be generated to select the appropriate insulin infusion. However, because of the requirement for additional experiments, this is rarely done.
The versatility provided by arterial catheterization extends to experimental approaches beyond euglycemic clamps. For example, hyperglycemic clamps, in which glucose is infused at a variable rate to maintain hyperglycemia relative to fasting glucose, can be used to assess endogenous pancreatic function in vivo2, 20, 21. The measurement of first phase insulin secretion during this test requires frequent acquisition of blood samples (i.e every 2-5 min), which is not feasible when obtaining samples from the tip of the tail. Furthermore, elevated catecholamines resulting from tail sampling can impair insulin secretion and enhance glucagon secretion22. The insulin clamp protocol can also be modified to allow for glucose levels to fall to relative hypoglycemia to assess counter-regulatory response2, 23, 24. Arterial catheterization can also be used to assess the dynamics of glucose metabolism during exercise25-30. This is significantly advantageous over conventional approaches conducted at single time points pre- and post-exercise or in isolated muscles ex vivo. The techniques presented here can also be used to assess not just glucose, but also fatty acid metabolism31.
The authors have nothing to disclose.
This work was supported by Grant 5-U24-DK059637-10 to the Vanderbilt Mouse Metabolic Phenotyping Center.