Postmaximal contraction blood volume responses are blunted in obese and type 2 diabetic subjects in a muscle-specific manner

Abstract

The purpose of this study was to determine whether there are differences in postisometric contraction blood volume and oxygenation responses among groups of type 2 diabetes mellitus (T2DM), obese, and lean individuals detectable using MRI. Eight T2DM patients were individually matched by age, sex, and race to non-T2DM individuals with similar body mass index (obese) and lean subjects. Functional MRI was performed using a dual-gradient-recalled echo, echo-planar imaging sequence with a repetition time of 1 s and at two echo times (TE = 6 and 46 ms). Data were acquired before, during, and after 10-s isometric dorsiflexion contractions performed at 50 and 100% of maximal voluntary contraction (MVC) force. MRI signal intensity (SI) changes from the tibialis anterior and extensor digitorum longus muscles were plotted as functions of time for each TE. From each time course, the difference between the minimum and the maximum postcontraction SI (ΔSI) were determined for TE = 6 ms (ΔSI₆) and TE = 46 ms (ΔSI₄₆), reflecting variations in blood volume and oxyhemoglobin saturation, respectively. Following 50% MVC contractions, the mean postcontraction ΔSI₆ values were similar in the three groups. Following MVC only, and in the EDL muscle only, T2DM and obese participants had ∼56% lower ΔSI₆ than the lean individuals. Also following MVC only, the ΔSI₄₆ response in the EDL was lower in T2DM subjects than in lean individuals. These data suggest that skeletal muscle small vessel impairment occurs in T2DM and body mass index-matched subjects, in muscle-specific and contraction intensity-dependent manners.

the delivery and distribution of blood through skeletal muscle vasculature is normally well regulated by a coordinated response that includes the release of local vasoactive substances, neural control, and local mechanical effects (49, 54). However, structural and functional alterations in the microvasculature are present in obese individuals (reviewed in Ref. 34), and there may be an independent effect of type 2 diabetic mellitus (T2DM) on the muscle microvasculature as well (30). These deficits have potential clinical significance, because skeletal muscle is the principal site for glucose uptake and storage in the body (6) and so inadequate skeletal muscle perfusion may act as a mechanism of whole body insulin resistance (11). Also, acute whole body exercise results in insulin-independent glucose uptake, and chronic exercise has favorable consequences for cardiovascular health and energy balance, but requires adequate muscle perfusion to support the increased metabolic demand. For these reasons, it is essential to evaluate small-vessel function in skeletal muscle in individuals with or at high risk of T2DM.

Recently, functional magnetic resonance imaging (MRI) methods, such as blood oxygenation-level dependent contrast, have been applied to skeletal muscle to infer microvascular function following isometric contractions (47, 59, 63), arterial occlusion (40), and the infusion of vasoactive compounds (65). Building on these studies, we have used a dual-gradient-recalled echo (GRE) MRI sequence to study blood volume (V_B) and oxyhemoglobin saturation (%HbO₂) changes in small vessels of the muscle tissue bed following isometric contractions noninvasively, and with high temporal resolution (1 s) (13) and sufficient repeatability (56). This method consists of measuring the MRI signal intensity (SI) at two echo times (TE): 6 and 46 ms. Following an isometric contraction, there are transient rises in the TE = 6 and 46 ms SI values (SI₆ and SI₄₆, respectively); we characterize these changes with the (maximum postcontraction SI) − (minimum postcontraction SI) difference (ΔSI) (56). While postcontraction V_B and %HbO₂ changes contribute to the SI changes at each TE (13, 62), the predominant determinants of ΔSI₆ and ΔSI₄₆ are V_B and %HbO₂, respectively (13). The approach is similar in principle to contrast-enhanced ultrasound (68), in that changes in an arbitrary image SI are interpreted to represent small-vessel function by applying an underlying biophysical model to the data.

The physiological basis for the SI₆ and SI₄₆ responses rests, in part, on the observation that, during an isometric contraction, intramuscular pressure increases. This pressure compresses the small vessels, except capillaries, along the muscle's fascicular lines (26). This vascular compression, in turn, ejects blood from the venous circulation and restricts blood flow into the arterioles, thereby decreasing V_B (15). Following a contraction, the collective action of neural, paracrine, and mechanical factors permits refilling of the small vessels and induces transient increases in blood flow and volume. There is a brief period during which the oxygen supply exceeds the oxygen demand, and so a transient hyperoxygenation occurs (12). The magnitude of the vascular response is proportional to the contraction intensity (5, 8, 70) and may further depend on the muscle's predominant fiber type and capillary density (38) and the physical fitness (63) and insulin sensitivity (28) of the individual. Because muscles with a high proportion of type I fibers are more insulin sensitive than those with a high proportion of type II fibers (27, 29), the relative impairment to muscle microvascular function in insulin-resistant individuals may be more pronounced in predominantly type II-fibered muscles. This possibility suggests the need to assess muscle microvascular function in a muscle-specific manner. Also, the measurements should be interpreted with respect to indexes of insulin sensitivity and properly measured chronic physical activity levels.

To investigate these issues, ΔSI₆ and ΔSI₄₆ following isometric dorsiflexion contractions were assessed in the tibialis anterior (TA) and the extensor digitorum longus (EDL) muscles in three groups of subjects: those with T2DM, body mass index (BMI)-matched (generally obese) nondiabetic subjects, and lean nondiabetic subjects. We studied the dorsiflexors because they are functionally important in postural control and locomotion, as revealed by the prior observation that dysfunction in the dorsiflexors may contribute to foot drop and the diabetic foot syndrome (1). We hypothesized that ΔSI₆ and ΔSI₄₆ would be lower in obese nondiabetic individuals than in lean nondiabetic individuals, and [because there may be an independent effect of T2DM on the muscle microvasculature (30)] we hypothesized that ΔSI₆ and ΔSI₄₆ would be further impaired in T2DM patients. We also hypothesized that these impairments would be greater in the EDL muscle [47% type I fibers (33)] than in the TA muscle [74% type I fibers (33, 52)]. Our final hypothesis was that, if insulin sensitivity depends on skeletal muscle small-vessel function, then ΔSI₆ and ΔSI₄₆ in response to muscle contractions would correlate to indexes of insulin sensitivity. While the largest responses (and, therefore, the greatest dynamic range for the detection of differences) should occur following maximal contractions (5, 8, 70), we also measured the responses to 50% maximal contractions, so that the full extent of any potential deficits in obese or T2DM individuals could be characterized.

METHODS

Subjects.

This study was approved by the Vanderbilt University Institutional Review Board. Eight subjects with physician-diagnosed T2DM (3 women) between 30 and 45 yr of age were individually matched by age (±2 yr), sex, and race to BMI-matched (±2 kg/m²) and lean individuals (BMI <25 kg/m²). Using the Centers for Disease Control and Prevention's BMI criteria (10), six of the eight T2DM subjects were obese, and the other two were overweight. The exclusion criteria are listed in Table 1. Medical approval was required for all T2DM patients, and every participant signed a written, informed consent form.

Table 1.

Exclusion criteria

Exclusion Criteria
All subjects
Smoking
Pregnancy or recent change in menstrual status
Ankle brachial index < 0.9
Limitation in the ability to exercise the dorsiflexor muscles
Contraindication to MRI
T2DM patients
Physician-diagnosed chronic disease other than T2DM
Chronic complications of diabetes (cardiomyopathy, proliferative retinopathy, neuropathy, or nephropathy)
Obese nondiabetic and lean nondiabetic subjects
Any physician-diagnosed chronic disease, as indicated by subject self-report

The existence of other chronic disease or diabetes complications was indicated by self-report and/or clinical records. Urinary albumin-to-creatinine (A/C) ratio > 300 μg albumin/mg creatinine was also used to exclude for nephropathy. MRI, magnetic resonance imaging; T2DM, type 2 diabetes mellitus.

Experimental protocol.

Each subject reported to the laboratory for orientation and testing sessions (Fig. 1). During the orientation session, the purposes and procedures of the study were explained, and the subject filled out health history, MRI screening, and physical activity questionnaires. Systolic blood pressure (BP) was measured using a 5-MHz Nicolet Vascular Doppler Elite 200 ultrasound device (Nicolet Vascular, Madison, WI) in the brachial and dorsalis pedis arteries to determine the ankle-brachial index (ABI). Resting heart rate (HR) was measured by palpation. During the interim period between the orientation and testing sessions, the subject continued all prescribed drugs, maintained regular physical activity habits, and wore a uniaxial accelerometer/pedometer (GT-1M ActiGraph, Pensacola, FL) (46) to measure average daily physical activity levels. To control for factors that affect vascular function, subjects were instructed to restrict certain behaviors during the 24 h before testing (Fig. 1).

Fig. 1.

Time sequence for the experimental procedures for each participant. After consent and health screening, an interim period between the orientation session and the testing day of at least 1 wk occurred. During the interim period, certain lifestyle behaviors were restricted, and physical activity was monitored. The bottom rectangle details the procedures during the testing day. MRI, magnetic resonance imaging; ABI, ankle-brachial index; A/C, urinary albumin-to-urinary creatinine ratio.

On the testing day, the subject arrived at the Vanderbilt General Clinical Research Center following an 8-h fast. Height, body mass, resting HR and systolic/diastolic BP (by auscultation), glycosylated hemoglobin (Hb_A1c), and fasting blood glucose concentration ([glucose]) and insulin concentration ([insulin]) values were determined. Fasting [glucose] and [insulin] were used to estimate steady-state β-cell function (%B) and insulin sensitivity (%S), as percentages of a normal reference population, using the Homeostasis Model Assessment method (67). Subjects collected urine samples on 2 consecutive days, and the mean urinary albumin-to-urinary creatinine ratio (A/C) was determined to assess capillary glomerular permeability to albumin as a marker of microvascular damage (2).

Caloric intake was limited on the testing day to account for the acute effects of blood glucose and free fatty acid levels on vascular function (39, 42). The subjects were provided with a breakfast consisting of 360 kcal (49% carbohydrate, 22% protein, and 29% fat). In addition, a 150-kcal snack (58% carbohydrate, 18% protein, and 24% fat) was offered following the exercise practice session. Before the MRI-exercise test, blood [glucose] was measured, and, if it was >180 mg/dl, the imaging session was rescheduled to prevent complications from acute hyperglycemia (57).

Exercise protocol.

Each subject performed isometric dorsiflexion contractions using his or her self-reported dominant foot to determine the maximal voluntary contraction (MVC) force and then practiced contractions at 50% of their MVC (50MVC). The exercise equipment and procedures have been reported in detail previously (43). Briefly, the subject was positioned supine on an exercise bench (for out-of-magnet studies) or the patient bed of the imager. The thigh and knee were slightly flexed, and the foot was strapped securely in an isometric dorxiflexion/plantarflexion apparatus. Contractions were performed using a custom-built, MRI-compatible device containing a load cell. Signal processing and recording included differential amplification and signal conditioning using a bridge amplifier, analog-to-digital conversion at 1,000 Hz, and recording on a laptop computer using a custom-written LabVIEW (National Instruments, Austin, TX) program that provided 20-Hz, real-time visual feedback concerning relative force production.

During the MRI exercise test, subjects initially performed three 3-s warm up MVCs, each separated by 1 min. Subjects then performed four 10-s isometric dorsiflexion contractions at 50MVC and four contractions at MVC, in random order. Four minutes of rest were allowed between each contraction to provide sufficient time to recover energy stores and prevent muscle fatigue (35). The subjects received verbal encouragement during the contractions and were instructed to remain as still as possible after the contraction ended to avoid motion artifacts in the magnetic resonance (MR) images.

MRI data acquisition.

MRI data were acquired on a 3.0-T/60-cm horizontal bore Philips Intera Achieva MR imager/spectrometer using an eight-channel sensitivity encoding knee coil. The exercise device was bolted to the patient bed of the imager. The subject lay supine on the patient bed with his or her foot strapped to the exercise device. During a light contraction, the borders of the TA muscle were identified by visual inspection, and the maximal girth was aligned with center of the knee coil. To restrict subject motion, padded inserts were placed around the leg in the coil. After advancing the subject into the magnet, 3-plane GRE scout images were acquired to locate the maximal cross-sectional area (CSA) of the TA muscle. A single slice at that location was chosen for the remainder of the imaging acquisitions. Before acquiring each set of images, second-order shimming was performed on a volume that was just thicker than the imaging slice and that included the imaged leg only. Following the scout images, a high-resolution T₁-weighted anatomical image was acquired using a fast spin-echo sequence with slice thickness = 10 mm, field of view = 18 × 18 cm, acquired matrix = 256 × 128 (reconstructed matrix = 256 × 256), repetition time = 500 ms, TE = 16 ms, and number of excitations = 2. Next, muscle functional MRI data were acquired using a dual-GRE, echo-planar imaging sequence for 15 s before, during, and for 125 s after a 10-s isometric contraction. The functional images were acquired using the same geometric parameters as for the anatomical image, acquired matrix = 64 × 64 (reconstructed matrix = 128 × 128), slice thickness = 7.5 mm, repetition time = 1,000 ms (defining the time resolution of the measurements), and TE = 6 or 46 ms.

Data analysis.

Images were analyzed by using custom-written Matlab version 7.0.1 (The Mathworks, Natick, MA) routines. The functional images were coregistered to the anatomical image using a rigid registration procedure based on the optimization of a mutual information metric (69). Then the anatomical image was used to identify the TA and EDL muscles and regions of interest (ROIs) were specified separately around the EDL and the TA muscles. Resolved connective tissues and blood vessels were excluded from the ROIs. At each TE, the mean functional image SI for the ROI was calculated as a function of time and divided by the corresponding mean precontraction SI value. An average SI time course from each TE was obtained by averaging the SI time course from the four trials from each intensity (SI₆ and SI₄₆), as previously described (56). The averaged SI time course was fitted to a polynomial function, and inflections in the derivative of this polynomial were used to identify the minimum and maximum postcontraction SI values. The differences between these values were determined for each TE (ΔSI₆ and ΔSI₄₆). The ΔSI₆ and ΔSI₄₆ values were interpreted as reflecting variations in V_B and %HbO₂, respectively, in all vessels in the muscle bed that were not resolvable, given the image's spatial resolution (13). To account for variations in subject size on the force measurements, the mean relative force was calculated for each contraction. MVC force was normalized by dividing the force by the combined CSA of the TA and the EDL muscles, as measured in the anatomical images. The accelerometry data were quantified as the average number of counts per day.

Statistical analysis.

Statistical analysis was performed using SPSS version 15.0 (SPSS, Chicago, IL). Values are presented as means ± SE. A one-sample t-test was used to determine if subjects attained target force values during the contractions. Because we used a matched-subjects design, a general linear model, within-subjects analysis was performed to test for differences in ΔSI as a main effect of group (T2DM, obese, and lean), separately for each intensity and muscle. Bonferroni's post hoc pairwise comparison was applied when indicated. Similar procedures were used to compare the mean values for the descriptive and clinical data. For tests of means, statistical significance was accepted at P < 0.05.

Pearson correlation coefficients were determined between 11 physiological variables (see below) and the post-MVC ΔSI₆ in the EDL muscle. The post-MVC ΔSI₆ in the EDL was chosen because it had most significant differences among groups (see results). The following variables were assumed to have a potential physiological or metabolic relationship to ΔSI₆ and were used as independent variables: BMI, resting HR, systolic and diastolic BP, daily physical activity, fasting and pre-MRI blood glucose, %B, %S, Hb_A1c, and A/C. The five independent variables having the highest correlations with ΔSI₆ were then used to develop a multiple linear regression model using the stepwise algorithm method. The entry P value was ≤0.05, and the removal P value was ≥0.10.

RESULTS

General characteristics.

There were no differences in the groups' mean values of age, resting HR, systolic and diastolic BP, and ABI (Table 2). The T2DM subjects were well controlled, had been diagnosed for an average of 4.0 ± 2.9 yr, and took an average of 1.6 ± 1.0 blood glucose control medications per subject. One subject did not require any blood glucose-lowering medication, and two subjects took insulin.

Table 2.

General descriptive characteristics of the groups

	Lean	Obese	T2DM
Age, yr	37.8 ± 1.4	37.4 ± 1.5	37.3 ± 1.5
BMI, kg/m2	22.9 ± 0.6*†	32.8 ± 1.8	33.9 ± 1.7
Resting HR, beats/min	72.9 ± 3.5	73.6 ± 3.2	71.7 ± 3.6
Diastolic BP, mmHg	74.1 ± 3.4	78.0 ± 3.3	72.5 ± 2.6
Systolic BP, mm/Hg	124.3 ± 4.6	127.3 ± 4.2	123.5 ± 2.6
ABI	1.02 ± 0.01	1.00 ± 0.01	0.99 ± 0.01
FH diabetes	3/8	5/8	3/8
FH CVD	3/8	3/8	2/8
FH HTA	2/8	5/8	1/8
Years with T2DM			4.0 ± 2.9
No. of medications			1.6 ± 1.0

Values are means ± SE; n = 8 subjects. The groups were matched for race, age, and sex; the T2DM and obese groups were matched in body mass index (BMI). Among the measured descriptive characteristics, the only significant difference among the groups was in BMI. HR, heart rate; BP, blood pressure; ABI, ankle-brachial index; FH, family history; CVD, cardiovascular disease; HTA, hypertension.

^*P < 0.05, lean vs. T2DM.

^†P < 0.05, lean vs. obese.

There were no differences among the groups in physical activity, homeostasis model assessment method measured %S, and A/C (Table 3). The fasting blood [glucose] was higher in the T2DM subjects than in the lean (P = 0.007) and obese (P = 0.03) groups. The mean pre-MRI blood glucose value was higher in the T2DM subjects than in the lean subjects (P = 0.02). β-Cell function was higher in the obese group than in the T2DM group (P = 0.004). All other intergroup comparisons were not statistically significant.

Table 3.

Physical activity and metabolic data of the groups

	Lean	Obese	T2DM
Physical activity, ×103 counts/day	250.9 ± 37	249.1 ± 61	218.1 ± 22
Glucose fasting, mg/dl	92.4 ± 3.3*	90.0 ± 6.4†	128.5 ± 8.9
Glucose pre-MRI, mg/dl	109.5 ± 6.5*	107.9 ± 7.5	140.9 ± 9.5
HbA1c, %	5.2 ± 0.1*	5.4 ± 0.1†	7.1 ± 0.4
%B	70.3 ± 5.2	149.2 ± 25.4†	65.9 ± 15.9
%S	142.2 ± 23.0	73.6 ± 14.9	93.2 ± 29.9
Urinary A/C	3.44 ± 0.4	1.68 ± 0.3	45.8 ± 31.3

Values are means ± SE. Hb_A1c, glycosylated hemoglobin; %B, β-cell function; %S, insulin sensitivity. Physical activity was monitored for an average of 11.8 ± 0.8 days per subject. The large variability for A/C in the T2DM group was due to one subject with a value of 258.6 g/l. Without this subject, the mean A/C for the T2DM group decreased to 15.4 ± 8.6 g/l, but still did not differ significantly from that of the obese or lean groups (P = 0.14).

^*P < 0.05, lean vs. T2DM.

^†P < 0.05, obese vs. T2DM.

Force measurements.

The mean values for the CSA-normalized MVC force did not differ among the subject groups (see Table 4; P = 0.11; n = 7 due to data loss for one T2DM subject). When expressed relative to the value measured during the pre-MRI testing, the contraction intensity during MRI testing did not differ among the groups at 50MVC, but was significantly lower in the T2DM group than in the lean group at MVC (P = 0.02). The T2DM group did not attain the pre-MRI testing MVC value during the MVC's performed in the MR imager. There were no differences in relative contraction intensity as functions of trial number for the 50MVC (P ≥ 0.27 for each group) and MVC contractions (P ≥ 0.08 for each group).

Table 4.

Normalized MVC force and relative contraction intensities during 50MVC and MVC contractions

	Lean	Obese	T2DM
Normalized MVC force, N/cm2	38.5 ± 2.8	31.9 ± 3.5	38.4 ± 2.7
Relative intensity, %pretesting MVC
50MVC	48.7 ± 0.6	51.2 ± 0.9	50.0 ± 0.9
MVC	102.5 ± 4.0	95.5 ± 3.0	88.2 ± 2.6*

Values are means ± SE. MVC, maximum voluntary contraction; 50MVC, 50% of MVC. There were no differences among the groups in the mean values of normalized force (P = 0.25). At MVC, the mean relative contraction intensity was lower for T2DM subjects than for the lean subjects.

^*P < 0.05, lean vs. T2DM.

Differences in ΔSI₆ and ΔSI₄₆ by intensity, muscle, and group.

Figure 2 provides representative post-MVC SI₆ and SI₄₆ time courses in matched lean, obese, and T2DM participants. Figure 2, A and D, shows, respectively, the SI₆ and SI₄₆ time courses for the lean subject. The spikes in SI at the times of −10 and 0 s reflect the beginning and end of the contraction. After the contraction, a pronounced rise and fall in SI are noted. In contrast, there are small or essentially nonexistent responses in the T2DM and obese subjects (Fig. 2, B, C, E, and F).

Fig. 2.

Time courses for signal intensity for echo times of 6 ms (SI₆; A–C) and 46 ms (SI₄₆; D–E) before, during, and following an isometric maximal voluntary contraction (MVC) in the extensor digitorum longus (EDL) muscle of a lean subject (A and D), the obese matching subject (B and E), and type 2 diabetes mellitus (T2DM) matching subject (C and F). The data for the lean subject reveal several characteristics of the typical responses: the large spikes are typical contraction-induced artifacts and reflect the beginning and end of the contraction (13, 47, 56, 63). Both SI₆ and SI₄₆ demonstrate a transient rise and fall in signal following the contraction. The maximum postcontraction and minimum postcontraction SI values are indicated by the points labeled a and b, respectively. Severely blunted responses are noted in the obese and T2DM subjects.

Figure 3, A and B, presents the mean ΔSI₆ values for the EDL and TA, respectively; Tables 5 and 6 present all of the ΔSI data in tabular form. At 50MVC, there were no differences in the mean ΔSI₆ values among the groups for the EDL muscle (P > 0.10) or the TA muscle (P > 0.20). However, the responses to more intense contractions differed by group and muscle. In the EDL, the mean MVC ΔSI₆ value was 5.0-fold higher than the mean 50MVC value in the lean group (P = 0.002; Fig. 3A), but the mean ΔSI₆ value failed to increase as a function of contraction intensity in the T2DM and obese groups (P = 0.57 and 0.99, respectively). The EDL's mean MVC ΔSI₆ value was smaller in the T2DM group than in the lean group (P = 0.02) and tended to be smaller in the obese group than in the lean group (P = 0.055).

Fig. 3.

Mean difference (Δ) in SI₆ (A and B) and ΔSI₄₆ (C and D) values in the EDL (A and C) and tibialis anterior (TA) muscles (B and D) by group and contraction intensity. Values are means ± SE. 50MVC, 50% of MVC. *P < 0.05, lean vs. T2DM; **P = 0.055, lean vs. obese; and #P < 0.05, MVC vs. 50MVC for the indicated group. In the EDL, the ΔSI₆ responses were similar among the groups following the 50MVC contractions, but deficits were noted following MVC in the obese and T2DM groups. In the TA, only the T2DM group failed to increase its mean ΔSI₆ value between 50MVC and MVC. In contrast to the ΔSI₆ data, the impairments in the obese and T2DM groups were not as great for ΔSI₄₆ as they were for ΔSI₆.

Table 5.

Values for signal intensity difference for 6-ms echo time in the EDL and TA muscles, presented by group and contraction intensity

	EDL		TA
	50MVC	MVC	50MVC	MVC
T2DM	0.74 ± 0.13	0.89 ± 0.23	1.33 ± 0.36	1.16 ± 0.29
Obese	0.91 ± 0.28	0.90 ± 0.19	0.80 ± 0.18	1.56 ± 0.22
Lean	0.40 ± 0.09	2.04 ± 0.31*†‡	0.78 ± 0.29	2.14 ± 0.44

Values are means ± SE. EDL, extensor digitorum longus; TA, tibialis anterior.

^*P = 0.02 vs. T2DM group at MVC.

^†P = 0.055 vs. obese group at MVC.

^‡P = 0.002 vs. 50MVC.

Table 6.

Values for signal intensity difference for 46-ms echo time in the EDL and TA muscles, presented by group and contraction intensity

	EDL		TA
	50MVC	MVC	50MVC	MVC
T2DM	4.99 ± 1.36	4.24 ± 0.73	4.78 ± 0.47	7.33 ± 0.53†
Obese	3.58 ± 0.72	6.27 ± 0.96†	4.19 ± 0.84	8.40 ± 1.21†
Lean	3.36 ± 0.55	7.89 ± 0.50*†	4.19 ± 1.37	9.21 ± 1.41†

Values are means ± SE.

^*P = 0.04 vs. T2DM group at MVC.

^†P < 0.05 vs. respective mean at 50MVC.

In the TA, the mean ΔSI₆ responses to MVCs were larger than the responses to 50MVC contractions in the lean and obese groups only (P < 0.001 and P = 0.01, respectively; P = 0.65 for the T2DM group; Fig. 3B). There were no differences among the groups in the TA's mean MVC ΔSI₆ values (P = 0.18).

Figure 3, C and D, compares the mean ΔSI₄₆ values in the EDL and the TA, respectively. At 50MVC, there were no differences in the mean ΔSI₄₆ responses among the subject groups for either muscle (P = 0.18 and 0.87 for the EDL and TA, respectively). In the EDL, the mean MVC ΔSI₄₆ values for the lean (P = 0.002) and obese (P = 0.02) groups increased by approximately twofold over the respective values for 50MVC, but the mean ΔSI₄₆ in the EDL did not increase as a function of contraction intensity in the T2DM group (P = 0.60) (Fig. 3C). The MVC mean ΔSI₄₆ in the EDL muscle was lower in the T2DM group than in the lean group (P = 0.04).

In the TA, the MVC ΔSI46 values increased by 1.5- to ∼2-fold over the values for 50MVC (P < 0.05 for all groups; Fig. 3D). There were no differences in the mean MVC ΔSI₄₆ values in the TA muscle among the groups (P = 0.56).

Associations among ΔSI₆ and selected physiological and metabolic variables.

Figure 4 shows the scatterplots for the three variables that were significantly correlated with MVC ΔSI₆ in the EDL: BMI (r² = 0.387), Hb_A1c (r² = 0.172), and resting HR (r² = 0.171). Among the other variables tested, only the correlations of physical activity (r² = 0.151, P = 0.061) and pre-MRI glycemia (r² = 0.141, P = 0.071) with MVC ΔSI₆ in the EDL were near significant. When these five variables were entered into the multiple regression model, only BMI significantly predicted the EDL's MVC ΔSI₆.

Fig. 4.

Scatterplots illustrating the relationships between the three variables having significant (P < 0.05) correlations with ΔSI₆ in the EDL muscle following MVC. A: body mass index (BMI). B: glycosylated hemoglobin (HbA1c). C: resting heart rate (HR). In the multiple-regression analysis, only the BMI remained as a significant predictor of the EDL's post-MVC ΔSI₆. bpm, beats/min.

DISCUSSION

A functional MRI method sensitive to postisometric contraction changes in V_B (as reflected by ΔSI₆) and %HbO₂ (as reflected by ΔSI₄₆) in the small vessels of skeletal muscles was used to assess differences among groups of T2DM, BMI-matched (generally obese), and lean individuals. In the EDL muscle, lean subjects had a larger post-MVC V_B response than did the T2DM and obese subjects (Figs. 2 and 3). These deficits in the T2DM and obese subject groups occurred in the absence of overt macro- and microvascular complications and were more strongly related to the presence of obesity than to the presence of T2DM. In addition to providing new information about the nature of skeletal muscle microvascular impairment in insulin-resistant conditions, such as obesity and T2DM, the findings also highlight the importance of evaluating small-vessel function in muscle- and contraction intensity-specific manners.

Variations in the post-MVC V_B response among groups.

As in our laboratory's previous studies (13, 56), we observed transient rises in SI₆ following brief isometric contractions and used the (maximum postcontraction) − (minimum postcontraction) SI difference to characterize their magnitude (Fig. 2). The magnitudes of the V_B responses following 50MVC contractions were similar in the T2DM, obese, and lean groups. In lean, healthy individuals, an increasing V_B response with contraction intensity is expected (56, 70), and this indeed occurred in both the EDL and TA muscles of these subjects (Fig. 3). However, a major finding of the present study was that, in well-controlled T2DM patients and obese subjects, the post-MVC V_B responses in the EDL muscle were similar in magnitude to the responses following 50MVC contractions, implying that these subjects' maximal vasodilatory capacity for a single contraction was reached at or below 50MVC. This resulted in a larger post-MVC V_B response in the lean group than in the other groups. Similar, but less pronounced, patterns were observed in the TA muscle: the post-MVC V_B response was larger than the post-50MVC V_B response for both the lean and obese groups, but the expected increase in the postcontraction hyperemia (as reflected by V_B) at the higher contraction intensity was not observed in the T2DM group (Fig. 3). This suggests that the impairment to skeletal muscle microvascular function occurs in a muscle-specific manner.

Before interpreting these findings with respect to the microvascular pathology of insulin-resistant states, we consider the influence of several potentially confounding variables. One of these is the absolute or relative contraction intensity (70). For example, reduced neuromuscular recruitment could lead to a lower post-MVC V_B response. However, reduced neuromuscular recruitment would cause the CSA-normalized force to be different among the groups, which was not the case (Table 4). Also, the relative contraction intensity did not vary as a function of contraction number, suggesting that there was not a differential sensitivity to muscle fatigue among the groups that could have confounded the data. While the relative contraction during the nominal MVC condition was lower in the T2DM group than the lean group, we do not feel that this accounted for the reduced ΔSI₆ responses in the T2DM participants, for several reasons. First, the correlation between relative contraction intensity at MVC and ΔSI₆ was not significant (r² = 0.09). Also, the pattern of relative contraction intensity depression (only the T2DM subjects had reduced values) did not match the pattern of ΔSI₆ depression (both the T2DM and obese subjects had reduced values). Significantly, Wigmore et al. (70) showed that there is a plateau in the magnitude of V_B-weighted SI transients between 60 and 100% MVC, and a preliminary report from our group using similar MRI and exercise methods as those used presently indicated that there is no difference in the ΔSI₆ responses between 85 and 100% MVC contractions (55). Finally, the apparently reduced contraction intensities reflect an a priori decision to normalized contraction intensities by the pre-MRI testing MVC value; however, all subjects attempted to perform maximal efforts during MRI testing. Collectively, these factors make it highly unlikely that the >50% reduction in the post-MVC V_B response in diabetic subjects can be explained by an ∼10% reduction in relative contraction intensity.

Another potentially confounding factor, daily physical activity (63), was not different among groups. In addition, the subjects were instructed not to perform vigorous physical activity 24 h before the testing session, and so it is not likely that differences in acute physical activity were responsible for the post-MVC V_B response differences among the groups either. Regarding the resting cardiovascular variables, resting HR was correlated with ΔSI₆, but the systolic and diastolic BP, resting HR, and ABI were not different among groups, and their mean values were within appropriate ranges for a healthy population. Finally, the urinary A/C was not different among groups and was not significantly correlated with ΔSI₆. Collectively, the points raised in this and the preceding paragraph support the conclusions that 1) there is a decreased post-MVC skeletal muscle V_B response that occurs in the EDL in obese and T2DM persons; and 2) these impairments occur in the absence of overt macrovascular and renal microvascular complications, as reflected in the ABI and urinary A/C values, respectively.

Physiological and pathological influences on the post-MVC V_B response.

As revealed by the patterns of group-mean differences in ΔSI₆ (Fig. 3, A and B) and the significant correlation of ΔSI₆ with only BMI in the multiple regression model, the post-MVC V_B response in the EDL was more strongly related to the presence or absence of obesity than the presence or absence of T2DM. There are a number of structural and functional alterations to the skeletal muscle microvasculature in insulin-resistant states that may account for the decreased postcontraction V_B responses. Structural changes to the microcirculation in the obese Zucker rat include thicker capillary basement membranes (37), lower capillary density (22, 37), a reduced capillary-to-fiber ratio (22), and a smaller terminal arteriole diameter (22). A reduced capillary density has also been observed in skeletal muscles of obese human subjects (25, 32).

Regarding potential functional deficits to the microcirculation, it is unlikely that differences in local mechanical effects on vascular function (such as the muscle pump) caused the differences among the groups, because the relative contraction intensities and CSA-normalized peak force levels were similar among the groups. However, endothelium-dependent deficits in the skeletal muscle resistance vessels and microcirculation occur in obesity and T2DM, including reduced responses to infused vasodilators (e.g., Refs. 3, 17, 18, 60) and shear stress (3). The potential mediators of these diminished responses may include altered serum levels of cytokines (18, 58), oxidative stress (21, 24), or nitric oxide quenching by advanced glycosylation products (9). Endothelium-independent vasodilation deficits may also occur, as indicated by reduced vasodilation on administration of the nitric oxide donor sodium nitroprusside, which has been observed in some (66), although not all (30, 44), studies. This would indicate impairment in the vasodilatory signaling cascade distal to the endothelium, perhaps due to increased adrenergic tone in the vascular smooth muscle and/or the narrowing of the arterioles (22, 23). Consistent with this, even though the mean resting HR values did not differ among the groups (Table 2), there was sufficient heterogeneity in this variable within the groups to allow detection of a general correlation between resting HR and ΔSI₆ (Fig. 4). This finding may suggest that a mechanism through which obesity influences muscle microvascular function is through an increase in sympathetic tone (19, 20). This study was not intended to distinguish between these many factors or to identify any of them as a specific pathological mechanism, but, clearly, these are important questions to be resolved in the future.

One question that arises is why the microvascular responses in the EDL were not more impaired in the T2DM group than in the obese group, especially given the correlation between ΔSI₆ and Hb_A1c (Fig. 4B). There are at least four factors that alone or in concert could have resulted in the similar responses in the T2DM and obese groups. First, the T2DM participants were, as a group, well controlled and had no other known microvascular complications. Second, metformin was prescribed in six out of the eight T2DM subjects. This medication improves microvascular and endothelial function in normoglycemic subjects with insulin resistance and in first-degree relatives of T2DM patients with metabolic syndrome, respectively (14, 36). Furthermore, metformin decreases the media-to-lumen ratio and the plasma level of the vasoconstrictor endothelin-1 in Goto-Kakizaki diabetic rats (53). Therefore, metformin may have prevented further impairment in the EDL's postcontraction V_B response in the T2DM patients. Third, the T2DM patients were a heterogeneous group with regard to family histories, medication usage, BMI, and physical activity. This heterogeneity among the potential influences on the ΔSI₆ response may have added variability to the T2DM data, reducing a potential additional effect of T2DM to below the level of detectability. Finally, we cannot exclude the possibility that severely involved T2DM patients may have voluntarily excluded themselves from the study on the basis of a limited exercise capacity.

There was a muscle-specific post-MVC V_B response such that the T2DM and obese subjects' V_B responses were impaired in the EDL only. Assuming that the EDL and TA muscles were both fully activated at MVC, this muscle-dependent impairment is most likely related to the differences in fiber-type composition between the muscles. Muscles with predominantly type I fibers have a higher capillary density (31), insulin sensitivity (41), and blood flow response to exercise (4) than muscles having predominantly type II fibers. In addition, the muscles of obese (61) and T2DM (45) subjects have a higher proportion of type II fibers than healthy subjects. Therefore, the higher proportion of less insulin-sensitive type II fibers in the EDL than in the TA may cause the EDL to be an earlier or more sensitive site of impairment than the TA. Longitudinal studies are necessary to test this possibility.

Finally, we note some similarities and differences with similarly themed published studies. Mohler et al. (48) observed a reduced ability of T2DM patients, with and without peripheral arterial disease, to expand V_B during treadmill walking, as observed by near-infrared spectroscopy; they attributed this to impaired vasodilation consequent to dysfunction in the vascular endothelium. In contrast to our results, however, Womack et al. (71) evaluated the postcontraction V_B response among lean subjects and two groups of obese T2DM subjects (with and without microvascular complications). Their study did not show any differences in the contraction-induced V_B response between lean and T2DM subjects without microvascular complications. The apparently discordant results in our study and the Womack et al. study may be attributable to differences in imaging methods (MRI vs. contrast-enhanced ultrasound), exercise protocols (single contractions vs. prolonged exercise), and/or the preparation of the subjects (continuation vs. discontinuation of prescribed medications in the pretesting period). Additionally, the muscle-specific V_B impairment pattern discussed above may explain the difference between the present findings and those of Womack et al.

Variations in %HbO₂ response among groups.

The postcontraction ΔSI₄₆ response, reflecting changes in %HbO₂, was lower in T2DM patients than in lean participants, but only at MVC and in the EDL muscle; there were no differences among groups in the TA muscle at either intensity. The data from the TA are similar to those in a previous preliminary report in which no differences in the magnitude of a post-1-s dorsiflexion MVC, blood oxygenation-level, dependent-weighted SI transient were observed between T2DM patients and BMI, physical activity, and age-matched controls (59). Considering the data from the EDL, because %HbO₂ reflects the balance between oxygen demand and supply, the lower postcontraction %HbO₂ response may be due to an increase in postcontraction oxygen extraction relative to the delivery, or a decrease in oxygen delivery relative to the extraction. Oxygen extraction is affected by the diffusion of oxygen and the microvascular Po₂ gradient, which may be amplified if mitochondrial function is increased. An increase in oxygen extraction is an unlikely mechanism, because mitochondrial function has been reported to be decreased (51) or normal (16) in insulin-resistant conditions compared with healthy individuals; furthermore, oxygen diffusion is impaired in T2DM rats (50). Thus a more plausible explanation is a decrease in oxygen delivery in the T2DM participants, consistent with the lower postcontraction V_B response in the EDL and the microvascular impairments discussed above. However, direct measurements of muscle perfusion and oxygenation would have to be performed to test this hypothesis.

Summary and conclusions.

The magnitudes of the V_B and %HbO₂ responses to 50% maximal and maximal isometric muscle contractions were compared among lean, obese, and T2DM groups using a noninvasive MRI technique. No differences among groups were observed following the submaximal contractions. In the predominantly type II-fibered EDL muscle, the V_B response to a maximal isometric contraction was lower in T2DM patients and obese subjects than in lean individuals, suggesting that structural and/or functional impairments to the skeletal muscle microcirculation can exist in both well-controlled T2DM patients and obese individuals without T2DM. In the predominantly type I-fibered TA, no significant differences in the V_B responses to isometric contractions were observed. The different patterns of impairment among the groups in the two muscles support the argument that impairments to the postcontraction V_B response are muscle specific and are probably indirectly related to muscle fiber type by virtue of differences in capillary density. Also, the balance between oxygen delivery and utilization following a maximal contraction is altered in T2DM subjects compared with lean individuals, which may be related to the decreased V_B response or other aspect of the postcontraction hyperemic response.

The V_B deficits occurred in the absence of overt macrovascular or microvascular complications and irrespective of the presence of well-controlled T2DM. Moreover, the multiple-regression analysis showed that (contrary to our hypotheses) the ΔSI₆ values were not uniquely correlated with any other measured clinical and physiological variables, except BMI. These findings suggest that measuring the post-MVC ΔSI₆ response reports on the pathology of insulin-resistant states in a manner unique from the other indexes of micro- or macrovascular dysfunction or glucose regulation that we measured. It is also important to note that, while the contractions that elicited the ΔSI₆ responses were maximal, the hyperemic response to a brief MVC is submaximal [the up to 6-fold increase in muscle blood flow over rest following a single high-intensity contraction (64) is lower than the up to 20-fold increase in flow during the steady state of intense exercise (54)]. Also, a previous study of human T2DM patients (7) suggests delayed initial muscle blood flow kinetics. Thus this exercise model, and the observed V_B deficits, may bear strong relevance to the hemodynamic conditions under which most activities are performed in daily life. Future research should be directed toward answering the questions of whether these deficits precede the onset of T2DM and/or T2DM complications in individual persons to identify the pathological and physiological bases for these deficits and to determine the reversibility of the pathophysiology.

GRANTS

This work was supported by UL1 RR024975 from the National Center for Research Resources, National Institutes of Health (NIH), and NIH/NIDDK R21 DK076880 from the National Institute of Diabetes and Digestive and Kidney Diseases, NIH.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).