INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

OVER TWO DECADES AGO, Roussos and Macklem (44) demonstrated that the diaphragm, the major inspiratory muscle of humans, can become fatigued. In 1988, the National Institutes of Health Respiratory Muscle Fatigue Workshop (41) concluded that patients with chronic obstructive pulmonary disease (COPD) represented a clinical cohort likely to develop diaphragmatic fatigue. However, a major determinant of whether severe COPD patients will develop diaphragmatic fatigue is the bioenergetic adaptations elicited by severe COPD in the human diaphragm. To provide insight into the adaptations elicited by severe COPD in the human diaphragm, many investigators have studied diaphragmatic adaptations elicited by animal models of severe COPD.

In seminal experiments, Supinski and Kelsen and colleagues (23, 24, 26, 50) and Farkas and Roussos (12-14) studied the response of the hamster diaphragm to elastase-induced emphysema; 6-18 mo after the intratracheal instillation of elastase, they noted twofold increases in total lung capacity and a 20% decrease in diaphragmatic fiber length [measured in situ at a lung volume approximating functional residual capacity (FRC)]. Although the major focus of these studies was on the manner in which the diaphragm adapts to this in situ shortening, Farkas and Roussos noted that strips from these emphysematous hamster diaphragms exhibited an increased fatigue resistance (FR) (12) as well as an increase in citrate synthase activity (12, 14). Subsequently, Lewis et al. (34, 35) also demonstrated that strips from the emphysematous hamster diaphragm exhibited an increased FR, and they noted that elastase-induced emphysema also elicited increases in activity of succinate dehydrogenase (SDH), a marker enzyme of mitochondrial oxidative activity, in both type I and II fibers of the hamster diaphragm. Importantly, these latter authors pointed out that these increases in mitochondrial oxidative activity might be causally related to the increases in fatigue resistance.

Our laboratory previously hypothesized that severe COPD elicits adaptations in the human diaphragm that increase FR (32, 33). At the present time, because of technical difficulties, a satisfactory technique does not exist for in vitro assessment of FR of human diaphragmatic muscle strips. However, previous work by Van der Laarse et al. (51) as well as by Watchko and Sieck (52) showed that the ratio of oxidative ATP-generating capacity to ATP utilization, as evaluated by the ratio of SDH to the ATPase activity of the myosin heavy chain (MHC) (i.e., SDH/mATPase, where mATPase is the Ca²⁺-activated myosin ATPase), is the best biochemical reflector of FR at the cellular level. Therefore, our hypothesis predicts that severe COPD will elicit an increase in the SDH/mATPase. To test this hypothesis, we carried out quantitative measurements of fiber type-specific SDH activity and mATPase activities on intraoperative diaphragm biopsies obtained from seven subjects with severe COPD and seven control subjects.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

Informed consent for diaphragm biopsies was obtained from each of the subjects, and our protocol was approved by the Institutional Review Boards of the Philadelphia Veterans Affairs Medical Center and the Hospital of the University of Pennsylvania.

Pulmonary Function Tests

Biopsies

General

Overview of methodology. On removal from the freezer, specimens were placed in a cryostat microtome (HM 505N, Microm) maintained at 20°C. Serial sections of known thickness (6 or 10 µm) were then cut, placed onto coverslips, and air dried (30 min) before being assayed by either qualitative or quantitative histochemical techniques. After these histochemical assays, images from the serial sections were digitized and subsequently analyzed for optical density (OD) and CSA, as previously described (7, 18, 27), by using public domain National Institutes of Health Image analysis software. Our image analysis system consisted of a Sony XC-77 charge-coupled device camera (Sony, Toyohashi, Japan) attached to an Olympus BH-2 microscope (Olympus, Tokyo, Japan) linked to a Macintosh Quadra-800 computer (Apple Computer, Cupertino, CA). Microdensitometric measurements of OD were carried out after calibration of the image gray scale level to the OD of a linear-step neutral density filter and correcting for uneven illumination of the camera field (3, 27).

Histochemical Techniques

Qualitative hATPase. For fiber typing, we used the techniques described by Brooke and Kaiser (5) and Guth and Samaha (19) as modified and carried out in our laboratory's previous work (7, 18, 27). Modifications to optimize the staining conditions for human skeletal muscles were made as follows. Briefly, we preincubated unfixed 10-µm sections for 7 min in a solution containing (in mM) 250 sodium acetate, 150 barbital sodium, and 100 HCl (pH 4.60). The sections were then rinsed for 1 min in a solution containing 18 mM CaCl₂ and 100 mM Tris (pH 7.8). After they were rinsed, sections were incubated in a reaction mixture containing 0.01 mM ATP (disodium salt) and 100 mM 2-amino-2-methyl-1-propanol (Sigma no. 221 buffer) (pH 9.4) for 45 min at 25°C. These sections then underwent the following sequential treatments: 1) two 4.5-min washes with 1% CaCl₂ (wt/vol); 2) placement in 2% CaCl₂ (wt/vol) for 3 min; 3) a quick rinse in distilled water; 4) two 1-min washes with 10 mM barbital sodium; 5) a 30-s rinse in distilled water; 6) placement in 1% ammonium sulfide for 45 s; 7) rinse for 4 min in running water; and 8) air drying and mounting in glycerin gelatin (pH 7.2).

Quantitative histochemical determination of SDH activity. The method of Blanco et al. (3) was used for measurements of SDH in individual fibers that had been typed by the hATPase technique, as previously described (7, 18, 27). The histochemical assays were carried out within the linear range of the reaction product formation and in complete darkness. Briefly, 10-µm sections were incubated in a reaction mixture containing 9.3 mM succinate and 5 mM EDTA in buffer A [i.e., 20 µM 1-methoxy-5-methylphenazine methyl sulfate, 1.2 mM nitro blue tetrazolium (NBT), 10 µM sodium azide, and 100 mM phosphate buffer, pH 7.6] for 10 min in the dark at 25°C.

1

1

Quantitative biochemical determination of calcium-activated mATPase activity. The method of Blanco and Sieck (4) was used for determination of mATPase in single fibers as previously described (6, 8, 18). Briefly, we incubated 6-µm serial sections in six solutions of different ATP concentrations (0, 0.5, 1, 2, 3, 4, and 5 mM). For each of the above-noted ATP concentrations, the incubations were carried out for 4 min at 25°C in a solution having the following composition: 65 mM calcium chloride (dihydrate), 16 mg/ml lead ammonium citrate-acetate complex, and 5 mM sodium azide prepared in 50 mM Tris-maleate buffer, pH 7.6 (4, 20). The use of incubations of different concentrations was necessary because, in contrast to the assays for SDH activity, each of the mATPase reactions was substrate limited. In this reaction, P_i is liberated by the ATPase and reacts with a lead ammonium citrate-acetate complex to form a lead-phosphate precipitate. This precipitate is converted to lead sulfide by reaction with sodium sulfide, and the change in OD per minute was measured at 570 nm. Subsequently, enzyme activity (expressed as mmol P_i · l¹ · min¹) was determined by use of the Lambert-Beer equation with a molar extinction coefficient for lead sulfide of 1,450 M¹ · cm¹. We then used the Michaelis-Menten model of enzyme kinetics to analyze the data, and the reciprocal of the y-intercept of the Lineweaver-Burk plot was taken as the maximum velocity of the mATPase.

Statistical Analysis

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Vital Statistics and Pulmonary Function Measurements

Fiber Type Proportions and Cross-Sectional Areas

View larger version (111K): [in this window] [in a new window]   Fig. 1.   Staining of serial diaphragmatic sections for fiber typing (A and B) and succinate dehydrogenase (SDH) activity (C and D). Labeled fibers in SDH sections are the same fibers labeled in the histochemical myosin ATPase (hATPase) sections. Calibration bars represent 50 µm. A and B: qualitative hATPase stain for fiber typing; representative sections are after preincubation at pH 4.60. Dark-stained fibers are type I fibers, light-stained fibers are type IIa fibers, and intermediate-staining fibers are type IIax fibers. C and D: SDH stain. The intensity of the stain is directly related to SDH activity; i.e., the darker the fiber, the greater the SDH activity.

Figure 2A indicates that control diaphragms contained the following proportions of types I, IIa, and IIax, respectively: 42 ± 2, 27 ± 3, and 31 ± 5%. In contrast, the figure indicates that COPD diaphragms contained the following proportions of types I, IIa, and IIax, respectively: 71 ± 5, 21 ± 3, and 8 ± 3%. These data indicate that COPD diaphragms exhibited a 29% increase (P = 0.0007) in the proportion of type I fibers and a 23% decrease (P = 0.0026) in the proportion of type IIax fibers; no statistically significant difference was noted between COPD and control diaphragms with respect to the proportion of type IIa fibers.

View larger version (26K): [in this window] [in a new window]   Fig. 2.   Comparisons of chronic obstructive pulmonary disease (COPD) and control diaphragms (n = 7 for each group) with respect to fiber type proportions, fiber type cross-sectional area (CSA), fiber type-specific SDH activity, and fiber type-specific quantitative Ca2+-activated myosin ATPase (mATPase) activity. Open bars represent measurements on control diaphragms, and solid bars indicate measurements on COPD diaphragms. Height of each bar indicates the mean for each group, and error bars show SE. Statistically significant differences between control and COPD fiber types: *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. A: fiber type proportions. COPD diaphragms exhibit a greater (P = 0.0007) percentage of type I fibers and a lesser (P = 0.0026) percentage of type IIax fibers. B: fiber type CSA. Type I fibers in COPD diaphragms have a lesser (P = 0.0009) CSA than control type I fibers. C: group effect whereby all COPD fiber types have greater (P = 0.0001) SDH activity than the corresponding fiber type in control diaphragms. D: type IIax fibers in COPD diaphragms have a lesser (P = 0.02) mATPase activity than IIax fibers in control diaphragms.

Figure 2B indicates that type I fibers in the COPD diaphragms had a 41% smaller CSA (P = 0.0009) than those in control diaphragms. No statistically significant differences were noted between COPD and control diaphragms with respect to the CSA of type IIa or IIax fibers.

To examine the combined effect of changes in fiber type proportions and CSA, we calculated the relative contribution of each fiber type to the CSA of the diaphragm (i.e., area fraction or %CSA) as previously described (7, 18, 27, 52). (Because both COPD and control tissue sections had the same thickness, these area fractions also represent volume fractions.) The area fraction of type I fibers was increased (48 ± 4 vs. 73 ± 4%, P = 0.001) in COPD diaphragms, whereas the area fraction of type IIax fibers was decreased (28 ± 6 vs. 6 ± 2%, P = 0.01). COPD elicited no change in the area fraction of type IIa fibers.

SDH Activity

Figure 2C indicates that COPD elicited a group effect whereby all COPD fiber types had higher (P < 0.0001) SDH activity than the corresponding fiber type in control diaphragms. In COPD diaphragms, SDH activity of fiber types I, IIa, and IIax was increased above control by 84, 114, and 130%, respectively.

mATPase Activity

SDH-to-mATPase Ratio

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Comparison of control and COPD diaphragms with respect to the fiber type-specific ratios of SDH activity to quantitative mATPase, i.e., SDH/mATPase. To normalize the variance of these ratios across groups and fiber types, the plot shows a logarithmic transformation of the data, and our statistical analysis was carried out on these log transformations. Our repeated-measures ANOVA showed a group-by-fiber type interaction; therefore, we used group t-tests to compare COPD and controls with respect to each of the fiber types. For each of the fiber types, COPD diaphragms had a greater SDH/mATPase than the control diaphragms. Statistically significant differences between control and COPD fiber types: **P < 0.01; ****P < 0.0001.

Additionally, we carried out repeated ANOVA on the nontransformed SDH/mATPase ratios. These analyses indicated that COPD elicited a group effect whereby all COPD fibers had higher (P = 0.03) SDH-to-mATPase ratios than the corresponding fiber type in control diaphragms. The SDH-to-mATPase ratios of COPD fibers expressed as percentage of control in types I, IIa, and IIax were 249, 295, and 357%, respectively.

Average Diaphragmatic Values for SDH, mATPase, and SDH/mATPase

View larger version (14K): [in this window] [in a new window]   Fig. 4.   Comparisons of control and COPD diaphragms with respect to average SDH activity, average mATPase activity, and average SDH/mATPase activity. As noted in the text, these enzyme activities and ratios represent values for a 1-compartment model (i.e., a single myofiber) representation of control and COPD diaphragms. See text for manner in which these activities and ratios were calculated. COPD diaphragms have higher activities than control for both average SDH activity (A) and average SDH/mATPase activity (C), whereas control diaphragms have a higher average mATPase activity (B). Statistically significant differences between control and COPD fiber types: *P < 0.05; ***P < 0.001.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Summary of New Findings

Critique of Methodology

Interpretation of Results

SDH activity. First, we comment on our observations that, in each of the fiber types, COPD fibers had a higher SDH activity than control fibers. We interpret this finding as indicating that severe COPD elicits an increased capacity to generate ATP through aerobic oxidative pathways in each of the fiber types in the human diaphragm.

mATPase activity. We noted that, for all fiber types, mATPase activity was less in COPD fibers than in control fibers, although this difference was only statistically significant for the IIax fibers. The precise mechanism underlying these adaptations is not known. However, we hypothesize that, for each of the diaphragmatic fiber types, COPD fibers have a lower MHC concentration than control fibers, and this decrease in MHC concentration accounts for the decrease in mATPase activity.

SDH/mATPase activity ratio. The relationship between fatigability and the SDH/mATPase warrants discussion. In the introduction, we noted that the work of Van der Laarse et al. (51) on single fibers from the lumbrical muscles of Xenopus laevis (a species of amphibian) and work of Watchko and Sieck (52) on the neonatal rat diaphragm indicate that a high degree of correlation was noted between the SDH/mATPase activity and FR. This concept is an intuitively attractive idea because it relates the FR of muscle to a balance between the quantitatively most important ATP-producing and -consuming reactions. However, neither of these papers provides a mechanistic interpretation of the manner in which an increase in SDH/mATPase increases FR. Therefore, we view the SDH/mATPase activity as being merely a biochemical indicator of FR at the myofiber level. Accordingly, we interpret the higher SDH/mATPase in our COPD diaphragms as suggesting that strips of COPD diaphragms would have a higher FR than those of control diaphragms.

Comparison of Our Human Work With That of Others

Observations on diaphragmatic adaptations elicited by COPD in humans. The literature contains a paucity of data on diaphragmatic adaptations of metabolic enzymes and fiber types associated with COPD. However, Sanchez and colleagues (45-47) in a series of papers showed that mild-to-moderate COPD [mean FEV₁ 60-70% of predicted normal (1)] elicited changes in neither mitochondrial oxidative enzyme activity nor fiber type proportions. Our unpublished observations also suggest that mild-to-moderate COPD does not elicit adaptations in either oxidative enzyme activities or fiber type proportions. However, Sanchez et al. (45, 47) reported that mild-to-moderate COPD elicited a decrease in CSA of both types I and II fibers. Additionally, Sanchez et al. (47) showed that decreases in their measure of CSA (i.e., least mean diameter) was positively correlated with the percent decrease in predicted body weight. Because Lewis et al. (35) and others (25) have demonstrated that nutritional depletion elicits decreases in CSA of both type I and type II fibers, the possibility exists that the decreases in CSA noted by Sanchez et al. were due to weight loss and not to some aspect of COPD per se.

Diaphragmatic Adaptations Elicited by Animal Models of COPD

Elastase-induced emphysema. Most work on experimental models of emphysema have utilized the response of the hamster diaphragm to elastase-induced emphysema (EIE). Using this model, Farkas and Roussos (14) demonstrated that EIE elicited a 20% increase in hamster diaphragmatic mitochondrial oxidative enzyme activity as assessed by biochemical determinations of citrate synthase. In subsequent work, Lewis et al. (34, 35) utilized quantitative histochemical techniques to demonstrate that EIE elicited a 28% increase in SDH activity of hamster diaphragmatic type I fibers whereas that of type II fibers increased 37%. At the present time, there appears to be general agreement that EIE in the hamster elicits increases in mitochondrial oxidative enzymes in all diaphragmatic fiber types, but the magnitudes of these increases are appreciably less than those noted in the diaphragms of patients with severe COPD. More importantly, both Farkas and Roussos (12) and Lewis et al. (34, 35) have demonstrated that EIE-induced increases in diaphragmatic mitochondrial oxidative activity are accompanied by increases in FR of muscle strips from these hamster diaphragms.

Relationship of Diaphragm Adaptations Noted in COPD to Endurance Training

Endurance exercise training of human limb muscles is usually associated with moderate increases in SDH activity as well as decreases in fiber CSA (9, 15, 16, 21, 37, 43); however, it is not associated with twofold increases in SDH activity and the fast-to-slow fiber type transformations that we observed in COPD diaphragms. Indeed, if the adaptations noted in the diaphragms of patients with severe COPD are related to this continuous increase in diaphragmatic energy expenditure (i.e., activity), this phenomenon would be unique for exercise-induced adaptations in mammalian skeletal muscle. However, some fundamental differences exist between the "diaphragm training" in severe COPD and limb muscle training protocols. Because the diaphragm is the only noncardiac striated muscle in humans that is continuously active, it carries out its training for 24 h/day, whereas limb muscle training protocols are limited to several hours per day. A second unique characteristic of the increased diaphragm work rate in severe COPD is that it probably occurs for at least several decades. Some exercise physiologists have suggested that endurance training may have the ability to elicit this fast-to-slow fiber type transformation, but "so far no lengthy human training study has unambiguously demonstrated such a shift" (1a). In closing, we respectfully suggest that the MHC and fiber type adaptations noted in our COPD diaphragms may represent an example of fast-to-slow fiber type transformations elicited by continuous long-term endurance training (i.e., several decades).