Basis for Increased Microtubules in Pressure-Hypertrophied Cardiocytes
Background We have shown on the levels of the sarcomere and the cardiocyte that a persistent increase in microtubule density accounts to a remarkable degree for the contractile dysfunction seen in pressure-overload right ventricular hypertrophy. In the present study, we have asked whether these linked phenotypic and contractile abnormalities are an immediate and direct effect of load input into the cardiocyte or instead a concomitant of hypertrophic growth in response to pressure overloading.
Methods and Results The feline right ventricle was pressure-overloaded by pulmonary artery banding. The quantity of microtubules was estimated from immunoblots and immunofluorescent micrographs, and their mechanical effects were assessed by measuring sarcomere motion during microtubule depolymerization. The biogenesis of microtubules was estimated from Northern and Western blot analyses of tubulin mRNAs and proteins. These measurements were made in control cats and in operated cats during and after the completion of right ventricular hypertrophy; the left ventricle from each heart served as a normally loaded same-animal control. We have shown that the alterations in microtubule density and sarcomere mechanics are not an immediate consequence of pressure overloading but instead appear in parallel with the load-induced increase in cardiac mass. Of potential mechanistic importance, both these changes and increases in tubulin poly A+ mRNA and protein coexist indefinitely after a new, higher steady state of right ventricular mass is reached.
Conclusions Because we find persistent increases both in microtubules and in their biosynthetic precursors in pressure-hypertrophied myocardium, the mechanisms for this cytoskeletal abnormality must be sought through studies of the control both of microtubule stability and of tubulin synthesis.
Recent work in this laboratory on load-induced cardiac hypertrophy has resulted in two novel findings.1 2 First, the contractile defects of the pressure-hypertrophied feline right ventricular cardiac muscle cell, or cardiocyte, are accounted for to a remarkable degree by an increased microtubule density, with normal contractile function being restored when the microtubules are depolymerized. Second, increased microtubules as well as nonpolymerized αβ-tubulin heterodimers are present as soon as hypertrophy is fully established and are persistent thereafter.
There are two questions of pathogenetic significance addressed in the present study that follow from these initial observations. First, is the increased microtubule density a direct result of load or instead a concomitant of the hypertrophy process? That is, there are both theoretical reasons3 and experimental observations4 that suggest that an extending force should rapidly shift the dynamic equilibrium between free and polymerized tubulin toward the polymerized form. Thus, does direct load input into the pressure-overloaded cardiocyte cause an immediate increase in microtubule density or do additional and/or alternative mechanisms cause a more gradual increase in microtubule density during the hypertrophic growth process? In this context, it should be noted that there must be specificity associated with pressure input per se, since an equivalent degree and duration of hypertrophy in response to a volume overload results neither in cardiocyte contractile dysfunction nor in microtubule changes.1 2 Second, since the cotranslational negative feedback control by both α- and β-tubulin of their own synthesis rates should downregulate α- and β-tubulin expression, why are the increases in microtubules, and especially in free tubulin, persistent? That is, it is established in other biological contexts that mRNA half-life and thus mRNA concentration for both α-tubulin and β-tubulin each decrease as the concentration of the respective protein in the cytoplasm increases.5 Given what we now know about free tubulin protein levels in cardiac hypertrophy, if the myocardial concentration of α-tubulin and β-tubulin mRNA were found to be increased during and after hypertrophic cardiac growth in response to a pressure overload, it would suggest that additional and/or alternative mechanisms must be responsible for the control of tubulin synthesis in the specific context of pressure-overload cardiac hypertrophy. Further, while the present study does not fully answer this second question, if tubulin mRNA and protein levels were found to be elevated concurrently, it would allow us in our future work to address this second question in terms of the control of tubulin mRNA expression in pressure-overload cardiac hypertrophy.
Long-term pressure-overload RV hypertrophy was induced by partially occluding the pulmonary artery with a 3.5-mm internal diameter band, just as we have described before.6 These cats were allowed to recover for 1 day to 6 months.
Short-term RV pressure overload was induced as we have described before by partial occlusion of the pulmonary artery with a specially fabricated balloon catheter.7 In this model, RV pressure was doubled for 4 hours, while systemic arterial pressure was unaltered.
Controls consisted of either normal cats or sham-operated cats submitted to thoracotomy and pericardiotomy without hemodynamic intervention. Since sham operation was without effect on any experimental variable, all control cats are considered throughout as a single group.
All operative procedures were carried out under full surgical anesthesia with meperidine (2.2 mg/kg IM), acepromazine maleate (0.25 mg/kg IM), and ketamine HCl (50 mg/kg IM). All procedures and the care of the cats were in accordance with institutional guidelines.
At the time of study, cats were lightly anesthetized with ketamine HCl (10 mg/kg IM). Echocardiographic data were obtained with a 7.5-MHz mechanical transducer (Hewlett-Packard; Sonos 1500). The heart was imaged from a parasternal window. Short-axis images were obtained in all cases. Right and left ventricular free wall thicknesses were measured from M-mode tracings with the ultrasonic beam directed at the chamber between the mitral valve echoes and the papillary muscle echoes. Thickness was measured from leading edge to leading edge. The same group of seven cats was studied before surgery and then at 4, 7, and 11 days and at 2, 3, and 8 weeks after pulmonary artery banding. Studies were recorded on high-resolution videotape for subsequent analysis.
At the time of terminal study, the control and long-term RV pressure-overloaded cats were anesthetized as above. Right heart and systemic arterial pressures were obtained as before,2 as was the arteriovenous difference in oxygen content, which was used as a measure of the adequacy of systemic perfusion.
For the short-term RV pressure-overload model, the cats were first anesthetized with meperidine (10 mg/kg IM) followed 15 minutes later by methohexital sodium (20 mg/kg IP), and followed after a further 15 minutes by α-chloralose (60 mg/kg IV). As we have described in detail before,7 RV pressure overload was created by passing a balloon-tipped catheter through the femoral vein into the pulmonary artery under fluoroscopic guidance and inflating the balloon in the pulmonary artery.
The methods that we use to obtain reproducible yields of calcium-tolerant, quiescent adult feline cardiocytes from the RV and LV have been described previously.8 9 10 11 After obtaining the RV and LV weights,11 the cardiocytes were isolated and then maintained for 1 hour at 37°C in collagenase-free 2.5 mmol/L Ca2+ buffer at pH 7.4 before contractile function was defined.
Cardiocyte length, width, and surface area were obtained as before12 from digitized photographs both of each living cell in which contractile function was characterized and of a random sampling of fixed, rod-shaped cells from the RV and LV of each cat.
The use of laser diffraction techniques for measuring sarcomere motion in isolated cardiocytes is well established.10 11 12 13 14 An outline of our method11 is as follows. An aliquot of isolated cardiocytes was added to 4 mL of the 2.5 mmol/L Ca2+ buffer in a well affixed to a glass slide. The cardiocytes came to rest on the bottom of this chamber, which was placed on the stage of an inverted microscope. The buffer was kept at 37±0.1°C by a thermostated heating stage. The cardiocytes were stimulated to contract between platinum wire electrodes by 0.25-Hz, 100-μA DC pulses of alternating polarity. When after 10 to 15 contractions the extent of shortening was stable, 10 contractions were sampled and averaged to yield a final profile of sarcomere length and velocity versus time during contraction. Changes in sarcomere length were measured from movement of the first-order diffraction pattern cast by a substage laser light passing through the sarcomeres of a given cardiocyte onto diametrically opposed optical sensors situated above the microscope stage. Each sensor was composed of a linear array of 256 photodiodes, which was interrogated at a frequency of 1 kHz. The distance between the first-order diffraction patterns at every millisecond was calculated by and stored in a computer.
Tubulin Protein and mRNA
In addition to documenting functional changes in RV cardiocytes after pulmonary artery banding, we wished to investigate the role of microtubules in these changes as well as the basis for any increases observed in free and polymerized tubulin. We therefore quantified free and polymerized tubulin protein levels as well as tubulin mRNA levels in RV and LV myocardium both from control cats and from short-term and long-term RV pressure-overloaded cats.
After completion of the hemodynamic studies, the cats were heparinized (1000 units IV) and placed on oxygen. A midline thoracotomy was performed, the pericardium was opened, and the heart was rapidly removed and weighed. The aorta was then cannulated, and the coronary arteries were gently flushed with microtubule stabilizing buffer.15 Two 0.25-g specimens were excised from both the RV and the LV free walls for tubulin protein isolation. The remaining myocardium was immediately immersed in ice-cold saline, after which additional tissue from the RV and LV free walls was excised and flash-frozen in liquid nitrogen for RNA isolation.
Immunoblots. For the immunoblot analysis, fresh 0.25-g specimens from the RV and LV of each cat were homogenized in 5 mL of microtubule stabilizing buffer15 and centrifuged at 100 000g, 25°C, for 15 minutes. The supernatants were saved as the free tubulin fractions, and the pellets were resuspended at 0°C in 4 mL of microtubule depolymerization buffer15 ; after 1 hour at 0°C they were centrifuged at 100 000g, 4°C, for 15 minutes, and the supernatants were saved as the polymerized tubulin fractions. Protease inhibitors16 were used throughout. For the subsequent 8% to 16% gradient SDS-PAGE, equal proportions of the free and polymerized samples were loaded onto the two lanes for each ventricle, and an equal amount of protein as determined by a bicinchoninic acid assay (Pierce) was loaded for the RV and LV samples. The samples were then transferred to polyvinylidene difluoride membranes (35 V, 75 minutes) and probed with a 1:500 dilution of a monoclonal antibody to either α-tubulin (DM1A; Amersham) or β-tubulin (DM1B; Amersham). The bound antibody was visualized with the use of a horseradish peroxidase–conjugated secondary antibody (Vector) and enhanced chemiluminescence (Amersham). In all cases, a single band at 55 kD with the same mobility as concurrently run bovine brain β-tubulin was detected. In addition, 0.25-g samples from the same RV and LV specimens were homogenized in depolymerization buffer15 to isolate total tubulin in the same RV and LV samples; these were run on the same gel as the free and polymerized fractions. Densitometric quantification of the immunoblots was carried out as described before.2
Analysis of the free tubulin fraction. The composition of the free tubulin fractions assayed in the immunoblots, and any possible contamination by higher-molecular-weight tubulin-containing polymers, were assessed by gel filtration analysis of the same RV and LV free tubulin fractions used for the immunoblots. For this purpose, 200-μL samples of the free tubulin fractions, with a protein concentration 2 mg/mL in microtubule stabilizing buffer15 containing 50% glycerol, were fractionated by gel filtration on a Superose 6 column (Pharmacia) after we equilibrated the column overnight at 4°C in the same buffer without glycerol. Protein samples collected as 1-mL fractions were concentrated to 50 μL with the use of an Amicon concentrator. The samples were then mixed with an equal volume of Laemmli sample buffer, boiled for 5 minutes, and analyzed by SDS-PAGE followed by immunoblotting as above with the same monoclonal antibody to β-tubulin.
Indirect immunofluorescence micrographs. For visualization of the appearance and density of the cardiocyte microtubule network, freshly isolated8 11 RV and LV cardiocytes were sedimented onto laminin-coated coverslips at 1g for 45 minutes, permeabilized for 1 minute by 1% Triton X-100 in stabilization buffer,17 washed twice in the same buffer, and fixed for 10 minutes with 3.7% formaldehyde. After blocking was done with 10% horse serum in 0.1 mol/L glycine, the cells were incubated overnight at 4°C with a 1:1000 dilution of the same antibody to β-tubulin used for the immunoblots, followed by a fluorescein-conjugated secondary antibody (Vector). They were mounted with 1% triethylenediamine and 50% glycerol in phosphate-buffered saline, and 0.7-μm optical sections were acquired by confocal laser microscopy (LSM GB-200; Olympus).
Quantification of total RNA. Total RNA was extracted by the method of Chirgwin et al.18 The RNA samples were dissolved in 500 μL of DEPC-treated water, quantified by spectrophotometric analysis at a 260/280-nm extinction coefficient, and both to assess RNA quality and to ensure equal loading of RV and LV RNA for subsequent Northern blots, 3- to 5-μg RNA samples were stained with ethidium bromide and run on 1% agarose check gels. We then electrophoresed 5- to 7-μg RNA samples on denaturing 2% formaldehyde/1% agarose gels, followed by 1.5 hours of pressure-driven blotting to a nylon membrane (Hybond-N; Amersham). The RNA was then immobilized on the nylon membrane by UV cross-linking (Stratalinker; Stratagene). The nylon membrane was prehybridized for 4 hours at 42°C in a solution containing 50% (vol/vol) deionized formamide, 0.2% (wt/vol) Ficoll, 0.02% (wt/vol) polyvinylpyrrolidone, 5× SSC, 10 mmol/L MOPS, pH 7.0, 2 mmol/L EDTA, 100 μg/mL denatured salmon sperm DNA, and 0.2% (wt/vol) SDS. The membrane was hybridized for 16 hours at 42°C in a solution containing 32P-radiolabeled probe (0.5 to 1.0×106 cpm/mL). The Northern blots were washed three times in 2× SSC, 0.1% SDS for 1.5 hours at 42°C, followed by a wash in 0.2× SSC, 0.1% SDS for 1 hour at 42°C, and then processed for autoradiography. As we have reported before,19 20 to quantify the amount of total RNA per lane, the blots were next reprobed with a clone of 28S rDNA that had been 32P-radiolabeled by nick translation (Amersham). The autoradiographic signals for tubulin mRNA were then normalized for RNA loading by using the autoradiographic signal for 28S rRNA in the same lane.
Quantification of poly A+ RNA by HPLC. The goal here was to allow the loading of equal amounts of poly A+ RNA on each lane of Northern and slot blots, such that the abundance of tubulin mRNA in the poly A+ RNA pool could be determined. Poly A+ RNA was first extracted using an mRNA isolation system (FastTrack; Invitrogen). To determine the concentration of poly A+ RNA in each sample, aliquots of each sample were hydrolyzed, and the amount of 3′-UMP was measured via HPLC. This method, which we have applied before to the measurement of rRNA synthesis,21 is illustrated in Fig 1⇓. The 3′-UMP was purified in a buffer containing 30 mmol/L H3PO4, 3.2 mmol/L MgSO4, and triethylamine, pH 6.5, using a linear gradient of 160 μmol/L MgSO4 per minute. UV absorbance was monitored at 254 nm, and the area under the 3′-UMP peak was integrated and compared with known standards (Baseline 810; Waters). The concentration of poly A+ RNA in each sample was then calculated. For the poly A+ Northern blots, equal amounts of poly A+ RNA were loaded on each lane, electrophoresed, and blotted onto a nylon membrane, just as was done for total RNA Northern blots. For the poly A+ slot blots, the same quantitative loading method was used to ensure that there were equal amounts of poly A+ RNA in each lane. The Northern blots and slot blots were hybridized and washed under the same conditions as those used for the total RNA Northern blots described above. Contamination by nonpolyadenylated RNAs was evaluated by hybridizing the Northern blots and the slot blots with a 32P-labeled rRNA probe for the 28S ribosomal subunit; any membrane in which hybidization with the rRNA probe was either substantial or not equivalent in all lanes was discarded. To ensure that poly A+ RNA was accurately quantified by Northern blots and that this quantification remained linear over a known range, specific amounts of poly A+RNA were electrophoresed, blotted onto a nylon membrane, and then quantified by densitometry.
Probe generation. A cDNA library was prepared from pooled cardiocytes isolated from several normal cats (Uni-ZAP XR; Stratagene). Feline-specific α-tubulin and β-tubulin cDNA clones were screened from the library with the use of human fibroblast α-tubulin and β-tubulin cDNAs as probes. The feline cDNA clones that were obtained were then sequenced by the chain-termination method to ensure their specificity. Selectivity between α-tubulin and β-tubulin was established by checking for the presence of cross-hybridization on Southern blots. Probes for feline-specific and α-tubulin– and β-tubulin–specific α-[32P]d-CTP–labeled cDNA were obtained by polymerase chain reaction amplification of each clone.
Quantification of Northern and slot blot signals by densitometry. The relative levels of mRNA signals were quantified from autoradiograms by optical densitometry; for this purpose, we used the same techniques that we had applied to a similar quantitative analysis of immunoblots in our earlier work.2 Linearity of the optical signals was determined over a specified range of known mRNA standards. Comparisons were limited to signals processed at the same time from a single blot. In this manner, variations in signal intensity caused by differing conditions or disparate probe specific activities were minimized. In all instances, two separate background readings were subtracted for each sample imaged. The densitometry readings were compared with known standards to demonstrate that the autoradiographic signals were proportional to the amount of mRNA examined and were linear across a known concentration range.
The mean value and the standard error of the mean are shown for each group of data. For the data in the tables, both nonparametric22 and parametric23 analyses were used. Where stated, group means were first compared by a one-way or two-way ANOVA, and if a difference was found, then each experimental mean was compared with that of the control and any other groups noted by the appropriate post hoc test23 as individually specified. All statistical evaluations were preceded by a one-sample Kolmogorov-Smirnov test to ascertain that the data comprising each of these values were normally distributed.
Characteristics of the Experimental Models
The goal of this study was to examine the interaction of cardiocyte mechanics and microtubules from the earliest stage of RV pressure-overload hypertrophy induction through the completion of the hypertrophic growth phase and then into an extended period of stable hypertrophy. The 4-hour balloon catheter cats were used as the nonsurvival model for early hypertrophy induction. Survival pulmonary artery band surgery was used as a model for the later time points of 1 day, 2 days, 1 week, 2 weeks, and 6 months of RV pressure overloading. The data in Fig 2⇓ formed the basis for the selection of these time points; they show that for the surgical pulmonary artery band model, RV hypertrophy as estimated from echocardiographic RV wall thickness is complete at 2 weeks and remains stable thereafter, a result consistent with earlier gravimetric data for this model.24
The major features of the surgical model used in this study are summarized in Table 1⇓. At 1 day after surgery, RV systolic pressure was almost doubled and showed a slight progressive increase thereafter. The ratios of RV weight to body weight and to tibial length were increased significantly and comparably by 1 week after RV pressure overloading but did not increase significantly thereafter. Thus, the data in Fig 2⇑, obtained from a group of 7 cats studied sequentially, show that a new steady state for an index of RV mass is reached at ≈2 weeks after a step increase in afterload, and the terminal study data in Table 1⇓ tend to bear this out. Body weight was similar in each group, and the ratio of LV weight to body weight did not differ among the groups, further precluding any effect of postoperative changes in body weight. In no group was there evidence for right heart failure in terms of either the presence of ascites and pleural effusion in any cat at the time of study or increases in AV-O2 difference, RV end-diastolic pressure, or the ratio of liver to body weight.
The major features of the RV and LV cardiocytes from the surgical model used in this study are summarized in Table 2⇓. By 1 week after RV pressure overloading, the width and surface area of the RV cells were each significantly greater than those of RV cells from control cats; by 2 weeks after RV overloading, cardiocyte length was significantly increased as well. The cells used for mechanical studies excluded trypan blue and were quiescent in 2.5 mmol/L Ca2+. Their average resting sarcomere length, which did not differ among groups, was 1.94 μmol/L and in no case was less than 1.85 μmol/L. These values are the same as those both in explanted superfused myocardium at slack length25 and in perfusion-fixed unloaded myocardium when the fixative is isosmotic and contracture is avoided.26
Cardiocyte Mechanics: Effects of Colchicine
We have shown that microtubule depolymerization, either by low temperature or by colchicine, restores the initially abnormal contractile performance of the pressure-hypertrophied feline RV cardiocyte to normal after hypertrophy is complete.1 2 The data in Fig 3⇓ first show under basal conditions before adding colchicine (time 0) that the extent and velocity of sarcomere shortening were identical in RV and LV cardiocytes from control cats and remained equivalent at 2 days after pulmonary artery banding. However, by 1 week after banding, when RV hypertrophy is ≈50% complete (Fig 2⇑), both of these indices of contractile function were quite reduced in cardiocytes from the pressure-overloaded RV, and after 2 weeks of pressure overloading they were each reduced to a level comparable to what we observed at 2 weeks after banding in our earlier study.1 2
Fig 3⇑ also shows the effects of microtubule depolymerization by colchicine on sarcomere mechanics for RV and same-animal normally loaded control LV cardiocytes during the development of RV pressure-overload hypertrophy. The four left panels of Fig 3⇑ show the maximum extent of sarcomere shortening, defined as initial sarcomere length minus minimum sarcomere length, at the indicated times after the addition of 10−6 mol/L colchicine to RV and LV cardiocytes from the same cats. All cells were sampled sequentially at the indicated times after drug exposure. The four right panels of Fig 3⇑ show the maximum velocity of sarcomere shortening, defined as the maximum positive rate of length change, for the same contractions summarized in the corresponding left panels. For both the control cats (Fig 3A⇑) and the 2-day pressure-overloaded cats (Fig 3B⇑), the contractile function of cardiocytes from the two ventricles was identical. Furthermore, colchicine caused only a small, statistically insignificant improvement in contractile function for both RV and LV cardiocytes during the first 30 minutes of drug exposure. For the 1-week (Fig 3C⇑) and 2-week (Fig 3D⇑) pressure-overloaded cats, the initial differences between RV and LV cardiocytes were no longer statistically significant 30 minutes after the addition of colchicine, and for the RV cardiocytes after 30 minutes there was a significant difference from their initial values for both sarcomere shortening and sarcomere shortening velocity. Thus, exposure of hypertrophied RV cardiocytes from these cats to colchicine essentially normalized the initially quite abnormal contractile function. As with baseline sarcomere mechanics, the response to colchicine at the 2-week time point was closely comparable to what we observed at 2 weeks after pulmonary artery banding in our earlier study.1 2 Furthermore, our earlier data1 2 demonstrate that these changes are persistent, and perhaps even progressive, as late as 6 months after feline RV pressure overloading.
Fig 4⇓ comprises immunofluorescence confocal micrographs of cardiocyte microtubules, with use of the same β-tubulin antibody as that used for the immunoblots, in RV cells from each of the four categories of cats shown in Fig 3⇑, as well as from a 6-month RV pressure-overloaded cat. The micrographic density of the microtubule network is alike for the control (Fig 4A⇓) and 2-day pressure-overloaded (Fig 4B⇓) RV cardiocytes. In comparison, however, an increased microtubule density is readily apparent in the 1-week (Fig 4C⇓), 2-week (Fig 4D⇓), and 6-month (Fig 4E⇓) pressure-overload hypertrophied RV cardiocytes.
Myocardial Free and Polymerized Tubulin
The top panel of Fig 5⇓ shows immunoblots of free (lanes 1 and 3) and polymerized (lanes 2 and 4) β-tubulin from paired RV and LV samples from control cats and from RV pressure-overloaded cats at the indicated times after surgery. It is important to note that in each of these five immunoblots the RV and LV samples from the same heart were run together, such that visual comparisons within that blot are valid; however, comparison of one immunoblot with another requires densitometric analysis using concurrently run β-tubulin standards, as was done in generating the data shown in the lower panel of the figure. Thus, the bottom panel of Fig 5⇓ provides summary data from these and additional blots for these time points, and for two earlier pressure-overload time points, in terms of the ratios of RV/LV free tubulin and RV/LV polymerized tubulin. No increase in the RV concentration of either protein fraction was detected through 2 days of pressure overloading, but there was a doubling of the RV concentration of both fractions thereafter.
Composition of the Free Tubulin Fraction
To be sure that the apparent increase in free tubulin seen in our immunoblots was authentic and to unambiguously define the relationship of free αβ-tubulin heterodimer concentration both to microtubule density and to the concentration of α-tubulin and β-tubulin mRNAs, it was important to ascertain that the free tubulin fractions assayed in our immunoblots consisted of αβ-tubulin heterodimers alone, without significant contamination either by microtubules or by microtubule fragments. Fig 6⇓ shows this to be the case both for the pressure-overloaded RV and for the normally loaded LV and both at 1 week after pulmonary artery banding during the active hypertrophic growth phase and at 14 weeks after pulmonary artery banding when hypertrophic growth is long-since complete (Fig 2⇑). The top panel of Fig 6⇓ shows immunoblots of free (lanes 1 and 3) and polymerized (lanes 2 and 4) α-tubulin and β-tubulin from paired RV and LV samples from two RV pressure-overloaded cats at the indicated times after surgery. There is comparable upregulation on the protein level of both α-tubulin and β-tubulin during and after the induction of pressure-overload RV hypertrophy: both free and polymerized α-tubulin and β-tubulin are greater in the RV than in the LV at both 1 week and 14 weeks of RV pressure overloading. Lanes 5 and 6 show total tubulin: both α-tubulin and β-tubulin are greater in the RV than in the LV at 1 week and at 14 weeks of RV pressure overloading. The gel filtration analysis in the bottom panel of this figure demonstrates that in the RV and LV myocardium of the same two cats used for the immunoblots shown above, the fractions containing β-tubulin are centered within the range of 44 to 158 kD, with no evidence for tubulin-containing protein species larger than the 110 kD αβ-tubulin heterodimer.
Relationship of Cardiocyte Microtubules to Cardiocyte Mechanics
Fig 7⇓ summarizes the essential features of the relationship of cardiocyte microtubules to cardiocyte mechanics during the development of RV pressure-overload hypertrophy. In Fig 7A⇓, it is seen that there are parallel decrements in both the velocity and the extent of sarcomere shortening in RV cardiocytes as hypertrophy progresses (Fig 2⇑) and the concentration of RV microtubules increases (Fig 5⇑). The attribution of this change in sarcomere mechanics to an increase in microtubule density is validated in Fig 7B⇓. That is, as the concentration of microtubules increases during RV hypertrophy, the specific ameliorative effect of microtubule depolymerization on the velocity and extent of sarcomere shortening in RV cardiocytes increases in a parallel fashion.
Myocardial Tubulin mRNA
The top panel of Fig 8⇓ shows Northern blots loaded on the basis of total RNA on the left and on the basis of poly A+ RNA on the right. Summary data for these two types of Northern blots are given in the lower panel of this figure. The summary data for the total RNA Northern blots show that while the amount of β-tubulin mRNA is the same in the RV and LV of control cats, it is clearly upregulated in the RV after 1 day of pressure overloading and markedly upregulated after 2 days. Thereafter, the level of β-tubulin mRNA in the two ventricles once again apparently becomes equivalent as the RV hypertrophy process reaches completion (Fig 2⇑).
There are several reasons, however, that conclusions about changes in mRNA levels based on standard Northern blots during a dynamic growth process such as cardiac hypertrophy may be uncertain. The first and more obvious reason is that RNA quantification by UV spectrophotometric absorbance is quite imprecise. The second and more basic reason is that the biologically relevant RNA species in question, the translatable pool of poly A+ RNA, is only a very small fraction of the total RNA pool that consists largely of rRNA, which we have shown increases markedly during cardiac hypertrophy.7 That is, of total cellular RNA, ≈1% to 2% consists of poly A+ RNA, and ≈90% consists of rRNA; clearly, substantive changes in the latter pool could easily alter standard Northern blots loaded on the basis of total RNA so as to obscure changes of interest in the translation-competent pool of poly A+ RNA. The usage of poly A+ Northern and slot blots, with quantification of gel loading of poly A+ RNA by HPLC (Fig 1⇑), addresses both of these concerns. Summary data for the poly A+ Northern blots in Fig 8⇑ again show that there is quite substantial early upregulation of β-tubulin mRNA. However, in contrast to the standard Northern blots, they show that upregulation of this pool of β-tubulin mRNA is persistent well after the hypertrophy process is complete (Fig 2⇑).
Because it was important for the purposes of this study to be sure that there is indeed a persistent increase in tubulin poly A+ RNA, we also examined this β-tubulin mRNA pool by the more quantitative technique of slot blotting. That is, total RNA and poly A+ RNA Northern blots can provide only semiquantitative data because of the unavoidable variability in gel loading and transfer. However, in all cases herein they displayed a single band of the appropriate size, thus validating the usage of slot blots. Fig 9⇓ shows examples of concurrently run β-tubulin slot blots for 3 cats at each of four time points spanning the RV hypertrophy process. Densitometric scans of these and other RV and LV slot blots prepared from cats at 4 hours, 1 day, and 1 week after pulmonary artery banding showed a fourfold increase in the ratio of RV/LV β-tubulin poly A+ RNA at 1 day after banding; this ratio reached a persistent steady state twofold RV/LV value at 1 week after banding. Thus, the major findings reached by this technique confirm the findings from the poly A+ Northern blots: there is an early increase in the translatable β-tubulin mRNA pool size in the hypertrophying RV, and this increased pool size persists well after the RV hypertrophy process is complete.
Finally, two further concerns needed to be addressed. First, for the increased tubulin mRNA to have significance for the greater microtubule density found in the hypertrophied RV, there would have to be coordinate upregulation of the mRNA pools of both α-tubulin and β-tubulin, since their protein products would be expected to be present in a 1:1 stoichiometric ratio in order to form the αβ-tubulin heterodimers from which microtubules are assembled. Fig 10A⇓, which gives summary data from total RNA Northern blots for α-tubulin and β-tubulin mRNA levels, shows this to be the case throughout the RV hypertrophy process. Second, the myocardium is a complex tissue composed of a number of cellular elements; it was therefore necessary to establish that tissue levels of α-tubulin and β-tubulin mRNA and their changes during hypertrophy are reflective of levels in the cardiocyte. Fig 10B⇓, which gives summary data from total RNA Northern blots for both the α-tubulin and β-tubulin mRNA levels in cardiocytes, shows this also to be the case.
This study was designed to answer two mechanistic questions about the increased microtubule density that we found in pressure-hypertrophied cardiocytes. The first of these questions was whether this increased microtubule density and the associated contractile defects are a direct result of stress loading or instead an indirect result of the hypertrophic growth resulting from stress as opposed to strain loading.1 2 That is, does exerting an extending force on the cardiocyte constitute an immediate cause for the increased microtubule density seen in this cell when it is pressure overloaded? The data in this study suggest that while this mechanism may play a contributory role, other factors acting in a less direct manner during and after hypertrophy must be of primary importance. The second of these questions was whether the persistent increase in free and polymerized tubulin seen in the pressure-hypertrophied cardiocyte1 2 represents an abrogation of the negative feedback control mechanism by which both α-tubulin and β-tubulin autoregulate their own rate of synthesis. The data in this study suggest that additional and/or alternative mechanisms for the regulation of tubulin synthesis must be operative in the pressure-hypertrophied cardiocyte.
The first of these two questions had its genesis in our original hypothesis1 that stress as opposed to strain loading of the cardiocyte might shift the equilibrium between free and polymerized tubulin toward the polymerized form. Were such a direct mechanical input into microtubule polymerization in the cardiocyte the sole mechanism responsible for increased microtubule density during pressure-overload cardiac hypertrophy, one might expect the increased microtubules to become apparent within minutes to hours. That is, the dynamic instability of microtubules in interphase cells results in a very rapid interchange between free tubulin heterodimers and microtubules, with a half-time for tubulin incorporation into the microtubules of cultured fibroblasts being less than 30 minutes.27 In contrast to these very rapid kinetics, Fig 2⇑ shows that the hypertrophy response to a step increase in afterload requires about 2 weeks to reach completion. Thus, a solely load-related shift in the equilibrium between free tubulin heterodimers and polymerized microtubules toward the polymerized form might be expected to result in increased microtubule density in the pressure-overloaded cardiocyte before the hypertrophy process is well under way. The summary data in Fig 11⇓ show this not to be the case. Instead, there is a parallel increase both in cardiac mass and in microtubule protein throughout the course of pressure-overload RV hypertrophy.
This finding does not exclude the possibility that load modulation of the set point of the tubulin-microtubule equilibrium may be partially responsible for the appearance and persistence of increased microtubule density in pressure-overload cardiac hypertrophy. It does, however, suggest that other mechanisms are likely to be operative. Several possibilities would seem to be attractive, and by no means mutually exclusive, candidate mechanisms. The first of these is posttranslational tyrosine phosphorylation of tubulin itself, which appears to favor microtubule formation.28 The second candidate mechanism would be an increased quantity and/or an altered phosphorylation state of the nonmotor fibrous MAPs. Here the most attractive candidate would be MAP 4, since this is the predominant MAP of striated muscle.29 This protein, in common with other MAPs, binds to and stabilizes interphase microtubules; its activity in this regard is strongly influenced by modification of its phosphorylation state by protein kinases and phosphatases.30 The third candidate mechanism would be an alteration in tubulin isoform expression during cardiac hypertrophy. While all isoforms of α-tubulin and β-tubulin are probably functionally equivalent in terms of microtubule assembly in most tissues, the isoform-variable carboxy-terminal domain of these isoforms may well affect MAP binding.31 That is, MAPs, including MAP 4, bind to the carboxy-terminal domain of both α- and β-tubulin, and the various tubulin isoforms have differential affinities for MAPs.32 33 A changed pattern of tubulin isoform expression during pressure-overload hypertrophy therefore could have an indirect effect on microtubule stability and thus density via differential MAP 4 binding. In summary, the negative finding shown in Fig 11⇑ with respect to a direct load causation of increased microtubule density during cardiac hypertrophy makes exploration of these three potential alternative mechanisms an important goal.
The second of these two questions had its genesis in our original observation1 of a persistent increase not only in microtubules but also in unpolymerized tubulin heterodimers in the pressure-hypertrophied cardiocyte, even after hypertrophic growth was complete and a new steady state of cardiac mass had been achieved. The first reason that this observation was of interest, especially that having to do with the persistent increase in free tubulin heterodimers, was that after hypertrophy was complete and cardiocyte volume had become fixed at a new, higher value, the dynamic instability model of microtubule behavior would predict that for a given number of microtubule nucleation sites, an increased concentration of free tubulin heterodimers should result in increased microtubule density.34 35 The second and more basic reason that the persistent increase in tubulin heterodimers seen here was of interest was that it suggested that the accepted mechanism for the regulation of tubulin synthesis was not operative either during the dynamic phase of pressure-overload hypertrophy or after this process was complete and a new steady state had been reached. The currently accepted mechanism for the autoregulation of tubulin synthesis had its origin in the initial observation that drug-induced depolymerization of microtubules in cultured cells, with a parallel increase in the concentration of free tubulin heterodimers, caused a rapid degradation of tubulin mRNA.36 After the observation that tubulin message stability was indeed regulated by the concentration of the protein product,37 it was shown that this mechanism was operative in enucleated cells, such that the control point was at the level of the cytoplasmic ribosomes.38 For β-tubulin, this was then shown to be a cotranslational process in which the mRNA is targeted as a substrate for destabilization via recognition of the initially translated amino terminal tetrapeptide as it exits the ribosome39 and that active translation must be under way for this targeted mRNA degradation to occur.40 For α-tubulin, despite the fact that the mRNA level varies inversely with that of the free protein, a different cotranslational mechanism for regulating message stability is operative.41
In contrast to this beautifully delineated mechanism, the data in Figs 2 through 10⇑⇑⇑⇑⇑⇑⇑⇑⇑ show quite clearly that at all times during and after the completion of pressure-overload cardiac hypertrophy there are parallel increases in cardiac mass and in both tubulin protein and tubulin mRNA, including the translation-competent poly A+ form. Further, the data in Fig 6⇑ show that the free tubulin fraction as assessed by gel filtration analysis is indeed composed almost entirely of tubulin heterodimers, both in the pressure-overloaded RV and in the normally loaded LV, and whether during hypertrophy or after hypertrophy is complete, such that the appropriate molecular species to exert autoregulatory control of α-tubulin and β-tubulin mRNA stability is present in the cardiocyte.
Nonetheless, these data do not prove that the autoregulatory control of tubulin synthesis seen in response to acute interventions in cultured cells is not operative in the long-term in vivo process of cardiac hypertrophy. At least in cultured cells, microtubule depolymerization in response to decreased external load42 as opposed to drug treatment43 results in a relatively long-term increase in the concentration of tubulin heterodimers because under this circumstance, the rate of tubulin protein degradation is greatly decreased. This control of tubulin concentration is superimposed upon rather than alternative to autoregulatory control of tubulin mRNA stability by tubulin protein concentration. In the context of cardiac hypertrophy, this thermodynamic input into the balance between tubulin monomer and polymer concentrations is a particularly attractive possibility, such that in our current work we are attempting to define tubulin mRNA stability in the pressure-hypertrophied cardiocyte.
While this study has answered to a large extent the two questions initially posed, there are major residual questions regarding the control of microtubule density and the control of tubulin synthesis in the pressure-hypertrophied cardiocyte. Insight into these closely interrelated problems will be essential if we are to understand the mechanisms responsible for the increased microtubule density and the associated contractile defects in pressure-overload cardiac hypertrophy.
Selected Abbreviations and Acronyms
|HPLC||=||high-performance liquid chromatography|
|LV||=||left ventricle (ventricular)|
|RV||=||right ventricle (ventricular)|
This study was supported by grants HL-37196 and HL-48788 from the National Heart, Lung, and Blood Institute and by research funds from the Department of Veterans Affairs. The authors thank Sebette Hamill, Heather Downs, and Mary Barnes for excellent technical assistance.
Reprint requests to George Cooper IV, MD, Cardiology Section, VA Medical Center, 109 Bee St, Charleston, SC 29401-5799. E-mail firstname.lastname@example.org.
- Received September 7, 1995.
- Revision received October 18, 1995.
- Accepted October 23, 1995.
- Copyright © 1996 by American Heart Association
Tsutsui H, Ishihara K, Cooper G. Cytoskeletal role in the contractile dysfunction of hypertrophied myocardium. Science. 1993;260:682-687.
Tsutsui H, Tagawa H, Kent RL, McCollam PL, Ishihara K, Nagatsu M, Cooper G. Role of microtubules in contractile dysfunction of hypertrophied cardiocytes. Circulation. 1994;90:533-555.
Joshi HC, Chu D, Buxbaum RE, Heidemann SR. Tension and compression in the cytoskeleton of PC 12 neurites. J Cell Biol. 1985;101:697-705.
Cooper G, Satava RM, Harrison CE, Coleman HN. Mechanism for the abnormal energetics of pressure-induced hypertrophy of cat myocardium. Circ Res. 1973;33:213-223.
Cooper G, Mercer WE, Hoober JK, Gordon PR, Kent RL, Lauva IK, Marino TA. Load regulation of the properties of adult feline cardiocytes: the role of substrate adhesion. Circ Res. 1986;58:692-705.
Mann DL, Kent RL, Cooper G. Load regulation of the properties of adult feline cardiocytes: growth induction by cellular deformation. Circ Res. 1989;64:1079-1090.
Kent RL, Mann DL, Urabe Y, Hisano R, Hewett KW, Loughnane M, Cooper G. Contractile function of isolated feline cardiocytes in response to viscous loading. Am J Physiol. 1989;257:H1717-H1727.
Mann DL, Urabe Y, Kent RL, Vinciguerra S, Cooper G. Cellular versus myocardial basis for the contractile dysfunction of hypertrophied myocardium. Circ Res. 1991;68:402-415.
Urabe Y, Mann DL, Kent RL, Nakano K, Tomanek RJ, Carabello BA, Cooper G. Cellular and ventricular contractile dysfunction in experimental canine mitral regurgitation. Circ Res. 1992;70:131-147.
Krueger JW, Forletti D, Wittenberg BA. Uniform sarcomere shortening behavior in isolated cardiac muscle cells. J Gen Physiol. 1980;76:587-607.
Drubin DG, Feinstein SC, Shooter EM, Kirschner MW. Nerve growth factor-induced neurite outgrowth in PC 12 cells involves the coordinate induction of microtubule assembly and assembly-promoting factors. J Cell Biol. 1985;101:1799-1807.
Schliwa M, van Blerkom J, Porter KR. Stabilization of the cytoplasmic ground substance in detergent-opened cells and a structural and biochemical analysis of its composition. Proc Natl Acad Sci U S A. 1981;78:4329-4333.
Kent RL, Rozich JD, McCollam PL, McDermott DE, Thacker UF, Menick DR, McDermott PJ, Cooper G. Rapid expression of the Na+-Ca2+ exchanger in response to cardiac pressure overload. Am J Physiol. 1993;265:H1024-H1029.
Johnson TB, Kent RL, Bubolz BA, McDermott PJ. Electrical stimulation of contractile activity accelerates growth of cultured neonatal cardiocytes. Circ Res. 1994;74:448-459.
McDermott PJ, Rothblum LI, Smith SD, Morgan HE. Accelerated rates of ribosomal RNA synthesis during growth of contracting heart cells in culture. J Biol Chem. 1989;264:18220-18227.
Siegel S. Nonparametric Statistics. New York, NY: McGraw-Hill; 1956:18-194.
Hocking RR. The Analysis of Linear Models. Monterey, Calif: Brooks/Cole; 1985:69-181.
Bassett AL, Gelband H. Chronic partial occlusion of the pulmonary artery in cats. Circ Res. 1973;32:15-26.
de Tombe PP, ter Keurs HEDJ. Sarcomere dynamics in cat cardiac trabeculae. Circ Res. 1991;68:588-596.
Schulze E, Kirschner MW. Microtubule dynamics in interphase cells. J Cell Biol. 1986;102:1020-1031.
Katagiri K, Katagiri T, Kajiyama K, Yamamoto T, Yoshida T. Tyrosine-phosphorylation of tubulin during monocytic differentiation of HL-60 cells. J Immunol. 1993;150:585-593.
Aizawa H, Emori Y, Mori A, Murofushi H, Sakai H, Suzuki K. Functional analyses of domain structure of microtubule-associated protein-4 (MAP-U). J Biol Chem. 1991;266:9841-9846.
Cleveland DW. The multitubulin hypothesis revisited. J Cell Biol. 1987;104:381-383.
Lu Q, Luduena RF. In vitro analysis of microtubule assembly of isotypically pure tubulin dimers. J Biol Chem. 1994;269:2041-2047.
Gay DA, Sisodia SS, Cleveland DW. Autoregulatory control of β-tubulin mRNA stability is linked to translation elongation. Proc Natl Acad Sci U S A. 1989;86:5763-5767.
Bachurski CJ, Theodorakis NG, Coulson RMR, Cleveland DW. An amino-terminal tetrapeptide specifies cotranslational degradation of β-tubulin but not α-tubulin mRNAs. Mol Cell Biol. 1994;14:4076-4086.
Mooney DJ, Hansen LK, Langer R, Vacanti JP, Ingber DE. Extracellular matrix controls tubulin monomer levels in hepatocytes by regulating protein turnover. Mol Biol Cell. 1994;5:1281-1288.
Caron JM, Jones AL, Kirschner MW. Autoregulation of tubulin synthesis in hepatocytes and fibroblasts. J Cell Biol. 1985;101:1763-1772.