In situ localization of cholesterol in skeletal muscleby use of a monoclonal antibody

MARK S. F. CLARKE,1,4 CHARLES R. VANDERBURG,2 MARCAS M. BAMMAN,3

ROBERT W. CALDWELL,4 AND DANIEL L. FEEBACK5

1Division of Space Life Sciences, Universities Space Research Association, and 5Life SciencesResearch Laboratories, National Aeronautics and Space Administration/Johnson Space Center,Houston, Texas 77058; 2Research Space Management Group, Massachusetts General Hospital,Boston, Massachusetts 02114; 3Division of Exercise Physiology, University of Alabama atBirmingham, Birmingham, Alabama 35294; and 4Department of Pharmacology and Toxicology,Medical College of Georgia, Augusta, Georgia 30912Received 4 February 2000; accepted in final form 15 March 2000

Clarke, Mark S. F., Charles R. Vanderburg, MarcasM. Bamman, Robert W. Caldwell, and Daniel L. Fee-back. In situ localization of cholesterol in skeletal muscle byuse of a monoclonal antibody. J Appl Physiol 89: 731–741,2000.—A common perception is that cholesterol, the majorstructural lipid found in mammalian membranes, is localizednearly exclusively to the plasma membrane of living cells andthat it is found in much smaller quantities in internal mem-branes. This perception is based almost exclusively on cellfractionation studies, in which density gradient centrifuga-tion is used for purification of discrete subcellular membranefractions. Here we describe a monoclonal antibody, MAb2C5-6, previously reported to detect purified cholesterol insynthetic membranes (Swartz GM Jr, Gentry MK, AmendeLM, Blanchette-Mackie EJ, and Alving CR. Proc Natl AcadSci USA 85: 1902–1906, 1988), that is capable of detectingcholesterol in situ in the membranes of skeletal muscle sec-tions. Localization of cholesterol, the dihydropyridine recep-tor of the T tubule, and the Ca21-ATPase of the sarcoplasmicreticulum (SERCA2) by means of double and triple immuno-staining protocols clearly demonstrates that cholesterol isprimarily localized to the sarcoplasmic reticulum membranesof skeletal muscle rather than the sarcolemmal or T tubulemembranes. The availability of this reagent and its ability tospatially localize cholesterol in situ may provide a greaterunderstanding of the relationship between membrane cho-lesterol content and transmembrane signaling in skeletalmuscle.

immunolocalization; sarcoplasmic reticulum

CHOLESTEROL HAS AT LEAST TWO fundamental and impor-tant functions in mammalian cellular membranes.First, cholesterol serves as the major structural lipidthat provides strength to the membrane bilayer be-cause of its ability to pack into the intermolecularspaces that would otherwise exist between the constit-uent phospholipid molecules (14, 44). Cholesterol hasbeen termed a “rigidifying” lipid, and as the choles-

terol-to-phospholipid ratio of a mammalian membraneincreases, the membrane fluidity decreases (13, 15, 31,33, 37). Paradoxically, however, the absence of mem-brane cholesterol results in a purely phospholipid bi-layer in which the solid-to-liquid crystalline phasetransition temperature is ;42°C, which is well abovethe normal physiological temperature (e.g., 37°C inhumans) required for optimal membrane function inmammals. However, the addition of cholesterol tomammalian cell membranes reduces the solid-to-liquidcrystalline phase transition temperature to ;25°C,which is well below the physiological temperature re-quired for normal membrane function in mammals(10, 44).

A second function of cholesterol in mammalian cel-lular membranes is to contribute to the formation ofdistinct membrane microdomains, which have beenreferred to as “detergent-resistant membranes” (10,35), “cholesterol-rich microdomains” (6), or “lipid rafts”(9). One apparent function of these lipid rafts is toprovide a means of partitioning certain membraneproteins into cholesterol-rich or cholesterol-poor re-gions of the membrane. Whether proteins are incorpo-rated directly into these cholesterol-rich domains/lipidrafts at the level of the Golgi apparatus or physicallypartitioned between cholesterol-rich and cholesterol-poor membrane regions dependent on lipid solubilityonce they enter the membrane is unclear (for reviewsee Ref. 10). However, numerous studies in vitro and invivo have demonstrated that the activity of certainmembrane proteins is dependent on the presence andthe relative concentration of cholesterol in the mem-brane in which the proteins reside (for review see Ref.6). For example, increased levels of free cytosolic Ca21

in otherwise intact cells can be due to a cholesterol-related disruption in the function of membrane pro-teins responsible for Ca21 handling. For example,

Address for reprint requests and other correspondence: M. S. F.Clarke, Universities Space Research Association, Div. of Space LifeSciences, 3600 Bay Area Blvd., Houston, TX 77058 (E-mail:mark.s.clarke1@jsc.nasa.gov).

The costs of publication of this article were defrayed in part by thepayment of page charges. The article must therefore be herebymarked ‘‘advertisement’’ in accordance with 18 U.S.C. Section 1734solely to indicate this fact.

J Appl Physiol89: 731–741, 2000.

http://www.jap.org 731

increased resting cytosolic Ca21 levels in arterial vas-cular smooth muscle cells (46), rat cardiac myocytes(5), platelets (38), monocytes (1), and erythrocytes (29)have been linked to an increase in Ca21 flux across theplasma membrane due to increased membrane choles-terol content. In vascular smooth muscle cells, theactivity of the L-type Ca21 channel of the plasmamembrane was enhanced as the membrane cholesterolcontent was increased (36). A similar effect was ob-served in cardiac myocytes, where Ca21 channel activ-ity was enhanced as membrane cholesterol content wasincreased (5). Conversely, the activity of the Ca21-AT-Pase pump found in isolated sarcolemmal vesicles fromcardiac tissue was inhibited by increased membrane cho-lesterol content, whereas it was enhanced by reducedmembrane cholesterol content (30, 32). Membrane cho-lesterol content also appears to be important with respectto insulin and insulin-like growth factor I (IGF-I) signal-ing in skeletal muscle, so that as myofiber membranecholesterol content increases, so too does the insulin andIGF-I insensitivity of the muscle (27, 28, 39).

The prevailing view in the mammalian system isthat the majority of membrane-associated cholesterolis located in the plasma membrane of the cell, withlittle or no cholesterol in intracellular membrane com-ponents such as endoplasmic reticulum membranes (2,44). This view has been derived nearly exclusively fromcell fractionation studies, in which linear sucrose gra-dients have been utilized to separate cellular mem-brane components on their respective densities, fol-lowed by biochemical analysis of the lipid constituentsof these membrane fractions. This approach in skeletalmuscle has led to the view that the majority of mem-brane-associated cholesterol is found in the sarco-lemma and T tubule membrane fraction of adult skel-etal muscle (amphibian and mammalian), with little orno cholesterol being detected in the sarcoplasmic retic-ulum (SR) membrane fraction (19, 20, 23, 34). Morerecent studies have challenged this finding with bio-chemical demonstration that significant amounts ofcholesterol are detectable in the SR membrane frac-tions of mammalian skeletal muscle (8, 43) and thatthe relative amount of cholesterol in these membranesincreases with increasing age (25). Because of the po-tentially far-reaching effects of increased membranecholesterol content on the biomechanical propertiesand the transmembrane signaling function of mem-brane proteins located within the lipid matrix of skel-etal muscle membranes, this study was undertaken todevelop an immunohistochemical method that wouldallow the precise in situ cellular localization of choles-terol in frozen sections of normal adult rat and humanskeletal muscle tissue.

MATERIALS AND METHODS

Antibodies

An anti-cholesterol mouse IgM monoclonal antibody (MAb)was derived from a hybridoma cell line purchased fromAmerican Type Culture Collection (HB-8995, designationMAb 2C5-6) and grown according to the supplier’s instruc-

tions. This antibody has been shown to detect cholesterol insynthetic phospholipid vesicles in a concentration-dependentfashion and to specifically bind to the surface of purifiedcrystalline cholesterol at the ultrastructural level by electronmicroscopy (3, 41). The anti-dihydropyridine (DHP) receptor(anti-pan a1-subunit) rabbit IgG polyclonal antibody wasobtained from Almone Labs (Jerusalem, Israel). The anti-SRCa21-ATPase mouse IgG MAb (SERCA2), which recognizesonly the SR Ca21-ATPase found in type I myofibers, wasobtained from Research Diagnostics (Flanders, NJ). The rho-damine-conjugated F(Ab9)2 fragment of goat anti-mouse IgM(m-chain) polyclonal antibody was obtained from ICN Phar-maceuticals (Aurora, CA). The Alexa 350-conjugated goatanti-mouse IgG and Alexa 488-conjugated goat anti-rabbitIgG antibodies were purchased from Molecular Probes (Eu-gene, OR). The aqueous antifade mounting medium Fluoro-Mount G was purchased from Electron Microscopy Supplies(Fort Washington, PA).

Frozen Sectioning

Frozen (10-mm-thick) sections of perfusion-fixed (formalde-hyde) or snap-frozen rat soleus and triceps muscle were ob-tained as previously described (12). Frozen (10-mm-thick) sec-tions of needle biopsy samples from the vastus lateralis ofhuman test subjects were obtained as previously described (4).

Immunostaining Protocol

All incubations and washes were carried out at roomtemperature (i.e., 25°C). Frozen sections were washed in PBScontaining 1 mM Ca21 and 1 mM Mg21 (D-PBS) (pH 7.2) andthen incubated with D-PBS (pH 7.2) containing 50 mM am-monium chloride for 4 h to reduce autofluorescent signal inthe tissue. The sections were again washed once with D-PBSover a period of 5 min and then incubated in blocking bufferconsisting of D-PBS containing 4% (vol/vol) heat-inactivatedgoat serum (HIGS) for 1 h. Alternatively, those sectionsdestined for localization of the DHP receptor were incubatedin D-PBS containing 0.5% (vol/vol) Triton X-100 for 30 minand then incubated for 1 h in blocking buffer. Sections werethen incubated in primary antibody diluted with D-PBScontaining 1% HIGS for 2 h as follows: the anti-cholesterolmouse IgM monoclonal antibody MAb 2C5-6 was used at a1:4 dilution of neat hybridoma supernatant, the anti-DHPreceptor (anti-pan a1-subunit) rabbit IgG polyclonal antibodywas used at a 1:60 dilution of manufacturer’s stock solution,and the anti-SR Ca21-ATPase mouse IgG monoclonal anti-body (SERCA2) was used at a 1:400 dilution of the manufac-turer’s stock solution. Those sections destined for double ortriple labeling of cholesterol, the DHP receptor, and/or the SRCa21-ATPase in the same section were incubated for 2 h in amixture of primary antibodies at the dilutions noted above.After primary antibody incubation, the sections were washedfive times over a period of 30 min with D-PBS and thenincubated for 1 h with the relevant secondary antibody aloneor a mixture of secondary antibodies, all of which were usedat a 1:400 dilution of the manufacturer’s stock solutionsdiluted in D-PBS containing 1% (vol/vol) HIGS. All primaryand secondary antibody staining solutions were centrifugedat 10,000 g for 2 min immediately before use to remove anyparticulate matter. The sections were then washed over aperiod of 30 min with five changes of D-PBS, mounted inFluoroMount G antifade aqueous mounting medium, cover-slipped, and sealed with clear nail polish before microscopicexamination. Immunostained sections were viewed using aZeiss Axioplan2 fluorescent microscope or a Zeiss scanningconfocal microscope. Epifluorescent images were captured

732 IMMUNOLOCALIZATION OF CHOLESTEROL IN MUSCLE

using a COHU charge coupled device camera attached to theAxioplan2 microscope controlled by MacProbe image capturesoftware (PSI, League City, TX). Confocal images were cap-tured using the Zeiss image analysis software integral to thescanning confocal microscope.

Demonstration of MAb 2C5-6 Specificity for Cholesterolin Frozen Sections

Cross-reaction of MAb 2C5-6 with skeletal muscle mem-brane-associated proteins. Crude membrane fractions wereprepared from rabbit skeletal muscle as previously described(18) (membrane fractions were kindly provided by Dr. SusanHamilton, Baylor College of Medicine, Houston, TX). Crudemembrane fractions were solubilized in standard Laemmlisolubilization buffer by boiling for 5 min at a concentration of1 mg of total protein per microliter of buffer. The solution wasthen centrifuged at 10,000 g to remove any insoluble mate-rial, and the resulting supernatant was collected. The super-natant was then dot blotted onto 0.2-mm supported nitrocel-lulose in 5-ml aliquots and allowed to dry. Subsequentaliquots of supernatant were reapplied to each dot blot tocreate a concentration gradient of solubilized membrane-associated protein of 5, 10, 15, and 30 mg/dot blot, respec-tively. Negative controls containing 1 mg of BSA per microli-ter of Laemmli buffer were applied in a similar fashion to thenitrocellulose membrane. The dot blot was blocked withD-PBS containing 4% (vol/vol) HIGS and 0.1% (vol/vol)Tween 20 and then incubated with a 1:8 dilution of MAb2C5-6 hybridoma supernatant made up in D-PBS containing1% (vol/vol) HIGS and 0.1% (vol/vol) Tween 20 for 1 h at roomtemperature with gentle agitation. The blot was washedthree times with D-PBS containing 0.1% (vol/vol) Tween 20over a period of 20 min and then incubated with a 1:1,000dilution of an alkaline phosphatase-conjugated F(Ab9)2 frag-ment of goat anti-mouse IgM (m-chain) polyclonal antibodymade up in D-PBS containing 1% (vol/vol) HIGS and 0.1%(vol/vol) Tween 20 for 1 h at room temperature with gentleagitation. The blot was again washed, and secondary anti-body binding was disclosed using 5-bromo-4-chloro-3-in-dolylphosphate-p-toluidine-nitro blue tetrazolium substratesolution (Vector Laboratories, Burlingame, CA).

Acetone extraction. Frozen sections were incubated in ace-tone (previously dehydrated with calcium chloride) for 1 h at4°C. This procedure has previously been shown to extract alllipids, including cholesterol, from fixed and snap-frozen tis-sues (24). The sections were then stained for cholesterol withMAb 2C5-6 as described above.

Cold ethanol-Triton X-100 extraction. Frozen sections wereincubated in four changes of pure ethanol containing 0.5%Triton X-100 for 1 h at 4°C. Triton X-100 and/or cold ethanolhave previously been shown to extract a variety of lipids andproteins, excluding cholesterol, from fixed and snap-frozentissues (10, 24). The sections were then stained for choles-terol with MAb 2C5-6 as described above.

Treatment with saponin. Frozen sections were incubated infour changes of Ca21- and Mg21-free PBS containing 0.1%saponin for 1 h at room temperature. Saponin treatment ofcells has previously been shown to “chelate” or “complex”cholesterol within the matrix of the membrane (7, 17, 45).The sections were then stained for cholesterol with MAb2C5-6 as described above.

Specificity of MAb 2C5-6 for Cholesterol on the Basisof Histological Criteria

The cells of Ito are cells that occasionally line the bilecanaliculi of the liver. These cells are in intimate contact with

bile fluid and contain large amounts of cholesterol. In addi-tion, hepatocytes synthesize and process large amounts ofcholesterol. Frozen sections of perfusion-fixed rat liver werestained with MAb 2C5-6, with and without the cholesterolextraction treatments described above before staining.

Preadsorption of Primary Antibody Against PurifiedCrystalline Cholesterol and Closely Related Compounds

Purified crystalline cholesterol (99.9% pure; Sigma Chem-ical, St. Louis, MO), dihydrocholesterol (cholesterol dimer,95% pure; Sigma Chemical), 4-cholesten-3-one (cholesteroloxidation product, 95% pure; Sigma Chemical), cholesteryloleate (cholesterol ester, 99% pure; Sigma Chemical), orbrain lipid extract (type VII from bovine brain, acetone pre-cipitated Folch extract of whole bovine brain containing allthe major brain phospholipids and glycolipids; Sigma Chem-ical) were each dissolved in 200 ml of chloroform at a concen-tration of 200 mg/ml and dispensed into 15-ml glass bottlesthat had been rinsed with chloroform. The glass bottles werethen rolled along their long axes to coat the internal surfaceof the bottles with the chloroform-lipid solution. As the chlo-roform evaporated from the bottles, a thin, even coating ofcrystalline lipid was deposited on the internal surface of theglass bottles. The bottles were dried in an oven at 65°C for1 h, filled with D-PBS containing 4% (vol/vol) HIGS (i.e.,frozen section blocking buffer), and then incubated for 1 h atroom temperature and gently washed with D-PBS. Afternonspecific protein blocking of the crystalline lipid layer, 4 mlof MAb 2C5-6 hybridoma supernatant [diluted 1:4 in D-PBScontaining 1% (vol/vol) HIGS] were added to each bottle. Thebottles were then gently rotated for 16 h at 4°C, therebyensuring efficient contact of the antibody solution with thecrystalline lipid layer. The preadsorbed antibody solutionswere removed from each bottle, centrifuged at 15,000 g for 15min to remove any particulate matter, and used instead ofprimary antibody to stain frozen sections as described above.

RESULTS

When frozen sections of rat soleus or human vastuslateralis were stained with MAb 2C5-6, which was pre-viously shown to recognize purified cholesterol in syn-thetic phospholipid vesicles and purified crystalline cho-lesterol (3, 41), distinct staining of intramyofibermembranes could clearly be seen (Fig. 1). In cross section,the MAb 2C5-6 staining pattern exhibited a “rosette”appearance around individual myofibrils that was mor-phologically reminiscent of the pattern expected if theterminal cisternae of the SR membrane component incross section had been stained. In longitudinal sections,the MAb 2C5-6 staining appeared as an intermittentbanding pattern around individual myofibrils. Surpris-ingly, no evidence of any staining at the light-microscopiclevel of the sarcolemma of individual myofibers in humanor rat muscle sections was observed. Because the prevail-ing view is that cholesterol is not found in the SR mem-branes, but in the sarcolemma/T tubule membranes ofskeletal muscle, the specificity of the MAb 2C5-6 stainingfor cholesterol in frozen tissue sections was verified byadditional experiments.

The first step in this process was to determinewhether MAb 2C5-6 cross-reacted with protein compo-nents of skeletal muscle membranes. When SDS-solu-bilized crude muscle membrane fractions were dot-

733IMMUNOLOCALIZATION OF CHOLESTEROL IN MUSCLE

blotted and immunostained with MAb 2C5-6, noreactivity was observed, indicating that this antibodydoes not recognize an epitope found in reduced skeletalmuscle membrane-associated protein samples. Thesecond step in this process was to determine whetherMAb 2C5-6 stained cellular structures definitivelyknown to contain cholesterol. With use of frozen sec-tions of perfusion-fixed rat liver, MAb 2C5-6 wasshown to heavily stain the cells that occasionally linethe bile canaliculi (Fig. 2). These cells, known as thecells of Ito, are responsible for the uptake and concen-tration of cholesterol from the surrounding hepatocytesbefore secretion of cholesterol into the bile canaliculi asa constituent of the bile fluid. In addition, MAb 2C5-6also stained membranous and nonmembranous compo-nents of the hepatocytes, the major cholesterol-produc-ing cells of the liver. When liver sections were ex-tracted with acetone before they were stained withMAb 2C5-6, all staining was abolished (Fig. 2B). Asimilar decrease in staining intensity was observed ifliver sections were stained with MAb 2C5-6 pread-sorbed with purified crystalline cholesterol (Fig. 2C).When liver sections were extracted with a mixture ofcold ethanol and Triton X-100 (neither of which extract

cholesterol at room temperature or below) before theywere stained with MAb 2C5-6, staining intensity of thecells of Ito and hepatocytes was enhanced (Fig. 2D). Ifliver sections were treated with saponin before theywere stained with MAb 2C5-6, the staining patternchanged from a discrete to a diffuse particulate pattern(Fig. 2E). This alteration in staining pattern is consis-tent with the cholesterol chelating or complexing ac-tion of saponin in mammalian membranes (7). No sig-nificant staining was observed in liver sections if amixture of fresh hybridoma growth medium containing1% (vol/vol) HIGS was substituted for primary anti-body solution (Fig. 2F).

Similar effects on MAb 2C5-6 staining were observedin human vastus lateralis sections (Fig. 3) when thesesamples were treated in a similar fashion to the liversections, namely, staining was abolished if sectionswere acetone extracted before they were stained (Fig.3B), staining was reduced if sections were stained withMAb 2C5-6 preadsorbed against purified crystallinecholesterol (Fig. 3C), staining was enhanced if sectionswere treated with cold ethanol-Triton X-100 (Fig. 3D),and MAb 2C5-6 staining was redistributed if the sec-tions were treated with saponin before they werestained (Fig. 3E). No significant staining was observedin skeletal muscle sections if a mixture of fresh hybrid-oma growth medium containing 1% (vol/vol) HIGS wassubstituted for primary antibody solution (Fig. 3F).Similar results were observed if frozen sections of per-fusion-fixed or snap-frozen rat soleus muscle weretreated in a similar manner (data not shown).

To further demonstrate the specificity of MAb 2C5-6for cholesterol, antibody supernatant was preadsorbedagainst a number of purified lipid molecules closelyrelated to cholesterol in structure, namely, dihydrocho-lesterol (cholesterol dimer), cholesteryl oleate (choles-terol ester), and 4-cholesten-3-one (cholesterol ketone,95% pure; Sigma Chemical). In addition, antibody su-pernatant was also preadsorbed against a complexmixture of lipids (i.e., brain extract, fraction VII) thatcontains all the major phospholipids and glycolipids ofbrain tissue but is cholesterol depleted by virtue ofbeing an acetone precipitation of a Folch extraction.When frozen sections of rat soleus muscle were stainedwith MAb 2C5-6 preadsorbed against these purifiedlipids, a significant reduction in staining was observedif the antibody was preadsorbed with cholesterol (Fig.4B) or its dimer dihydrocholesterol (Fig. 4C) but notwith cholesteryl oleate (Fig. 4D), 4-cholesten-3-one(Fig. 4E), or brain extract (fraction VII; Fig. 4F).

On the basis of previously published results thatdemonstrated the specificity of MAb 2C5-6 for choles-terol in synthetic membranes, as well as at the elec-tron-microscopic level for purified crystalline choles-terol (3, 41), combined with the immunofluorescentstaining results obtained in frozen liver and musclesections reported above, we conclude that MAb 2C5-6can be used to detect cholesterol biochemically, as haspreviously been reported (3, 41), but also to specificallylocalize cholesterol immunohistochemically at the cel-lular level in situ.

Fig. 1. Fluorescent micrograph of MAb 2C5-6 staining in a frozencross section of human vastus lateralis (A) and a longitudinal sectionof rat soleus muscle (B). Note the “rosette” appearance of the stainingpattern (A) around individual myofibrils in cross section and theintermittent banding pattern (B) around individual myofibrils inlongitudinal section. Scale bars, 50 and 10 mm in A and B, respec-tively.

734 IMMUNOLOCALIZATION OF CHOLESTEROL IN MUSCLE

As illustrated in Fig. 1, the cellular distribution ofcholesterol in skeletal muscle appears on morphologi-cal criteria at the light-microscopic level to be localizedalmost exclusively to the SR membrane component. Toverify this result, frozen sections of rat soleus musclewere double immunolabeled for cholesterol and eitherthe a1-subunit of DHP receptor (40) found exclusivelyin the sarcolemma/T tubule membranes (Fig. 5) or theCa21-ATPase of type I myofibers (SERCA2) (22) foundexclusively in SR membranes (Fig. 6). Cholesterolstaining (red, Fig. 5A) and DHP receptor staining(green, Fig. 5B) do not exhibit a similar pattern withinthe same section. When a composite image of choles-terol and DHP receptor staining in the same section isviewed at a higher magnification (Fig. 5C), the DHPreceptor staining is spatially oriented above or besidethe membrane component stained for cholesterol, indi-cating that cholesterol does not colocalize at the light-microscopic level with a protein (i.e., DHP receptor)found exclusively in the sarcolemma/T tubule mem-branes of skeletal muscle. Conversely, if the section isdouble immunolabeled for cholesterol (red, Fig. 6A)and the SERCA2 protein (green, Fig. 6B), the choles-

terol and SERCA2 staining patterns at the light-micro-scopic level are identical. When a composite image ofcholesterol and SERCA2 staining in the same section isviewed at a higher magnification (Fig. 6C), both anti-bodies recognize antigens localized to the same mem-brane structures within an individual myofiber. Con-focal microscopy of frozen cross sections of humanvastus lateralis stained for cholesterol and the DHPreceptor showed a clear spatial delineation of the Ttubule marker (i.e., the DHP receptor) and cholesterolstaining (Fig. 7). When longitudinal frozen sections ofrat soleus were triple immunolabeled for cholesterol(red, Fig. 8A), the DHP receptor (green, Fig. 8B), andSERCA2 (blue, Fig. 8C), the composite overlay image(Fig. 8D) clearly indicates that DHP receptor staining(green) could be distinguished from the cholesterol andSERCA2 staining signal, which appeared as a magentacolor (red-and-blue mix).

DISCUSSION

The most commonly perceived function for the mem-branes of living cells is that of a physical barrier

Fig. 2. Fluorescent micrographs ofMAb 2C5-6 staining in perfusion-fixedfrozen sections of rat liver tissue afterMAb 2C5-6 staining of liver tissuewithout pretreatment (A), MAb 2C5-6staining of liver tissue after acetoneextraction (B), MAb 2C5-6 staining ofliver tissue after preadsorption of theprimary antibody against purified crys-talline cholesterol immobilized on asolid support (C), MAb 2C5-6 stainingof liver tissue after extraction with amixture of ethanol and 1% (vol/vol)Triton X-100 (D), MAb 2C5-6 stainingof liver tissue after saponification ofthe tissue (E), and immunostaining as-sociated with secondary antibody bind-ing only (F). All images were acquiredunder identical photographic conditions.Note the intense staining of the cells ofIto after ethanol-Triton X-100 extrac-tion (D) and the redistribution of theMAb 2C5-6 staining pattern to a finegranular pattern after saponification ofthe tissue (E). Scale bars, 10 mm.

735IMMUNOLOCALIZATION OF CHOLESTEROL IN MUSCLE

between the extracellular and intracellular environ-ment (i.e., plasma membrane) or as “partitioning”structures between specific regions of the intracellularenvironment (e.g., nuclear membrane and endoplasmicreticulum). Historically, cellular membranes havebeen considered an inert “lipid scaffold structure,” inwhich the elements essential for transmembrane sig-naling (i.e., proteins) are housed and supported whilethey carry out their signal transduction function (10).However, it seems unlikely that the large heterogene-ity observed in the lipid components of the plasmamembrane or intracellular membranes would be re-quired if they simply served only a physical barrierfunction. It is the growing body of experimental evi-dence which suggests that the lipid composition of amembrane, specifically its cholesterol content, can al-ter the biomechanical properties of the membrane andthe transmembrane signaling function of membraneproteins embedded within the three-dimensional ma-trix of the membrane that has driven this study (forreview see Refs. 9, 6, 44).

Utilizing a novel immunostaining procedure for de-tecting cholesterol in frozen sections, we have localized

cholesterol, at the fluorescent light-microscopic level,exclusively to the SR membrane of normal skeletalmuscle rather than to the sarcolemma/T tubule mem-brane. This is not to say that cholesterol is completelyabsent from other myofiber membranes (i.e., sarco-lemma/T tubule or mitochondrial membranes) butrather indicates that the vast majority of cholesterolfound in skeletal muscle is present in the SR mem-brane component. Utilizing a series of selective solventand/or detergent extraction protocols (i.e., acetone ex-traction of all lipids including cholesterol, ethanol-Triton X-100 extraction of some lipids and some proteinswithout extraction of cholesterol, and chelation/complex-ing of cholesterol using saponin) and preadsorption of theprimary antibody against purified antigen and closelyrelated compounds (i.e., purified crystalline choles-terol, dihydrocholesterol, cholesteryl oleate, 4-cho-lesten-3-one, and a cholesterol-depleted complex lipidmixture), we have demonstrated that the epitope rec-ognized by MAb 2C5-6 is specific to cholesterol and itsdimer and that it is capable of recognizing nativecholesterol residing in skeletal muscle membranes offrozen sections in situ. In addition, utilizing known

Fig. 3. Fluorescent micrographs ofMAb 2C5-6 staining in snap-frozencross sections of human vastus latera-lis after MAb 2C5-6 staining withoutpretreatment (A), MAb 2C5-6 stainingafter acetone extraction (B), MAb2C5-6 staining after preadsorption ofthe primary antibody against purifiedcrystalline cholesterol immobilized ona solid support (C), MAb 2C5-6 stain-ing after extraction with a mixture ofethanol and 1% (vol/vol) Triton X-100(D), MAb 2C5-6 staining after saponi-fication of the tissue (E), and immuno-staining associated with secondary an-tibody binding only (F). All imageswere acquired under identical photo-graphic conditions. Scale bars, 10 mm.

736 IMMUNOLOCALIZATION OF CHOLESTEROL IN MUSCLE

histological criteria, we have demonstrated that MAb2C5-6 recognizes cellular structures, membranous andotherwise, definitively known to contain large quanti-ties of cholesterol in vivo, namely, hepatocytes andcells of Ito in frozen liver sections. Furthermore, utiliz-ing double and triple immunolabeling protocols withantibodies that recognize proteins already shown to beexclusively localized to the sarcolemma/T tubule mem-branes (i.e., the DHP receptor) (40) or the SR mem-branes (i.e., SR Ca21-ATPase pump) (22), we havedemonstrated at the light-microscopic level that cho-lesterol is localized to the same membrane componentas the SR Ca21-ATPase pump (i.e., the SR membrane)rather than to the sarcolemma/T tubule component inhuman and rat skeletal muscle.

Because MAb 2C5-6 does not significantly stainother membrane structures within the same musclesection (i.e., sarcolemma, T tubule, or mitochondrialmembranes) or show significant cross-reactivity with acholesterol-depeleted complex lipid mixture containingphospholipids and glycolipids (Fig. 4F), it appears un-likely that MAb 2C5-6 recognizes the phospholipids orglyolipids presumably common to all mammalian mem-branes. In addition, extraction of skeletal muscle sec-tions with cold acetone abolishes MAb 2C5-6 staining(Figs. 2 and 3). Inasmuch as acetone extracts onlyhydrophobic lipids, rather than hydrophilic lipids,phospholipids and glycolipids being hydrophilic lipids(24), these data, in combination with a lack of cross-

reactivity with a complex lipid mixture (Fig. 4), confirmthat MAb 2C5-6 recognizes neither phospholipids norglycolipids in addition to cholesterol. It is possible thatMAb 2C5-6 may recognize several other hydrophobiclipid species that may be present in skeletal muscle SRmembranes, including free fatty acids, terpenes, neu-tral fats, waxes, cholesterol esters, and steroid hor-mones that have a hydrophobic nature similar to thatof cholesterol (i.e., are soluble in acetone). However,the lack of cross-reactivity of MAb 2C5-6 with purifiedcholesteryl oleate or 4-cholesten-3-one (Fig. 4) indi-cates that MAb 2C5-6 recognizes an epitope common tocholesterol and its dimer, but not its ester or ketone.

The localization of the majority of cholesterol (andpotentially its dimer dihydrocholesterol) present in skel-etal muscle to the SR membranes rather than to thesarcolemmal/T tubule membranes in vivo raises manyinteresting physiologically relevant questions as to theimportance of relative membrane cholesterol content andthe localization of cholesterol in a particular membrane.For example, we and others have previously shown thata large increase in membrane cholesterol content (asoccurs in endothelial cells after exposure to serum con-taining hypercholesterolemic levels of low-density li-poproteins) increases cell susceptibility to mechanicallyinduced membrane damage or “membrane wounding”(13, 42). These effects have been attributed to cholester-ol’s ability to increase membrane order (i.e., decreasemembrane fluidity or increase membrane rigidity), re-

Fig. 4. Fluorescent micrographs of MAb 2C5-6 stainingin snap-frozen longitudinal sections of rat soleus mus-cle after preadsorption of the primary antibody againsta range of lipids immobilized on glass: MAb 2C5-6staining after preadsorption of the primary antibody tochloroform-washed glass (A), MAb 2C5-6 staining afterpreadsorption of the primary antibody to immobilizedcrystalline cholesterol (B), MAb 2C5-6 staining afterpreadsorption of the primary antibody to immobilizedcrystalline dihydrocholesterol (C), MAb 2C5-6 stainingafter preadsorption of the primary antibody to immo-bilized crystalline cholesteryl oleate (D), MAb 2C5-6staining after preadsorption of the primary antibody toimmobilized crystalline 4-cholestene-3-one (E), andMAb 2C5-6 staining after preadsorption of the primaryantibody to immobilized brain lipid extract (F). Allimages were acquired under identical photographicconditions. Note the significant reduction in stainingcompared with control levels when primary antibodywas preadsorbed against cholesterol (B) and cholester-ol’s dimer dihydrocholesterol (C), but not with cho-lesteryl oleate (D), 4-cholestene-3-one (E), or brain ex-tract (a cholesterol-depleted complex lipid mixture, F).No significant staining was observed with use of sec-ondary antibody alone. Scale bar, 10 mm.

737IMMUNOLOCALIZATION OF CHOLESTEROL IN MUSCLE

Fig. 5. Double immunofluorescent labeling of snap-frozen longitudi-nal sections of rat soleus with MAb 2C5-6 [tetramethylrhodamineisothiocyanate (TRITC)-linked secondary antibody, red; A] and ananti-dihydropyridine (DHP; pan a1-chain) antibody (Alexa 488-linked secondary antibody, green; B) and a composite overlay of bothstaining patterns (C). MAb 2C5-6 and the anti-DHP (a1-chain) anti-body do not exhibit a similar staining pattern within the samesection, and, in a composite overlay of the same section viewed at ahigher magnification (C), DHP receptor staining is spatially orientedabove or beside the MAb 2C5-6 staining. Scale bars, 10 mm.

Fig. 6. Double immunofluorescent labeling of snap-frozen longitudi-nal sections of rat soleus with MAb 2C5-6 (TRITC-linked secondaryantibody, red; A) and an anti-SERCA2 antibody (Alexa 488-linkedsecondary antibody, green; B) and a composite overlay of both stain-ing patterns (C). MAb 2C5-6 and the anti-SERCA2 antibody exhibita similar staining pattern within the same section, and, in a com-posite overlay image of the same section viewed at a higher magni-fication (C), it appears that the SERCA2 and MAb 2C5-6 stainingpatterns are identical. Scale bars, 10 mm.

738 IMMUNOLOCALIZATION OF CHOLESTEROL IN MUSCLE

sulting in a membrane that is less capable of withstand-ing the imposition of mechanical force (15). Evidence tosupport this contention comes from the rapid (i.e., 2 min)reversal of this effect by the addition of a membrane-

disordering lipid to the cholesterol-enriched endothelialcell membrane (13).

In addition, cholesterol-induced increases in mem-brane order can affect the transmembrane signalingfunction of membrane proteins in a variety of differentsystems. First, increased amounts of cholesterol in theplasma membrane of a variety of different cell types, invitro and in vivo, result in a concomitant increase inmembrane order and alteration in the fidelity of nu-merous transmembrane signaling events, includingthose associated with neuronal transmission and intra-cellular Ca21 regulation (reviewed in Ref. 6). Second,increased cholesterol content and the subsequent in-crease in membrane order of skeletal muscle sarco-lemma detected after feeding rats a high-fat diet re-sults in a reduction in the sensitivity of this tissue tocirculating anabolic factors, such as insulin and IGF-I(27, 28, 39). Decreased sensitivity of skeletal muscle tothese growth factors has been associated with the in-ability of the growth factor receptor to undergo theappropriate conformational changes required for effi-cient transmembrane signaling of the growth factorstimulus. This effect has been referred to as “molecularfreezing” of the receptor protein in the three-dimen-sional lipid matrix of the sarcolemma. This phenome-non is directly related to a number of alterations in thelipid matrix of the plasma membrane, including in-creased cholesterol content, increased saturated fattyacid content, and increased cross-linking of the lipidmembrane matrix due to free radical-induced lipidperoxidation, all of which result in an increase inmembrane order (11, 27, 39). More importantly, thedecrease in skeletal muscle sensitivity to insulin andIGF-I observed after an increase in membrane orderdue to cholesterol enrichment of the sarcolemma was

Fig. 7. A composite confocal image of double immunofluorescentlabeling of a snap-frozen cross section of a single myofiber of humanvastus lateralis with MAb 2C5-6 (TRITC-linked secondary antibody,red) and an anti-DHP (pan a1-chain) antibody (Alexa 488-linkedsecondary antibody, green). Note the clear spatial delineation of themembrane network containing the T tubule marker (i.e., the DHPreceptor, green) and the membrane structures stained with MAb2C5-6.

Fig. 8. Triple immunofluorescent label-ing of snap-frozen longitudinal sectionsof rat soleus with MAb 2C5-6 (TRITC-linked secondary antibody, red; A), ananti-DHP (pan a1-chain) antibody (Alexa488-linked secondary antibody, green;B), and an anti-SERCA2 antibody (Alexa350-linked secondary antibody, blue; C)and a composite overlay image of all 3staining patterns (D). MAb 2C5-6 andthe anti-DHP (a1-chain) antibody do notexhibit a similar staining pattern withinthe same section, whereas MAb 2C5-6and the anti-SERCA2 antibody do. In acomposite overlay of all 3 staining pat-terns within the same section (D), theDHP receptor staining (green) is spa-tially oriented above or beside the colo-calized MAb 2C5-6/anti-SERCA2 stain-ing (magenta). Scale bars, 10 mm.

739IMMUNOLOCALIZATION OF CHOLESTEROL IN MUSCLE

reversed if the animals were fed a diet high in n–3fatty acids. The n–3 fatty acids were shown to substi-tute into the sarcolemma and bring about a decrease inmembrane order (i.e., an increase in membrane fluid-ity) that was paralleled by a return to normal levels ofinsulin and IGF-I sensitivity in the muscle (27, 28).Aged muscle has been reported to contain elevatedlevels of cholesterol compared with young muscle (25,43), and Ca21 uptake is decreased in skeletal muscleSR vesicles isolated from aged rodent (16, 26) andhuman (21) muscle.

Historically, the prevailing view has been that themajority of cholesterol found in skeletal muscle waslocated in the sarcolemma and contiguous T tubulemembrane system. This view has been supported bylipid analysis of skeletal muscle membranes isolatedand purified by differential centrifugation by use oflinear sucrose gradients (19, 23, 34). These studiessuggested that the majority of cholesterol was associ-ated with the sarcolemma and T tubule membranefractions rather than the SR membrane fractions ofpurified skeletal muscle membranes. However, severalstudies utilizing the same technology have challengedthis view, suggesting that skeletal muscle SR mem-branes contain a significant amount of cholesterol (8,25, 43) and that incomplete separation/purification ofsarcolemma/T tubule and SR membrane fractions bydensity centrifugation may have led to this discrep-ancy.

Our results utilizing MAb 2C5-6 to spatially localizecholesterol in frozen sections indicate that SR mem-branes contain the vast majority of cholesterol presentin rat and human skeletal muscle. However, the pos-sibility that cholesterol may behave as a “cryptic” an-tigen in natural and synthetic membranes has beensuggested theoretically and partially demonstrated ex-perimentally. The inability of anti-cholesterol antibod-ies to react with the surface of synthetic liposomescontaining ,50% cholesterol, coupled to the theoreticalanalysis of the size of an antibody molecule relative tothe membrane-embedded antigen (i.e., cholesterol in-tercalated into a phospholipid bilayer) (reviewed inRef. 3), could potentially explain why we did not ob-serve cholesterol staining with MAb 2C5-6 in sarco-lemma or T tubule membranes. Specifically, the rela-tive cholesterol concentration in the sarcolemma or Ttubule membranes was below the “detectable” limit forour antibody because of antigen-antibody inaccessibil-ity, whereas the relative concentration of cholesterol inthe SR membranes was above the detectable limit byutilization of immunofluorescent localization techniques.However, unlike a phospholipid liposome where the cho-lesterol molecule is embedded within the lipid matrixand may be inaccessible to an antibody below a certainconcentration threshold, the experiments detailed herewere carried out on transversely sectioned membranestructures in situ. As such, exposure of the centralregions of sarcolemma/T tubule and SR membranes byfrozen sectioning will reduce antigen-antibody accessi-bility problems to a minimum. Our results suggest thatif cholesterol is present in membrane components of

normal skeletal muscle other than SR membranes,such as sarcolemma/T tubules or mitochondrial mem-branes, then the absence of staining at the immunoflu-orescent microscopic level is due to the presence ofcholesterol in very small amounts relative to theamount found in SR membranes.

The ability to spatially immunolocalize cholesterol infrozen sections by use of MAb 2C5-6 will provide auseful tool in the study of the cellular localization ofthis important molecule in normal and disease statesand will further our knowledge of the specific spatialinteractions of cholesterol with proteins embedded inthe three-dimensional matrix of the membrane.

REFERENCES

1. Aepfelbacher M, Hrboticky N, Lux I, and Weber PC. Cho-lesterol modulates PAF-stimulated Ca21 mobilization in mono-cytic U937 cells. Biochim Biophys Acta 1074: 125–129, 1991.

2. Alberts B, Bray D, Lewis J, Raff M, Roberts K, and WatsonJD. Molecular Biology of the Cell (2nd ed.). New York: Garland,1989.

3. Alving CR and Swartz GM Jr. Antibodies to cholesterol,cholesterol conjugates and liposomes: implications for athero-sclerosis and autoimmunity. Crit Rev Immunol 10: 441–453, 1991.

4. Bamman MM, Clarke MS, Feeback DL, Talmadge RJ,Stevens BR, Lieberman SA, and Greenisen MC. Impact ofresistance exercise during bed rest on skeletal muscle sarcopeniaand myosin isoform distribution. J Appl Physiol 84: 157–163,1998.

5. Bastiaanse EM, Atsma DE, Kuijpers MM, and Van derLaarse A. The effect of sarcolemmal cholesterol content onintracellular calcium ion concentration in cultured cardiomyo-cytes. Arch Biochem Biophys 313: 58–63, 1994.

6. Bastiaanse EML, Hold KM, and Van der Laarse A. Theeffect of membrane cholesterol content on ion transport pro-cesses in plasma membranes. Circ Res 33: 272–283, 1997.

7. Bomford R. Studies on the cellular site of action of the adjuvantactivity of saponin for sheep erythrocytes. Int Arch Allergy ApplImmunol 67: 127–131, 1982.

8. Borchman D, Simon R, and Bicknell-Brown E. Variation inthe lipid composition of rabbit muscle sarcoplasmic reticulummembrane with muscle type. J Biol Chem 257: 14136–14139, 1982.

9. Brown DA and London E. Functions of lipid rafts in biologicalmembranes. Annu Rev Cell Dev Biol 14: 111–136, 1998.

10. Brown DA and London E. Structure and origin of orderedlipid domains in biological membranes. J Membr Biol 164: 103–114, 1998.

11. Choi JH and Yu BP. Brain synaptosomal aging: free radicalsand membrane fluidity. Free Radic Biol Med 18: 133–139, 1995.

12. Clarke MS, Khakee R, and McNeil PL. Loss of cytoplasmicbasic fibroblast growth factor from physiologically wounded myo-fibers of normal and dystrophic muscle. J Cell Sci 106: 121–133,1993.

13. Clarke MSF, Pritchard KA, Medows MS, and McNeil PL.An atherogenic level of native LDL increases endothelial cellvulnerability to shear-induced plasma membrane wounding andconsequent FGF release. Endothelium 4: 127–139, 1995.

14. Demel RA and De Kruyff B. The function of sterols in mem-branes. Biochim Biophys Acta 457: 109–132, 1976.

15. Evans E and Rawicz W. Entropy-driven and bending elasticityin condensed-fluid membranes. Phys Rev Lett 64: 2094–2097,1990.

16. Gafni A and Yuh KC. A comparative study of the Ca21-Mg21

dependent ATPase from skeletal muscles of young, adult and oldrats. Mech Ageing Dev 49: 105–117, 1989.

17. Gogelein H and Huby A. Interaction of saponin and digitoninwith black lipid membranes and lipid monolayers. Biochim Bio-phys Acta 773: 32–38, 1984.

18. Hamilton SL, Alvarez RM, Fill M, Hawkes MJ, Brush KL,Schilling WP, and Stefani E. [3H]PN200-110 and [3H]ryano-

740 IMMUNOLOCALIZATION OF CHOLESTEROL IN MUSCLE

dine binding and reconstitution of ion channel activity withskeletal muscle membranes. Anal Biochem 183: 31–41, 1989.

19. Headon DR, Barrett EJ, Joyce NM, and O’Flaherty J.Cholesterol in muscle membranes. Mol Cell Biochem 17: 117–123, 1977.

20. Hidalgo C. Lipid phase of transverse tubule membranes fromskeletal muscle. An electron paramagnetic resonance study. Bio-phys J 47: 757–764, 1985.

21. Hunter SK, Thompson MW, Ruell PA, Harmer AR, ThomJM, Gwinn TH, and Adams RD. Human skeletal sarcoplasmicreticulum Ca21 uptake and muscle function with aging andstrength training. J Appl Physiol 86: 1858–1865, 1999.

22. Jorgensen AO, Arnold W, Pepper DR, Kahl SD, Mandel F,and Campbell KP. A monoclonal antibody to the Ca21-ATPaseof cardiac sarcoplasmic reticulum cross-reacts with slow type Ibut not with fast type II canine skeletal muscle fibers: an immu-nocytochemical and immunochemical study. Cell Motil Cytoskel-eton 9: 164–174, 1988.

23. Kidwai AM, Radcliffe MA, Lee EY, and Daniel EE. Isolationand properties of skeletal muscle plasma membrane. BiochimBiophys Acta 298: 593–607, 1973.

24. Kiernan JA. Histological and Histochemical Techniques: The-ory and Practice (2nd ed.). New York: Pergamon, 1990.

25. Krainev AG, Ferrington DA, Williams TD, Squier TC, andBigelow DJ. Adaptive changes in lipid composition of skeletalsarcoplasmic reticulum membranes associated with aging. Bio-chim Biophys Acta 1235: 406–418, 1995.

26. Larsson L and Salviati G. Effects of age on calcium transportactivity of sarcoplasmic reticulum in fast- and slow-twitch ratmuscle fibres. J Physiol (Lond) 419: 253–264, 1989.

27. Liu S, Baracos VE, Quinney HA, and Clandinin MT. Di-etary v-3 and polyunsaturated fatty acids modify fatty acylcomposition and insulin binding in skeletal-muscle sarcolemma.Biochem J 299: 831–837, 1994.

28. Liu S, Baracos VE, Quinney HA, Le Bricon T, and Clan-dinin MT. Parallel insulin-like growth factor I and insulinresistance in muscles of rats fed a high fat diet. Endocrinology136: 3318–3324, 1995.

29. Locher R, Neyses L, Stimpel M, Kuffer B, and Vetter W.The cholesterol content of the human erythrocyte influencescalcium influx through the channel. Biochem Biophys Res Com-mun 124: 822–828, 1984.

30. Mas-Oliva J and Santiago-Garcia J. Cholesterol effect onthermostability of the (Ca21, Mg21)-ATPase from cardiac musclesarcolemma. Biochem Int 21: 233–241, 1990.

31. Needham D and Nunn RS. Elastic deformation and failure oflipid bilayer membranes containing cholesterol. Biophys J 58:997–1009, 1990.

32. Ortega A and Mas-Oliva J. Cholesterol effect on enzyme ac-tivity of the sarcolemmal (Ca21 1 Mg21)-ATPase from cardiacmuscle. Biochim Biophys Acta 773: 231–236, 1984.

33. Pritchard KA, Schwarz SM, Medow MS, and StemermanMB. Effect of low-density lipoprotein on endothelial cell mem-brane fluidity and mononuclear cell attachment. Am J PhysiolCell Physiol 260: C43–C49, 1991.

34. Rosemblatt M, Hidalgo C, Vergara C, and Ikemoto N.Immunological and biochemical properties of transverse tubulemembranes isolated from rabbit skeletal muscle. J Biol Chem256: 8140–8148, 1981.

35. Schroeder RJ, Ahmed SN, Zhu Y, London E, and BrownDA. Cholesterol and sphingolipid enhance the Triton X-100insolubility of glycosylphosphatidylinositol-anchored proteins bypromoting the formation of detergent-insoluble ordered mem-brane domains. J Biol Chem 273: 1150–1157, 1998.

36. Sen L, Bialecki RA, Smith E, Smith TW, and Colucci WS.Cholesterol increases the L-type voltage-sensitive calcium chan-nel current in arterial smooth muscle cells. Circ Res 71: 1008–1014, 1992.

37. Song J and Waugh RE. Bending rigidity of SPOC membranescontaining cholesterol. Biophys J 64: 1967–1970, 1993.

38. Sorisky A, Kucera GL, and Rittenhouse SE. Stimulatedcholesterol-enriched platelets display increased cytosolic Ca21

and phospholipase A activity independent of changes in inositoltrisphosphates and agonist/receptor binding. Biochem J 265:747–754, 1990.

39. Storlien LH, Pan DA, Kriketos AD, O’Connor J, CatersonID, Cooney GJ, Jenkins AB, and Baur LA. Skeletal musclemembrane lipids and insulin resistance. Lipids 31: S261–S265,1996.

40. Striessnig J, Glossmann H, and Catterall WA. Identificationof a phenylalkylamine binding region within the a1-subunit ofskeletal muscle Ca21 channels. Proc Natl Acad Sci USA 87:9108–9112, 1990.

41. Swartz, GM Jr, Gentry MK, Amende LM, Blanchette-Mackie EJ, and Alving CR. Antibodies to cholesterol. ProcNatl Acad Sci USA 85: 1902–1906, 1988.

42. Tomeczkowski J, Ludwig A, and Kretzmer G. Effect ofcholesterol addition on growth kinetics and shear stress sensi-tivity of adherent mammalian cells. Enzyme Microb Technol 15:849–853, 1993.

43. Williams KD and Smith DO. Cholesterol conservation in skel-etal muscle associated with age- and denervation-related atro-phy. Brain Res 493: 14–22, 1989.

44. Yeagle PL. Cholesterol and the cell membrane. Biochim Bio-phys Acta 822: 267–287, 1985.

45. Yu BS and Jo IH. Interaction of sea cucumber saponins withmultilamellar liposomes. Chem Biol Interact 52: 185–202, 1984.

46. Zhou Q, Jimi S, Smith TL, and Kummerow FA. The effect ofcholesterol on the accumulation of intracellular calcium. Bio-chim Biophys Acta 1085: 1–6, 1991.

741IMMUNOLOCALIZATION OF CHOLESTEROL IN MUSCLE