- Email: [email protected]
The effect of prefixation on the quality of vascular corrosion casts of rat heart
ELSEVIER
The Effect of Prefixation on the Quality of Vascular Corrosion Casts of Rat Heart P. A. Reddy, MD: J. E. Douglas, MD,* M. Schulte, MD,* and F. E. Hossler, PhD,+ Deparrments of *Internal Medicine and +Anatomy, J. H. Quillen College of Medicine, East Tennessee State University, Johnson City, Tennessee
To help define the optimal conditions for the preparationof vascularcorrosion castsof rat heart, we examinedthe effect of prefixation with aldehydefixatives on the perfusion rates of rat heart and on the quality of vascular casts.For these studies,beating hearts were removedfrom rats, cannulated via the aortic stump, arrestedwith KCI, perfusedretrogradewith buffered saline or fixative, and infused with resin to preparecorrosion casts.Fixatives usedwere 2.5 % glutaraldehydeor 2% paraformaldehyde,and the castingresin consistedof a Mercox-methylmethacrylatemixture (4:l). All perfusion pressureswere monitored at 80 to 100mm Hg using a mercury manometer.The perfusion rate of control hearts was 13 to 14 mL/min. Prefixation with glutaraldehyde and paraformaldehydereduced perfusion to 8.5 and 8.1 mL/min, respectively. Cast quality was observedgrossly and with the scanningelectronmicroscope.Control heartsyielded high quality, completecastswith 2570 capilIarieslmm2.Castsfrom prefixed heartsexhibited areasof incomplete vesselfilling and resistedcomplete tissue maceration, leaving tissue remnants adhering to the vessel replicas. Casts from glutaraldehyde-fixedheartswere of very poor quality. Our results indicate that prefixation is an unnecessarystep in the preparation of vascular castsof rat heart and is inconsistent with cast quality. -
The need to understand the myocardial circulation at the microvascular level in clarifying the causesand effects of heart disease is clear (1). Myocardial perfusion of the left ventricle is unique, occurring primarily during diastole. The extensive capillary network that closely invests each myocyte is com-
pressedduring myocardialcontraction. In a sense,during systole the left ventricle throttles its own perfusion. Myocardial ischemia and reperfusion in the “thrombolytic era” are conjoined topics of profound interest to clinical cardiologists.
Measuringcapillary responsesto regionalandlocal ischemia, and to reperfusion with and without various pharmacologic
agents(suchasantioxidants,Ca+ + channelblockers,vasodilators, and membranestabilizers), however,is fraught with many procedural hazards.The very processof observing or attempting to documenta perturbation and its consequences risks disrupting the events being examined. Manuscript received June 3, 1994; accepted December 8, 1994. Address for reprints: Dr. John E. Douglas, Department of Internal Medicine, Box 70622, J. H. Quillen College of Medicine, East Tennessee State University, Johnson City, TN 37614; telephone: (615) 926-1171, ext. 2462, fax: (615) 926-1171, ext. 2449. Cardiovascular Pathology Vol. 4, No. 2, April-June 1995:133-140 o 1995 by Elsevier Science Inc. 655 Avenue of the Americas, New York, NY 10010
The applications of vascular corrosion casting in understandingthe microvascularanatomyin health anddiseaseare many (2), and various modifications of the castingtechnique havebeenemployedto improve the quality and completeness of the castsobtained (3). In previous reports we described a methodfor viewing the three-dimensionalmicrovasculararchitecture of the rat heart using vascular corrosion casting (4) anddemonstratedthat quantitativemeasurementson those castsclosely resemblesimilar measurementsmade in vitro (5, 6). Theseresults prompted us to try to develop a model for evaluatingtheeffectsof ischemia,reperfusion,andpharmacologic interventions on capillary integrity and morphology. One controversyregardingmethodologythat hasnot been adequatelyaddressed,however,is whether or not to fix tissueprior to vascularcasting (3,7). Aldehyde fixation causes tissuecontraction (shrinkage),probably asa result of the ability of aldehydesto cross-link proteins (8, 9). In somecases prefixation hasbeenrecommendedfor the preservationof endotbelial imprints (3), to maintain tissueintegrity (lo), to prevent leakagewith low-viscosity resins, or for the preparation of complete casts (7). 1054-8807/95/$9.50 SSDI 1054-8807(94)00047-U
134
REDDY ET AI.. PREFIXATION AND VASCCI.AR
CASTS OF RAT HEART
However, for many applications, especially those involving contractile tissues (e.g., heart and skeletal muscle), completeness of the casts seems to be hampered by tissue contraction induced by fixation (authors’ unpublished observation). In the course of prefixing a potassium-arrested rat heart, we have noted that the moment the fixative enters the coronary arterial bed, the heart appears to contract. Does this indicate that the process of aldehyde fixation could actually be distorting the very capillary bed we would want undisturbed in our ischemiareperfusion model? The general view of many investigators is that the inclusion of an additional fixation step in the casting procedure is unnecessary and greatly lengthens maceration time (discussion at platform session, “Microcorrosion Casting Techniques and Applications” symposium, Scanning Electron Microscopy Society meeting, New Orleans, LA. 1986). In the present study we compare the quality of vascular corrosion casts of rat heart with and without prefixation with aldehydes in an effort to better define the optimal conditions for the preparation of vascular casts of this tissue. This information is preliminary to projected corrosion cast studies to quantitate the myocardial microvasculature following ischemia and reperfusion.
Materials and Methods Heart preparations. Twenty Sprague-Dawley rats (males and females, 200-500 g) were anesthetized with nembutal (13-30 mg) and anticoagulated with heparin (2,000-5,000 U) by intraperitoneal injection 30 minutes before use. Beating hearts were removed through a midsternal incision and immersed in warm saline (0.9% NaCl, 10 mM phosphate, and 15-20 U heparin/mL, pH 7.3,30”(Z). A flared cannula (usually 1.14mm i.d.) attached to a syringe needle (18 gauge) was inserted into the aortic stump and secured with a ligature. The heart was suspended in saline and allowed to beat several more minutes, pumping fluid from the needle’s hub. A syringe barrel was attached to the Luer fitting on the needle, and the heart was flushed by gentle manual pressure, first with 10 mL of heparinized saline and then with 10 mL isotonic KC1 (to arrest the heart). The cannula with the heart attached was secured to a perfusion apparatus fitted with a fluid reservoir, a mercury manometer, and a pressure bulb, (4. 5), and was suspended over a catch basin for further perfusion or corrosion casting as indicated below. perfusion fixation. Heart preparations were perfused with saline for five minutes at a monitored pressure of SO-100 mm Hg and divided randomly into three groups. Group 1 hearts served as unfixed controls. Group 2 hearts were fixed by perfusion with buffered saline containing 2.5 % glutaraldehyde (pH 7.3,368 mosm/L; 37°C) for five minutes. Group 3 hearts were fixed by perfusion with buffered saline containing 2% paraformaldehyde (freshly prepared and filtered before use; pH 7.3, 650 mosm/L: 37°C) for five minutes. As a control for the possible impact of the high osmolarity encountered in the paraformaldehyde fixative, a fourth group of hearts was
i-dlLilOi,i~~ Zaihoi ‘j,“! -: ‘b. ?pr;? -!iinc 1995 1 *) :-Li;
perfused with buffered saline with an ctsmolarity c)( 656 mosm/L Measurement of perfusion rates. Perfusion rates for al i hearts were determined by collecting the effluent for 1 min ute at a monitored pressure of 80 to 100 mm Hg at rime 7c1‘i, (baseline, before beginning tixation), I minute, 3 minuie.. and 5 minutes. Perfusion rates were expressed as mX..:mio Drop counts indicated that there was no difference in pcrfu. sion rates between low- and high-osmolarny saline-perfu:;ed controls. Preparation of vascular corrosion costs. C‘ori-osion cast:, of control and perfusion-fixed hearts were prepared by mfus ing resin as previously described (4). Briefly, low-viscosity resin, prepared just before use. was infused at a pressun: ot 80 to 100 mm Hg until the onset of polymerization i&our 5 -8 minutes) using the perfusion apparatus dercribed above The resin consisted of 4 mL Mercox CL2B (Ladd Reszarclb Industries, Burlington, VT). 1 mL Sevriton dental 3ealan; (Dentsply Ltd.. Weybridge, Surrey, England) or tnrthy methacrylate monomer (Polysciences, Inc., Warrington. PA 1. and 0.2 mL catalyst. Addition of Sevriton or methylmethacry. late monomer to Mercox reduces resin viscosity by about 5C)t’c (4). Resin-filled hearts were immersed in hot water (50. 60°C’ , for 30 minutes to complete resm polymerization. ;eit~oned with razor blades, and cleared of tissue by repeated maccriation in 10% NaOH or KOH, alternated with rinses In dib tilled water. Complete maceration required about one week for unfixed tissue. After maceration. casts were rinsed that oughly in distilled water and air dried or lritical-point d&f from liquid CO? (Samdri-PVT-3B. Tousinus Research Clorp ~ Rockville. MD). Scanning electron microscopy and quantitation ofcorrosion casts. Casts were mounted on stubs with silver past< and sputter coated with gold-palladium (Desk-l sputter coatcr. Denton Vacuum, Inc., Cherry Hill, NJ) for routine scanning electron microscopy using a S-430 (Hitachi Scientific Instru merits, Inc., Mountain View, CA) or a DSM 490 (Carl Z&x. Inc., Thornwood, NY) scanning electron microscope. Capiilary density determinations were made directly from the 1171 croscope screen (with the aid of a measuring cursori. from negatives (with the aid of a photographic enlarger). ckrfront computer prints using a digital image transfer system (Carl Zeiss, Inc.) and Harvard Graphics 3.0 software. Te t&litatc counting, casts were oriented so that capillary beds were viewed in cross section. Five fields were chosen at random from the interventricular septum and left ventricular free wall of each heart for capillary counting and usually ranged in size from 0.05 to 0.3 mm. When necessary, dried casts could he embedded in polyethylene glycol. sectioned on a sliding microtome, cleared of the polyethylene glycol in hot water (50”&?. and mounted for scanning electron microscopy. Statistical methods. Means and standard deviations foi capillary densities and perfusion flow rates for each group at baseline. 1, 3, and 5 minutes of perfusion with saline, p-formaldehyde, or glutaraldehyde. and analysis of variance among the groups were computed with :?ibstat 4 iii
Cardiovasc Pathol Vol. 4, No. 2 April-June 1995:133-140
REDDY ET AL. PREFIXATION AND VASCULAR CASTS OF RAT HEART
135
Table 1. Effect of Aldehyde Fixation on Perfusion Rates of Rat Hearts Perfusion Rate (mL/min -f- SD) Baseline
1 min
3 min
5 min
Saline (5)
14.2 f 0.7
-
-
13.4 * 0.9 @ = 0.13)
P-formaldehyde (5)
13.7 * 0.7
Condition (IV)
10.6 f 1.6
@= Glutaraldehyde (5)
14.1 f 0.5
Results Because perfusion rates are indicative of the patency of the vasculature of an organ, the effect of aldehyde fixation on the perfusion rates of rat heart was measured at a constant perfusion pressure of 80 to 100 mm Hg. Hearts generally swelled very slightly during saline perfusion as the result of filling of the vasculature. After 4 to 6 minutes perfusion rates stabilized and remained constant throughout the Sminute period
0.0043) 9.4 f 1.4 @ = 0.0014)
@=
i 1.4 0.0003) 9.4 f 1.4
@ < 0.ooo1)
@ < O.ocKIl)
@ < 0.ooo1)
10.8
8.5 8.1
i 1.0 *
0.6
of monitoring. Perfusion rates were very similar at 13 to 14 mL/min in all hearts tested prior to fixation. Perfusion rates remained constant in saline-treated hearts during the Sminute test period but continued to decrease in both groups of aldehyde-fixed hearts (Table 1). However, hearts appeared to contract slightly and effluent rates decreased noticeably immediately after beginning the infusion of fixative (Table 1). In paraformaldehyde-perfused hearts, flow rates were 77 % (10.6 mL/min) and 62% (8.5 mL/min) of baseline at 1 and
Figure 1. Vascular corrosion cast preparation of saline-perfused heart prior to maceration (anterior view). Some of the larger, resinfilled coronary vessels (arrows) are evident on the heart surface. Only a few, small dark patches (arrowheads) indicative of incomplete vessel filling can be identified on an otherwise completely blanched surface (Original magnification x 6.)
Figure 2. Vascular corrosion cast preparation of a paraformaldehyde-perfused heart prior to maceration (anterior view). Dark patches (arrowheads) on the heart surface indicate areas of incomplete vessel filling. (Original magnification x 6.)
136
REDDY ET Al.. PREFIXATION AND VASCULAR
CASTS OF RAT HEART
Figure 3. Vascular corrosion cast preparation of glutaraldehyde-
perfusedheart prior to maceration(anterior view). Numerousdark patches(arrowhe&) indicative of incompletevesselfilling are evident on the heart surface. (Original magnification ~6.)
5 minutes, respectively, after beginning fixation. Glutaraidehyde-perfused hearts were 66% (9.4mLimin) and 57% (8.1 mL/min) of baseline at 1 and 5 minutes. respectively. after beginning perfusion. Saline controls remained essentially- un. changed during 5 minutes of perfusion. Fixation with aldehydes also compromised vascuiar car-rosion casting of rat hearts. When saline-perfused hearts were infused with casting resin, a uniform blanching of the myocardium was evident as the resin filled the microvascuiature, changing the normally red-brown tissue to white. The percrstence of red-brown patches, most often near the apex of’ the heart, was indicative of areas of poor resin fiiling. These were rare with saline-perfused hearts, more common but usually widely spacedin paraformaldehyde-perfused hearts, and abundant in glutaraldehyde-perfused hearts (Figs. 1. 2, and 3) Typically, about 1 week was required for complete maccr ation and cleaning of casts from saline-perfused hearts, whereas casts from fixed hearts often required repeated hydroxide-water treatments over a 4-week period for tissue: removal. With glutaraldehyde-fixed hearts tissue remnants were commonly seen adhering to the casts in the scanning electron microscope even after 5 weeks of maceration, Grossly, castsof saline-perfused hearts often appearedcorn plete with few unfilled vessels (Fig. 4), paraformaldehyde-fixed heart casts exhibited some areas of incompiete casting, and glutaraldehyde-fixed heart casts were commonly incomplete, often breaking into small pieces on handling. Microscopically, sections of saline-perfused hearts exhibited dense fields of capillaries within which were interspersed branches of the larger coronary vessels (Figs. 5 and 6). Casts from paraformaldehyde-fixed hearts closely resembled those of saline perfused hearts (Fig. 71, but casts of hearts prefixed with glutaraldehyde were characterized by sparse capillary beds
Figure 4. Midsection through a vascular cast of a saline-perfusedheart after tissue maceration and drving. Note the densemicrovasculature-inall areasindicating essentially complete vessel filling. RVFW. right ventricular free wall; IVS, interventricular septum;LVFW, left ventricular free wall. (Original magnification X 8.2.:
Cardiovasc Pathol Vol. 4, No. 2 April-June 1995:133-140
PREFIXATION
AND VASCULAR
REDDY ET AL. CASTS OF RAT HEART
137
Fire 5. Cross section of a vascular cast of left ventricular free wall from a saline perfusedheart. Shown are capillary beds(C) suppliedby arterioles and venules (arrowheads). (Original magnification X 2 10.)
/
with incomplete filling of vessels and the presence of tissue remnants (Fig. 8). Prefixation with paraformaldehyde reduced density of the capillaries seen on the casts by 14% (Table 2). Prefixation with glutaraldehyde reduced capillary density on the casts much more so, but poor cast quality and integrity caused casts to fall apart on handling and precluded accurate determinations of capillary density.
Fiie 6. High magnificationof vascular castof left ventricular free wall from saline perfused heart. Arrowheads:capillary casts.(Original magnification X 608.)
Discussion The present study demonstrates clearly that prefixation with either of the two most common aldehyde fixatives, paraformaldehyde and glutaraldehyde, reduces the perfusion rate of the heart and compromises the preparation of quality vascular corrosion casts of rat hearts using our procedures. The effects of glutaraldehyde were much more dramatic
138
REDDY ET AL. PREFIXATION AND
VASCULAR
CASTS
OF KM-
HEAR1
Figure 7. Vascular cast of left ventricular free wall from paraforrnaldehyde-perfusedheart. Densecapillary beds are evident throughout the wall, while largercoronaryvesselsare mostprevalentin theepicardium(EP) EN: endocardium;Tc: vasculatureot trabecula carneae. (Original magnification X41.)
Figure 8. High magnification of vascular castof left ventricular free wall from glutaraldehyde-perfusedheart. C: capillary casts; arrowheadsindicateincompletelytigestedtissueremnants.(Or&al magnification X832.1
than those of paraformaldehyde. Casts from paraformaldehyde-fixed hearts were intermediate in quality between those perfused with only saline and those prefixed with glutaraldehyde but were far superior to the latter. This is in good agree-
Table 2. Effect of Fixation on Capillary Corrosion
Casts
Condition (NJ Saline (25) P-formaldehyde (24) Glutaraldehyde
Capillaries/mm’
Density -____i SD
2570.8 + 495.4 2218.3 + 352.8
in % of Control 100% 86%
ment with similar studies on the effects of fixation on corrosion casts of lung by Schraufnagel and Schmid (7). Results with formaldehyde were intermediate between those seen with glutaraldehyde and unfixed controls with regard to resin leakage and vessel density and were explained by the differential ability of the two fixatives to cross-link proteins. This is, in fact, believed to be the basis for the advantage glutaraldehyde has over formaldehyde as a fixative for electron microscopy (9). That is to say, cast quality is inversely related to the degree of tissue cross-linking in these two studies. Although it was difficult to quantitate, we did observe con traction (or shrinkage) of the hearts during glutaraldehyde per-
Cardiovasc Path01Vol. 4, No. 2 April-June 1995:133-140
fusion. Because each myocardial cell is surrounded on the average by four or five capillaries (11, 12, 13), cross-linking of the abundant contractile proteins in these cells might result in compression of the capillaries and the observed reduction in perfusion rate and increase in resistance to resin filling. Prefixation also greatly extended the time required for alkali to clear tissue from the casts (perhaps fivefold). Although this has not been reported to be a problem for many tissues (3, 14), clearing of lung casts following glutaraldehyde fixation, for example, was found to be incomplete even after 100 days of maceration (7). In the latter cases, the authors suggested that this may be attributable to the trapping of tissue in glutaraldehyde polymers formed under alkaline conditions (8). Glutaraldehyde-induced increases in maceration time are thus not limited to the heart. Although Hodde et al. (14) could detect no difference in corrosion time between fixed and unfixed tissue, their corrosion protocol included treatment with Triton X-100 detergent, which may have facilitated the alkali digestion. One advantage of prefixation with glutaraldehyde is that it may prevent vessel leakage when low-viscosity resins are used (7) or when tissues are especially fragile (3). Most previous corrosion cast studies of the microvasculature of heart have utilized unfixed tissue with satisfactory results (15-21). In fact, the especially lucid casts obtained by Potter and Groom (21), for example, were prepared by injecting resin into beating hearts, the contractions of which were believed to enhance vessel filling. However, a few studies using prefixed tissue also demonstrated satisfactory corrosion casts of myocardium. Phillips et al. (22) prefixed hearts with a combination of paraformaldehyde and glutaraldehyde, but they did report some problems with incomplete filling of vessels and noted that successful resin injections were not routine. In the present study the effect of prefixation with combinations of aldehydes was not tested. In efforts to identify functional capillaries in rabbit and rat hearts, Sage and Gavin (12, 13, 23) prefixed hearts with glutaraldehyde and prepared casts using low-viscosity L. R. White resin. In rabbits they calculated 53 % to 57% capillary filling and, in rats, 62% filling. In the latter study, vasodilation with procaine prior to fixation increased the filling to 95 % . Because it was important to view casted vessels in intact myocardium, these investigators did not remove surrounding tissue. If a capillary density of 3,345/mm2 is accepted for rat heart (11,24), then the present study obtained 77 % vessel filling in unfixed hearts and 66 % filling in paraformaldehydefixed hearts. This latter result is in the range of that reported by Sage and Gavin (12) using glutaraldehyde in the absence of vasodilators. However, the two studies are not easily compared because we did not test the effect of vasodilators, recognizing and reserving this as a potential intervention to be used in ischemia and reperfusion studies. In addition, we could not obtain a sufficient number of intact casts suitable for vessel counting following glutaraldehyde fixation. We conclude that prefixation before corrosion casting is an extra and unnecessary procedure that may be detrimental
REDDY ET AL. PREFIXATION AND VASCULAR CASTS OF RAT HEART
139
to the quality of casts, at least in the case of myocardium, and should be omitted unless special requirements (e.g., preservation of tissue integrity or prevention of leakage of lowviscosity resins; seeLametschwandtner et al. [3]) of the study dictate otherwise. The authors wish to exuress their anoreciation to Dr. Steven D. Richards for his assistance with higital image-processing and statistical analysis. -
References 1. Horan MJ, Steinberg GM, Dunbar JB, Hadley EC. Summary of NIH workshop on blood pressure regulation and aging. Hypertension 1986;8:178-180. 2. Konerding MA. Scanning electron microscopy of corrosion casting in medicine. Scanning Microsc 1991;5:851-865. 3. Lametschwandtner A, Lametschwandtner U, Weiger T. Scanning electron microscopy of vascular corrosion casts-technique and applications: updated review. Scanning Microsc 1990;4:889-941. 4. Hossler FE, Douglas JE, Douglas LE. Anatomy and morphometry of myocardial capillaries studied with vascular corrosion casting and scanning electron microscopy: a method for rat heart. Scan Electron Microsc 1986;4:1469-1475, 5. Hossler FE, Douglas JE, Verghese A, Neal L. Microvascular architecture of the elastase emphysemic hamster lung. J Electron Microsc Tech 1991;9:406-418. 6. Hossler FE. Some quantitative applications of vascular corrosion casting. In: Bailey GW, Bentley J, Small JA, eds. Proceedings of the 50th Annual Meeting of the Electron Microscopy Society of America. 1992:738-739. 7. Schraufnagel DE, Schmid A. Microvascular casting of the lung: effects of various fixation protocols. J Electron Microsc Tech 1988;8: 185-191. 8. Hayat MA. Principles and Techniques of Electron Microscopy: Biological Applications (3rd ed). Boca Raton, FL: CRC Press. 1989:24-37. 9. Sabatini DD, Bensch K, Barrnett RJ. Cytochemistry and electron microscopy: the preservation of cellular structure and enzymatic activity by aldehyde fixation. J Cell Biol 1963;17:19-58. 10. Kratky RG, Zeindler CM, Lo DKC, Roach MR. Quantitative measurements from vascular casts. Scanning Microsc 1989;3:937-943. 11. Gerdes AM, Callas G, Kasten FH. Differences in regional capillary distribution and myocyte sizes in normal and hypertrophic rat hearts. Am J Anat 1979;156:523-532. 12. Sage MD, Gavin JB. Morphological identification of functional capillaries in the myocardium. Anat Ret 1984;208:283-289. 13. Sage MD, Gavin JB. Scanning electron microscopy of heart muscle freeze-dried from dimethylsulfoxide for simultaneous demonstration of cell morphology and microvascular function. Stain Technol 1986; 61~261-267. 14. Hodde KC, Steeber DA, Albrecht RM. Advances in corrosion casting methods. Scanning Microsc 1990;4:693-704. 15. Anderson BG, Anderson WD. Microvasculature of the canine heart demonstrated by scanning electron microscopy. Am J Anat 1980; 158:217-227. 16. Anderson BG, Anderson WD. Myocardial microvasculature studied by microcorrosion casts. Biomed Res 1981;2(Suppl):209-217. 17. Irino S, Ono T, Shimohara Y. Microvascular architecture of the rabbit ventricular walls: a scanning electron microscopic study of corrosion casts. Scan Electron Microsc 1982;4: 1785-1792. 18. Kajihara H, Yamahara M, Malliwah JAM, Maeda Y. Three-dimensional vascular architecture of dog heart as revealed by injection replica scanning electron microscopy. Hiroshima J Med Sci 1985;34:165-171. 19. Lametschwandtner A, Mohl W. The microcirculatory vascular bed of the dog’s heart: a scanning electron microscopy study of vascular cor-
140
REDDY ET AL. PREFIXATION AND VASCULAR
rosion casts. In: Mohl W. ed. The Coronary Sinus. Darmstadt: SpringerVerlag, 1984:26-32. 20. Ono T, Shimohara Y, Okada K, Irino S. Scanning electron microscopic studies on microvascular architecture of human coronary vessels by corrosion casts: normal and focal necrosis. Scan Electron Microsc 1986; 1:263-270. 2 1. Potter RF, Groom AC. Capillary diameter and geometry in cardiac and skeletal muscle studied by means of corrosion casts. Microvasc Res 1983;25:68-84. 22. Phillips Sl, Rosenberg A, Meir-Levi
f’ard~ovasc l’athul Vol. -1. No. 2 April -1une 19W111.140
CASTS OF RAT HEART
D, Pappas E. Visualization of the
coronary microvascular bed by light and scanning electron mictoucop) and x-ray in the mammalian heart. Scan Electron Micros? !979. 3~735-742.
23. Sage MD, Gavin JB. Microvascular function ar the margms of earl) experimental myocardial infarcts in isolated rabbit hearts. Heart Vcs. sels 1986;2:81-86. 24. Henquell L, Odoroff CL, Honig CR. Intercapillary distance and cap& lary reserve in hypertrophied rat hearts beating in situ. Circ Res 1977: 41:400-408.
The Effect of Prefixation on the Quality of Vascular Corrosion Casts of Rat Heart P. A. Reddy, MD: J. E. Douglas, MD,* M. Schulte, MD,* and F. E. Hossler, PhD,+ Deparrments of *Internal Medicine and +Anatomy, J. H. Quillen College of Medicine, East Tennessee State University, Johnson City, Tennessee
To help define the optimal conditions for the preparationof vascularcorrosion castsof rat heart, we examinedthe effect of prefixation with aldehydefixatives on the perfusion rates of rat heart and on the quality of vascular casts.For these studies,beating hearts were removedfrom rats, cannulated via the aortic stump, arrestedwith KCI, perfusedretrogradewith buffered saline or fixative, and infused with resin to preparecorrosion casts.Fixatives usedwere 2.5 % glutaraldehydeor 2% paraformaldehyde,and the castingresin consistedof a Mercox-methylmethacrylatemixture (4:l). All perfusion pressureswere monitored at 80 to 100mm Hg using a mercury manometer.The perfusion rate of control hearts was 13 to 14 mL/min. Prefixation with glutaraldehyde and paraformaldehydereduced perfusion to 8.5 and 8.1 mL/min, respectively. Cast quality was observedgrossly and with the scanningelectronmicroscope.Control heartsyielded high quality, completecastswith 2570 capilIarieslmm2.Castsfrom prefixed heartsexhibited areasof incomplete vesselfilling and resistedcomplete tissue maceration, leaving tissue remnants adhering to the vessel replicas. Casts from glutaraldehyde-fixedheartswere of very poor quality. Our results indicate that prefixation is an unnecessarystep in the preparation of vascular castsof rat heart and is inconsistent with cast quality. -
The need to understand the myocardial circulation at the microvascular level in clarifying the causesand effects of heart disease is clear (1). Myocardial perfusion of the left ventricle is unique, occurring primarily during diastole. The extensive capillary network that closely invests each myocyte is com-
pressedduring myocardialcontraction. In a sense,during systole the left ventricle throttles its own perfusion. Myocardial ischemia and reperfusion in the “thrombolytic era” are conjoined topics of profound interest to clinical cardiologists.
Measuringcapillary responsesto regionalandlocal ischemia, and to reperfusion with and without various pharmacologic
agents(suchasantioxidants,Ca+ + channelblockers,vasodilators, and membranestabilizers), however,is fraught with many procedural hazards.The very processof observing or attempting to documenta perturbation and its consequences risks disrupting the events being examined. Manuscript received June 3, 1994; accepted December 8, 1994. Address for reprints: Dr. John E. Douglas, Department of Internal Medicine, Box 70622, J. H. Quillen College of Medicine, East Tennessee State University, Johnson City, TN 37614; telephone: (615) 926-1171, ext. 2462, fax: (615) 926-1171, ext. 2449. Cardiovascular Pathology Vol. 4, No. 2, April-June 1995:133-140 o 1995 by Elsevier Science Inc. 655 Avenue of the Americas, New York, NY 10010
The applications of vascular corrosion casting in understandingthe microvascularanatomyin health anddiseaseare many (2), and various modifications of the castingtechnique havebeenemployedto improve the quality and completeness of the castsobtained (3). In previous reports we described a methodfor viewing the three-dimensionalmicrovasculararchitecture of the rat heart using vascular corrosion casting (4) anddemonstratedthat quantitativemeasurementson those castsclosely resemblesimilar measurementsmade in vitro (5, 6). Theseresults prompted us to try to develop a model for evaluatingtheeffectsof ischemia,reperfusion,andpharmacologic interventions on capillary integrity and morphology. One controversyregardingmethodologythat hasnot been adequatelyaddressed,however,is whether or not to fix tissueprior to vascularcasting (3,7). Aldehyde fixation causes tissuecontraction (shrinkage),probably asa result of the ability of aldehydesto cross-link proteins (8, 9). In somecases prefixation hasbeenrecommendedfor the preservationof endotbelial imprints (3), to maintain tissueintegrity (lo), to prevent leakagewith low-viscosity resins, or for the preparation of complete casts (7). 1054-8807/95/$9.50 SSDI 1054-8807(94)00047-U
134
REDDY ET AI.. PREFIXATION AND VASCCI.AR
CASTS OF RAT HEART
However, for many applications, especially those involving contractile tissues (e.g., heart and skeletal muscle), completeness of the casts seems to be hampered by tissue contraction induced by fixation (authors’ unpublished observation). In the course of prefixing a potassium-arrested rat heart, we have noted that the moment the fixative enters the coronary arterial bed, the heart appears to contract. Does this indicate that the process of aldehyde fixation could actually be distorting the very capillary bed we would want undisturbed in our ischemiareperfusion model? The general view of many investigators is that the inclusion of an additional fixation step in the casting procedure is unnecessary and greatly lengthens maceration time (discussion at platform session, “Microcorrosion Casting Techniques and Applications” symposium, Scanning Electron Microscopy Society meeting, New Orleans, LA. 1986). In the present study we compare the quality of vascular corrosion casts of rat heart with and without prefixation with aldehydes in an effort to better define the optimal conditions for the preparation of vascular casts of this tissue. This information is preliminary to projected corrosion cast studies to quantitate the myocardial microvasculature following ischemia and reperfusion.
Materials and Methods Heart preparations. Twenty Sprague-Dawley rats (males and females, 200-500 g) were anesthetized with nembutal (13-30 mg) and anticoagulated with heparin (2,000-5,000 U) by intraperitoneal injection 30 minutes before use. Beating hearts were removed through a midsternal incision and immersed in warm saline (0.9% NaCl, 10 mM phosphate, and 15-20 U heparin/mL, pH 7.3,30”(Z). A flared cannula (usually 1.14mm i.d.) attached to a syringe needle (18 gauge) was inserted into the aortic stump and secured with a ligature. The heart was suspended in saline and allowed to beat several more minutes, pumping fluid from the needle’s hub. A syringe barrel was attached to the Luer fitting on the needle, and the heart was flushed by gentle manual pressure, first with 10 mL of heparinized saline and then with 10 mL isotonic KC1 (to arrest the heart). The cannula with the heart attached was secured to a perfusion apparatus fitted with a fluid reservoir, a mercury manometer, and a pressure bulb, (4. 5), and was suspended over a catch basin for further perfusion or corrosion casting as indicated below. perfusion fixation. Heart preparations were perfused with saline for five minutes at a monitored pressure of SO-100 mm Hg and divided randomly into three groups. Group 1 hearts served as unfixed controls. Group 2 hearts were fixed by perfusion with buffered saline containing 2.5 % glutaraldehyde (pH 7.3,368 mosm/L; 37°C) for five minutes. Group 3 hearts were fixed by perfusion with buffered saline containing 2% paraformaldehyde (freshly prepared and filtered before use; pH 7.3, 650 mosm/L: 37°C) for five minutes. As a control for the possible impact of the high osmolarity encountered in the paraformaldehyde fixative, a fourth group of hearts was
i-dlLilOi,i~~ Zaihoi ‘j,“! -: ‘b. ?pr;? -!iinc 1995 1 *) :-Li;
perfused with buffered saline with an ctsmolarity c)( 656 mosm/L Measurement of perfusion rates. Perfusion rates for al i hearts were determined by collecting the effluent for 1 min ute at a monitored pressure of 80 to 100 mm Hg at rime 7c1‘i, (baseline, before beginning tixation), I minute, 3 minuie.. and 5 minutes. Perfusion rates were expressed as mX..:mio Drop counts indicated that there was no difference in pcrfu. sion rates between low- and high-osmolarny saline-perfu:;ed controls. Preparation of vascular corrosion costs. C‘ori-osion cast:, of control and perfusion-fixed hearts were prepared by mfus ing resin as previously described (4). Briefly, low-viscosity resin, prepared just before use. was infused at a pressun: ot 80 to 100 mm Hg until the onset of polymerization i&our 5 -8 minutes) using the perfusion apparatus dercribed above The resin consisted of 4 mL Mercox CL2B (Ladd Reszarclb Industries, Burlington, VT). 1 mL Sevriton dental 3ealan; (Dentsply Ltd.. Weybridge, Surrey, England) or tnrthy methacrylate monomer (Polysciences, Inc., Warrington. PA 1. and 0.2 mL catalyst. Addition of Sevriton or methylmethacry. late monomer to Mercox reduces resin viscosity by about 5C)t’c (4). Resin-filled hearts were immersed in hot water (50. 60°C’ , for 30 minutes to complete resm polymerization. ;eit~oned with razor blades, and cleared of tissue by repeated maccriation in 10% NaOH or KOH, alternated with rinses In dib tilled water. Complete maceration required about one week for unfixed tissue. After maceration. casts were rinsed that oughly in distilled water and air dried or lritical-point d&f from liquid CO? (Samdri-PVT-3B. Tousinus Research Clorp ~ Rockville. MD). Scanning electron microscopy and quantitation ofcorrosion casts. Casts were mounted on stubs with silver past< and sputter coated with gold-palladium (Desk-l sputter coatcr. Denton Vacuum, Inc., Cherry Hill, NJ) for routine scanning electron microscopy using a S-430 (Hitachi Scientific Instru merits, Inc., Mountain View, CA) or a DSM 490 (Carl Z&x. Inc., Thornwood, NY) scanning electron microscope. Capiilary density determinations were made directly from the 1171 croscope screen (with the aid of a measuring cursori. from negatives (with the aid of a photographic enlarger). ckrfront computer prints using a digital image transfer system (Carl Zeiss, Inc.) and Harvard Graphics 3.0 software. Te t&litatc counting, casts were oriented so that capillary beds were viewed in cross section. Five fields were chosen at random from the interventricular septum and left ventricular free wall of each heart for capillary counting and usually ranged in size from 0.05 to 0.3 mm. When necessary, dried casts could he embedded in polyethylene glycol. sectioned on a sliding microtome, cleared of the polyethylene glycol in hot water (50”&?. and mounted for scanning electron microscopy. Statistical methods. Means and standard deviations foi capillary densities and perfusion flow rates for each group at baseline. 1, 3, and 5 minutes of perfusion with saline, p-formaldehyde, or glutaraldehyde. and analysis of variance among the groups were computed with :?ibstat 4 iii
Cardiovasc Pathol Vol. 4, No. 2 April-June 1995:133-140
REDDY ET AL. PREFIXATION AND VASCULAR CASTS OF RAT HEART
135
Table 1. Effect of Aldehyde Fixation on Perfusion Rates of Rat Hearts Perfusion Rate (mL/min -f- SD) Baseline
1 min
3 min
5 min
Saline (5)
14.2 f 0.7
-
-
13.4 * 0.9 @ = 0.13)
P-formaldehyde (5)
13.7 * 0.7
Condition (IV)
10.6 f 1.6
@= Glutaraldehyde (5)
14.1 f 0.5
Results Because perfusion rates are indicative of the patency of the vasculature of an organ, the effect of aldehyde fixation on the perfusion rates of rat heart was measured at a constant perfusion pressure of 80 to 100 mm Hg. Hearts generally swelled very slightly during saline perfusion as the result of filling of the vasculature. After 4 to 6 minutes perfusion rates stabilized and remained constant throughout the Sminute period
0.0043) 9.4 f 1.4 @ = 0.0014)
@=
i 1.4 0.0003) 9.4 f 1.4
@ < 0.ooo1)
@ < O.ocKIl)
@ < 0.ooo1)
10.8
8.5 8.1
i 1.0 *
0.6
of monitoring. Perfusion rates were very similar at 13 to 14 mL/min in all hearts tested prior to fixation. Perfusion rates remained constant in saline-treated hearts during the Sminute test period but continued to decrease in both groups of aldehyde-fixed hearts (Table 1). However, hearts appeared to contract slightly and effluent rates decreased noticeably immediately after beginning the infusion of fixative (Table 1). In paraformaldehyde-perfused hearts, flow rates were 77 % (10.6 mL/min) and 62% (8.5 mL/min) of baseline at 1 and
Figure 1. Vascular corrosion cast preparation of saline-perfused heart prior to maceration (anterior view). Some of the larger, resinfilled coronary vessels (arrows) are evident on the heart surface. Only a few, small dark patches (arrowheads) indicative of incomplete vessel filling can be identified on an otherwise completely blanched surface (Original magnification x 6.)
Figure 2. Vascular corrosion cast preparation of a paraformaldehyde-perfused heart prior to maceration (anterior view). Dark patches (arrowheads) on the heart surface indicate areas of incomplete vessel filling. (Original magnification x 6.)
136
REDDY ET Al.. PREFIXATION AND VASCULAR
CASTS OF RAT HEART
Figure 3. Vascular corrosion cast preparation of glutaraldehyde-
perfusedheart prior to maceration(anterior view). Numerousdark patches(arrowhe&) indicative of incompletevesselfilling are evident on the heart surface. (Original magnification ~6.)
5 minutes, respectively, after beginning fixation. Glutaraidehyde-perfused hearts were 66% (9.4mLimin) and 57% (8.1 mL/min) of baseline at 1 and 5 minutes. respectively. after beginning perfusion. Saline controls remained essentially- un. changed during 5 minutes of perfusion. Fixation with aldehydes also compromised vascuiar car-rosion casting of rat hearts. When saline-perfused hearts were infused with casting resin, a uniform blanching of the myocardium was evident as the resin filled the microvascuiature, changing the normally red-brown tissue to white. The percrstence of red-brown patches, most often near the apex of’ the heart, was indicative of areas of poor resin fiiling. These were rare with saline-perfused hearts, more common but usually widely spacedin paraformaldehyde-perfused hearts, and abundant in glutaraldehyde-perfused hearts (Figs. 1. 2, and 3) Typically, about 1 week was required for complete maccr ation and cleaning of casts from saline-perfused hearts, whereas casts from fixed hearts often required repeated hydroxide-water treatments over a 4-week period for tissue: removal. With glutaraldehyde-fixed hearts tissue remnants were commonly seen adhering to the casts in the scanning electron microscope even after 5 weeks of maceration, Grossly, castsof saline-perfused hearts often appearedcorn plete with few unfilled vessels (Fig. 4), paraformaldehyde-fixed heart casts exhibited some areas of incompiete casting, and glutaraldehyde-fixed heart casts were commonly incomplete, often breaking into small pieces on handling. Microscopically, sections of saline-perfused hearts exhibited dense fields of capillaries within which were interspersed branches of the larger coronary vessels (Figs. 5 and 6). Casts from paraformaldehyde-fixed hearts closely resembled those of saline perfused hearts (Fig. 71, but casts of hearts prefixed with glutaraldehyde were characterized by sparse capillary beds
Figure 4. Midsection through a vascular cast of a saline-perfusedheart after tissue maceration and drving. Note the densemicrovasculature-inall areasindicating essentially complete vessel filling. RVFW. right ventricular free wall; IVS, interventricular septum;LVFW, left ventricular free wall. (Original magnification X 8.2.:
Cardiovasc Pathol Vol. 4, No. 2 April-June 1995:133-140
PREFIXATION
AND VASCULAR
REDDY ET AL. CASTS OF RAT HEART
137
Fire 5. Cross section of a vascular cast of left ventricular free wall from a saline perfusedheart. Shown are capillary beds(C) suppliedby arterioles and venules (arrowheads). (Original magnification X 2 10.)
/
with incomplete filling of vessels and the presence of tissue remnants (Fig. 8). Prefixation with paraformaldehyde reduced density of the capillaries seen on the casts by 14% (Table 2). Prefixation with glutaraldehyde reduced capillary density on the casts much more so, but poor cast quality and integrity caused casts to fall apart on handling and precluded accurate determinations of capillary density.
Fiie 6. High magnificationof vascular castof left ventricular free wall from saline perfused heart. Arrowheads:capillary casts.(Original magnification X 608.)
Discussion The present study demonstrates clearly that prefixation with either of the two most common aldehyde fixatives, paraformaldehyde and glutaraldehyde, reduces the perfusion rate of the heart and compromises the preparation of quality vascular corrosion casts of rat hearts using our procedures. The effects of glutaraldehyde were much more dramatic
138
REDDY ET AL. PREFIXATION AND
VASCULAR
CASTS
OF KM-
HEAR1
Figure 7. Vascular cast of left ventricular free wall from paraforrnaldehyde-perfusedheart. Densecapillary beds are evident throughout the wall, while largercoronaryvesselsare mostprevalentin theepicardium(EP) EN: endocardium;Tc: vasculatureot trabecula carneae. (Original magnification X41.)
Figure 8. High magnification of vascular castof left ventricular free wall from glutaraldehyde-perfusedheart. C: capillary casts; arrowheadsindicateincompletelytigestedtissueremnants.(Or&al magnification X832.1
than those of paraformaldehyde. Casts from paraformaldehyde-fixed hearts were intermediate in quality between those perfused with only saline and those prefixed with glutaraldehyde but were far superior to the latter. This is in good agree-
Table 2. Effect of Fixation on Capillary Corrosion
Casts
Condition (NJ Saline (25) P-formaldehyde (24) Glutaraldehyde
Capillaries/mm’
Density -____i SD
2570.8 + 495.4 2218.3 + 352.8
in % of Control 100% 86%
ment with similar studies on the effects of fixation on corrosion casts of lung by Schraufnagel and Schmid (7). Results with formaldehyde were intermediate between those seen with glutaraldehyde and unfixed controls with regard to resin leakage and vessel density and were explained by the differential ability of the two fixatives to cross-link proteins. This is, in fact, believed to be the basis for the advantage glutaraldehyde has over formaldehyde as a fixative for electron microscopy (9). That is to say, cast quality is inversely related to the degree of tissue cross-linking in these two studies. Although it was difficult to quantitate, we did observe con traction (or shrinkage) of the hearts during glutaraldehyde per-
Cardiovasc Path01Vol. 4, No. 2 April-June 1995:133-140
fusion. Because each myocardial cell is surrounded on the average by four or five capillaries (11, 12, 13), cross-linking of the abundant contractile proteins in these cells might result in compression of the capillaries and the observed reduction in perfusion rate and increase in resistance to resin filling. Prefixation also greatly extended the time required for alkali to clear tissue from the casts (perhaps fivefold). Although this has not been reported to be a problem for many tissues (3, 14), clearing of lung casts following glutaraldehyde fixation, for example, was found to be incomplete even after 100 days of maceration (7). In the latter cases, the authors suggested that this may be attributable to the trapping of tissue in glutaraldehyde polymers formed under alkaline conditions (8). Glutaraldehyde-induced increases in maceration time are thus not limited to the heart. Although Hodde et al. (14) could detect no difference in corrosion time between fixed and unfixed tissue, their corrosion protocol included treatment with Triton X-100 detergent, which may have facilitated the alkali digestion. One advantage of prefixation with glutaraldehyde is that it may prevent vessel leakage when low-viscosity resins are used (7) or when tissues are especially fragile (3). Most previous corrosion cast studies of the microvasculature of heart have utilized unfixed tissue with satisfactory results (15-21). In fact, the especially lucid casts obtained by Potter and Groom (21), for example, were prepared by injecting resin into beating hearts, the contractions of which were believed to enhance vessel filling. However, a few studies using prefixed tissue also demonstrated satisfactory corrosion casts of myocardium. Phillips et al. (22) prefixed hearts with a combination of paraformaldehyde and glutaraldehyde, but they did report some problems with incomplete filling of vessels and noted that successful resin injections were not routine. In the present study the effect of prefixation with combinations of aldehydes was not tested. In efforts to identify functional capillaries in rabbit and rat hearts, Sage and Gavin (12, 13, 23) prefixed hearts with glutaraldehyde and prepared casts using low-viscosity L. R. White resin. In rabbits they calculated 53 % to 57% capillary filling and, in rats, 62% filling. In the latter study, vasodilation with procaine prior to fixation increased the filling to 95 % . Because it was important to view casted vessels in intact myocardium, these investigators did not remove surrounding tissue. If a capillary density of 3,345/mm2 is accepted for rat heart (11,24), then the present study obtained 77 % vessel filling in unfixed hearts and 66 % filling in paraformaldehydefixed hearts. This latter result is in the range of that reported by Sage and Gavin (12) using glutaraldehyde in the absence of vasodilators. However, the two studies are not easily compared because we did not test the effect of vasodilators, recognizing and reserving this as a potential intervention to be used in ischemia and reperfusion studies. In addition, we could not obtain a sufficient number of intact casts suitable for vessel counting following glutaraldehyde fixation. We conclude that prefixation before corrosion casting is an extra and unnecessary procedure that may be detrimental
REDDY ET AL. PREFIXATION AND VASCULAR CASTS OF RAT HEART
139
to the quality of casts, at least in the case of myocardium, and should be omitted unless special requirements (e.g., preservation of tissue integrity or prevention of leakage of lowviscosity resins; seeLametschwandtner et al. [3]) of the study dictate otherwise. The authors wish to exuress their anoreciation to Dr. Steven D. Richards for his assistance with higital image-processing and statistical analysis. -
References 1. Horan MJ, Steinberg GM, Dunbar JB, Hadley EC. Summary of NIH workshop on blood pressure regulation and aging. Hypertension 1986;8:178-180. 2. Konerding MA. Scanning electron microscopy of corrosion casting in medicine. Scanning Microsc 1991;5:851-865. 3. Lametschwandtner A, Lametschwandtner U, Weiger T. Scanning electron microscopy of vascular corrosion casts-technique and applications: updated review. Scanning Microsc 1990;4:889-941. 4. Hossler FE, Douglas JE, Douglas LE. Anatomy and morphometry of myocardial capillaries studied with vascular corrosion casting and scanning electron microscopy: a method for rat heart. Scan Electron Microsc 1986;4:1469-1475, 5. Hossler FE, Douglas JE, Verghese A, Neal L. Microvascular architecture of the elastase emphysemic hamster lung. J Electron Microsc Tech 1991;9:406-418. 6. Hossler FE. Some quantitative applications of vascular corrosion casting. In: Bailey GW, Bentley J, Small JA, eds. Proceedings of the 50th Annual Meeting of the Electron Microscopy Society of America. 1992:738-739. 7. Schraufnagel DE, Schmid A. Microvascular casting of the lung: effects of various fixation protocols. J Electron Microsc Tech 1988;8: 185-191. 8. Hayat MA. Principles and Techniques of Electron Microscopy: Biological Applications (3rd ed). Boca Raton, FL: CRC Press. 1989:24-37. 9. Sabatini DD, Bensch K, Barrnett RJ. Cytochemistry and electron microscopy: the preservation of cellular structure and enzymatic activity by aldehyde fixation. J Cell Biol 1963;17:19-58. 10. Kratky RG, Zeindler CM, Lo DKC, Roach MR. Quantitative measurements from vascular casts. Scanning Microsc 1989;3:937-943. 11. Gerdes AM, Callas G, Kasten FH. Differences in regional capillary distribution and myocyte sizes in normal and hypertrophic rat hearts. Am J Anat 1979;156:523-532. 12. Sage MD, Gavin JB. Morphological identification of functional capillaries in the myocardium. Anat Ret 1984;208:283-289. 13. Sage MD, Gavin JB. Scanning electron microscopy of heart muscle freeze-dried from dimethylsulfoxide for simultaneous demonstration of cell morphology and microvascular function. Stain Technol 1986; 61~261-267. 14. Hodde KC, Steeber DA, Albrecht RM. Advances in corrosion casting methods. Scanning Microsc 1990;4:693-704. 15. Anderson BG, Anderson WD. Microvasculature of the canine heart demonstrated by scanning electron microscopy. Am J Anat 1980; 158:217-227. 16. Anderson BG, Anderson WD. Myocardial microvasculature studied by microcorrosion casts. Biomed Res 1981;2(Suppl):209-217. 17. Irino S, Ono T, Shimohara Y. Microvascular architecture of the rabbit ventricular walls: a scanning electron microscopic study of corrosion casts. Scan Electron Microsc 1982;4: 1785-1792. 18. Kajihara H, Yamahara M, Malliwah JAM, Maeda Y. Three-dimensional vascular architecture of dog heart as revealed by injection replica scanning electron microscopy. Hiroshima J Med Sci 1985;34:165-171. 19. Lametschwandtner A, Mohl W. The microcirculatory vascular bed of the dog’s heart: a scanning electron microscopy study of vascular cor-
140
REDDY ET AL. PREFIXATION AND VASCULAR
rosion casts. In: Mohl W. ed. The Coronary Sinus. Darmstadt: SpringerVerlag, 1984:26-32. 20. Ono T, Shimohara Y, Okada K, Irino S. Scanning electron microscopic studies on microvascular architecture of human coronary vessels by corrosion casts: normal and focal necrosis. Scan Electron Microsc 1986; 1:263-270. 2 1. Potter RF, Groom AC. Capillary diameter and geometry in cardiac and skeletal muscle studied by means of corrosion casts. Microvasc Res 1983;25:68-84. 22. Phillips Sl, Rosenberg A, Meir-Levi
f’ard~ovasc l’athul Vol. -1. No. 2 April -1une 19W111.140
CASTS OF RAT HEART
D, Pappas E. Visualization of the
coronary microvascular bed by light and scanning electron mictoucop) and x-ray in the mammalian heart. Scan Electron Micros? !979. 3~735-742.
23. Sage MD, Gavin JB. Microvascular function ar the margms of earl) experimental myocardial infarcts in isolated rabbit hearts. Heart Vcs. sels 1986;2:81-86. 24. Henquell L, Odoroff CL, Honig CR. Intercapillary distance and cap& lary reserve in hypertrophied rat hearts beating in situ. Circ Res 1977: 41:400-408.