JOURNAL OF ELECTRON MICROSCOPY TECHNIQUE 14:39-45 (1990)
Morphology of Rat Exocrine Pancreas Prepared by Anhydrous Cryo-Procedures NORBERT ROOS, ULLA KINDE, AND JOHN A. MORGAN Electroi,microscopical Unit for Biological Sciences, University o f Oslo, Blindern N-0316 Oslo 3, Norway (N.R., U.K.) and Department of Zoology, University College, Cardiff C F l I X L , United Kingdom (J.A.M.1
KEY WORDS substitution
Exocrine pancreas, Cryofixation, Cryomicrotomy, Freeze-drying, Freeze-
ABSTRACT This paper describes and compares the morphology of a relatively complex tissue, the exocrine pancreas, prepared by state-of-the-art anhydrous cryoprocedures. Cryopreparative procedures are being used increasingly for a wide range of applications, for example, electron-probe x-ray microanalysis and immunocytochemical localization of antigenic molecules, because they preserve the composition of the specimen better than procedures involving aqueous media. Some doubts have remained concerning the morphology of cryosections and the precise identification of subcellular structures. We show that thin and sufficiently large cryosections of fresh biological tissues can be produced using commercially available hardware. The freeze-dried cryosections display high intrinsic contrast, are stable under the beam, and allow identification of intracellular fine structure. INTRODUCTION The advantages of cryopreparative procedures for the morphological and subcellular locational study of elements have been advocated persuasively by several authors (See Zierold, 1984, for references). An improved understanding of the physical processes involvec. in freezing (Dubochet et al., 1982; Escaig, 1982; Ryan et, al., 1987), coupled with the burgeoning commercial availability of essential hardware such a s rapid-freezing devices, cryomicrotomes, transfer devices, microscope cold-stages, etc., has meant that examination of ultrathin freeze-dried andlor frozenhydrated cryosections soon may be a routine event in several independent laboratories. In other words, the proper conjunction of available resources can, in principle at least, ensure maintenance of the structural and chemical integrity of biological specimens, with minimal compromise throughout the rigours of a fastidous cryopreparation chain. To clate, most studies employing ultracryomicrotomy with the expressed objective of measuring subcellular compclsition by microprobe analysis have been performed on model systems, cell cultures, or highly oriented muscles (Hagler and Buja, 1984; Somlyo et al., 1979; Wendt-Gallitelli and Wolburg, 1984; Zierold, 1984; Warley, 1986). Many of these workers have utilized for their studies items of hardware that they have innovatively constructed and modified in their own laboratories and that are of necessity, therefore, unique to those laboratories. In this paper, we describe the morphology of a relatively complex tissue-the exocrine pancreas-as i t appears in ultrathin cryosections prepared entirely by State-of-the-art, standard commercial equipment. The exocrine pancreas was chosen for this technical appraisal because i t has attracted several detailed ultrastructural studies using conventional methods
(Q 1990 ALAN
R. LISS, INC.
(Herzog and Reggio, 1980; Palade, 1975; Romanogli, 1984) and also serves a s a model for investigating diverse aspects of protein synthesisisecretion and membrane traffic (Romagnoli, 1985). Furthermore, and despite the fact that the exocrine cells are structurally polarized, for processing purposes, the tissue may be considered a reasonably representative soft tissue. In the context of electron probe x-ray microanalysis, the question that arises is: Does the morphological appearance of the exocrine cells in unfixed, unstained thin cryosections, given the inevitable intrusion of freezing and sectioning artifacts (Chang et al., 1983; Dubochet and McDowell, 1984; Frederik et al., 1984; McDowell et al., 1983) impose contraints on the positive identification, and subsequent analysis, of certain subcellular structural components? A second aim of this paper is to compare the morphology of cryosectioned pancreas with the morphology as seen in anhydrous “compromise” preparations, i.e., freeze-driediplastic embedded and freeze-substituted preparations. The latter procedures may offer some advantage over conventional non-cryotechniques involving sequential exposure to aqueous media (Dudek et al., 1982; Linner et al., 1986; Wroblewski and Wroblewski, 19861, although they may irrevocably redistribute certain solutes from their intracellular in vivo domains (Roos and Barnard, 1986).
Received November 10, 1988; accepted in revised form March 1, 1989. Address reprint requests to Norbert Roos, Electronmicroscopical Unit for Biological Sciences, University of Oslo, P.O. Box 1062, Blindern, N-0316 Oslo 3, Norway.
40
N. ROOS ET AL.
ANHYDROUS CRYOPROCEDURES
MATERIALS AND METHODS Freezing Male rats (200 g) were anaesthetized with Nembutal, dissected, and perfused via the aorta descendens with Krebs-Ringer phosphate buffer (KRPB) pH 7.4 for 10 minutes. Although the hazards due to the stress of dissect ion and prefreezing delays have been highlighted (Hall and Gupta, 1982), seldom are serious attempts made to overcome the problems (e.g., Somlyo et al., 1985). Therefore, the pancreas was carefully dissected along the duodenum, thus minimizing widespread trauma or stress, taken out a s a n entire organ and placed in a buffer bath kept a t 310K and mounted on a simple plastic frame (Roos, 1988). The pancreas was frozen in one of two ways. First, the sample was quench-frozen against a helium vapour-cooled copper block (the helium slammer designed and constructed by Dr. J. Escaig a t the C.N.R.S. in Paris; Escaig, 1982), quickly transferred to and stored in liquid nitrogen. (Liquid nitrogen and liquid helium cooled “slammers” have recently become commc:rcially available from Balzer, Reichert and Hexland.) Most of the cryosectioned and all of the resinembedded samples were frozen by this method to permit. valid comparative assessments. Second, some specimens were frozen in situ by the simple Hagler (Hagler and Buja, 1984) “cold pliers” method. (An automatic version of the pliers, the “cryosnapper” is produced by Gatan). Freeze-drying,fixation, embedding, and sectioning The freeze-drying unit was built by the “in-house” workshop. Small frozen pieces of tissue are placed in a specimen holder, the temperature of which can be recorded and regulated from 130K-400K. The specimen holder is placed in a glass cylinder, which is immersed in liquid nitrogen, thus acting as a cooling trap or condenser for the subliming water. The distance between specimen and condenser is 1.5 cm. The vacuum system of a n old SIEMENS ELMISKOP IA provided a vacuum better than 10 exp -5 Torr a t the gauge The freeze-drying schedule followed was proposed by Ingram and Ingram (1984) and is described in detail elsewhere (Roos and Barnard, 1984). Fixation was achieved by heating paraformaldehyde powder to 333K and osmium vapor fixation by introducing OsO4 crystals. For all of these operations, it is not necessary to break the vacuum. To introduce the resin, however, the vacuum was broken by leaking dry nitrogen gas
Fig. 1. Transmission micrograph of an ultrathin section of quenchfrozen, freeze-substituted, plastic-embedded rat exocrine pancreas. Section poststained with uranyl acetateilead citrate. Arrowheads indicate the freezing front; arrows indicate nuclear pores. M, mitochondria: MV, microvillar projections; N, nucleus; rER, rough endoplasmic reticulum; 2, zymogen granule (reproduced and modified from, Roos and Barnard, 1985). Fig. :2. Transmission micrograph of a n ultrathin section of the same specimen as in Figure 1 about 50 km deep in the tissue (measu-ed from the freezing front). M, mitochondria; N, nucleus; rER, rough ~ndoplasmicreticulum; Z, zymogen granule.
41
into the unit. Spurr low viscosity resin was used as embedding medium. The blocks were cured overnight at 333K and sectioned a t a nominal thickness of 100-200 nm using a dry knife and a Sorvall MT 5000 or a LKB ultratome V ultramicrotome. The sections were transferred with a n eyelash and mounted on a Formvar coated 75 mesh copper grids.
Freeze-substitution Small frozen tissue fragments were placed on frozen dry acetone containing 2% osmium a t 90K. The vials were placed on dry ice in a n insulated box. The dry ice was allowed to sublime overnight, leading to a slow warm-up and dehydration of the samples. Dehydrated tissue blocks were embedded in Spurr low viscosity resin and processed as described for frozen-dried samples. For morphological evaluations, the sections were post-stained with uranyl acetate and lead citrate. This protocol was adopted for morphological examination only; it is not suggested that it is appropriate for quantitatively measuring subcellular electrolyte distribution. Cryosectioning Freezing of the tissue against two cool metal surfaces (pliers) or the helium-cooled copper block (heliumslammer) results in a thin sample with two flat surfaces, both well frozen in case of plier freezing and one well frozen in the case of helium-slammer freezing. The thin wafers were mounted over liquid nitrogen a t approximately llOK transversely to the knife in a vice-type chuck. Cryosections were cut with a Sorvall MT 5000 and a FS 1000 cryoattachment, a t 200-250 nm nominal thickness using glass knives. The temperature of the knife and specimen was 138-128K and 126-116K for the chamber atmosphere. The hydrated sections were transferred to 75 mesh copper grids previously coated with Formvar using a n eyelash. The grids were moved to the grid holder and sandwiched under a second Formvar film. A precooled GATAN cold transfer stage was used to transfer the gridholder assembly into the microscope and to freeze-dry the sections a t 193K under the microscope vacuum. RESULTS AND DISCUSSION Through a rigorous understanding and application of the theory describing the properties of frozen water, it is possible to prepare frozen-hydrated sections of vitrified fresh biological tissues (Chang e t al., 1983; Dubochet and McDowell, 1984; McDowell e t al., 1983). The vitrified state of these ultrathin sections has been established unequivocally by electron diffraction: however, for certain biological applications, frozen-dried cryosections offer a number of practical advantages. In particular, they possess higher intrinsic contrast than frozen-hydrated sections, thus permitting higherresolution imaging and microprobe applications, and they are less susceptible to electron-beam damage (Hall and Gupta, 1984; Zglinicki et al., 1987; Zierold, 1984,1986). Subcellular electrolyte gradients can be maintained in frozen-dried cryosections (e.g.,Somlyo et al., 1985; Zglinicki and Bimmler, 1987), although such
42
N. ROOS ET AL.
Fig. 3. Transmission micrograph of an ultrathin section of quench-frozen, freeze-dried, osmiumvapour-fixed, plastic-embedded rat exocrine pancreas. Section poststained with uranyl acetate/lead citrate. M, mitochondria; N, nucleus; rER, rough endoplasmic reticulum; Z, zymogen granule; ES, extracellular space
preparations cannot be used for analysing free-fluid extracellular spaces (Hall and Gupta, 1984). Good cryofixation is now recognised to be a fundamental prerequisite to the maintenance of the ultrastructural and chemical integrity of biological softtissue specimens; the objective is to freeze (> 10,OOOKis) the specimen rapidly, so that ice-crystal growth and solute displacement are minimized. A number of freezing agents and methods have been described whereby this crucial requirement can be achieved consistently in a wide variety of different specimen types (Escaig, 1982; Plattner and Bachmann, 1982; Robards and Sleytr, 1985). The freezing methods used in the present study permit identification of many of the known structures within individual exocrine cells, although structural artefacts attributable to the freezing process were prevalent. Remarkably, the slamming procedure did not incur any obvious physical damage even a t the surface of the specimen that impinged against the metal block. This was evident from the freeze-substituted sections (Fig. l ) ,where the microvillar projections from the encapsulating endothelial cells were not flattened into the plane of the freezing front. Freeze-substituted and freeze-dried (Figs. 2,3), resin embedded sections indicated that slamming can provide specimens in which the microscopically discernible ice damage does not occur within about 100 km of the surfaces, which
corresponds with several exocrine cell diameters. Ice nucleation in these preparations was most conspicuous within the nuclei (Fig. 21, even in cells which were otherwise relatively free of gross ice damage. This confirms the findings of Linnen et al. (1986), which they suggest was due to relatively high nuclear freewater content. Wroblewski and Wroblewski (1986) found that they could reduce the problem of compression in Lowicryl and Araldite-embedded freeze-dried and freezesubstituted tissues by sectioning below the glasstransition temperatures of the resins. Compression was not apparent in our freeze-substituted and freeze-dried preparations sectioned a t ambient temperatures. (Figs. 2,3). The morphology of the acinar cells in both of the resin-embedded anhydrous preparations compared favourably with that described previously in conventional ultrastructural studies (Herzog and Reggio, 1980; Palade, 1975; Romagnoli, 1984). Euchromatin and heterochromatin were dispersed within all of the nuclei, whereas freeze-substituted nuclei possessed several discrete interruptions of their enclosing membranes, which probably represent nuclear pores (Fig. 1).Within the exocrine cytoplasm, a n extensive rough endoplasmic reticulum with resolvable ribosomes, dense mitochondria, and numerous zymogen granules were prominent features of the contrasted preparations (Figs. 1-
ANHYDROUS CRYOPROCEDURES
43
Fig. 4. Transmission micrograph of a thin, frozen-dried cryosection of quench-frozen r a t exocrine pancreas. M, mitochondria; N, nucleus; rER, rough endoplasmic reticulum; Z, zymogen granule.
3). Christae could be resolved in the freeze-substituted, but not in freeze-dried, mitochondria. Chiovetti et al. (1986) were also unable to preserve intramitochondrial detail in arterial smooth muscle prepared by a freezedrying protocol similar to ours; however, when the tissue wa:j quenched, frozen-dried, and embedded directly in Low cry1 K4M without osmication, the christae were preserx.ecl (Chiovetti et al., 1986). Ziercdd (1984) observed a linear shrinkage of 16% during the freeze-drying of cryosectioned yeast cells. Others have noted clear “shrinking zones,” particularly around the dense mitochondria, in freeze-dried preparations (Chiovetti et al., 1986; Coulter and Terracio, 1977; Dudek et al., 1981) Some differential shrinkage, in addition to the unidirectional compressive efi’ects of sectioning on zymogen granules could also be seen in our plastic-embedded freeze-dried sections (Fig.3), but not in the freeze-substituted sections (Fig. 11.In the freeze-dried cells, the cisternae of the endoplasmic reticulum were considerably narrower, with a much denser intracisternal matrix than the same structure in the substituted preparations (Figs. 1, 3). Similarly, the extracisternal “gaps” were wider and more electron-lucent in the freeze-dried cells. The discontinuous nature of the extracellular matrix in some regions of the freeze-dried sections (Fig. 3) also suggestx that the cells themselves suffered some shrinkage.
Freeze-dried cryosections of large surface area and relatively uniform thickness could be produced routinely by the commercial hardware and manipulative procedures adopted by us (Figs. 4,5). These sections traverse several exorcrine functional units or acini. Close examination reveals that the sections possess many of the freezing, cutting, and drying artefacts identified and described by previous authors (Chang et al., 1983; Dubochet and McDowell, 1985; Zierold, 1984, 1986).Some are identical to those described above in the freeze-dried, resin-embedded sections. Particularly noteworthy are: the susceptability of the nuclei to ice damage; the expanded electron-lucent extracisternal gaps, and compressed cisternae of the stacked endoplasmic reticulum; some evidence of shrinkage haloes around the dense mitochondria; and the unidirectional compression of the zymogen granules (Fig. 6). The intrinsic contrast of the cryosections frozen by both methods was more than adequate to reveal a subcellular morphology of comparable quality to that observed in the resin-embedded preparations. Zierold (1986) suggested that ice-crystal damaged cryosections exhibit greater contrast than undamaged sections. This comment is not doubted, although i t is worth emphasizing that detailed morphology can be resolved even in those structurally coherent regions of the acinar cytoplasm adjacent to the cryogenic surface (Fig. 5).
Fig. 5. Transmission micrograph of a thin, frozen-dried cryosection of quench-frozen rat exocrine pancreas. M, mitochondria; N, nucleus; rER, rough endoplasmic reticulum; Z, zymogen granule.
Fig. 6. Transmission micrograph of a thin, frozen-dried cryosection of quench-frozen rat exocrine pancreas. M, mitochondria; N, nucleus; rER, rough endoplasmic reticulum; Z, zymogen granules. Arrowheads indicate the freezing front.
ANHYDROUS CRYOPROCEDURES
CONCLUSION
45
Escaig, J. (1982) New instruments which facilitate rapid freezing at 83K and 6K. J. Microsc., 126:221-229. Frederik, P.M., Busing, W.M., and Persson, A. (1984) Surface defects The optimal preparation of biological materials for on thin cryosections. Scanning Electron Microsc 1:433-443. electro n microscopic examination should maintain the Hagler, H.K., and Buja, L.M. (1984) New techniques for the preparation of thin freeze dried cryosections for x-ray microanalysis. In: ultrastructural molecular integrity of cells and the Science of Biological Specimen Preparation. J.-P. Revel, T. Barnard, distribution of diffusible solutes. Cryopreparative proand G.H. Haggis, eds. Scanning Electron Microscopy Inc., AMF cedure:; are being used increasingly and routinely for a O'Hare, Chicago, pp. 161-166. wide range of microprobe and immunocytochemical Hall, T.A., and Gupta, B.L. (1982) Quantification for the x-ray microanalysis of cryosections. J. Microsc., 126:333-345. applications because they more faithfully preserve the T.A., and Gupta, B.L. (1984) The application of EDXS to the compositional fidelity of cells and tissues than proce- Hall, biological sciences. J . Microsc., 136:193-208. dures involving exposures to aqueous media and res- Herzog, V., and Reggio, H. (1980) Pathways of endocytosis from ins. Some doubts have remained concerning the morluminal plasma membrane in rat exocrine pancreas. Eur. J . Cell Biol., 21:141-150. pho1og:y of cryosections and whether or not it is possible F.D., and Ingram, M.J. (1984)Influences of freeze-drying and to identify subcellular structures with sufficient preci- Ingram, plastic embedding on electrolyte distributions. In: The Science of sion to undertake localization studies. The current Biological Specimen Preparation. J.-P. Revel, T. Barnard, G.H. paper contributed to the emerging view that for many Haggis, eds. Scanning Electron Microscopy Inc., AMF O'Hare, Chicago, pp. 167-174. purpos 3s the morphology of cryosections does not imJ.G., Bennett, S.C., Harrison, D.S., and Steiner, A.L. (1986) pose serious constraints, other than those introduced Linner, Cryopreparation of tissue for electron microscopy. In: The Science of by the freezing process itself. Biological Specimen Preparation, M. Mueller, R.P. Becker, A. Therefore, i t is not now tenable to state that ". . . . Boyde, and J.J. Wolosewick, eds. Scanning Electron Microscopy Inc., AMF O'Hare, Chicago, pp. 165-174. due to the poor morphology of freeze-dried thin cryoA.W., Chang, J.-J.,Freeman, R., Walter, C.A., and Dusections, only analysis of major cell compartments, e.g. McDowell, bochet, J. (1983) Electron microscopy of frozen hydrated sections of nuclei, cytoplasm, secretory granules, in a few cells per vitreous ice and vitrified biological samples. J. Microsc., 131:l-9. section and specimen can be performed" (Wroblewski Palade, G. (1975) Intracellular aspects of the process of protein secretion. Science, 189:447-358. and Wroblewski, 1986). Commercially available hardH., and Bachmann, L. (1982) Cryofixation: a tool in biologware c m produce thin cryosections of fresh biological Plattner, ical ultrastructural research. Int. Rev. Cytol., 79:237-304. tissues whose intrinsic contrast and behaviour under Robards, A.W., and Sleytr, U.B. (1985) Low Temperature Methods in electron irradiation are sufficiently favourable to proBiological Electron microscopy. Elsevier, Amsterdam, New York, Oxford. vide the capability of examining truly heterogenous P. (1984)The Golgi apparatus and lysosomes of rat acinar cell populations and to discern intracellular fine struc- Romagnoli, cells following refeeding. Histochem. J., 16:855-868. ture. Therefore, the only obvious advantage of plastic Romagnoli, P. (1985) The physiology of pancreatic acinar cells: embedded cryopreparations is the possibility of storing questions and perspectives on the secretory process. BioEssays, 2:68-71. sections indefinitely for future manipulative proceN. (1988) A possible site of calcium regulation in rat exocrine dures and examination: Many pathological applica- Roos, pancreas cells: An x-ray microanalytical study. Scanning Microsc. tions, for example, benefit from this facility, providing 2:323-329. the accompanying compromises are not seriously lim- Roos, N., and Barnard, T. (1984) Aminoplastic standards for quantitative x-ray microanalysis of thin sections of plastic-embedded iting. In all other respects, cryosections are to be biological material. Ultramicroscopy, 15:277-286. favoured. Roos, N., and Barnard, T. (1986) Preparation methods for quantitative electron probe x-ray microanalysis of rat exocrine pancreas: a review. Scanning Electron Microsc, 2:703-711. REFERENCES Ryan, K.P., Purse, D.H., Robinson, S.G., and Wood, J.W. (1987) The relative efficiency of cryogens used for plunge-cooling biological Chang, J-J., McDowell, A.W., Lepault, J., Freemann, R., Walter, specimens. J. Microsc., 14589-96. C.A., :.nd Dubochet, J. (1983) Freezing, sectioning and observation artefacts of frozen hydrated sections for electron microscopy. J. Somlyo, A.P., Somlyo, A.V., and Shuman, H. (1979) Electron probe analysis of vascular smooth muscle: composition of mitochondria, Microsc.. 132:109-123. nuclei and cytoplasm. J. Cell Biol., 81:316-335. Chiovetti, R., Little, S.A., Brass-Dale, J., and McGuffee (1986) A new approz ch to low temperature embedding: Quick freezing, freeze- Somlyo, A.P., Bond, M., and Somlyo, A.V. (1985) Calcium content of mitochondria and endoplasmic reticulum in liver frozen rapidly in drying and direct infiltration in Lowicryl K4M. In: Science of vivo Nature, 314:622-625. Biological Specimen Preparation. M. Mueller, R.P. Becker, A. Boyde and J.J. Wolosewick, eds. SEM Inc., AMF OHare, Chicago, Warley, A. (1986) Elemental concentrations in isolated rat thymocytes prepared for cryofixation in the presence of different media. J . pp. 15.5-164. Microsc., 144:183-191. Coulter, H.D., and Terracio, L. (1977) Preparation of biological tissue Wendt-Gallitelli, M.F., and Wolburg, H. (1984) Rapid freezing, cryofor electron microscopy by freeze-drying. Anat. Rec., 187:477-493. sectioning, and x-ray microanalysis of cardiac muscle preparations Dubochet, J., Lepault, J., Freeman, R., Berriman, J.A., and Homo, in defined functional states. J. Electron Microsc. Techn., 1:151-174. J.-C. 11982) Electron microscopy of frozen water and aqueous Wroblewski, J., and Wroblewski, R. (1986) Why low temperature soluticNns. J . Microsc., 128:218-237. embedding for microanalytical investigations? A comparison of Dubochet, J., and McDowell, A.W. (1984) Frozen hydrated sections. recently used preparation methods. J . Microsc., 142:351-362. In: Science of Biological Specimen Preparation. J.-P. Revel, T. Zglinicki, T., and Bimmler, M. (1987) The intracellular distribution of Barnard, and G.H. Haggis, eds. Scanning Electron Microscopy Inc., ions and water in rat liver and heart muscle. J . Microsc., 146:77-85. AMF O'Hare, Chicago, pp. 147-152. Zglinicki, T., Bimmler, M., and Krause, W. (1987) Estimation of Dudek, R.W., Boyne, A.F., and Freinkel, N. (1981) Quick-freeze organelle water fractions from frozen-dried cryosections. J. Mifixaticn and freeze-drying of isolated rat pancreatic islets: applicacrosc., 146:67-75. tion to the ultrastructural localization of inorganic phosphate in the Zierold, K. (1984) The morphology . -. of ultrathin cryosections. Ultramipancreatic beta cell. J . Histochem. Cytochem., 29:321-325. croscopy, 14:201-210. Dudek, ILW., Childs, G.V., and Boyne, A.F. (1982) Quick-freezing and Zierold, K. (1986) Preparation of cryosections for biological microanalysis. In: Science of Biological Specimen Preparation. M. freeze. drying in preparation for high quality morphology and Mueller, R.P. Becker, A. Boyde, and J.J. Wolosewick, eds. Scanning immuiiocytochemistry at the ultrastructural level: application to Electron Microscopy Inc., AMF O'Hare, Chicago, pp. 119-127. pancreatic beta cells. J. Histochem. Cytochem., 30:129-138.
Morphology of Rat Exocrine Pancreas Prepared by Anhydrous Cryo-Procedures NORBERT ROOS, ULLA KINDE, AND JOHN A. MORGAN Electroi,microscopical Unit for Biological Sciences, University o f Oslo, Blindern N-0316 Oslo 3, Norway (N.R., U.K.) and Department of Zoology, University College, Cardiff C F l I X L , United Kingdom (J.A.M.1
KEY WORDS substitution
Exocrine pancreas, Cryofixation, Cryomicrotomy, Freeze-drying, Freeze-
ABSTRACT This paper describes and compares the morphology of a relatively complex tissue, the exocrine pancreas, prepared by state-of-the-art anhydrous cryoprocedures. Cryopreparative procedures are being used increasingly for a wide range of applications, for example, electron-probe x-ray microanalysis and immunocytochemical localization of antigenic molecules, because they preserve the composition of the specimen better than procedures involving aqueous media. Some doubts have remained concerning the morphology of cryosections and the precise identification of subcellular structures. We show that thin and sufficiently large cryosections of fresh biological tissues can be produced using commercially available hardware. The freeze-dried cryosections display high intrinsic contrast, are stable under the beam, and allow identification of intracellular fine structure. INTRODUCTION The advantages of cryopreparative procedures for the morphological and subcellular locational study of elements have been advocated persuasively by several authors (See Zierold, 1984, for references). An improved understanding of the physical processes involvec. in freezing (Dubochet et al., 1982; Escaig, 1982; Ryan et, al., 1987), coupled with the burgeoning commercial availability of essential hardware such a s rapid-freezing devices, cryomicrotomes, transfer devices, microscope cold-stages, etc., has meant that examination of ultrathin freeze-dried andlor frozenhydrated cryosections soon may be a routine event in several independent laboratories. In other words, the proper conjunction of available resources can, in principle at least, ensure maintenance of the structural and chemical integrity of biological specimens, with minimal compromise throughout the rigours of a fastidous cryopreparation chain. To clate, most studies employing ultracryomicrotomy with the expressed objective of measuring subcellular compclsition by microprobe analysis have been performed on model systems, cell cultures, or highly oriented muscles (Hagler and Buja, 1984; Somlyo et al., 1979; Wendt-Gallitelli and Wolburg, 1984; Zierold, 1984; Warley, 1986). Many of these workers have utilized for their studies items of hardware that they have innovatively constructed and modified in their own laboratories and that are of necessity, therefore, unique to those laboratories. In this paper, we describe the morphology of a relatively complex tissue-the exocrine pancreas-as i t appears in ultrathin cryosections prepared entirely by State-of-the-art, standard commercial equipment. The exocrine pancreas was chosen for this technical appraisal because i t has attracted several detailed ultrastructural studies using conventional methods
(Q 1990 ALAN
R. LISS, INC.
(Herzog and Reggio, 1980; Palade, 1975; Romanogli, 1984) and also serves a s a model for investigating diverse aspects of protein synthesisisecretion and membrane traffic (Romagnoli, 1985). Furthermore, and despite the fact that the exocrine cells are structurally polarized, for processing purposes, the tissue may be considered a reasonably representative soft tissue. In the context of electron probe x-ray microanalysis, the question that arises is: Does the morphological appearance of the exocrine cells in unfixed, unstained thin cryosections, given the inevitable intrusion of freezing and sectioning artifacts (Chang et al., 1983; Dubochet and McDowell, 1984; Frederik et al., 1984; McDowell et al., 1983) impose contraints on the positive identification, and subsequent analysis, of certain subcellular structural components? A second aim of this paper is to compare the morphology of cryosectioned pancreas with the morphology as seen in anhydrous “compromise” preparations, i.e., freeze-driediplastic embedded and freeze-substituted preparations. The latter procedures may offer some advantage over conventional non-cryotechniques involving sequential exposure to aqueous media (Dudek et al., 1982; Linner et al., 1986; Wroblewski and Wroblewski, 19861, although they may irrevocably redistribute certain solutes from their intracellular in vivo domains (Roos and Barnard, 1986).
Received November 10, 1988; accepted in revised form March 1, 1989. Address reprint requests to Norbert Roos, Electronmicroscopical Unit for Biological Sciences, University of Oslo, P.O. Box 1062, Blindern, N-0316 Oslo 3, Norway.
40
N. ROOS ET AL.
ANHYDROUS CRYOPROCEDURES
MATERIALS AND METHODS Freezing Male rats (200 g) were anaesthetized with Nembutal, dissected, and perfused via the aorta descendens with Krebs-Ringer phosphate buffer (KRPB) pH 7.4 for 10 minutes. Although the hazards due to the stress of dissect ion and prefreezing delays have been highlighted (Hall and Gupta, 1982), seldom are serious attempts made to overcome the problems (e.g., Somlyo et al., 1985). Therefore, the pancreas was carefully dissected along the duodenum, thus minimizing widespread trauma or stress, taken out a s a n entire organ and placed in a buffer bath kept a t 310K and mounted on a simple plastic frame (Roos, 1988). The pancreas was frozen in one of two ways. First, the sample was quench-frozen against a helium vapour-cooled copper block (the helium slammer designed and constructed by Dr. J. Escaig a t the C.N.R.S. in Paris; Escaig, 1982), quickly transferred to and stored in liquid nitrogen. (Liquid nitrogen and liquid helium cooled “slammers” have recently become commc:rcially available from Balzer, Reichert and Hexland.) Most of the cryosectioned and all of the resinembedded samples were frozen by this method to permit. valid comparative assessments. Second, some specimens were frozen in situ by the simple Hagler (Hagler and Buja, 1984) “cold pliers” method. (An automatic version of the pliers, the “cryosnapper” is produced by Gatan). Freeze-drying,fixation, embedding, and sectioning The freeze-drying unit was built by the “in-house” workshop. Small frozen pieces of tissue are placed in a specimen holder, the temperature of which can be recorded and regulated from 130K-400K. The specimen holder is placed in a glass cylinder, which is immersed in liquid nitrogen, thus acting as a cooling trap or condenser for the subliming water. The distance between specimen and condenser is 1.5 cm. The vacuum system of a n old SIEMENS ELMISKOP IA provided a vacuum better than 10 exp -5 Torr a t the gauge The freeze-drying schedule followed was proposed by Ingram and Ingram (1984) and is described in detail elsewhere (Roos and Barnard, 1984). Fixation was achieved by heating paraformaldehyde powder to 333K and osmium vapor fixation by introducing OsO4 crystals. For all of these operations, it is not necessary to break the vacuum. To introduce the resin, however, the vacuum was broken by leaking dry nitrogen gas
Fig. 1. Transmission micrograph of an ultrathin section of quenchfrozen, freeze-substituted, plastic-embedded rat exocrine pancreas. Section poststained with uranyl acetateilead citrate. Arrowheads indicate the freezing front; arrows indicate nuclear pores. M, mitochondria: MV, microvillar projections; N, nucleus; rER, rough endoplasmic reticulum; 2, zymogen granule (reproduced and modified from, Roos and Barnard, 1985). Fig. :2. Transmission micrograph of a n ultrathin section of the same specimen as in Figure 1 about 50 km deep in the tissue (measu-ed from the freezing front). M, mitochondria; N, nucleus; rER, rough ~ndoplasmicreticulum; Z, zymogen granule.
41
into the unit. Spurr low viscosity resin was used as embedding medium. The blocks were cured overnight at 333K and sectioned a t a nominal thickness of 100-200 nm using a dry knife and a Sorvall MT 5000 or a LKB ultratome V ultramicrotome. The sections were transferred with a n eyelash and mounted on a Formvar coated 75 mesh copper grids.
Freeze-substitution Small frozen tissue fragments were placed on frozen dry acetone containing 2% osmium a t 90K. The vials were placed on dry ice in a n insulated box. The dry ice was allowed to sublime overnight, leading to a slow warm-up and dehydration of the samples. Dehydrated tissue blocks were embedded in Spurr low viscosity resin and processed as described for frozen-dried samples. For morphological evaluations, the sections were post-stained with uranyl acetate and lead citrate. This protocol was adopted for morphological examination only; it is not suggested that it is appropriate for quantitatively measuring subcellular electrolyte distribution. Cryosectioning Freezing of the tissue against two cool metal surfaces (pliers) or the helium-cooled copper block (heliumslammer) results in a thin sample with two flat surfaces, both well frozen in case of plier freezing and one well frozen in the case of helium-slammer freezing. The thin wafers were mounted over liquid nitrogen a t approximately llOK transversely to the knife in a vice-type chuck. Cryosections were cut with a Sorvall MT 5000 and a FS 1000 cryoattachment, a t 200-250 nm nominal thickness using glass knives. The temperature of the knife and specimen was 138-128K and 126-116K for the chamber atmosphere. The hydrated sections were transferred to 75 mesh copper grids previously coated with Formvar using a n eyelash. The grids were moved to the grid holder and sandwiched under a second Formvar film. A precooled GATAN cold transfer stage was used to transfer the gridholder assembly into the microscope and to freeze-dry the sections a t 193K under the microscope vacuum. RESULTS AND DISCUSSION Through a rigorous understanding and application of the theory describing the properties of frozen water, it is possible to prepare frozen-hydrated sections of vitrified fresh biological tissues (Chang e t al., 1983; Dubochet and McDowell, 1984; McDowell e t al., 1983). The vitrified state of these ultrathin sections has been established unequivocally by electron diffraction: however, for certain biological applications, frozen-dried cryosections offer a number of practical advantages. In particular, they possess higher intrinsic contrast than frozen-hydrated sections, thus permitting higherresolution imaging and microprobe applications, and they are less susceptible to electron-beam damage (Hall and Gupta, 1984; Zglinicki et al., 1987; Zierold, 1984,1986). Subcellular electrolyte gradients can be maintained in frozen-dried cryosections (e.g.,Somlyo et al., 1985; Zglinicki and Bimmler, 1987), although such
42
N. ROOS ET AL.
Fig. 3. Transmission micrograph of an ultrathin section of quench-frozen, freeze-dried, osmiumvapour-fixed, plastic-embedded rat exocrine pancreas. Section poststained with uranyl acetate/lead citrate. M, mitochondria; N, nucleus; rER, rough endoplasmic reticulum; Z, zymogen granule; ES, extracellular space
preparations cannot be used for analysing free-fluid extracellular spaces (Hall and Gupta, 1984). Good cryofixation is now recognised to be a fundamental prerequisite to the maintenance of the ultrastructural and chemical integrity of biological softtissue specimens; the objective is to freeze (> 10,OOOKis) the specimen rapidly, so that ice-crystal growth and solute displacement are minimized. A number of freezing agents and methods have been described whereby this crucial requirement can be achieved consistently in a wide variety of different specimen types (Escaig, 1982; Plattner and Bachmann, 1982; Robards and Sleytr, 1985). The freezing methods used in the present study permit identification of many of the known structures within individual exocrine cells, although structural artefacts attributable to the freezing process were prevalent. Remarkably, the slamming procedure did not incur any obvious physical damage even a t the surface of the specimen that impinged against the metal block. This was evident from the freeze-substituted sections (Fig. l ) ,where the microvillar projections from the encapsulating endothelial cells were not flattened into the plane of the freezing front. Freeze-substituted and freeze-dried (Figs. 2,3), resin embedded sections indicated that slamming can provide specimens in which the microscopically discernible ice damage does not occur within about 100 km of the surfaces, which
corresponds with several exocrine cell diameters. Ice nucleation in these preparations was most conspicuous within the nuclei (Fig. 21, even in cells which were otherwise relatively free of gross ice damage. This confirms the findings of Linnen et al. (1986), which they suggest was due to relatively high nuclear freewater content. Wroblewski and Wroblewski (1986) found that they could reduce the problem of compression in Lowicryl and Araldite-embedded freeze-dried and freezesubstituted tissues by sectioning below the glasstransition temperatures of the resins. Compression was not apparent in our freeze-substituted and freeze-dried preparations sectioned a t ambient temperatures. (Figs. 2,3). The morphology of the acinar cells in both of the resin-embedded anhydrous preparations compared favourably with that described previously in conventional ultrastructural studies (Herzog and Reggio, 1980; Palade, 1975; Romagnoli, 1984). Euchromatin and heterochromatin were dispersed within all of the nuclei, whereas freeze-substituted nuclei possessed several discrete interruptions of their enclosing membranes, which probably represent nuclear pores (Fig. 1).Within the exocrine cytoplasm, a n extensive rough endoplasmic reticulum with resolvable ribosomes, dense mitochondria, and numerous zymogen granules were prominent features of the contrasted preparations (Figs. 1-
ANHYDROUS CRYOPROCEDURES
43
Fig. 4. Transmission micrograph of a thin, frozen-dried cryosection of quench-frozen r a t exocrine pancreas. M, mitochondria; N, nucleus; rER, rough endoplasmic reticulum; Z, zymogen granule.
3). Christae could be resolved in the freeze-substituted, but not in freeze-dried, mitochondria. Chiovetti et al. (1986) were also unable to preserve intramitochondrial detail in arterial smooth muscle prepared by a freezedrying protocol similar to ours; however, when the tissue wa:j quenched, frozen-dried, and embedded directly in Low cry1 K4M without osmication, the christae were preserx.ecl (Chiovetti et al., 1986). Ziercdd (1984) observed a linear shrinkage of 16% during the freeze-drying of cryosectioned yeast cells. Others have noted clear “shrinking zones,” particularly around the dense mitochondria, in freeze-dried preparations (Chiovetti et al., 1986; Coulter and Terracio, 1977; Dudek et al., 1981) Some differential shrinkage, in addition to the unidirectional compressive efi’ects of sectioning on zymogen granules could also be seen in our plastic-embedded freeze-dried sections (Fig.3), but not in the freeze-substituted sections (Fig. 11.In the freeze-dried cells, the cisternae of the endoplasmic reticulum were considerably narrower, with a much denser intracisternal matrix than the same structure in the substituted preparations (Figs. 1, 3). Similarly, the extracisternal “gaps” were wider and more electron-lucent in the freeze-dried cells. The discontinuous nature of the extracellular matrix in some regions of the freeze-dried sections (Fig. 3) also suggestx that the cells themselves suffered some shrinkage.
Freeze-dried cryosections of large surface area and relatively uniform thickness could be produced routinely by the commercial hardware and manipulative procedures adopted by us (Figs. 4,5). These sections traverse several exorcrine functional units or acini. Close examination reveals that the sections possess many of the freezing, cutting, and drying artefacts identified and described by previous authors (Chang et al., 1983; Dubochet and McDowell, 1985; Zierold, 1984, 1986).Some are identical to those described above in the freeze-dried, resin-embedded sections. Particularly noteworthy are: the susceptability of the nuclei to ice damage; the expanded electron-lucent extracisternal gaps, and compressed cisternae of the stacked endoplasmic reticulum; some evidence of shrinkage haloes around the dense mitochondria; and the unidirectional compression of the zymogen granules (Fig. 6). The intrinsic contrast of the cryosections frozen by both methods was more than adequate to reveal a subcellular morphology of comparable quality to that observed in the resin-embedded preparations. Zierold (1986) suggested that ice-crystal damaged cryosections exhibit greater contrast than undamaged sections. This comment is not doubted, although i t is worth emphasizing that detailed morphology can be resolved even in those structurally coherent regions of the acinar cytoplasm adjacent to the cryogenic surface (Fig. 5).
Fig. 5. Transmission micrograph of a thin, frozen-dried cryosection of quench-frozen rat exocrine pancreas. M, mitochondria; N, nucleus; rER, rough endoplasmic reticulum; Z, zymogen granule.
Fig. 6. Transmission micrograph of a thin, frozen-dried cryosection of quench-frozen rat exocrine pancreas. M, mitochondria; N, nucleus; rER, rough endoplasmic reticulum; Z, zymogen granules. Arrowheads indicate the freezing front.
ANHYDROUS CRYOPROCEDURES
CONCLUSION
45
Escaig, J. (1982) New instruments which facilitate rapid freezing at 83K and 6K. J. Microsc., 126:221-229. Frederik, P.M., Busing, W.M., and Persson, A. (1984) Surface defects The optimal preparation of biological materials for on thin cryosections. Scanning Electron Microsc 1:433-443. electro n microscopic examination should maintain the Hagler, H.K., and Buja, L.M. (1984) New techniques for the preparation of thin freeze dried cryosections for x-ray microanalysis. In: ultrastructural molecular integrity of cells and the Science of Biological Specimen Preparation. J.-P. Revel, T. Barnard, distribution of diffusible solutes. Cryopreparative proand G.H. Haggis, eds. Scanning Electron Microscopy Inc., AMF cedure:; are being used increasingly and routinely for a O'Hare, Chicago, pp. 161-166. wide range of microprobe and immunocytochemical Hall, T.A., and Gupta, B.L. (1982) Quantification for the x-ray microanalysis of cryosections. J. Microsc., 126:333-345. applications because they more faithfully preserve the T.A., and Gupta, B.L. (1984) The application of EDXS to the compositional fidelity of cells and tissues than proce- Hall, biological sciences. J . Microsc., 136:193-208. dures involving exposures to aqueous media and res- Herzog, V., and Reggio, H. (1980) Pathways of endocytosis from ins. Some doubts have remained concerning the morluminal plasma membrane in rat exocrine pancreas. Eur. J . Cell Biol., 21:141-150. pho1og:y of cryosections and whether or not it is possible F.D., and Ingram, M.J. (1984)Influences of freeze-drying and to identify subcellular structures with sufficient preci- Ingram, plastic embedding on electrolyte distributions. In: The Science of sion to undertake localization studies. The current Biological Specimen Preparation. J.-P. Revel, T. Barnard, G.H. paper contributed to the emerging view that for many Haggis, eds. Scanning Electron Microscopy Inc., AMF O'Hare, Chicago, pp. 167-174. purpos 3s the morphology of cryosections does not imJ.G., Bennett, S.C., Harrison, D.S., and Steiner, A.L. (1986) pose serious constraints, other than those introduced Linner, Cryopreparation of tissue for electron microscopy. In: The Science of by the freezing process itself. Biological Specimen Preparation, M. Mueller, R.P. Becker, A. Therefore, i t is not now tenable to state that ". . . . Boyde, and J.J. Wolosewick, eds. Scanning Electron Microscopy Inc., AMF O'Hare, Chicago, pp. 165-174. due to the poor morphology of freeze-dried thin cryoA.W., Chang, J.-J.,Freeman, R., Walter, C.A., and Dusections, only analysis of major cell compartments, e.g. McDowell, bochet, J. (1983) Electron microscopy of frozen hydrated sections of nuclei, cytoplasm, secretory granules, in a few cells per vitreous ice and vitrified biological samples. J. Microsc., 131:l-9. section and specimen can be performed" (Wroblewski Palade, G. (1975) Intracellular aspects of the process of protein secretion. Science, 189:447-358. and Wroblewski, 1986). Commercially available hardH., and Bachmann, L. (1982) Cryofixation: a tool in biologware c m produce thin cryosections of fresh biological Plattner, ical ultrastructural research. Int. Rev. Cytol., 79:237-304. tissues whose intrinsic contrast and behaviour under Robards, A.W., and Sleytr, U.B. (1985) Low Temperature Methods in electron irradiation are sufficiently favourable to proBiological Electron microscopy. Elsevier, Amsterdam, New York, Oxford. vide the capability of examining truly heterogenous P. (1984)The Golgi apparatus and lysosomes of rat acinar cell populations and to discern intracellular fine struc- Romagnoli, cells following refeeding. Histochem. J., 16:855-868. ture. Therefore, the only obvious advantage of plastic Romagnoli, P. (1985) The physiology of pancreatic acinar cells: embedded cryopreparations is the possibility of storing questions and perspectives on the secretory process. BioEssays, 2:68-71. sections indefinitely for future manipulative proceN. (1988) A possible site of calcium regulation in rat exocrine dures and examination: Many pathological applica- Roos, pancreas cells: An x-ray microanalytical study. Scanning Microsc. tions, for example, benefit from this facility, providing 2:323-329. the accompanying compromises are not seriously lim- Roos, N., and Barnard, T. (1984) Aminoplastic standards for quantitative x-ray microanalysis of thin sections of plastic-embedded iting. In all other respects, cryosections are to be biological material. Ultramicroscopy, 15:277-286. favoured. Roos, N., and Barnard, T. (1986) Preparation methods for quantitative electron probe x-ray microanalysis of rat exocrine pancreas: a review. Scanning Electron Microsc, 2:703-711. REFERENCES Ryan, K.P., Purse, D.H., Robinson, S.G., and Wood, J.W. (1987) The relative efficiency of cryogens used for plunge-cooling biological Chang, J-J., McDowell, A.W., Lepault, J., Freemann, R., Walter, specimens. J. Microsc., 14589-96. C.A., :.nd Dubochet, J. (1983) Freezing, sectioning and observation artefacts of frozen hydrated sections for electron microscopy. J. Somlyo, A.P., Somlyo, A.V., and Shuman, H. (1979) Electron probe analysis of vascular smooth muscle: composition of mitochondria, Microsc.. 132:109-123. nuclei and cytoplasm. J. Cell Biol., 81:316-335. Chiovetti, R., Little, S.A., Brass-Dale, J., and McGuffee (1986) A new approz ch to low temperature embedding: Quick freezing, freeze- Somlyo, A.P., Bond, M., and Somlyo, A.V. (1985) Calcium content of mitochondria and endoplasmic reticulum in liver frozen rapidly in drying and direct infiltration in Lowicryl K4M. In: Science of vivo Nature, 314:622-625. Biological Specimen Preparation. M. Mueller, R.P. Becker, A. Boyde and J.J. Wolosewick, eds. SEM Inc., AMF OHare, Chicago, Warley, A. (1986) Elemental concentrations in isolated rat thymocytes prepared for cryofixation in the presence of different media. J . pp. 15.5-164. Microsc., 144:183-191. Coulter, H.D., and Terracio, L. (1977) Preparation of biological tissue Wendt-Gallitelli, M.F., and Wolburg, H. (1984) Rapid freezing, cryofor electron microscopy by freeze-drying. Anat. Rec., 187:477-493. sectioning, and x-ray microanalysis of cardiac muscle preparations Dubochet, J., Lepault, J., Freeman, R., Berriman, J.A., and Homo, in defined functional states. J. Electron Microsc. Techn., 1:151-174. J.-C. 11982) Electron microscopy of frozen water and aqueous Wroblewski, J., and Wroblewski, R. (1986) Why low temperature soluticNns. J . Microsc., 128:218-237. embedding for microanalytical investigations? A comparison of Dubochet, J., and McDowell, A.W. (1984) Frozen hydrated sections. recently used preparation methods. J . Microsc., 142:351-362. In: Science of Biological Specimen Preparation. J.-P. Revel, T. Zglinicki, T., and Bimmler, M. (1987) The intracellular distribution of Barnard, and G.H. Haggis, eds. Scanning Electron Microscopy Inc., ions and water in rat liver and heart muscle. J . Microsc., 146:77-85. AMF O'Hare, Chicago, pp. 147-152. Zglinicki, T., Bimmler, M., and Krause, W. (1987) Estimation of Dudek, R.W., Boyne, A.F., and Freinkel, N. (1981) Quick-freeze organelle water fractions from frozen-dried cryosections. J. Mifixaticn and freeze-drying of isolated rat pancreatic islets: applicacrosc., 146:67-75. tion to the ultrastructural localization of inorganic phosphate in the Zierold, K. (1984) The morphology . -. of ultrathin cryosections. Ultramipancreatic beta cell. J . Histochem. Cytochem., 29:321-325. croscopy, 14:201-210. Dudek, ILW., Childs, G.V., and Boyne, A.F. (1982) Quick-freezing and Zierold, K. (1986) Preparation of cryosections for biological microanalysis. In: Science of Biological Specimen Preparation. M. freeze. drying in preparation for high quality morphology and Mueller, R.P. Becker, A. Boyde, and J.J. Wolosewick, eds. Scanning immuiiocytochemistry at the ultrastructural level: application to Electron Microscopy Inc., AMF O'Hare, Chicago, pp. 119-127. pancreatic beta cells. J. Histochem. Cytochem., 30:129-138.
