Combined high resolution electron microscopy and electron probe X

Micron, 1971, 2:390-404

390

with I plate

Combined high resolution electron microscopy and electron probe X-ray microanalysis and its applications to medicine and biology B. A. WEAVERS Consultant Laboratory, A E I Scientific Apparatus Ltd., P.O. Box No. 1, Edinburgh Way, Harlow, Essex, England, U.K.

Manuscript received June 10, 1971

The basic principles employed in combining high resolution electron microscopy with electron probe X-ray microanalysis are outlined and a short description given of the only commercially available instrument, E M M A 4. Methods of preparing specimens and the advantages of using ultra-thin frozen sectionsfor chemical analyses are briefly considered. The applications of the technique to various branches of medicine and biology, based on the use of E M M A 4, are discussed. The fine structure of cells and tissues can be correlated in situ with both a qualitative and quantitative analysis of the chemical elements they contain with an accuracy that cannot be obtained with any other form of instrumentation. The technique has obvious implications in investigative medicine as well as in otherfields of medical and biological research. Les principes de base employdspour la combinaison de la microscopieglectronique gthaute rgsolution et la microanalyse aux rayons X ~ sonde glectronique sont exposgs dans leurs lignes ggngrales et une brkve description est faite de l'gtude du seul appareil disponible commercialement, E M M A 4. Des mgthodes pour la prgparation des spgcimens et les avantages dgrivgs de l'utilisation des sections congeldes ultra fines sont brikvement envisagds. L'application de cette technique gzdiverses branches de la mgdecine et de la biologie, basge sur l' emploi du E M M A 4 est discutge. La fine structure des cellules et des tissus peut ~tre mise en correlation in situ moyennant une analyse tant qualificative que quantitative des llgments chimiques qu'ils contiennent avec une prgcision qui ne peut ~tre obtenuepar l' emploi de n' importe quel autre instrument scientifique. La techniqueposs~de des portges indiscutables dans la mgdecine investigative ainsi que dans les autres domaines des recherches mgdicaleset biologiques. Prinzipien, die der Kombinierung von hochaufl6senden Elektronenmikroskopen mit Elektronensonden-Roentgenstrahlen-Mikroanalysatoren zugrundeliegen, und die I(onstruktion des einzig erha'ltlichen Instruments E M M A 4 werden kurz beschrieben. Die Herstellung von Prizparaten und die Vorteile der Verwendung gefrorener Ultradiinnschaittenfiir chemische Analyser werden kurz erOrtert. Anwendungsm6glichkeiten solcher Geritte besonders E M M A 4 in der Medizin und der Biologie werden besprochen. Mit diesem Verfahren ist es m6glich qualitative und quantitative Analysen chemischer Elemente mit der Feinstruktur yon Zellen und Gewebe in Beziehung zu bringen. Die Analyse kann mit einer Priizision durchgefiihrt werden, die mit keinen anderen Instrumenten erreichbar ist. Das Verfahren ist yon Bedeutung sowohlfiir die Diagnostik als auch fiir andere Zweige medizinischer und biologischerForschung.

INTRODUCTION The all important advantage of using a single instrument with combined facilities for high resolution transmission electron microscopy and electron X-ray probe microanalysis, is that a precise correlation can be obtained between the fine structures of cells and tissues and their basic chemical composition as determined by a quantitative and qualitative determination of the elements they contain. The chemical analysis depends upon the characteristic X-ray spectrum excited by the electrons passing

391

WEA\:ERS

through the specimen; this spectrum is analysed either by X-ray wavelength dispersion using crystal spectrometers, or by non-dispersive energy discrimination using solidstate or gas-flow counters. The technique is currently based on the design and use of the AEI Electron Microscope Micro-Analyser, designated E M M A 4 (Cooke and Duncumb, 1967; Cooke and Openshaw, 1970), and the preparation of specimens by methods usually employed for transmission electron microscopy. Further methods of preparation are being developed, in particular, the use of ultra-thin frozen sections. T H E I N S T R U M E N T AND ITS USE EMMA 4 E M M A 4, shown schematically in Text Figure 1, is a development from a standard transmission electron microscope column, by the substitution of a specially designed objective lens and the addition of a third condenser 'mini-lens', two X-ray crystal spectrometers and a non-dispersive detector plus the associated spectrometer drives and counting nucleonics. The instrument has a conventional high resolution imaging system consisting of the objective and two projector lenses. The overall design, therefore, makes it possible both to examine the specimen at high resolution and to select an area for chemical analysis. The illumination system, which comprises the two standard condenser lenses and the 'mini-lens', produces a probe as small as 200nm in diameter, which, with the beam deflector system can be positioned over any part of the specimen. In this way a precisely defined area of the specimen can be accurately selected and analysed since the X-ray emission arises only from the area under the probe. The X-ray emission is detected at a take-off angle of 45 ° to the specimen plane by the two crystal spectrometers and the non-dispersive analyser which are isolated from the high vacuum of the column by X-ray transparent plastic windows. Each crystal spectrometer contains a choice of four diffracting crystals to allow the most efficient collection of X-rays over the complete range of elements in the periodic table from sodium to uranium. With developments already planned it will be possible to detect elements lighter than sodium, for example, by the use of thin spectrometer windows and the appropriate choice of crystal. The two crystal spectrometers allow two elements to be detected simultaneously for accurate comparative analysis. The non-dispersive detector can be used either for an initial survey of the chemical elements present in the selected area of the specimen, or to determine the 'mass thickness' of the specimen by detecting the 'white' radiation at the same time as the dispersion analysis is being performed by the two crystal spectrometers (Hall and Werba, 1968). Analysis of specimens whose chemical content is not known Having found an area of likely interest in the specimen, the 'mini-lens' is excited and the illumination spot diameter reduced until, with the aid of the beam deflector system, the electron probe is located exactly on the selected area. The non-dispersive detector is then used to detect all the X-ray emission from the selected area and the spectrum of X-ray energy levels is displayed as a series of pulses on a cathode ray oscilloscope (see Text Figure 2). Pulse height analysis then gives an initial indication of the chemical elements which are present. Subsequent analysis of X-ray spectra

COMBINED TEM AND EPMA--EMMA 4

392

ELECTRON GUN

DOUBLE CONDENSER LENS .

BEAM DEFL COILS

SPECTROMETER

SPECTROMETER

SPECIMEN AIRLOCK

MINI LENS OBJECTIVE LEI

FIRST PROJECTOR_ LENS

GEARBOX FOR SERVO CONTROL FAST AND SLOW SCAN SPECTROMETER CONTROLS

SECOND PROJECTOR LENS

CAMERA

VIEWING CHAMBER

Text Figure 1. Schematic diagram of the column of E M M A 4. The instrument combines the facilities for high resolution transmission electron microscopy with those for electron probe X-ray microanalysis.

393

'~ EAVERS

taken by scanning the crystal spectrometers over the wavelengths indicated by the initial non-dispersive analysis then accurately discriminates between the indicated elements.

EMMA 4

lW

Text Figure 2. Schematic diagram of the X-ray detector nucleonics of EMMA 4. The analytical data is presented directly on an osdUoscope (pulse monitor) for pulse height analysis of a particular element, and also on a digital readout linked to a scaler/timer. A permanent record is obtained via a strip chart recorder. Analysis of specimens with a known or suspected chemical composition The detection of elements suspected or known to be present in the specimen is performed by selecting the optimum crystal in each spectrometer and setting the calibrated drive controls to the appropriate wavelength, or the cursor to the correct position on the wavelength-element scale (see Text Figure 3). In addition, the spectrometers may be 'tuned' to exactly the optimum setting by the use of pure elements as standards. Each spectrometer may be 'we-tuned' to six different elements which may then be selected automatically by servo control mechanisms. The X-ray emission at selected wavelengths, generated by the electron probe, is detected by gas-flow proportional counters, the amplified signals from each spectrometer being analysed by a ratemeter and scaler/timer (see Text Figure 2). For quantitative studies, the X-ray emission rate can be related to the amount of the element present in the selected area of the specimen while the scaler/timer permits a count of the X-ray quanta emitted over a pre-selected time.


394

Quantitative interpretation T h e c o m b i n a t i o n of dispersive spectrometers and non-dispersive detector allows accurate q u a n t i t a t i o n o f the analytical data. T h e c o u n t i n g rate o b t a i n e d at a charac-

Text Figure 3. Illustration of one of the front cases of the two spectrometers of E M M A 4. The cases are illuminated and marked in wavelengths and K~ and L~ emission. The spectrometer setting is indicated by a cursor. Four crystals can be accommodated in each spectrometer with a choice usually amongst the following: lithium flouride (LiF), ammonium dihydrogen phosphate (ADP), gypsum (GYPM), mica (MICA), and potassium acid phthalate (KAP).

395

WEAVERS

teristic wavelength indicates the concentration of a particular element. To eliminate the effects of 'white' radiation emitted at the same wavelengths, the counting rate is first taken with the spectrometer geometry set for the maximum collection of the X-radiation characteristic of the element being examined. A background reading is then taken with the spectrometer set to collect X-rays away from the peak. This background count is subtracted from the peak count, thereby ensuring that only X-rays from the element of interest are used. The data obtained in this way needs further correction before it is possible to determine the exact amount of the element that is present. To correct this data, it is necessary to know the 'mass thickness' of the area being analysed (Hall, 1968). To do this, counts are obtained from the nondispersive detector at the same time as those taken from the dispersive analyser. The non-dispersive detector collects a proportion of the total X-ray spectrum emitted from the area being analysed which is, therefore, proportional to the total mass of the area being analysed. If the count rate obtained from the element under investigation (that is, the peak count less that due to the background) is divided by the 'white' count, a measure of the weight fraction of the element within the selected area is obtained. The analysis of thin specimens with E M M A 4 is both accurate and unambiguous because of the minimal spread of the illumination within the specimen. Corrections for fluorescence and absorption effects, common in microprobe analysis of bulk specimens, are small enough to be ignored for the majority of analyses of ultra-thin sections.

Specimenpreparation The procedure at present follows the same general pattern as used when preparing specimens for conventional transmission electron microscopy. Hence the specimens are fixed, dehydrated and finally embedded in one of the plastic embedding media, usually an epoxide resin such as Araldite. The sections are then cut in the usual way and mounted on copper grids. However, certain precautions are necessary. Thus for chemical analyses, and depending upon the amount of the various elements present, there may be advantages in using somewhat thicker sections (100nm or more) than used for morphological studies only. Also the use of any chemicals with an X-ray spectrum close to the elements about to be investigated should be avoided if at all possible. For this reason, it is advisable to avoid glutaraldehyde when looking for sulphur since this element is present in this particular fixative. Similarly, the use of heavy metals for staining purposes, e.g. lead and uranium, should be avoided since the X-ray emission may mask the presence of other elements. Again, there may be advantages in using one embedding medium, e.g. methacrylate, in preference to another, e.g. Araldite. When specimens are prepared as described above, certain problems arise common to most histochemical techniques, namely the diffusion of substances from their original sites and the partial or total loss of the more soluble ones such as various salts of sodium and potassium (Hodson, 1968). In order to overcome problems of this kind several investigators have explored the use of freezing techniques (Bernhard and Leduc, 1967; Christensen, 1967; Appleton, 1968; Hodson and Marshall, 1970; Persson, 1970; Seveus, 1970). Also in recent times, commercial apparatus has become available for the rapid freezing of fresh tissue and the cutting of ultra-thin frozen

C O M B I N E D T E M AND E P M A - - E M M A 4

396

sections. While progress is being made in this direction, there is no doubt that the use of freezing methods raises other problems, including the well-known one of the difficulty of obtaining optimum preservation for morphological studies. Nevertheless rapid freezing and cryo-ultramicrotomy will provide the most reliable information, both qualitative and quantitative, for analytical studies. With further developments it may be possible to maintain the morphological integrity of biological specimens and thus to combine, in one type of preparation, fine structure with optimum retention of its constituent chemical elements. SOME A P P L I C A T I O N S IN MEDICINE AND BIOLOGY

Histo- and Cyto-pathology The value of transmission electron microscopy in the study of medical problems has now been well established. In more recent years the technique has proved to be of importance not only for the study of fundamental problems, but also in the early diagnosis of disease, for example, in investigations of renal failure and hepatic dysfunction (Lynn, Martin, Carson, Willis and France, 1968). However, the latter application is severely limited by the fact that, for the most part, it depends on the determination of differences between the fine structure of the suspected tissue as compared with that seen in normal, healthy material. This requires considerable experience and is beset by the usual problem that only a relatively minute fraction of the whole tissue can be examined. These limitations do not apply to the same extent in light microscopy where interpretation is helped by the use of a variety of staining techniques ranging from the routine use of haematoxylin and eosin to precise histochemical tests. Some staining techniques have been developed for the electron microscope which are of diagnostic value inasmuch as they can be correlated with those used for light microscopy. Thus the technique of Seligman, Wasserkrug, Deb and Hanker (1968), makes it possible to distinguish between the more acid and basic components of cells and in this respect resembles the traditional use of haematoxylin and eosin of light microscopy. However, the reagents used in the former technique (osmiummercapto succinic acid and osmium-tetramethylenediamine) are extremely expensive while, in common with various other staining methods for the electron microscope, the procedure is applied to the tissue before embedding and the whole process is tedious and time-consuming. The ideal type of electron microscope for histopathological studies should be one which, in addition to providing morphological data, should be capable of providing the same type of chemical information that can be obtained with various forms of light microscopy. For example, histochemical methods used in conjunction with the conventional light microscope can be supplemented by immunofluorescent studies using ultra-violet light. Similarly, tissue with a mineral or crystalline content can be examined using polarised light. The analytical electron microscope, E M M A 4, provides a way of obtaining chemical information without recourse to a variety of staining methods or different instruments. Further, and unlike most staining methods, it can be used to provide both qualitative and quantitative data. In addition to this, the detailed structure of the tissue can be determined from the same specimen and an exact correlation obtained between morphology and chemical constituents.

397

W EA VERS

To date little is known of the process of protein uptake by mammalian cells (Ryser, 1968) though this field of investigation is important in relation to the initiation of infection (Holland, McLaren and Syverton, 1959; Alexander, Koch, Mountain and Van Damme, 1958) and the process of malignant transformation (Ito, 1960). It is anticipated that it may be possible to monitor physiological events such as the uptake of proteins by pinocytosis using E M M A 4. Such experiments could be conducted by using proteins 'labelled' with a detectable element and pinocytosis induced by means of a polyampholyte.

Histochemistry It might be considered that since most histochemical techniques are for the detection of compounds, that X-ray analysis is of very limited value since it only reveals elements. However, the presence of an individual element is, of course, indicative of any compound which contains that element. In this respect, the presence of several elements, particularly in known proportions, can be most informative. T h e n again, using this approach it may be possible to utitise an established histochemical technique. Thus one of the methods for demonstrating mucin is the alcian blue method of Steedman (1950). Alcian blue is a copper phthalocyanin dye belonging to the chloromethyl substituted group in which the chloromethyl groups may be reacted with thiourea or alkyl thioureas to give an isothiouronium derivative (Pearse, 1968). This derivative contains three elements copper, sulphur and chlorine, all of which are easily detected by E M M A 4. Since quite different compounds can contain the same elements it is apparent that difficulties of interpretation can arise, particularly in the absence of quantitative data. Here again, it may be possible to overcome the problem by utilising known histochemical tests. Thus if sulphated mucopolysaccharides or sulphated proteins are present, they can be identified provisionally by detecting the presence of sulphur. However, on this basis alone, it is not possible to distinguish between them. To do this a suitable histochemical technique can be employed, that is, one which depends upon the presence of an element as distinct from a reactive group. For example, the ferric-ferricyanide method of Ch~vremont and Frederic (1943) as employed by Yao (1949) where the presence of iron indicates protein. Alternatively, the DDD method (dihydroxy-dinaphthyl-disulphide) could be employed, although after this reaction the only detectable element is sulphur (Barrnett and Seligman, 1952, 1954; Barrnett, 1953). The identification of sulphated protein groups can also be carried a stage further, since the presence of sulphur does not distinguish between SH (cysteine) and SS (cystine) linkages. For SS groups the preferred method of differentiation is that of Olszewska, Wronski and Fortak (1967), as following this reaction the SS group can be determined by the presence of selenium. Finally it is apparent that the combination of high resolution transmission electron microscopy and X-ray microanalysis offers the opportunity for absolute quantitation that has so far been unobtainable with most, if not all, histochemical techniques.

Immunology At present the detection sites of antigen-antibody reactions by electron microscopy depend on the presence or formation of an electron dense 'marker'. These techniques,


398

although well established, are rather complex and extensive controls are necessary to ensure the correct interpretation of results. Of the methods using labelled immunoglobulin fractions, the ferritin conjugation method is preferred by Singer and Schick (1961). Immunoenzyme techniques have been described by Ram, Nakane, Rawlinson and Pierce (1966); Nakane, Ram and Pierce (1966); Seligman (1966); Hanker, Seamen, Weiss, Ueno, Borgman and Seligman (1964); Hanker, Katzoff, Rosen, Seligman, Ueno and Seligman (1964); Holt and Hicks (1962). Many methods have also been described employing labelled, purified antibodies (Sternberger, 1967). Quantitation is not possible with any of these methods, with perhaps the exception

IODINE

so

70 60 50

/

~

LITHIUM

TI

FLUORIDE

e

CRYSTAL

40 30 20 10 •

•

l

i

•

3.125 3.142 3.158 3.1"/4 3.191A° Text Figure 4. A micrograph illustrating one of the applications of EMMA 4 as referred to in the text (see over) and as used by Clarke, Salsbury and Willoughby (1970) in investigating the transfer of antigenic material from macrophage to lymphocyte. The antigen was labelled with iodine which was detected in the bridges shown in this figure (see also Text Figure 5). M, macrophages; L, lymphocyte. Text Figure 5. Spectrum of iodine using the non-dispersive system of EMMA 4. The spectrum was taken from one of the bridges formed between the macrophages and lymphocyte~ similar to those shown in Text Figure 4. The element was detected ufing a lithium fluoride crystal at a wavelength of 3.158A of the immunoferritin technique (Sternberger, 1967). E M M A 4, however, offers not only the capability of qualitative elemental analysis and thereby chemical localisation, but also the facility for quantitative analysis by monitoring the mass thickness of the areas analysed. Clarke, Salsbury and Willoughby (1970) have described the use of

399

WEAVERS

E M M A 4 in monitoring the transfer of antigenic material. The results obtained support the hypothesis that transfer of antigenic material occurs from macrophages to lymphocytes via intercellular bridges under certain conditions (see Text Figure 4). In this work, iodine labelled antigen was used and electron microscope microanalysis demonstrated that the intercellular bridges contained iodine (Text Figure 5), thereby indicating the presence of antigen. The failure of other analytical techniques to accurately locate elements has previously made this kind of investigation impossible. This work has opened up new possibilities for the use of E M M A 4 in immunochemical investigations at the ultrastructural level.

Biochemistry SubceUular analysis of organelles using E M M A 4 has shown that the data obtained can be correlated to the known biochemistry of the system, even though the diameter of the electron probe is insufficiently small to allow analysis of areas o f macromolecular dimensions. This work opens up new fields in which quantitative electron microscopy can be applied. The monitoring of biochemical systems directly in thin sections can add to the knowledge of structural-functional interrelationships of many cellular systems. If the correct models are chosen for such work, it is feasible that gross biochemical changes can be determined in relation to ultrastructural changes, these in turn being related to known biochemical conditions. For instance, the ultrastructural changes shown by transmission electron microscopy to occur in isolated mitochondria in various metabolic steady states (Hackenbrock, 1966, t968), may be studied. Such mitochondria may be analysed in specific areas and the detection of marker elements such as iron in the cristae will indicate the presence of the cytochrome system, and will monitor any changes occurring in various metabolic states.

Physiology: detection of sodium and potassium ions The distribution and movement of sodium and potassium ions is of special interest to many physiologists. As pointed out in the section on Histochemistry, these elements are difficult to stain using conventional methods of preparing specimens for the electron microscope and it is necessary to turn to the use of freezing techniques. The merit of using freezing techniques in conjunction with X-ray analysis is also underlined b y the doubts cast on the available histochemical techniques. Thus, according to Lane and Martin (1969), the techniques of Komnick and Komnick (1963), and Hartmann (1966), are of questionable specificity. The specificity of the Komnicks' pyroantimonate technique for sodium, as modified by Spicer, Hardin and Greene (1968), has recently been critically examined by Herman, Sato and Weavers (1971) using E M M A 4. Following the application of the technique, thin sections of pancreatic islet tissue were analysed for antimony and also for five cations that reacted in vitro with potassium pyroantimonate. The X-ray analysis showed that the pyroantimonate reaction is not specific for sodium and that the electron dense precipitates seen by transmission electron microscopy (Figures 1 and 2) are composed of sodium, calcium, potassium, magnesium and manganese antimonates. This investigation has also confirmed the view of Hodson and Marshal

COMBINED TEM AND EPMA--EMMA4

400

(1970) that most of the sodium is removed from the tissue during the embedding process. CONCLUSIONS The combination of high resolution transmission electron microscopy and X-ray microanalysis as found in E M M A 4, makes it possible to correlate in situ, the fine structure of cells and tissues with both a qualitative and quantitative determination of the chemical elements they contain. The development of techniques for cutting ultra-thin frozen sections of biopsy material, when used in conjunction with this type of analytical electron microscope, will allow both a rapid and accurate assessment of morphological and chemical content of specimens for diagnostic purposes. In biological research, E M M A 4 will provide further information on the relationship between structure and cellular dynamics. Hence this type of instrument will find an increasing use both in medicine and biology. ACKNOWLEDGEMENTS Thanks are due to my colleagues in the Electron Microscope Consultant Laboratory for m a n y helpful discussions, and to AEI Scientific Apparatus Limited for permission to publish this paper. REFERENCES ALEXANDER, H. E., KOCH, G., MOUNTAIN, I. M. and VAN DAMME, O., 1958. Infectivity of ribonucleic acid from Poliovirus in human cell monolayers. 07. Exp. Med., 108: 934-936. APPLETON, T., 1968. Ultrathin frozen sections for electron microscopy. Preliminary Report, LKB Instruments, Rockville, Maryland, U.S.A. BARRNETT,R. J., and SELIGMAN,A. M., 1952. Demonstration of protein bound sulphydryl and disulphide groups by two new histochemical methods. 07. nat. Cancer Inst., 13: 215-216. BARRNETT, R. J., 1953. The histochemical distribution of protein bound sulphydryl groups. 07. nat. Cancer Inst., 13: 905-927. BARRNETT, R.J. and SELIGMAN,A. M., 1954. Histochemical demonstration ofsulphydryl and disulphide groups of protein. 07. nat. Cancer Inst., 14: 769-803. BERNHARD, W. and L~DUC, E. H., 1967. Ultrathin frozen sections I. Methods and ultrastructural preservation. 07. Cell. Biol., 34: 757-772. Crt~.VREMONT, M. and FREDERIC, J., 1943. Une nouvelle mdthode histochimique de mise en ~vidence des substances ~ fonction sulfydrile. Arch. Biol., $4: 589-605. CHRISTE~SEN, K., 1967. A simple way to cut frozen thin sections of tissue at liquid nitrogen temperature. Anat. Rec., 157: 227. CLARKE,J. A., SALSBURV,A.J., and WILLOUOHBV,D. A., 1970. Application of electron probe microanalysis and electron microscopy to the transfer of antigenic material. Nature, Lond., 277: 69-71. COOKE,C.J. and DUNCUMB,P., 1968. Performance analysis of a combined electron microscope and electron probe microanalyser--EMMA. In : Proc. Fifth Int. Cong. on X-ray Optics and Microanalysis, Mollenstedt, G. and Gaulker, K. H., (eds.), Springer-Verlag, Berlin, 245-247. COOKE, C. J. and OPENSAW, I. K., 1970. Combined high resolution electron microscopy and X-ray microanalysis. In: Proc. 28th Annual Meeting EMSA, Arceneaux, C. J., (ed.), Claitors, Baton Rouge, Louisana, U.S.A., 552-553.

~1

WEAVERS

HACKENBROCK,C. R., 1966. Ultrastructural bases for metabolically linked mechanical activity

in mitochondria I. Reversible ultrastructural changes with change in metabolic steady state in isolated liver mitochondria. J. Cell. Biol., 30: 269-297. HACKENBROCK, C. R., 1968. Uhrastructural bases for metabolically linked mechanical activity in mitochondria II. Electron transport linked uhrastructural transformations in mitochondria. J. Cell. Biol., 37: 345-369. HALL, T. A. and WERBA, P., 1968. The measurement of total mass per unit area and elemental weight-fractions along line scans in thin specimens. In : Proc. 5th International Congress on X-ray Optics and Microanalysis, Mollenstedt, G. and Gaukler, K. H., (eds.), Springer-Verlag, Berlin, 93-98. HALL,T. A., 1968. Some aspects of microprobe analysis of biological specimens. In: Quantitative electron probe microanalysis. Heinrieh, K. F.J., (ed.), National Bureau of Standards Special Publication 298, Washington D.C., U.S.A., 269-299. HANKER,J. S., SEAMEN,A. R., W~.xss, L. P., UENO, H., BOROMAN,R. A. and SELmMAN,A. M., 1964. Osmiophilic reagents: New cytochemical principle for light and electron microscopy. Science, 1416: 1039-1043. HANKER,J. S., KATZOFF, L., ROSEN, H. R., SELIO~AN, M. L., UENO, H. and SELmMAN,A. M., 1966. Design and synthesis of thiolesters for the histochemical demonstration of esterase and lipase via the formation ofosmiophilic diazo thiothers. 07. Med. Chem., 9: 288-291. HARTMANN,J. F., 1966. High sodium content of cortical astroyetes. Archiv. Neurol., Chicago., U.S.A., 15: 633-642. HF.RMAN,L., SATO,T. and WEAVERS,B. A., 1971. In : 29th Annual Proc. EMSA (in press). HODSON, S., 1968. Inadequacy of aldehyde fixation in preserving the ultrastructure of corneal endothelium in rabbit and monkey. Exp. Eye Res., 7: 221-224. HODSON, S. and MARSHALL,J., 1970. Tissue-sodium and potassium: Direct detection in the electron microscope. Experentia, 26: 1283-1284. HODSON, S. and MARSHALL,J., 1970. Ultracryotomy: A technique for cutting uhrathin sections of unfixed frozen biological tissue for electron microscopy. 07. Microscop., 91: 105-117. HOLLAND,J. J., McLAREN, L. C. and SYVERTON,J. T., 1959. The mammalian cell virus relationship. J. Exp. Med., 116: 65-80. HOLT, S. J. and HICKS, R. M., 1962. Staining methods for enzyme localisation. Brit. rned. Bull., 18: 214-219. ITO, Y., 1960. A tumor producing factor extracted by phenol from papillomatous tissue (Shope of Cotton tail rabbits). Virology, 12: 596-601. KOMNICK, H. and KOMNmK, U., 1963. Electronen Mikroskopische Untersuchungen zur Funktionellen Morphologie des Iontransportes in der Salzdruese von Larus Argentatus. Z. Zellf°rsch, 6@: 163-170. LANk, B. P. and MARTIN, E., 1969. Electron probe analysis of cationic species in pyroantimonate precipitates in epon embedded tissue. 07. Histochem. Cytochem., 17: 102-106. NAKANE, P. K., RAM,J. S. and PmRCE, G. B., 1966. Enzyme labelled antibodies for the ultra structural localisation of antigens. In: Proc. 6th International Congress on Electron Microscopy, Ryozi Uyeda (ed.), Maruzen Co. Ltd., Tokyo, 2: 51-52. OLszEwsv~, M.J., WRONSKI,M. and FORTAIqW., 1967. A histochemical method for revealing disulphide bonds by means of hydrogen selenide. Fotia Histochem. Cytochem., S: 7-14. PEARSE, A. G. E., 1968. Histochemistry, Theoretical and Applied. J. and A. Churchill Ltd., London, 3rd Ed., 1. PERSSON, A., 1970. Cryo-ultramicrotomy. In: Proc. 7th International Congress on Electron Microscopy. Favard, P., (ed.), Socifitfi Franqais de Microscopic Electronique, Paris, France, 1: 421-422. RAM, J. S., NAKANE, P. K., RAWHNSON,D. G. and PIERCE, G. B., 1966. Enzyme labelled antibodies for ultrastructural studies. Fed. Proc., 25" 732-738. RVSER, H., 1968. Uptake of protein by mammalian cells: An undeveloped area. Science, ISS: 390-396.


402

SELIGMAN,A. M., 1966. New methods in electron microscopic cytochemistry. J. Histochem. Cytochem., 14: 745-746. SELIGMAN,A. M., WASSERKRUO,H. L., DEB, C. and HANRER,J. S., 1968. Osmium containing compounds with multiple basic or acidic groups as stains for ultrastructure. 07. Histochem. Cytochem., 16: 87-101. SEVERS, L., 1970. Frozen ultra-thin sections. In: Proc. 7th International. Congress on Electron Microscopy. Favard, P., (ed.), Soci6t6 Franqais de Microscopie Electronique, Paris, France, 1: 423-424. SINOER, S.J. and SenICK, A. F., 1961. The properties of specific stains for electron microscopy prepared by the conjugation of antibody molecules with ferritin. 07. Biophys. Biochem. Cytol., 9: 519-537. SPICER, S. S., HARDIN,J. H. and GREENE, W. B., 1968. Nuclear precipitates in pyroantimonate--osmium tetroxide fixed tissues. J. Cell Biol., 39: 216-221. STEEDMAN, H. F., 1950. Alcian blue 8GS: A new stain for mucin. Quart. 07. Micr. Sci., 91: 477-479. STERNBERGER, L. A., 1967. Electron microscopic immunocytochemistry. A Review. 07. Histochem. Cytochem., 15: 139-158. YAO, T., 1949. Cytochemical studies on the embryonic development of Drosophila melanogaster. Quart. ft. Micr. Sci., 90:401-409.

403

WEAVERS

FIGURES 1 and 2

Evaluation of the pyroantimonatetechniqueby X-ray microanalyszs Figure 1. Lower power micrograph of pancreatic islet cells following the use of the Komnicks' (1963) pyroantimonate technique for sodium, as modified by Spicer, Hardin and Greene (t968). When such sections were examined under EMMA 4 the dense precipitate was found to contain sodium, calcium, potassium, magnesium and manganese antimonate (from Herman, Sato and Weavers, 1971). × 8,000.

Figure 2. High power micrograph of pancreatic islet cells processed by the technique referred to in Figure 1. The dense deposit along the cell membrane and in the various organetles is not specific for sodium but contains a variety of elements. × 25,000.


404