Combined light and electron microscope in routine ... · Sections stained for light microscopy and viewed in an electron microscope present new problems. Ultramicroscopic stain deposits,

J Clin Pathol 1982;35:425-429

Combined light and electron microscope in routinehistopathologyS JONES, SK CHAPMAN, PR CROCKER, GINA CARSON, DA LEVISON

From the Department of Histopathology, St Bartholomew's Hospital and Medical College, LondonECIA 7BE

SUMMARY We report our experience with a prototype combined light and electron microscope(the LEM 2000) with particular reference to its application to routine surgical histopathology. Wefound its major advantages over conventional transmission electron microscopes were due to thelarge grid size (7 mm diameter), low magnification capacity ( x 250), and the built-in microprocessorfor recording areas of interest. These features combihe to reduce sampling errors and greatlyfacilitate orientation and relocation of fields of diagnostic importance.

A new type of transmission electron microscope, thecombined light and electron microscope (LEM 2000)has recently been developed by InternationalScientific Instruments Ltd. The HistopathologyDepartment of this hospital has used a prototypeLEM for three months, this paper being an evalua-tion of the instrument and associated techniques ofspecimen preparation. We have specifically testedpossible applications of the LEM to routine surgicalhistopathology. Initial details of the instrumentsuggested that its main advantage over othertransmission electron microscopes would be theopportunity to observe the same specimen in boththe light and the transmission electron modes,combining selective colour staining with highresolution microscopy in the electron mode. Whathad not been foreseen was that the larger grid size(7mm diameter as opposed to the conventional 3 mmdiameter) combined with the low magnificationcapacity ( x 250) and the microprocessor forrecording areas of interest, would prove so useful inovercoming sampling difficulties and problems withorientation and relocation of key fields. Althoughother instruments have similar low magnificationcapacity, no other instrument, as yet, has theability to accept a large grid.

Material and methods

Specimens were prepared using a range of tissues ofvarying sizes measuring from 1 mm square to 5 mmsquare by 1 mm thick. Tissues were processedaccording to the general principles involved in

Accepted for publication 13 July 1981

preparing specimens for electron microscopy. Timeschedules were adjusted to allow for larger pieces oftissue. The schedule used for a typical specimen isshown in Table 1. Tissues were embedded using theSorvall JB4 embedding system.'

Table 1 Processing schedule

Step Time (h)

1 Fix 4% glutaraldehyde in cacodylate sucrose buffer. 1-22 Wash cacodylate sucrose. 1-23 1 % osmium tetroxide in cacodylate buffer. 1-24 Wash cacodylate sucrose buffer. 0 35 Saturated uranyl acetate (aqueous). 1-26 70% alcohol. 0-57 90% alcohol. 0-58 Absolute alcohol (three changes). 0 5 each9 Propylene oxide: Taab resin (1:1). 1

*10 Taab r_sin (three changes). 24-4811 Embed. Fresh resin and polymerise. 70°C. 24

Agitation of tissue if possible at steps 9 and 10.*Infiltration of resin carried out at 40°C.

Sections were cut on a JB4 A microtome using450 angled knives. Section thicknesses ranged from1-0 ,um to 0-1 ,im. If possible, block faces approxi-mately 6 mm in diameter were cut including a borderof excess resin. It was found that excess resin wasnecessary to aid adherence of the section to the grid.Sections were cut on to distilled water and stretchedby holding over them a brush soaked in xylene.Sections were picked up by bringing the 7 mmdiameter LEM grid from underneath at an angle of450. Picking up sections by laying the grid on top andusing surface tension was found not to work becauseof the increased weight of the large sections. Dryingof the sections was facilitated by placing a piece of

425

on May 22, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.35.4.425 on 1 April 1982. D

ownloaded from

Jones, Clhapman, Crocker, Carson, Levison

filter paper against the side of the grid and thenholding over a hot plate. Wet grids could not belayed directly on to filter paper because sections oftensagged through the grid spaces, sticking to the paper.Alternatively, sections could be cut on to smaller con-ventional 3 mm grids as the LEM can accommodateboth sizes. Sections of various thickness werestained with Reynold's lead citrate, various lightmicroscopical stains, or with both techniques.Before some light staining techniques could becarried out the resin was first etched with sodiumethoxide for one to three seconds depending onsection thickness. With basic dyes such as toluidineblue, methylene blue and basic fuchsin, this was notnecessary. If the sodium ethoxide treatment lastedtoo long sections fell apart in the electron beam.Sections from tissue fixed in osmium tetroxide werealso treated with 1% periodic acid for 10 min toimprove light microscopic staining. Staining tech-niques attempted were toluidine blue, methylviolet, haematoxylin and phloxine, periodic acidSchiff, alcian blue, Giemsa, Jones' hexamine silver.

Sections stained for light microscopy and viewedin an electron microscope present new problems.Ultramicroscopic stain deposits, undetectable bylight microscopy, can appear as large obstructions atelectron microscopic level, impeding viewing of thespecimen. To avoid this, all stains were carefullyfiltered.

Staining techniques were carried out in micronembedding capsules or alternatively on a drop ofstain on a glass slide. Using the latter method,heating of the stain was easily facilitated by placingthe slide on a hot plate. Drying of sections afterstaining was carried out as before. Human materialwe examined in the LEM included renal biopsies,liver biopsies, lymph nodes, and lymphomas andvarious tumours-particularly some thought likelyto contain neurosecretory granules or contractilefilaments.

Results

Post-fixation and staining with osmium tetroxide anduranyl acetate was found to give the best detail offine structure and contrast for visualisation. Speci-mens fixed only in glutaraldehyde were more difficultto examine even when stainedwithuranyl acetate. Theblock was stained with uranyl acetate rather than thecut section because of ease, and also because wefound that it is virtually impossible to stain withtoluidine blue a section already stained with uranylacetate.

Sections more than 0 5 ,um thick gave a very clearlight microscopical picture although electron micro-scopy was difficult to interpret. Reducing section

thickness to 0 5 ,um-0 25 um made light microscopicstaining more difficult: some techniques gaveinsufficient density of staining for interpretation.Toluidine blue was the most informative stain at thisthickness. Electron microscopic examination ofsections in this thickness range produced a clearenough image for all cytoplasmic organelles to berecognised.With the reduction in section thickness to 0-25 Htm-

01 pum (Figs. 1 to 4) a considerable increase inelectron optical resolution was achieved, sacrificingsome light optical contrast. Limited light microscopycould be carried out at the thicker end of this rangewith the toluidine blue technique (Figs. 1 and 3). Oneway of obtaining acceptable light microscopy andgood electron microscopy was found to be to mount0 25 pm and 0-1 ,um thick sections side by side on thegrid and then stain by the combined toluidine blue/lead citrate method (Fig. 3). We could achieve goodlight microscopy on the thicker section and goodelectron microscopy on the thinner section. Thelimit of electron optical magnification with thisrange of thickness was x 20 000. One other findingof some interest at this section thickness was thatpretreatment with 0-5% potassium permanganate(acid) followed by 1% oxalic acid enhanced thetoluidine blue staining. However, on subsequentexamination in the electron microscope, there seemedto be some loss of clarity. Possibly the lead had beenwashed out.

I

<.wm .:.l.. r

7aD:J\.

.tzA,6 A

Ii m4

Fig. 1 Light microscopy ofO-2 um thick sectionprocessed as in Table I and stainedby toluidine blue/leadcitrate. Renal biopsy, membranousglomerulonephritis.Photo ofsection on the grid x 70.

.:

426

on May 22, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.35.4.425 on 1 April 1982. D

ownloaded from

Combined light and electron microscope in routine histopathology

Fig. 2 Transmission electron micrograph ofpart of thesame glomerulus in the same section illustrated in Fig. 1.Despite the unconventional thickness of the section thecharacteristic diagnostic epimembranous deposits ofmembranous glomerulonephritis are obvious x 1720.

Fig. 3 Light micrograph of thick (0-25 Am) and thinner(0-1 ,m) serial sections ofa renal biopsy mounted side byside in the 7 mm diameter grid. Sections stained bycombined toluidine blue/lead citrate method x 3.

Some conventional electron microscopy wascarried out with sections thinner than 01 um (Fig. 5).The instrument was capable of studies to levels ofresolution equal to that considered to be conventional

Fig. 4 Low power transmission electron micrograph(0-2 ,um thick, stained by combined toluidine blue/leadcitrate method) offollicular lymphoma. The follicleoccupies the upper right half of the field. Such areaswere easily found in the 5 mm square sections, but verydifficult to find in I mm square sections of the samespecimen x 600.

i... .T

Fig. 5 Conventional thickness (0 05 pAm) and staining(uranyl acetate/lead citrate) of section of spleen inHairy Cell Leukaemia showing typical Hairy Cell. TheLEMfunctioning as a conventional transmission electronmicroscope x 6980.

for thin sections of tissue-that is, 1/10 of thesection thickness (Cosslett's law).2 Suggested uses ofsections of different thicknesses are summarised inTable 2.

427

on May 22, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.35.4.425 on 1 April 1982. D

ownloaded from

428

Table 2 Specimen thickmess evaluation

Microscopy Section thickness

07-05 ,um 05-0-25 jAm 0-250-1 gAm < 01 gsm

LM Suitable for Suitable for Only the Not suitable.a wide fewer blue stainsrange of stains. or silver arestains. suitable.Sodiumethoxideetching maybe required.

EM up to up to up to up tox 3000 x 7000 x 20 000 x 45 000

It was noticed that under the electron beam thelightmicroscopy stain was "burned off" leaving areas ofunstained resin when re-examined in the lightmicroscope. The section could not be restained due tofurther polymerisation of the resin in the heat of thebeam. Silver staining techniques could not be carriedout with the section on the grid as the solutionattacked and dissolved the copper grid. The sectionhad to be first stained and then mounted.The main advantage of the LEM in diagnostic

histopathology only became apparent when we beganto concentrate on using the instrument as a trans-mission electron microscope, exploiting its large gridsize (7 mm as opposed to the conventional 3 mm),low magnification capacity (x 250) and micro-processor for recording areas of interest. We found,for example in focal glomerulonephritis, that thelarger number of glomeruli available for examinationand easily reidentifiable made electron microscopicdiagnosis of this condition much easier and morereliable. In the electron microscopic examination oflymphomas we found orientation very much less of aproblem than with conventional grids and infollicular lymphomas we could easily find andexamine neoplastic follicles (Fig. 4).

In the investigation of the histogenesis of tumoursthought likely to be rhabdomyosarcomas or likely tocontain neurosecretory granules-a larger section,with low power scanning capacity and the ability torecord and relocate areas of interest, obviouslyimproved the chances of making a positive diagnosis.The examination of pieces of intestinal mucosa andpancreas was also greatly facilitated on larger piecesof tissue-orientation was much easier and isletssimple to find. Pieces of liver could be examinedfaster and low power scans for HBsAg and HBcAgin hepatocyte cytoplasm and nucleus respectivelywere much more readily performed.

In other words, sampling and orientation problemswere greatly reduced in the LEM. When we wished tohave a second look at a section at a later date,because the 7 mm diameter grid would only fit intothe grid holder in one position the section was

Jones, Chapman, Crocker, Carson, Levison

identically orientated every time it was placed in theinstrument. With the co-ordinates automaticallyrecorded on every photograph, areas of interest werereadily relocated.

Discussion

In the short time we had the use of the prototypeLEM we certainly did not study its potentialexhaustively. We probably underestimated theadvantages of its capacity to view the same sectionby light and electron microscopy. For example, wedid no histochemical nor immunoperoxidase workand as the reaction products of these techniques canbe visible by both light and electron microscopy, thiswould seem an ideal area in which to use theinstrument. Also, the examination of brain tissuewith the use of silver staining techniques would seemanother area in which much might be achieved.However, our aim was to develop specimen

preparation techniques and to evaluate the instrumentas a tool in diagnostic histopathology. The varioustechnical problems encountered and our solutions tothese are reported. Many of the techniques we triedhad to be modified as the requirements and thepotential of the instrument were new to us, but noneof the problems seemed insurmountable and mostwere easily solved. We personally think that themain advantages of the LEM over conventionaltransmission electron microscopes in the field ofdiagnostic histopathology stem from the large gridsize. Furthermore, the capacity of the instrument toview larger sections could be useful if only a verylimited amount of tissue were available for micro-scopic examination. In such circumstances it isobviously an advantage not to have to divide aspecimen into parts for election microscopy andlight microscopy. The whole specimen can be em-bedded and a section from the whole face of thespecimen can be examined in the LEM.The microprocessor system for relocating areas of

interest is certainly a convenience although 3 mmdiameter "finder" grids perform the same functionin conventional electron microscopes. The large LEMgrid is a finder grid, but we found the relocatingsystem invaluable in dealing with this larger area.We know there are those who would prefer to retainseparate dedicated instruments for light and electronmicroscopy. One is compromising to some extentwith a combined instrument such as the LEM, butthe only compromise as far as the electron micro-scopist is concerned is the upper limitation inmagnification ( x 45 000). This, however, has to bebalanced against advantages such as the potential oflarge sections and the ability to look at exactly thesame section by light and electron microscopy.

on May 22, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.35.4.425 on 1 April 1982. D

ownloaded from

Combined light and electron microscope in routine histopathology

With the current difficulties of both justifying andfinancing electron microscopes within the healthservice, the introduction of the LEM offers a newview. Semithin microtomes can be used for sectioncutting; ultramicrotomes are not essential-somaking specimen preparation cheaper and simpler.The operation of the LEM does not require a highdegree of manipulative skill and it does not require adarkened environment.

In summary, we find the LEM 2000 to function as a

good transmission electron microscope. Its capacityto take large grids and sections reduces samplingerror and helps with orientation. The built-inmicroprocessor records and relocates areas ofinterest and photographs are all automaticallynumbered and have the co-ordinates of the fieldmarked on them. One has the opportunity to observethe same section by light and electron microscopy, or

to examine parallel mounted serial thick and thinsections on the same grid to optimise light andelectron definition.

We should like to thank ISI Ltd of Newmarket forthe opportunity to work with the LEM. We areindebted to Mrs Sue Knott for typing the manuscript.

References

Green GH, Kurreim F. Glycol methacrylate embedding ingeneral histopathology. ACP Broadsheet 97, 1981.

2 Meek GA. Practical electron microscopy for biologists.Wiley-Interscience, 1970:82.

Requests for reprints to: Dr DA Levison, Department ofHistopathology, St Bartholomew's Hospital, LondonECIA 7BE, England.

429

on May 22, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.35.4.425 on 1 April 1982. D

ownloaded from