Technical method

Technical method

A simple procedure for thepreparation of rosetted cells forelectron microscopyCM PAYNE AND VF SATTERFIELD Department ofPathology, University of Arizona Health SciencesCenter, Tucson, Arizona 85724, USA

Examination of surface markers on lymphocytesfrom the normal peripheral circulation has revealeda heterogeneity of lymphocyte populations.' Manyof the surface marker tests performed in a routineclinical immunology and pathology laboratoryutilise specific marker particles which adhere toreceptors on the cell surface, forming a rosette whichis visible by light microscopy. The ultrastructuraldetail of the central rosette-forming cell (CRFC)and the mode of interaction between the CRFC andits marker particles have been investigated in only alimited number of situations. In addition, the neces-sity for examining lymphocytes at the ultrastructurallevel has received new impetus after Payne andGlasser2 determined that lymphocytes which containspecific organelles called parallel tubular arraysrepresent a distinct subpopulation of lymphocytes.We have examined many rosetted preparations at

the ultrastructural level3 4 and have encounteredtwo major technical difficulties in preparing themfor transmission electron microscopy. The firstinvolved the undesirable packing of unrosetted cellsaround the rosettes after centrifugation. The secondinvolved the dissolution of the pellet during sub-sequent dehydration steps, often resulting in aninadequate specimen for evaluation. A survey of theliterature revealed that most authors did not discussthese problems in their methodology sections.5-12Most authors simply state that 'cell pellets werefixed, dehydrated, and embedded' for electronmicroscopy. We therefore present a simple techniquefor routinely preparing rosettes for electron micro-scopy which overcomes these two major technicaldifficulties and may be incorporated into any clinicalpathology laboratory set-up.

Material and methods

Rosettes were prepared by mixing a suspension ofwhite blood cells (1 x 106 cells/ml) with an equal

Received for publication 12 September 1979

volume of a red blood cell suspension (1-5 x 107cells/ml). The volume of the final rosetted prepar-ation prepared for electron microscopy variedbetween 0-2 and 4 ml. An equal volume of 3%glutaraldehyde in 0*1 M phosphate buffer (pH 7-2)was then added to the rosetted cell suspension andgently mixed by drawing the suspension up anddown in a glass pipette. Any pellet present was like-wise resuspended. The rosetted preparations wereallowed to pre-fix in the diluted glutaraldehyde for1 hour at room temperature. The suspension wasthen gently centrifuged at 1000 rpm in a SorvallGLC-1 swinging bucket clinical centrifuge for 5minutes. The pellet was resuspended in a smallaliquot of human plasma and transferred to a large(17 x 100 mm) round-bottomed plastic centrifugetube (No. 2059 tube; Falcon, Oxnard, Calif). Thesuspension was again centrifuged at 1000 rpm in theswinging bucket clinical centrifuge for 5 minutes.The supernatant was removed and replaced with

cold 3% glutaraldehyde in 0 1 M phosphate buffer.The pellet was allowed to fix for 1 hour at 4°C. Thefixative was then decanted and replaced with 0-1 Mphosphate buffer. After two changes of buffer(15 minutes each), the pellet was post-fixed for 14hours in 1 % osmium tetroxide in 0-1 M phosphatebuffer and dehydrated through a graded series ofethanol. The pellet was then loosened from thebottom of the plastic tube with a thin wedge-shapedwooden stick. The ethanol was replaced with a 1:1mixture of 100% ethanol and Spurr's epoxy'3 andallowed to infiltrate for 2-4 hours. The pellet wasbroken up into smaller pieces, transferred to a newvial of pure epoxy after blotting on filter paper, andallowed to infiltrate for approximately 20-24 hoursat room temperature. The small pieces were thentransferred to Beem capsules containing freshlymixed Spurr's epoxy and allowed to polymerise in a70°C oven for approximately 20 hours.One-micron sections were cut and stained with

toluidine blue and examined with the light micro-scope. Those blocks containing numerous rosetteswere selected, and ultrathin sections were cut with adiamond knife on a Sorvall MT2-B ultramicrotomeand mounted on uncoated 200-mesh copper grids.The ultrathin sections, each of which almost coveredthe entire grid surface, were stained with 5%aqueous uranyl acetate and lead citrate and thenlightly carbon-coated before examination with aHitachi HU-12 electron microscope.The above procedure was used with (1) human

lymphocytes which form spontaneous rosettes with505

505

on 24 May 2018 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.33.5.505 on 1 May 1980. D

ownloaded from

Technical method

.~~~~~~~ ~~~~~~~~~~ ......... ....S~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~. .'' '; . ^.'': '' :.:: : ..:~~~~~~~~~~~~~~~~~~~..fi...e.. l . .|Kll~~~f An-' .mk A

Low-power electronmicrograph of an 'EA' rosettepreparation. Rosetted cells can beeasily distinguished fromunrosetted cells since there is nopacking of red cells around therosettes. Three rosettes can beseen in the field. Thehomogeneous, slightly electrondense material which surroundsthe cells represents the plasmain which the cells wereresuspended. (uranyl acetate,lead citrate x 2900)

sheep erythrocytes ('E' rosettes), (2) human mono-nuclear cells which form rosettes with human Rh( +)erythrocytes reacted with anti-D antibody ('EA'rosettes), and (3) human mononuclear cells whichform rosettes with sheep erythrocytes reacted withsheep cell hemolysin and fresh human AB, Rh(-)serum as a source of complement ('EAC' rosettes).

Results

In every attempt to prepare 'E', 'EA', and 'EAC'rosettes for electron microscopy using the methoddescribed herein, the pellets remained intact duringthe dehydration and infiltration steps.

In all preparations examined, rosetted cells couldbe easily distinguished from unrosetted cells sincethere was no packing of red cells around the rosettes(Figure). This separation of cells was achievedirrespective of the volume of the initial rosettedsuspensions which varied between 0-2 and 4 ml.The separation of cells was due to two factors:

the use of round-bottomed tubes and the re-suspension of cells in plasma. A comparison of thearea formed by the pellet of cells at the bottom of thetubes revealed a larger area in the rosette preparationsthat were resuspended in plasma than in thosewithout.

Single ultrathin sections covered the entire grid

506

... ............ ..... ... .. ....................

on 24 May 2018 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.33.5.505 on 1 May 1980. D

ownloaded from

Technical method 507

surface and allowed one to examine numerousrosettes under the electron microscope at one time.

Discussion

The procedure outlined in this paper represents asimple technique which we use routinely whenpreparing rosettes for electron microscopy. Onemajor technical difficulty which was overcome wasthe undesirable packing of unrosetted cells aroundthe rosetted cells after centrifugation. This wasaccomplished by centrifuging the cell suspensions,which were resuspended in plasma, in large, round-bottomed plastic centrifuge tubes in a swingingbucket centrifuge which allowed the cells to spreadout in a near monolayer in some areas at the bottomof the tube. A comparison of the area formed by thepellet of cells at the bottom of the tube revealed alarger area in the rosette preparations that wereresuspended in plasma than in those without. Cellswhich are surrounded by a protein solution appar-ently do not pack as closely as those which aresurrounded by buffer or fixative. Spinning the cells insmaller, round-bottomed tubes or in conical tubescaused unrosetted cells to pack too closely around therosettes, thereby interfering with the visualisation ofthe true rosetted cells. The rosettes were prefixedbefore centrifugation to stabilise them during thecentrifugation steps. In an ultrastructural study ofhuman lymphocyte-sheep erythrocyte rosettes,Kataoka et al.'4 lightly agitated their tubes beforefixation to minimise the mechanical contact of thecells. This is an undesirable step since fragile rosettescould easily be disrupted. Reyes et al.'5 used thetechnique of micromanipulation to separate therosettes from unrosetted cells, which is very timeconsuming. Lay et al.16 used chamber-slides andinverted gelatin capsules of epon directly on to theslides. Chamber slides are not routinely used in aclinical immunology laboratory, and, additionally,the technique involves washing off non-adherentcells which would also result in washing off somerosettes. The inverted capsule technique is also time-consuming and is not applicable to the low-viscosityepoxy resin used here.'3The second major technical difficulty that was

overcome was the dissolution of the cell pelletduring the dehydration and infiltration steps whichcould result in an inadequate specimen for evalua-tion. The object was to replace the medium surround-ing the cells with a protein-rich matrix which wouldcross-link upon contact with glutaraldehyde andbehave as a piece of tissue. Three different sources ofprotein were utilised including human plasma, a 7%bovine serum albumin (BSA) solution, and a 1 %agar solution. Human plasma was superior to the

other solutions and formed a firm gel uponcontact with 3% glutaraldehyde, thereby bindingthe cells to each other while at the same timemaintaining their spatial orientation. The 7% BSAsolution, although a similar protein concentrationto human plasma,'7 did not form as firm a gel withglutaraldehyde as the plasma and, in addition, wasdifficult initially to get into solution; it is alsorelatively expensive. The 1 % agar solution is a solidat room temperature and must be heated in orderto resuspend the cells. Care must also be exercisedto prevent premature cooling and trapping of cellsin the gel before they can be spread at the bottomof the centrifuge tube. Because of these difficultiesin working with BSA and agar, human plasma waspreferred.Abramson et al.18 used a technique developed by

Anderson and Doane'9 for the ultrastructuralexamination of rosettes which employs a com-mercially available vinyl cup. No details were givenfor preventing loss of material during the fixationand dehydration steps. In addition, Anderson andDoane state that difficulty in stripping the vinyl cupfrom the polymerised plastic was encountered whenthe temperature of polymerisation exceeded 60°C.The Spurr's epoxy resin we use requires polymeris-ation at 70°C. Therefore, our technique forpreparing rosettes is simple and reproducible andcan be used with any epoxy or other plastic resin.One advantage, however, in using Spurr's epoxy isthe large block face which can be thin-sectioned,enabling many rosettes to be examined under theelectron microscope at the same time. Embeddingthe thin cell pellet in Beem capsules as opposed toflat embedding moulds is recommended in order toobtain large surface areas for cutting.We feel that this technique can be easily adapted

to any clinical or research laboratory setting and willenable valuable material to be consistently preservedfor ultrastructural evaluation.

We thank Dr Jack Layton (head, Department ofPathology), whose continuing support has beeninvaluable, and Sherry Fogleman for typing themanuscript.

References

Ross GD. Surface markers of B and T cells. Recenttechnical developments reveal a heterogeneity oflymphocyte subpopulations. Arch Pathol Lab Med1977;101 :337-41.

2 Payne CM, Glasser L. The effect of steroids onperipheral blood lymphocytes containing paralleltubular arrays. Amer J Pathol 1978 ;92 :611-8.

3 Payne CM, Jones JF, Sieber OF, Jr, Fulginiti VA.

on 24 May 2018 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.33.5.505 on 1 May 1980. D

ownloaded from

508 Technical method

Parallel tubular arrays in severe combined immuno-deficiency disease: an ultrastructural study of peri-pheral blood lymphocytes. Blood 1977 ;50:55-64.

4 Payne CM, Nagle RB. Complement receptors onnormal human lymphocytes containing paralleltubular arrays. Amer J Pathol 1980, in press.

5 McKenna RW, Parkin J, Gajl-Peczalska KJ, Kersey JH,Brunning RC. Ultrastructural, cytochemical, andmembrane surface marker characteristics of theatypical lymphocytes in infectious mononucleosis.Blood 1977;50:505-15.

6Lo Buglio AF, Cotran RS, Jandl JH. Red cells coatedwith immunoglobulin G: Binding and sphering bymononuclear cells in man. Science 1967;158:1582-5.

7Douglas SD, Huber H. Electron microscopic studiesofhuman monocyte and lymphocyte interaction withimmunoglobulin- and complement-coated erythro-cytes. Exp Cell Res 1972;70:161-72.

8 Elson CJ, Bradley J, Howells RE. The mechanism ofrosette formation between Rh(D)-positive erythro-cytes and peripheral blood lymphocytes from Rhisoimmunized individuals. The role of surface micro-projections. Immunology 1972;22:1075-86.

i Chen LT, Eden A, Nussenzweig V, Weiss L. Electronmicroscopic study of the lymphocytes capable ofbinding antigen-antibody-complement complexes.Cell Immunol 1972;4:279-88.

10 Bentwich Z, Douglas SD, Siegal FP, Kunkel HG.Human lymphocyte-sheep erythrocyte rosette form-ation: some characteristics of the interaction. ClinImmunol Immunopath 1973 ;1 :511-22.

Galey FR, Prchal JT, Amromin GD, Jhurani Y."Hairy" B cells and "smooth" T cells. (Letter.)

N EnglJ Med 1974;290:690.12 Marchalonis JJ, Bucana C, Hoyer L, Warr GW,

Hanna MG, Jr. Visualization of a guinea pig Tlymphocyte surface component cross-reactive withimmunoglobulin. Science 1978;199:433-5.

13 Spurr AR. A low viscosity epoxy resin embeddingmedium for electron microscopy. J Ultrastruct Res1969;26:31-43.

14 Kataoka K, Minowada J, Pressman D. Electronmicroscope study on human lymphocyte-sheeperythrocyte rosettes. J Natl Can Inst 1975 ;55 :1323-5.

15 Reyes F, Le Go A, Delrieu F, Bach JF. Ultrastructureof cells binding immunoglobulin-coated erythrocytesin rheumatoid arthritis. Clin Exp Immunol 1974;17:533-46.

16 Lay WH, Nussenzweig V. Receptors for complementon leukocytes. J Exp Med 1968;128:991-1009.

17 Bell GH, Davidson JN, Scarborough H. Textbook ofPhysiology and Biochemistry, 5th ed. Edinburgh:Livingstone. Baltimore: The Williams and WilkinsCompany, 1961.

18 Abramson N, Lo Buglio AF, Jandl JH, Cotran RS.The interaction between human monocytes and redcells. Binding characteristics. J Exp Med 1970;132:1191-206.

19 Anderson N, Doane FW. Epoxy embedding of thin-layer material in commercially available vinyl cupsfor light and electron microsocpy. Stain Tech 1967;42:169-73.

Requests for reprints to: Dr Claire M Payne, Departmentof Pathology, University of Arizona Health SciencesCenter, Tucson, Arizona 85724, USA.

Letters to the Editor

Use of Intralactam for detection of/3-lactamase production by Neisseriagonorrhoeae

In your Journal of July 1979 (p 738), DBWheldon and Mary PE Slack reported anevaluation of Intralactam for the detectionof ,B-lactamase production by Haemophilusinfluenzae. At the end of their paper theycommented that it might also be used todetect gonococcal f3-lactamase.We have used this acidometric method

to detect /3-lactamase production inprimary isolates of Neisseria gonorrhoeaefor the past year. Initially we preparedpaper strips impregnated with penicillinand indicator, as described by Slack et al.,'but then began using the commerciallyavailable Intralactam.

All confirmed isolates of N. gonorrhoeaehave minimum inhibitory concentrations(MIC) determined to penicillin, spectino-mycin, and tetracycline. At first, testingfor /-lactamase production was limitedto isolates having an MIC to penicillin of> 2-0 mg/I. (Exact quantification of theMIC value above 2 mg/I was notperformed.) Of 2357 N. gonorrhoeaeisolated in this laboratory from clinicalspecimens in 1978-79, 38 organisms (1-6 %)had an MIC to penicillin of > 2-0 mg/l.These 38 organisms had a ,B-lactamasedetermination performed, the majorityusing Intralactam strips, but a smallnumber using our own prepared stripswith penicillin and neutral red incor-porated, as described by Slack et al.' Ofthese 38 organisms, 36 were identified as

f-lactamase producers. However, as MICresults and identification of ,-lactamaseproduction were not usually available tothe clinicians until four days after primaryisolation, we decided to perform f-lactamase determination on all isolatesusing Intralactam strips at the same timeas a presumptive diagnosis of N. gonor-rhoeae was made.We now provide clinicians with a pre-

sumptive diagnosis of N. gonorrhoeae ifoxidase-positive, Gram-negative diplo-cocci are present on primary isolationplates after 18-24 hours' incubation, andat the same time report the presence orabsence of fl-lactamase production. Theprimary isolation medium used is modi-fied New York City Medium.2We have used Intralactam strips to

on 24 May 2018 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.33.5.505 on 1 May 1980. D

ownloaded from