Experience Microscopy the Differential Diagnosis Smallpox · KENNETH L. HERRMANN, AND BERNARD LOURIE Nationial Communiiiiicable Disease Ce/iter, Atliwita, Georgica 30333 Received

APPLIED MICROBIOLOGY, Sept. 1970, p. 497-504Copyright P) 1970 Am11er-icaIn Society for Microbiology

Vol. 20, No. 3Pri,ited ill U.S.A.

Experience with Electron Microscopy in theDifferential Diagnosis of Smallpox

GARY W. LONG,1 JOHN NOBLE, JR.,2 FREDERICK A. MURPHY,KENNETH L. HERRMANN, AND BERNARD LOURIE

Nationial Communiiiiicable Disease Ce/iter, Atliwita, Georgica 30333

Received for publication 23 March 1970

The usefulness of negative-contrast electron microscopy in the rapid differentialdiagnosis of poxvirus and herpesvirus exanthems is described in this study of 301specimens from patients with vesicular exanthematous diseases. Specimens frompatients with smallpox, various forms of vaccination complications, varicella,zoster (shingles), and herpes simplex are included in this evaluation. Electronmicroscopy, when applied to the study of lesion material, was found to be more

sensitive than the classical techniques of virus isolation in the diagnosis of bothpoxvirus and herpes/'varicella virus infections. However, since specific identificationof a virus within a group cannot be made morphologically by electron microscopy,it is recommended that both electron microscopy and virus isolation methods be em-

ployed for the routine differential diagnosis of vesicular exanthematous diseases inthe reference diagnostic laboratory.

Electron microscopy of lesion material in thedifferential diagnosis of poxvirus and herpesvirusexanthems was first employed in 1947 during anoutbreak of smallpox in New York (7). However,not until the present decade, with the advent ofnegative-contrast techniques involving the use ofphosphotungstate salts (1), has electron micros-copy become generally accepted as a practicaldiagnostic method. Its use has increased as thenecessary facilities have become available. Therapidity, sensitivity, and limitations of negative-contrast electron microscopy have been well de-scribed by several investigators (2, 5, 6, 8). How-ever, each of the reported studies has dealt withspecimens from fewer than 75 patients.The problems encountered by a regional or in-

ternational reference diagnostic laboratory relatedto quantity and quality of specimens and ade-quacy of accompanying clinical information arequite different from those of a research labora-tory. The Vesicular Disease Laboratory of theNational Communicable Disease Center is thesmallpox diagnostic reference laboratory for theUnited States and provides reference diagnosticsupport to the World Health Organization. Nega-tive-contrast electron microscopy has been one ofthe routine diagnostic techniques employed by theVesicular Disease Laboratory since 1966. Our re-

' Present address: Department of Pathology, University ofNew Mexico School of Medicine, Albuquerqlue, N.M. 87106.

2 Present address: Middlesex County Hospital, Watltham,Mass. 02154.

port presents the results obtained by this labora-tory on specimens submitted since 1 January1967. All smallpox and some nonsmallpox speci-mens were from patients in West Africa andSoutheast Asia. Other nonsmallpox specimenswere from patients in the United States.

MATERIALS AND METHODS

Electron microscopic observations were madebetween 1 January 1967 and 30 April 1969 on 301specimens from 287 patients. Specimens were sub-mitted to this laboratory because either a poxvirusor a herpesvirus was presumed to be responsible forthe patient's illness. The poxvirus specimens for thisstudy were from patients with smallpox (variola),vaccinia, or paravaccinia; the herpesvirus specimenswere from patients with chickenpox (varicella), zoster,or herpes simplex. The clinical information whichaccompanied the specimens was frequently inadequateand raised serious doubts regarding the validity ofthe primary clinical diagnosis. A specimen was in-cluded in the study only if at least one laboratorytest confirmed the presence of a virus, either byisolation in an appropriate host system or by electronmicroscopic identification. Accordingly, the 301specimens reported here were selected from a muchlarger group.

Usually the specimens submitted were dried smearsor swabs of lesion exudate, lesion fluid in capillarytubes, or crusts from desiccated lesions. However,viral isolates or tissues obtained from autopsy werealso submitted. Most specimens were shipped un-refrigerated. Upon arrival, they were either examinedimmediately or frozen.

497

on March 9, 2020 by guest

http://aem.asm

.org/D

ownloaded from

APPL. MICROBIOL.

For virus identification by electron microscopy, thespecimen was suspended in Mcllvaine's phosphatebuffer, pH 7.2. The volume of buffer used was de-pendent upon the quantity of the specimen.

Smears. Smears were suspended in 2 to 3 dropsof buffer for grid preparation; the specimen suspensionwas then further diluted to facilitate inoculation intothe appropriate host systems.

Swabs. Swabs were soaked and then vigorouslyagitated in 0.5 to 1.0 ml of buffer.

Fluid specimens. Fluid specimens were diluted withan equal volume of buffer. High-quality grids couldnot be prepared from undiluted lesion fluid becauseit contained excessive protein and cellular debris.

Crusts. Crusts were ground in thick-walled glasstissue grinders with sufficient buffer to produce an

opalescent suspension.Virus isolates from tissue cultures. Chorioallantoic

membranes were homogenized by vigorous agitationwith glass beads. No additional preparatory proce-dures were performed.

Autopsy tissues. Autopsy tissues were groundwithout an abrasive in mortars with an equal volumeof buffer. The suspension was clarified by low-speedcentrifugation.A drop of the specimen suspension was placed on

a paraffin-coated microscope slide alongside a dropof 2% sodium phosphotungstate, pH 7.0. Two Form-var-coated grids were prepared from each specimenby touching them successively to the drop of specimensuspension and phosphotungstate. Excess liquid wasremoved by touching the edge of each grid to filterpaper. The grids were then disinfected by exposureto ultraviolet light for 30 min before they were re-moved from the containment facilities of the labora-tory to the electron microscope area. The grids wereexamined until a viral particle with diagnostic mor-phology was observed. The grids were routinelyscanned at a magnification of 18,500, and a highermagnification was used to confirm morphologicaldetail. Examination for 1 min or less would frequentlyreveal diagnostic poxvirus particles in a positivespecimen; finding herpesvirus particles of diagnosticquality was usually more difficult and required longerperiods of examination. Each of the two grids wasexamined for at least 5 min before the specimen was

considered negative.Virus isolation was attempted in embryonated

eggs, in tissue cultures, or in both. The chorioallantoicmembranes (CAM) of 12-day-old embryonatedhens' eggs (3) were inoculated if variola, vaccinia,cowpox, or herpes simplex was suspected. Tissuecultures (human diploid lung fibroblast cells) were

inoculated only if the specimen was frozen whenreceived and if the clinical history suggested a varicellazoster infection.

RESULTS

The 301 specimens were separated into threegroups according to primary clinical diagnoses.Thirty-three (34.0%) were submitted with thediagnosis of a herpesvirus infection (Table 1).Electron microscopy failed to make an identifica-

tion in only one specimen when isolation studieswere successful. No poxvirus particles were foundin this group of specimens.

Table 2 summarizes the diagnostic experiencewith 61 specimens in which herpesvirus was iso-lated or identified (sum of herpesvirus specimensfrom Tables 1 and 3). Herpesvirus was isolated in16 (26.2%) and identified by electron microscopyin 60 (98.4%) of the specimens. Only one false-negative diagnosis was made by electron micros-copy, as compared with 45 false-negative diag-noses by isolation procedures.

Tlhe second group was composed of 268 speci-mens submitted with the primary clinical diag-nosis of a poxvirus infection (Table 3). Twenty-eight (10.4%) were found to contain herpesvirusparticles despite the fact that many were obtainedfrom patients in endemic areas for smallpox.Both CAM cultures and electron microscopy werepositive for poxvirus with 201 (83.8%) of the re-maining 240 poxvirus specimens. Eighteen false-negative diagnoses were obtained by electron mi-croscopy and 19 false-negative diagnoses by CAMculture (Table 4). Two specimens of paravacciniawere excluded from these figures comparing theaccuracy of electron microscopy and culturetechniques because these viruses cannot be cul-tured on the CAM or in human fibroblast cells.These two paravaccinia cases illustrate that thelaboratory diagnosis of some infrequently en-

TABLE 1. Virus isolationsa and identificationisb for33 specimens submitted with a primary clinical

diagnosis of a herpesvirus infection

Result NNo. ofspecimens

Herpesvirus isolated 15Herpesvirus identified .................. 32Herpesvirus isolated and identified...... 14Herpesvirus isolated or identified........ 33Poxvirus isolated or identified.................. 0

a Isolations by culture on the chorioallantoicmembrane of embryonated hens' eggs, in humandiploid lung fibroblast tissue culture, or in both.

b Identification accomplished by electron mi-croscopy.

TABLE 2. Summary of laboratory diagnoses for 61herpesvirus specimens

Result No. ofspecimens

Herpesvirus isolated.................... 16Herpesvirus identified 60Herpesvirus isolated and identified 15Herpesvirus isolated or identified ....... 61

498 LONG ET AL.

on March 9, 2020 by guest

http://aem.asm

.org/D

ownloaded from

DIFFERENTIAL DIAGNOSIS OF SMALLPOX

TABLE 3. Virus isolationsSa antd ide,ntificationlsb for268 specimenis submitted with a primary clintical

diagntosis of a poxviruts inifectionz

No. ofResult speci-

mens

Poxvirus isolated 219Poxvirus identified 222Poxvirus isolated and identified 201Poxvirus isolated or identified ........... 240

Herpesvirus isolated 1.....................Herpesvirus identified 28Herpesvirus isolated and identified1.......Herpesvirus isolated or identified ........ 28

a Isolations by culture on the chorioallantoicmembrane of embryonated hens' eggs, in humandiploid lung fibroblast tissue culture, or in both.

b Identification accomplished by electron mi-croscopy.

TABLE 4. Electronl microscopic (EM) anid chorio-allanitoic membranie (CAM) culture results

obtained for 240 poxviruis specimens

No. of specimens

Virus CAMI CANI CAMpositive, positive, negative,EM EM ENI

positive negative positive

Poxvirusa .............. 201 18 10(Variola) (179) (10) -

(Vaccinia) (22) (8)(Cowpox) (0) (0) -

Paravacciniab.. (2)c

Totals.. 201 18 21

a Refers to the genus Poxvirus which comprisesthe variola-vaccinia-cowpox group.

b Synonyms are milkers' nodes or pseudocow-pox.

c Members of this group of poxviruses do notproduce pocks on the CAM; therefore, a negativeresult does not represent a failure of the CAMculture technique.

countered poxviruses (molloscum contagiosum,Milker's nodules, orf, and Yaba monkey virus)may depend on electron microscopy when appro-priate tissue culture systems are not available.The distinct morphological differences between

particles of paravaccinia and those of the variola-vaccinia-cowpox group can be seen by comparingFig. 1 and 2. The particles in both preparationsare superficially penetrated by phosphotungstate,demonstrating the character and pattern of surfacefilaments. The vaccinia particle (Fig. 1) was ob-tained from a highly purified, washed suspension,

and phosphotungstate did not aggregate aroundit. The virion appears rectangular or brick-shapedwhen viewed lengthwise and circular when viewedon end (Fig. 1 inset). The surface filaments areshort and a round central body was frequentlyseen. The paravaccinia particle (Fig. 2) was ob-tained directly from unpurified specimen material.Phosphotungstate may be seen aggregated aroundthis particle. It appears cylindrical with a spiralarrangement of the surface filaments; a centralbody is not visible.

Poxvirus particles deeply penetrated by phos-photungstate were frequently encountered in di-agnostic specimens (Fig. 3). The surface filamentsare not visible in such preparations. Instead, thevirion consists ofa sharply defined, relatively densecore surrounded by several concentrically lami-nated zones of differing densities. The innermostzone, immediately adjacent to the core, and theoutermost zone are of low density and appearlighter. Between them lies a denser, darker zone.Frequently, the middle and outermost zones havea crinkled appearance. Those poxvirus virionswhich are superficially penetrated with phospho-tungstate are slightly smaller than their deeplypenetrated counterparts (Fig. 4). Poxvirus virionsare much smaller than bacilli and significantlylarger than herpesvirus virions (Fig. 5; a compo-site micrograph).The spectrum of viral particle degeneration was

broad and extended from nearly typical particles(Fig. 1 and 3-6) through degenerated, yet identi-fiable, virus particles (Fig. 7 and 8) to entitieswhich were nondiagnostic (Fig. 9). Some partiallydegenerated particles (Fig. 7) were considereddiagnostic because sufficient morphological detailwas evident. Particles such as the one in the lowerright of Fig. 4 and the one in the uper left of Fig. 8were too degenerated to be considered diagnosticin themselves, despite the fact that their size,shape, and immediate proximity to a nearly typi-cal particle would suggest that they were a part ofpoxvirus aggregate. Such particles, had they beenfound individually, would have presented muchthe same diagnostic problem as the one illustratedin Fig. 9. This last particle, although nondiag-nostic as previously stated, was sufficiently sug-gestive that additional time was spent searchingthe grids for particles of diagnostic quality. Fre-quently, suspicious but nondiagnostic objects werefound in crust specimens at low, scanning magni-fications (far left of Fig. 4). Further examinationof such particles at higher magnification failed toreveal diagnostic characteristics of the poxvirusgroup.Although herpesvirus particles were signifi-

cantly smaller than those of poxviruses (Fig. 5)

VOL. 20, 1970 499

on March 9, 2020 by guest

http://aem.asm

.org/D

ownloaded from

APPL. MICROBIOL.

I

'el

I)

FIG. 1. Purified poxvirus (variola-vaccinia-cowpox group), phosphotungstate stain2 (superficial penietration).Note the short surface filamenits antd the round cenitral body within the rectanigular virion. Iniset provides anl onl-enldview. X 150,000.

FIG. 2. Paravaccinia virus from specimen material, phosphotunigstate stain (superficial penietrationi). Note thesmaller size, cylindrical shape, spiral arrangement of surface filamenits and absenice ofa central body. X 150,000.

FIG. 3. Poxvirus (variola-vaccinia-cowpox group) from specimeni material, phosphotungstate staini. The deeppenetrationi of the staini permits observationi of three conicenitrically laminiated zonies around the core. X 150,000.

500 LONG ET AL.

on March 9, 2020 by guest

http://aem.asm

.org/D

ownloaded from

DIFFERENTIAL DIAGNOSIS OF SMALLPOX

FIG. 4 Vairiola viruis, phosphotlungstate staini. Sutperficiail/j penetraited variolai viruis particles aippear slightlysmaller thian the deeply penetrated onie. Object at thte left is a firequently eiicouiiitered inonviriis artifiict. Phiospho-tugstate stain of specimeii materiail. X 65,000.

they could be detected with relative ease at thelow magnifications used for scanning specimensfor pzxviruses. Indeed, the 28 herpesvirus diag-noses established by electron microscopy(Table 3) were made under such conditions. Fullyenveloped herpesvirus particles (Fig. 10 and 12)were most frequently found in specimens of vesic-ular fluid, whereas naked capsids (Fig. 11 and 14)were usually found in crust specimens. However,numerous enveloped particles were occasionallyfound in crust specimens.The morphological detail of herpesvirus par-

ticles was usually not as well preserved as that ofthe poxviruses. Enveloped particles as illustratedin Fig. 5 and 10 and naked capsids as in Fig. 14were fairly representative of the virus quality en-countered in specimens submitted to a referencediagnostic laboratory. Although less than ideal,the quality ofsuch a single particle was sufficient topermit a positive diagnosis. Occasionally, speci-mens contained only more severely degeneratedparticles (Fig. 15 and 16). Such a specimen wasconsidered positive only if three or more particleswere found.

DISCUSSIONNegative-contrast electron microscopy is well

established as a rapid means of detecting both

poxvirus and herpesvirus particles in diagnosticspecimen materials. The technique is rapid andsensitive, and can be used to detect inactivatedvirus. However, the technique has several short-comings, the major one being a lack of specificity.Differentiation is possible between the virusgroups (e.g., poxvirus versus herpesvirus) but notbetween the viruses within a group (e.g., variolaversus vaccinia). For this reason, specific identifi-cation of the virus by isolation techniques shouldbe attempted. The hazard of smallpox to a sus-ceptible population is of such magnitude that spe-cific identification of any poxvirus is essential.Another shortcoming of electron microscopy is

the relatively high concentration of virus particlesnecessary in a specimen. Macrae (4) estimatedthat 1 million to 10 million virus particles permilliliter must be present for electron micros-copy to give successful results. It is possible thateven higher concentrations of herpesvirus par-ticles may be necessary, because they appear todegenerate more quickly under adverse conditionsof transit and storage than do poxviruses.Although virus isolation is more specific, it is

slower in yielding an answer and it does not ap-pear to be any more sensitive than electron micro-scopy in the diagnosis of poxviruses. Even thehardy poxvirus may be inactivated in a small but

VOL. 20, 1970 501

on March 9, 2020 by guest

http://aem.asm

.org/D

ownloaded from

APPL. MICRO310L.

3_ 8

FiG. 5. C~omposite micrograph for size comparisoni ofpoxvirus (top), bacillus (cenzter), and lherpesvirius (bottomi).Phosphotuiigstate stain of specimeii materials. X 150,000.

FIG. 6-8. Variola virus particles in different degrees of degenterationi All particlici (except for the onie in theuipper left of Fig. 8),-etain sufficieuti morphological detail to permit diaignostic idenitificationt Phosphotioiigstate stainiof specime. material X 150 000

significant portion of specimens. The marked la-bility of the herpesviruses is a definite factor inaccounting for the superiority of electron micros-copy over isolation techniques.

This study is rather unique in that it was con-cerned with specimens which in large part wereobtained, stored, and shipped under adverseconditions to a geographically distant reference

502 LONG ET AL.

on March 9, 2020 by guest

http://aem.asm

.org/D

ownloaded from

DIFFERENTIAL DIAGNOSIS OF SMALLPOX

9

FIG. 9. Pairticles with a degree of degenterationt makinig diagnlostic idenitificationt impossible. Pliosphotunigstatestaini of specimen material. X 150,000.

FIG. 10. Eniveloped herpesvirus particles are commont in specimenis of vesicular flutid, but rare int crusts. Phlos-photunlgstate staint. X 150,000.

FIG. 11 and 12. Naked anzd enzveloped herpesvirus particles of exceptionial preservationz for specimen material.Phosphotunlgstate staill. X 150,000.

laboratory. The specimens of lesion material sub-mitted were, in order of increasing frequency,smears, fluids, and crusts.

Several generalizations may be made concern-ing the adequacy of the various types of specimensfor diagnostic procedures. For poxviruses, speci-mens of lesion fluid and crusts were found to besatisfactory for both virus isolation and identifica-tion, but poor results were obtained with smearspecimens. Herpesvirus hominus or Herpesvirusvaricellae was successfully isolated and identifiedfrom most lesion fluid specimens. Crust specimenswere usually satisfactory for electron microscopicidentification but not for isolation of herpes-viruses. As Cruickshank (2) has observed, herpes-virus particles in crust specimens are usually notnumerous and frequently are present as nakedcapsids with poorly preserved fine structure. As aresult of these limitations of crusts, electron mi-

croscopy presents the only practical means of ob-taining diagnostic information from such speci-mens.The importation of contagious diseases into

nonendemic countries is a current problem re-sulting from the use of jet aircraft for interna-tional travel. Outbreaks of imported contagiousdisease have already occurred in several countries,emphasizing the need for adequate surveillanceand diagnostic facilities. Smallpox has been one ofthe diseases involved in several such outbreaks ofimported disease. Negative-contrast electron mi-croscopy is of great value in the rapid differentialdiagnosis of vesiculating viral exanthems. Thetechniques of electron microscopic identificationand virus isolation are complementary, as eachcorrects the major diagnostic deficiencies of theother.

503VOL. 20, 1970

on March 9, 2020 by guest

http://aem.asm

.org/D

ownloaded from

LONG ET AL. APPL. MICROBIOL.

14

FIG. 13 and 14. Herpesvirias particles in varionis states of degeneration. Naked capsids are most frequentlyeuicoaiitered in cra.st specimens. Plhosphotungstate staini of specimen materiail. X 150,000.

FIG. 15 and 16. Severely degenierated herpesviruis particles which arce liot acdeqaiate fbr diagnostic identifica-tioni. Plhosphiotilagstate stainl of .specimen material. X 150,000.

LITERATURE CI'I'ED

1. Brenner-, S., and R. W. Horne. 1959. A negative stainingmethod for highl r-esolution electron microscopy of viruses.Biochim. Biophys. Acta 34:103-110.

2. Cruickshank, J. G., H. S. Bedson, and D. H. Watson. 1966.Electr-on microscopy in the rapid diagnosis of smallpox.Lancet 2:527-530.

3. Downie, A. W., and K. R. Dumbell. 1947. Isolation andcultivation of variola virus on the chorioallantois of chickembryos. J. Pathol. Bacteriol. 59:189-198.

4. Macrae, A. D. 1967. Laboratory diagnosis of smallpox. Mon.Bull. Min. Health Public Health Lab. Serv. 26:189-191.

5. Nagington, J. 1964. Electron imiicroscopy in differential diag-nosis of poxvirus infections. Brit. Med. J. 2:1499-1500.

6. Nagington, J., and A. D. Macrae. 1965. Smallpox diagnosisand electron microscopy. Mon. BuLll. Min. Health PublicHealth Lab. Serv. 24:382-383.

7. Nagler, F. P. O., and G. Rake. 1948. The use of the electronmicroscope in diagnosis of variola, vaccinia, and varicellai.J. Bacteriol. 55:45-51.

8. Williamiis, M. G., J. D. Almeida, and A. F. Howatson. 1962Electron microscope studies on viral skin lesions. Arch.Deriotatol. 86:290-297.

504

13

<*..

on March 9, 2020 by guest

http://aem.asm

.org/D

ownloaded from