Feb. Vol. for Pathogenesis ofHerpetic Neuritis ... · in the peripheral nervous system constitutes the present report (an abstract ofthis material has been presented, Bacteriol, Proc

INFECTION AND IMMUNITY, Feb. 1973, p. 272-288Copyright ( 1973 American Society for Microbiology

Vol. 7, No. 2Printed in U.S.A.

Pathogenesis of Herpetic Neuritis andGanglionitis in Mice: Evidence for

Intra-Axonal Transport ofInfection

MARGERY L. COOK AND JACK G. STEVENS

Reed Neurological Research Center and Department of Medical Microbiology and Immunology,University of California at Los Angeles School of Medicine, Los Angeles, California 90024

Received for publication 25 September 1972

The pathogenesis of acute herpetic infection in the nervous system has beenstudied following rear footpad inoculation of mice. Viral assays performed onappropriate tissues at various time intervals indicated that the infection pro-gressed sequentially from peripheral to the central nervous system, with infec-tious virus reaching the sacrosciatic spinal ganglia in 20 to 24 hr. The infectionalso progressed to ganglia in mice given high levels of anti-viral antibody. Im-munofluorescent techniques demonstrated that both neurons and supportingcells produced virus-specific antigens. By electron microscopy, neurons werefound to produce morphologically complete virions, but supporting cells repli-cated principally nucleocapsids. These results are discussed in the context ofpossible mechanisms by which herpes simplex virus might travel in nervetrunks. They are considered to offer strong support for centripetal transport inaxons.

Despite the volume of published work con-cerning the interaction between herpes simplexvirus (HSV) and nervous tissue, several aspectsrelating to the pathogenesis of this infection re-main undefined. Among the most importantare: (i) The reaction of various cell types in thenervous system to viral infection, (ii) the phe-nomenon of viral latency as it relates to nervoustissue, and (iii) the mechanism(s) by which vi-rus is transported in the nervous system.With these aspects in mind, an experimental

model in the mouse has been studied in detail.In this model, when virus is inoculated into arear footpad of a mouse, the infection pro-gresses sequentially through the peripheral tothe central nervous system (12, 29). We previ-ously showed that HSV establishes a latent in-fection in the sacrosciatic spinal ganglia ofmice inoculated in this fashion (26). A defini-tion of the pathogenesis of the acute infectionin the peripheral nervous system constitutesthe present report (an abstract of this materialhas been presented, Bacteriol, Proc. 71:198,1971). The experiments were designed andare discussed with particular emphasis placedon a definition of the route(s) by which virustravels in the peripheral nervous system.

272

MATERIALS AND METHODS

Virus. The Maclntyre strain of HSV, a prototype1 strain (7), was kindly supplied by Matthew A.Bucca of the Viral Reagents Unit, CommunicableDisease Center, Public Health Service, Atlanta, Ga.It was serially passed 17 to 23 times in vivo in mousebrains before use in these experiments. A typicalpool possessed a titer of 4 x 104 RK,3 cell plaque-forming units (PFU)/ml; this corresponds to 2 x 105mouse intracerebral mean lethal doses per ml.

Mice. Four-week-old SJL mice obtained from TheJackson Laboratories, Bar Harbor, Me., were usedin all experiments.Method for inoculation of mice. Prior injection

of a hypertonic saline solution at the same site en-hances the neuropathogenicity of HSV inoculatedinto the footpads of adult mice (19, 29). Our standardmethod of infection is based on that observation. A0.1-ml amount of 10% aqueous NaCl solution was in-jected into the left, hind footpad of an anesthetized4-week-old SJL mouse. This was followed in 6 hr byabrasion of the edematous area of the foot with 20light strokes from a fine emory board. This procedureremoves the stratum corneum but causes little, ifany, bleeding. From a 26-gauge needle, 1 drop of thevirus suspension was placed on the abraded areaand gently rubbed with the shank of the needle (10strokes).

Tissue assays. Tissues were removed from in-

on July 20, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

http://iai.asm.org/

HERPETIC NEURITIS AND GANGLIONITIS IN MICE

fected mice at appropriate times and frozen at-70 C. When all tissues were accumulated, similarspecimens from four or five mice were pooled andground in Ten Broeck homogenizers to make a 20%(w/w) suspension in minimal essential medium(MEM) and 5% fetal calf serum. Smaller sampleswere prepared as 1% suspensions. These prepara-tions were then centrifuged for 10 min at 12,000 x gin a Sorvall RC-2B refrigerated centrifuge, and thesupernatant fluid was used for viral assays on RK,3cells. The spinal chord was removed by the methodoriginally described by Wildy (29). After the animalhad been sacrificed and the brain removed, the dor-sal skin was reflected, and a 20-gauge needle at-tached to a syringe filled with MEM was insertedinto the spinal canal at an appropriate level. Quickpressure on the plunger forced the spinal cordthrough the foramen magnum. For removal of dorsalroot ganglia, the sciatic nerve and its branches wereexposed and followed to the vertebral column undera low-power dissecting microscope. At this point,using fine forceps and extreme care, it was possibleto break the spinal column near the ganglia and re-move them intact. Employing this technique, five tosix ganglia associated with the sciatic nerve can rou-tinely be removed from one or both sides of the cord.For removal of dorsal and ventral roots, the ventralsurface of the vertebral column was cut with finescissors; the roots were then exposed and removed.

Cells. The RK13 rabbit kidney cell line was usedfor propagation of the viral stocks of cell culture ori-gin and for quantitative assays of mouse tissues. Themethods employed for maintenance and use of thiscell line for viral assays have been described previ-ously (27).

For studies of the viral growth cycle in vitro, mouseembryo fibroblasts derived from embryos processedat the 19th day of gestation were used. They wereprepared and maintained by employing methods de-veloped for chicken embryo fibroblasts (24).

Detection of neutralizing antibody in mousesera. Heart blood was removed from those mice sac-rificed for viral assay of tissues. Blood samples fromall mice sacrificed on a given day were pooled, andsera were removed, stored at -70 C, and inactivatedat 56 C for 30 min before use. They were then em-ployed at increasing dilutions (beginning at 1:5)against 20,000 PFU/0.2 ml of virus at 37 C for 1 hr.The antisera-virus mixture was then diluted 1:100and assayed on RK,3 cells.

Light microscopy. Tissue samples were takenfrom five infected mice on days 1 to 9, 11, 13, 15, 18,25, and 32. They were then fixed in buffered Forma-lin, sectioned, and stained with hematoxylin andeosin.

Immunofluorescent microscopy. The indirectimmunofluorescent method was used to detectherpes simplex antigens in frozen sections of appro-priate tissues. Sections of mouse tissue were cut on acryostat at 6 Mm and fixed in acetone prior to stain-ing. Specific antiserum to the Maclntyre strain ofHSV was raised in New Zealand White rabbits byinjection of maximally infected RK, 3 cells inFreunds complete adjuvant (eight biweekly injec-tions over a 4-week period using 107 PFU per injec-

tion). The pooled sera from six rabbits possesseda neutralization constant of 72 ml/min at 37 C (1).Using standard procedures (21), both pre- and post-immunization sera were employed in the primaryreaction, and fluorescein-labeled goat anti-rabbitgamma globulin antiserum (Antibodies Inc., Davis,Calif.) was used in the secondary reaction. Stainedsections were examined with a Leitz fluorescencemicroscope. Control sections consisted of infectedtissues exposed to preimmune serum and normaltissues exposed to immune serum prior to treat-ment with fluorescene-conjugated antiserum.

Electron microscopy. Tissues from infected andcontrol mice to be examined by electron microscopywere either removed after perfusion of the mousewith fixative (1.5% glutaraldehyde in 0.1 M sodiumcacodylate buffer at pH 7.3) or they were bathed infixative during the dissection and quickly taken fromanesthetized mice. For perfusion, four infected micefor each time period examined were anesthetized,and the heart was exposed. The tip of the left ven-tricle was amputated, and a cannula, connected to areservoir containing fixative, was inserted into theleft ventricle and held in place with a small curvedforceps. An incision was made in the right auricularappendage for efflux of blood and perfusate. The per-fusion was usually started within 30 sec of the thora-cotomy and was continued for 5 to 10 min. Blanchingof tissues indicated a successful perfusion. Whiletissues were being excised, they were bathed andthen collected in fixative. After all tissues were col-lected and held in fixative for a minimum of 1 hr,they were rinsed in buffer alone, postfixed in osmicacid (1% in cacodylate buffer at pH 7.3), and pro-cessed for electron microscopy (4). Thick sections(1 Mm) of these specimens were cut, stained withtoluidine blue, and examined by light microscopy forselection of areas to be examined by electron micros-copy.

RESULTSClinical observations. In all experiments,

mice developed acute inflammatory lesions inthe footpad which healed completely by the14th day after infection. A flaccid paralysis de-veloped in the ipsilateral leg of approximately90% of mice by the 8th day after infection; theremainder did not develop clinically apparentdisease. Of those paralyzed, approximately30% died with encephalitis by day 12. Thisevent was often preceeded by paralysis in thecontralateral leg. About 40% of paralyzed micerecovered within 3 weeks without sequelae,whereas the remainder were left permanentlyparalyzed.

Viral assays. To establish the route andkinetics of viral passage to the brain during thisacute disease, pertinent tissues were homoge-nized and assayed for virus at intervals afterinfection. The results of a representative ex-periment are shown in Fig. 1. By comparisonof the characteristics of viral growth cycles in

273VOL. 7, 1973


.org/D

ownloaded from

http://iai.asm.org/

COOK AND STEVENS

Ic

FIXvarioirear )reprecouldlymp(sensity <becaifecti(thesemaxiceptiinfec

perijpeakpriorsigniIt isbe r(or dr

)7 _ . neutralizing antibody became detectable inT8 sera by the 6th day of infection. (A 1:5 dilu-

/\F<ETF tion of serum neutralized 20 of 100 PFU of virus* m \at 6 days; a 1:25 dilution neutralized > 75

of 100 PFU by 10 days). We consider these>6 _ Z . * \ data to be support for the concept that se-l.6*\ quential infection of the nervous system is the

pathway by whichHSV passes from the foot tothe brain of the mouse.Light microscopy of acute disease. Path-

)5 - ological changes in the feet and peripheral ner-vous system of infected mice during the acute

l : X i \ infection were defined by conventional histo-I!/\l \ logical techniques and by observation of thick

sections at the time that suitable areas were)4 selected for observation by electron micros-

1.'l\l * \ copy. The feet of infected mice demonstrated. l/ \ t \acute necrotizing lesions starting on day 1. This

increased until day 4 when epidermal cells.;6.7\\l were recruited into giant cells containing intra-

)3 h6gOOTS G nuclear inclusions characteristic of herpetic inROOTS i ~~GANG/L/A fection.

SCIATIC An extensive infiltration of inflammatoryNERVES SPINAL -I cells (Fig. 2 and 3) was observed in the sciatic

COROS,i nerve and the associated spinal ganglia. These)2 - 'BRAINS *. infiltrates, predominately composed of macro-

phages and lymphocytes, began to appear/ , *9about 2 days after virus was first detected in

I-t-il -; these tissues and reached maximal intensity inthe next 2 to 3 days. By day 6, neurons in theganglia became vacuolated and were sur-rounded by nests of round cells, while someswelling of myelin was observed in the nervetrunk. Increased numbers of inflammatory cellswere associated with the blood vessels in both

c I I ganglia and nerve. This inflammatory response2 4 6 8 10 12 was maintained but in diminishing amounts

1.DAountofYhrpesimpleviruthrough at least the 32nd day after infection.G. 1. Amount of herpes simplex virus present In Imuoureen miospy factius tissues after intradermal inoculation of one Immunofluorescent mieroscopy of acutefootpad of SJL mice. The squares on all curves infection. To determine whether differingXsent levels of sensitivity for the assay. Virus cell types produced virus-specific antigens,i not be recovered at anytime from draining tissues from the peripheral nervous system re-)h nodes (sensitivity < 50 PFUIg), blood serum moved during the acute infection were exam--itivity < 10 PFU/ml), or ventral roots (sensitiv- ined by immunofluorescent techniques. In500 PFU/g). In addition, neutralizing antibody nerve trunks, virus-specific fluorescence wasme detectible in sera by the 6th day after in- observed, but it was not possible to identify theon. cell types involved (Fig. 4). In sacrosciatice tissues, particularly the times at which spinal ganglia, neurons and satellite cells ex-imal viral titers are reached, with one exhibited virus-specific fluorescence (Fig. 5), andion there is a sequential progression of the unidentifiable antigen-containing cells weretion from the foot to the brain through the observed in the dorsal roots (Fig. 6). In all?heral and central nervous systems. The cases, control sections of normal tissues; viral titer observed in dorsal root ganglia stained with immune serum or virus-infectedr to that in the associated sciatic nerve is tissues stained with pre-immune serum dem-ificant and will be discussed in detail later. onstrated only insignificant background fluo-also important to note that virus could not rescence. It is important to note that no posi-ecovered from either serum, ventral roots, tive fluorescence could be detected in eitherraining lymph nodes at any time, and that nerves or ganglia after the 6th day of infection.

IC

IC

1Cw

U~)

(.9

LLa.

274 INFECT. IMMUNrry


.org/D

ownloaded from

http://iai.asm.org/


Ri ~~~~~~~~~~~~~~~3FIG. 2. Micrograph of sciatic nerve from a mouse 5 days postinfection showing focal and diffuse round cell

infiltration. x 120.FIG. 3. Micrograph of a sciatic spinal ganglion from a mouse 4 days postinfection exhibiting a heavy infil-

tration of round cells. x 120.

As shown earlier (Fig. 1), this is also the time atwhich infectious virus can no longer be de-tected.

Electron microscopy of acute disease.From the experiments involving immunofluo-rescent methods, it was clear that neurons andsatellite cells were producing viral antigens.Ultrastructural methods were employed to es-tablish whether these and other cells were

replicating morphologically complete virions.Here, unless otherwise stated, all specimenswere processed at 4 days postinfection.Neurons. An important feature revealed in

ganglia was the productive infection of neu-

rons. As can be seen in Fig. 7, the cytoplasm ofthis neuron contains many mature viruses sur-

rounded by vacuolar membranes. The nucleusdemonstrates marginated chromatin and con-

tains a nucleocapsid as well as a virion associ-ated with membranous material. That this in-tranuclear membranous material is probablyderived from the inner nuclear membrane isseen in Fig. 8. Here, the nucleus contains a

loop of membranes derived in part from the

inner nuclear membrane. The cytoplasm alsocontains such a loop, but it is clearly seen tobe derived from both the inner and outer nu-clear membranes. Both membranous loops con-tain viral capsids. These membranous ex-tensions are considered to represent the nu-clear membrane proliferation and reduplica-tion often seen in various cell types infectedwith herpesviruses (17). That the envelope ofthe virus can be derived from the neuronal in-ner nuclear membrane, a common feature inthe morphogenesis of many herpesviruses (5)is seen in Fig. 9A and B. In both figures, con-tinuity between the inner nuclear membraneand the viral envelope is demonstrated. In ad-dition, the outer nuclear membrane surroundsthese virions. Dense, granular bodies oftenassociated with developing viral forms werefrequently encountered in nuclei of infectedneurons. Two of these structures and viralcapsids in the nucleus of a neuron are shownin Fig. 10.An observation made repeatedly with in-

fected neurons was the presence of virions be-

VOL. 7, 1973 275

_',, ,, .4.


.org/D

ownloaded from

http://iai.asm.org/

COOK AND STEVENS

tween the neuronal plasma membranes andadjacent plasma membranes of satellite cells(Fig. 11). With rare exceptions, all extracellularviruses seen were observed in this location. Re-tention of viral products in this area certainlymust contribute to the bright staining often ob-served to be outlining neurons in infected gan-glia examined by immunofluorescent tech-niques.

Several examples of intra-axonal virions wereseen. In all cases, they were in ganglia and inclose proximity, but not always adjacent, toinfected perikarya. An example is seen in Fig.12. Both the infected neuron and satellite cellcontain intranuclear capsids. An axon contain-ing a virion inside a vacuole is seen in the cyto-plasm of this satellite-Schwann cell. This areais seen at a higher magnification in Fig. 13.From this micrograph it is impossible to decidewhether the morphologically defective viruses

in the vicinity of the axonal membrane are en-tering the supporting cell from the axon orwhether the reverse is true. However, there ap-pear to be remnants of envelopes here andsince satellite and Schwann cells envelopevirus only inefficiently, (see below), these vi-ruses most probably come from the axon. Weinterpret all these virions to represent retro-grade infection (an infection progressing cen-trifugally) in axons since they were only seen inganglia containing heavily infected neuronalsomas.These observations constitute a direct dem-

onstration that HSV can be found inside axonsand imply that they may also travel intra-axonally. In summary, the results presented inthis section show that morphologically maturevirions are replicated in neurons and can befound in axons.

Satellite cells. The two most significant ob-

- - - U W_ -w .

FIG. 4. Micrograph of a sciatic nerve from a mouse 6 days postinfection stained by indirect immunofluroes-cent methods first employing an antiserum specific for herpes simplex antigens. Subsequently, sections werestained with fluorescein-labeled goat antiserum to rabbit gamma globulin. Some unidentified cells exhibitherpes-specific immunofluorescence. x430.

FIG. 5. Micrograph of a sciatic spinal ganglion from a mouse 4 days postinfection. Specific immunofluores-cent staining of herpes simplex virus antigens can be seen in a neuron (Ne) and adjacent satellite cells (S).Other stained and unstained neurons are shown in this section. x430.

FIG. 6. Micrograph of a dorsal root from a mouse 5 days postinfection. Specific immunofluorescent stainingof herpes virus antigens are present in unidentified cells. x430.

276 INFECT. IMMUNITY

I on July 20, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

http://iai.asm.org/

*.. *1l

.t sr 4v r.b. r+v

.a If;-1 .N 4w t

.j*. 4

a

....

-t I

-' a,

et

.\ a

..

, 5b

sr.

V

at *

r 4_o

F. -

A'

4B~~~~~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~

t; < s s/ 4,.i{t

A~ ~ ~

_ t t t jj[ 2 '>4V_i

FIG. 7. A productively infected ganglionic neuron containing herpes simplex virus. The nucleus (N) containsa nucleocapsid (arrow) and a virion associated with membranous material. Many mature viruses (V) surroundedby vacuolar membranes are seen in the cytoplasm (Cy). x40,000.

277

.I.


.org/D

ownloaded from

http://iai.asm.org/

. t.';., 4¢ }

i r.- 1 t

we * b *:i

,v.~~~~~~~~~~~~~~~~~~~~~U

4~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~4

#e -..r

* U~~~~~~~~~~~~-

FIG. 8. Proliferating nuclear membrane of a ganglionic neuron. Loops containing herpes simplex virus cap-sids project into the nucleus (N) and cytoplasm (Cy). x51,000.

FIG. 9. Examples depicting the derivation of the herpes simplex virus envelope from neuronal nuclear mem-branes. Continuity between the inner nuclear membrane and the viral envelope can be seen (N, nucleus; Cy,cytoplasm). x64,000.

FIG. 10. Portion of an infected neuronal nucleus containing dense granular bodies associated with develop-ing viral forms. x43,000.

FIG. 11. Example of herpes simplex virions present between the plasma membranes of neurons (Ne) anda satellite cell (S). x26,000.

278

4.

i....I.


.org/D

ownloaded from

http://iai.asm.org/

mij

--

FIG. 12. Survey micrograph of an infected neuron (Ne) and an adjacent supporting cell (S). Herpes simplexvirus capsids (arrows) are present in the nuclei of both cells, and at least one virion can be seen in the cyto-plasm (Cy) of the neuron. A virion is also seen in an axon (Ax) surrounded by the supporting cell cytoplasm.x 17,000.

FIG. 13. Higher magnification of the axon present in Fig. 12. A virion surrounded by a vacuolar membrane(arrow) is considered to be in the cytoplasm of a neuron. Whether morphologically defective viruses along theaxonal membrane (*) are entering or leaving the supporting cell cannot be determined in this micrograph.X67,000.

279


.org/D

ownloaded from

http://iai.asm.org/

COOK AND STEVENS

servations made in infected satellite cells were(i) an extraordinary proliferation of membranesinside nuclei and (ii) a paucity of morphologi-cally mature virions. Examples of these phe-nomena are shown in Fig. 14. Here, a satellitecell is adjacent to the cytoplasm of an infectedneuron. The satellite cell nucleus includes atleast three areas of membrane proliferationand a few scattered viral capsids and nucleo-capsids, one of which is in the satellite cell cy-toplasm. No mature virions can be seen. Al-though virions can be found in these cells, theyare observed only rarely. The neuron nucleusin the upper right contains viral capsids andnucleocapsids. Infected neurons surroundedby uninfected satellite cells were observed; theconverse was not true. This implies that theneuron is the cell initially infected in the gan-glion.One of the areas of membrane proliferation in

the satellite cell is depicted at a higher powerin Fig. 15. In this micrograph, it is clearly seenthat extensions from the inner and outer nu-clear membrane are invaginating the nucleusof the satellite cell and forming membranouswhorls. Free ribosomes and a viral capsid arelying inside the pore. An area of increaseddenseness between pairs of membranes is alsoseen. This denseness appears to be analogousto phenomena associated with cell fusionnoted by others in HeLa cells (18). Theseauthors suggested that fusion followed the ac-cumulation of viral antigen at surfaces of pro-liferated membranes.Schwann cells associated with myeli-

nated axons. The most significant aspect ofthose Schwann cells associated with myeli-nated axons was the almost total lack of evi-dence for viral infection. These Schwann cellswere observed to contain viral products only inganglia (never in the nerve or nerve root) and inthis location only twice. Figure 16 representsone of these cells. Here, the Schwann cell nu-cleoplasm contains a few scattered viral cap-sids. No extraneous membranes or mature vir-ions were ever seen in these cells.

If intra-axonal movement of virus in nervetrunks is the principal mechanism by whichthis virus travels in the nervous system (seeDiscussion), then these cells could be less per-missive because of the myelin barrier betweenaxon and the Schwann cell cytoplasm proper.However, myelinating Schwann cells in virus-producing ganglionic organ cultures derivedeither from latently infected mice or from nor-mal ganglia infected in vitro were never seen tocontain viral products (unpublished observa-tions). These latter observations would suggest

that the myelinating Schwann cell is refractoryto infection in this model.Schwann cells associated with nonmyeli-

nated axons. The reaction to infection of non-myelinating Schwann cells found in the sciaticnerve, ganglia, and dorsal root was identical tothat of satellite cells in ganglia. That is, infec-tion was accompanied by excessive nuclearmembrane proliferation and a paucity of ma-ture viruses.A ganglionic Schwann cell enclosing nonmye-

linated axons is seen in Fig. 17. The nucleus ofthis cell contains marginated chromatin and amembranous whorl, effects which are seen innonmyelinating Schwann cells in sciatic nervesand ganglia. It is of interest to note here thatsimilar whorls were seen in an oligodendrocytefrom a case of herpetic encephalitis in man (3).Viral capsids are associated with the whorls.A suggestion that the nuclear membrane is theorigin of those excess membranes can be seenin this micrograph. In addition, the increaseddenseness occasionally seen between pairs ofthese membranes is demonstrated.

Finally, Fig. 18 also depicts a nonmyelinat-ing, ganglionic Schwann cell. This cell wouldappear to be in a late stage of infection sincethere is extensive membrane proliferation asso-ciated with immature viral forms in the nu-cleus. Numerous capsids are also present inthe cytoplasm, presumably as a result either ofpores at the origin of the membranes at the nu-clear membrane or of direct breaks in the con-tinuity of the nuclear membrane.

Although the previous two figures are typi-cal of nonmyelinating Schwann cells in the gan-glion, an exception is worth mentioning. Thenucleus of one Schwann cell found in a ganglionwas without excess membranous material andcontained numerous viral forms, many of whichwere mature. This exceptional observationdemonstrates that nonmyelinating Schwanncells can, in extremely rare instances, be pro-ductively infected.Other supporting and infiltrating cells.

Fibroblasts, lymphocytes, macrophages,plasma cells, and polymorphonuclear leuko-cytes, when observed in either nerve or gan-glion, appeared not to contribute significantlyto the productive infection. Cells presump-tively identified as fibroblasts or lymphocyteswere the only infiltrating or supporting cellsever seen to contain any visible viral products.Infected cells considered to be lymphocyteswere observed only in the sciatic nerve andhere only in one nerve specimen 6 days postin-fection. These cells contained both mature andimmature virus particles. Infected fibroblasts



.org/D

ownloaded from

http://iai.asm.org/

30~~~~~~~~~~~~~~~~~~~~~~1

AN4 \ #+

14 A~~~~~~~~~~~~~~

%

14.~ ~ ~ , 'Jb ~

~V

to.'

~~~~; I)~~~~~~~~~~.~j i~q'A -.. ~~4 ~~~~ ~~~~~ ~4~ Ai.

-c~~~~~~~~~~~~~4.

o++:44~~~~ ~ ~ ~~~~~~~~~-

Ai~~~~~~~YPIt~ ~ ~ .,

WAth ~ ~ ~ ;~#4 ~IX +,~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~FIG.14.Satellitcell(5infectd with hrpes simlex viru adjacet to an nfected euron (N). Threaresfmmbaneprlieraio an satere cpsis ndnuceoapsds(arow) re reen inth saelitcellnuclus.nencleoapsdiscontinedin te cyoplsm o thesatelitecell(arow) nd aporton o thneuonancles n te ppe rghtcotais vra prduts arows. 1700

N.,~ ~ ~ 28


.org/D

ownloaded from

http://iai.asm.org/

'1

,

k

.:4

qia .r

..x, %. ri4.

I i

I,.

°lil

.4op I 1.4p-,4 .,k.'e .,;1.

Ot. '. , 41 . .1%, .

.t

! I ..-:. -k

.:' lllh.- .. .

r.~~~~~.ts~ rg,2

*r,~~~~~~~r

; ~ ~ , t ' '- Q

*>

e

.,

FIG. 15. Area of membrane proliferation enlarged from Fig. 14. Extensions from the inner and outer nuclearmembrane invaginate at a pore, forming whorls in the nucleus. An area of increased denseness is seen betweenpairs of membranes (large arrow). In addition, Capsids (small arrows) are associated with the whorls, whilefree ribosomes (R) and another viral capsid (*) are lying inside the pore. x 76,000.

FIG. 16. Schwann cell (S) associated with a myelinated axon (Ax). In this rare observation, a few scatteredviral capsids are seen in the nucleoplasm. x21,000.

282

~1W0*"'

A Ar

A.t,t..s E,Z

t " >;'1IQ,,

di. jffaLs

JtitH

,k..

A>x

I

,.

II.

O_

"

.,;i : *! s t

' -

'.-V

;

.spiYl,IL)-

t W 'S

t

I' .,.'.


.org/D

ownloaded from

http://iai.asm.org/

tth~ ~ ~ aX '4Ct'4r<|

s_Y. . i-* w

v-.a Pt ti1

''.9.,:4 ' ', ,

~~~~~~~~~~~;J

4 ~ ~ ~ V tj a/k .-- Wi "2

L}y'~~~thXstt..,'~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~F

FIG. 17. Nonmyelinating Schwann cell (S) with intranuclear viral capsids and a membranous whorl. In addi-tion, there is a suggestion that the excess membranes are derived from the nuclear membrane (arrow). An areaof increased denseness is seen between a pair of membranes (*). (Ax, axons; ES, extracellular space). x30,OOO.

FIG. 18. A nonmyelating Schwann cell (S). Capsids, nucleocapsids, and excess membranes are seen in thenucleus. Viral products are also present in the cytoplasm. x26,000.

283


.org/D

ownloaded from

http://iai.asm.org/

COOK AND STEVENS

were observed only twice, and neither cell con-tained mature viruses, and only a few intranu-clear capsids were observed. It may be of im-portance to note that virus was found only oncein the extracellular space in sciatic nerves de-rived from infected mice during the acute in-fection. Finally, virus was never observed eitherin vessles or in the endothelial cells of vessels.These ultrastructural observations coupled

with viral assays demonstrate that infection ofthe peripheral nervous system of the mouse byHSV in this model results in an early, produc-tive infection of neurons in the sacrosciaticspinal ganglia. Other cell types in the nerve

trunk, ganglia, and dorsal root are either aber-rantly or infrequently infected. Since the con-tribution to the infectious process of theselatter cell types appeared to be minimal, itseemed unlikely that they could be the onlymeans of viral transport in nerves. This obser-vation led to the following attempts to definethe mechanism by which HSV travels in thenervous system of mice.

Kinetics of viral transport to spinal gan-glia. Previous experiments in this laboratoryresulted in the establishment of a techniquewhereby intact sacrosciatic spinal ganglia la-tently infected with HSV could be maintainedand induced to produce virus when coculti-vated with RK13 cells (26). This technique wasemployed to determine the time interval be-tween initial infection of the mouse foot andarrival of virus in the dorsal root ganglion asso-ciated with the sciatic nerve. In these experi-ments, mice were infected in both feet usinghigh titer RK,3 cell-passed HSV (107 PFU/ml).At specific periods after this procedure, fourmice were sacrificed, and all the intact gangliaassociated with the sciatic nerve of each ani-mal were placed in a French square bottle pre-viously seeded sparcely with RK,3 cells andthen incubated at 37 C. These cultures wereobserved over the next 21-day period andscored for herpetic cytopathic effects. Thissystem, then, served as an amplificationmethod to detect the presence of smallamounts of virus in the ganglion. It can be con-cluded from three such experiments (Table 1)that the time interval required for virus totravel to the ganglia from feet is between 20and 24 hr.

This short time interval required for passageof virus from foot to ganglion suggested that vi-rus passed principally by some conduit. Tofurther investigate this possibility, the lengthof the viral replication cycle in neuronal sup-porting cells was determined. Since this couldnot be determined directly, the experiment

was performed using mouse embryo fibroblastsmaintained in vitro.One-step growth cycle of HSV in mouse

embryo fibroblasts. The mouse embryo fibro-blasts were infected in French square bottleswith RK,3 grown HSV at multiplicities suffi-cient to infect all mouse cells. After 1 hr of ad-sorption, the inocula were removed by threewashes with MEM, media was readded, andthe cultures were incubated as 37 C. At 2-hrintervals, duplicate cultures were removed andfrozen at -70 C. When cultures for all time in-tervals had been collected, frozen, and thawedone additional time, they were titrated onRK13 monolayers.As shown in Fig. 19, the one-step growth

cycle for HSV-MacIntyre in MEF has aneclipse period of about 4 hr and a rise periodof about 6 hr. If this 4-hr period is taken as theestimate of the eclipse period in supportingcells of the peripheral nervous system, then,at most, six adjacent cells could be infectedbetween the foot and the ganglion of mice. Thisresult, when coupled with the time interval re-quired for virus to travel from foot to ganglion,strongly suggests that virus does not travel bydirect passage from supporting cell to sup-porting cell.

Effect of passive antibody on passage ofHSV. The previous experiments support theconcept that virus reaches ganglia by way of anendoneural conduit. Two obvious routes areendoneural lymph channels or bloodstream. To

TABLE 1. Recovery of herpes simplex virus fromsacrosciatic spinal ganglia following rear

footpad inoculation in miceaTime interval between

Exp no. infection and Mice positive/explantation of mice testedganglia (hr)

12 0/41 24 4/4

48 4/4

12 0/42 16 0/4

20 2/424 4/4

16 0/43 20 3/4

24 4/4

aMice were infected in both hind feet with high-ti-tered RK,, passed HSV. At each time period, the spinalganglia associated with both sciatic nerves from each of fourmice were removed and cocultivated with RK ,s cellsmaintained in monolayer. These cultures were maintainedand observed daily for virus-specific cytopathic effects for21 days. Positive cultures appeared between the 2nd and17th day of cocultivation.



.org/D

ownloaded from

http://iai.asm.org/


PFU/ml

105- l

O 2 4 8 1 1 14TIME, hours

FIG. 19. One-step growth cycle of herpes simplexvirus in mouse embryo fibroblasts.investigate these possibilities, mice were in-fected in both rear footpads with high-titeredRK,3 cell passed HSV and then injected intra-venously with 0.5 ml of rabbit hyperimmuneanti-HSV antiserum (K = 72 ml/min) 5 or 10 hrlater. At 48 hr after infection, the sacrosciaticspinal ganglia were removed and co-cultivatedwith RK,, cells. These cultures were then ob-served for virus-specific cytopathic effects.As can be seen in Table 2, virus reached the

ganglia of every mouse given antibody at 5 or

10 hr after viral inoculation. A 1: 500 dilution ofsera taken from these passively immunizedmice at the time of sacrifice neutralized 120 of150 PFU of HSV, indicating that significantamounts of antibody persisted. Since HSVrequires 20 to 24 hr to travel from foot to gan-glion (Table 1), progression of herpetic infec-tion in the presence of antibody given at 5 or 10hr suggests strongly that a route other thanneuronal lymph channels or bloodstream is theroute of passage. Our reasons for preferringaxons are given in the next section.

DISCUSSIONAfter infection with HSV, neurons, satellite

cells, and nonmyelinating Schwann cells in theperipheral nervous system of the mouse pro-

duce virus-specific products. As assessed bythe ultrastructural techniques employed here,viral replication in the neurons of spinal ganglia

has many morphological similariteis to herpessimplex infections studied in various permis-sive cells maintained in cell culture (cf. 17).Numerous virions were seen in neuronal peri-karya, and several were observed in axons.

Supporting cells, when compared to neurons,exhibited consistent differences in the patternof viral replication. Fibroblasts and Schwanncells associated with myelinated axons onlyrarely demonstrated evidence of infection,and they were never observed to produce mor-

phologically complete virions. Satellite cellsand nonmyelinating Schwann cells reacted toinfection by producing numerous capsids andnucleocapsids and an excessive proliferationof nuclear membranes. The inflammatory cells,principally lymphocytes, were only rarely ob-served to contain herpesviruses. This differ-ence in viral susceptibility between neurons

and supporting cells has very recently been ob-served independently by Dillard et al. (6).However, other reports of electron microscopestudies (14, 22, 25) differ with respect to viralreplication patterns in cells of the peripheralnervous system. Different experimental con-

ditions have no doubt contributed to these di-vergent results.Wildy (29), studying a system very similar to

ours, has furnished strong evidence that HSVtravels centripetally via the sciatic nerve to thecentral nervous system, a conclusion which isnow generally accepted. The results presentedhere, including the kinetics of viral replicationin elements of the peripheral nervous system,our inability to recover virus from either sera or

draining lymph nodes during the acute infec-tion, and the progression of infection in thepresence of anti-viral antibody also supportthis concept. Although it is reasonably well es-tablished that HSV travels in nerves, the endo-

TABLE 2. Recovery of herpes simplex virus fromsarcrosciatic spinal ganglia after passive

administration of anti-viral sera tomicea

Interval between time ofinfection and administration miceposte!

of serum (hr)

5 5/55 5/510 6/610 6/6

aAt least 10 sacrosciatic spinal ganglia from each mouse(five from each side) were removed at 48 hr after infection,pooled, and cocultivated with monolayer cultures of RK,,cells. Virus-induced cytopathic effects in the RKI 3 cellswere observed beginning on the third day of cocultivation.All cultures demonstrated virus-specific cytopathic effectsby the 8th day of cocultivation.

285VOL. 7, 1973


.org/D

ownloaded from

http://iai.asm.org/

COOK AND STEVENS

neural routes involved have not been so clearlydefined. Blood, axonal cylinders, tissuespaces, neural lymphatics, and sequential in-fection of endoneural cells (13, 20) have all beensuggested, with the latter route currently re-ceiving the strongest direct support (12, 22, 25,30).

Since the infection progresses in the pres-ence of significant levels of neutralizing anti-body (Table 2), we tentatively conclude thatendoneural blood and lymph vessels are not in-volved in transport of virus to ganglia. Thecentrifugal flow in neural lymphatics (8) alsomitigates against this latter route. Of possibleroutes and combinations of routes remaining,three are reasonable when one remembers thatvirus reaches ganglia in 20 to 24 hours (Table1). Thus, virus could pass either by way of sup-porting cells coupled with intermittent passagein the extracellular space (or axons) by the ex-tracellular space alone, or via axons alone.With data now available, we cannot unequiv-

ocally differentiate between these alternatives,and we have not yet been able to devise exper-iments to directly answer the question. How-ever, when taken together, the following indi-rect arguments offer strong support for axonalpassage as the route by which HSV travelsmost efficiently in peripheral nerves. (i) Ifcharacteristics of the replication cycle of HSVin mouse embryo fibroblasts is representativeof that in other mouse cells, then cell-to-cellpassage of infectious virus in supporting cellswould take longer than the time interval ob-served for virus to reach the sacrosciaticspinal ganglia. As a maximum, only six cellscould have been sequentially infected in the 20-to 24-hr time interval required for virus totravel from foot to ganglia. Therefore, if sup-porting cells contribute to the process, inter-mittent passage of infectious virus from onesupporting cell to a conduit to another sup-porting cell must be postulated. In this regard,it is important to reemphasize that morpho-logically mature virions were only rarely seenin supporting cells and that extracellular nu-cleocapsids are generally considered to pos-sess greatly reduced infectivity (17, 23).

(ii) Maximal titers of virus were consistentlyobserved in ganglia earlier than in sciaticnerves. This result is undoubtedly related tothe relative permissiveness of neurons com-pared to supporting cells. In addition, it islikely that a retrogressive infection induced byvirus traveling centrifugally in axons from themore productively infected neurons contrib-utes significantly to the quantity of infectiousvirus recovered from nerves. Virions observed

in the axons of neurons in which the nuclei andperikarya also contain viral products supportsthis latter conclusion. Thus, virions can exist,and very likely move centrifugally, in axons.

(iii) If the extracellular space were importantas a conduit for viral transport, fibroblasts inthis area would be exposed to infectious parti-cles and thus would be expected to demon-strate significant levels of infection. As wasstated earlier, this was not observed. It is alsoimportant to note that viruses were almostnever seen to be free in the extracellular space(with the exception of virions between neuronaland satellite cell plasma membranes in gan-glia), but again, they were observed in axons. Inaddition, infected satellite cells were invari-ably seen to surround infected neurons only.This latter observation would be unlikely if theextracellular space were important in viraltransport; the converse would be the expectedfinding.

Evidence which also supports the intra-ax-onal transport of herpesviruses has been ob-tained by others. The initial, classic experi-ments were performed by Goodpasture andTeague (10) and Goodpasture (9) who con-cluded in 1925 that HSV probably traveled inaxons in a centripetal manner because of thedistribution of lesions in the central nervoussystem following peripheral inoculation of rab-bits. Wildy (29) showed that, after footpad in-fection of mice, HSV appeared in dorsal rootganglia prior to the sciatic nerve and concludedthat virus reached the sensory ganglia via thesciatic nerve trunk. More recently, Baringerand Griffith (2) inoculated the cornea of rabbitswith HSV and produced a primary infection ofneurons with lesions restricted to the ophthal-mic division of the trigeminal tract. They sug-gested that the virus is capable of traveling inaxonal cylinders. Kristensson (14) concludedfrom the distribution of lesions that axonaltransport was the most likely method of trans-port of HSV in nerve trunks after footpad andintramuscular inoculation of mice. Later,Kristensson et al. (15) found infectious parti-cles in the spinal cord before the sciatic nerveafter mice were infected in the footpad withHSV. The spread of infection could be pre-vented by treating the nerve with colchicine, aprocedure which was claimed to "collapse"axons while sparing the integrity of endoneuralspaces. Finally, in a very recent report, Hill etal. (11) also showed that HSV can exist in axonsof sciatic nerves.

Experimental evidence that sequential infec-tion of endoneural cells is the mechanism bywhich HSV travels in the peripheral nervous



.org/D

ownloaded from

http://iai.asm.org/


system rests primarily upon the experimentsreported by Johnson (12). Using the fluores-cent-antibody staining technique on infectedsuckling mice, he showed that endoneural cellscontained viral antigens and suggested thatvirus could be transmitted in peripheral nervesthrough sequential infection of endoneuralcells. That Schwann cell infection might mir-ror aberrant progression of virus within theaxon was ruled against by the absence of fluo-rescence in ganglionic cells until the "tide" ofinfected endoneural cells reached them. Byusing organ culture techniques, we have foundinfectious virus in ganglia at a time (20-24 hrafter infection) when no viral antigens in anyganglionic cells could be stained by immuno-fluorescent procedures. Specific fluorescentcells were found, however, by 48 hr after in-fection (unpublished observations). Thus,organ culture techniques represent moresensitive viral assay procedures than do immu-nofluorescent procedures. In addition, unin-fected neurons were never seen to be sur-rounded by infected supporting cells; theopposite was true.

Historically, the strongest argument whichmitigates against axons as routes for centrip-etal passage of viruses has been the initial, welldocumented finding that axonal flow is awayfrom the neuronal perikaryon (28). However,in more recent years, reports from several lab-oratories have appeared which indicate thatthere is also significant long-range centripetalmovement of molecules and cellular organellesin the axons of neurons populating both pe-ripheral and central nervous systems (16, 31).Thus, there appears to be no physiological bar-rier to centripetal movement of viruses inaxons. In conclusion, our results, when takenwith these findings, constitute strong supportfor the concept that HSV can travel in neuronsand their processes in the peripheral nervoussystem of mice.

ACKNOWLEDGMENTSWe thank D. and H. Porter for introducing us to the tech-

niques of immunofluorescence and P. Cancilla for advice con-cerning the perfusion of animals for electron microscopy.The technical assistance of E. Scott and V. Bastone is grate-fully acknowledged.

This invesitgation was supported by Public Health Servicegrant AI-06246 from the National Institute of Allergy and In-fectious Diseases, and by grant NS-8711. This work was ini-tiated while one of us (M. L. C.) was a National MultipleSclerosis Society Postdoctoral Fellow.

LITERATURE CITED1. Adams, M. H. 1959. Bacteriophages, p. 463-466. Inter-

science Publishers, Inc., New York.2. Baringer, J. R., and J. F. Griffith. 1970. Experimental

herpes simplex encephalitis: early neuropathologicchanges. J. Neuropathol. Exper. Neurol. 29:89-104.

3. Chou, S. M., and J. D. Cherry. 1967. Ultrastructure ofCowdry type A inclusions. I. In human herpes simplexencephalitis. Neurology 17:575-586.

4. Cook, M. L., and J. G. Stevens. 1970. Replication ofVaricella-Zoster virus in cell culture: an ultrastructuralstudy. J. Ultrastruct. Res. 32:334-350.

5. Darlington, R. W., and L. H. Moss, III. 1968. Herpes-virus envelopment. J. Virol. 2:48-55.

6. Dillard, S. H., W. J. Cheatham, and H. L. Moses. 1972.Electron microscopy of zosteriform herpes simplexinfection in the mouse. Lab. Invest. 26:391-402.

7. Dowdle, W. R., A. J. Nahmias, R. W. Harwell, and F. P.Pauls. 1967. Association of antigenic type of Herpes-virus hominis with site of viral recovery. J. Immunol.99:974-980.

8. Field, E. J., and J. B. Brierley. 1948. The lymphaticdrainage of the spinal nerve roots in the rabbit. J. Anat.82:198-206.

9. Goodpasture, E. W. 1925. The axis-cylinders of periph-eral nerves as portals of entry to the central nervoussystem for the virus of herpes simplex in experimen-tally infected rabbits. Amer. J. Pathol. 1:11-28.

10. Goodpasture, E. W., and 0. Teague. 1923. Transmissionof the virus of herpes febrilis along nerves in experi-mentally infected rabbits. J. Med. Res. 44:139-184.

11. Hill, T. J., H. J. Field, and A. P. C. Roome. 1972. Intra-axonal location of herpes simplex virus particles. J.Gen. Virol. 15:253-255.

12. Johnson, R. T. 1964. The pathogenesis of herpes virusencephalitis. I. Virus pathways to the nervous systemof suckling mice demonstrated by fluorescent antibodystaining. J. Exp. Med. 119:343-356.

13. Johnson, R. T., and C. A. Mims. 1968. Pathogenesis ofviral infections of the nervous system. New Engl. J.Med. 278:23-30.

14. Kristensson, K. 1970. Morphological studies of the neu-ral spread of herpes simplex virus to the central ner-vous system. Acta Neuropathol. 16:54-63.

15. Kristensson, K., E. Lycke, and J. Sjostrand. 1971.Spread of herpes simplex virus in peripheral nerves.Acta Neuropathol. 17:44-53.

16. LaVail, J. H., and M. M. LaVail. 1972. Retrograde ax-onal transport in the central nervous system. Science176:1416-1417.

17. Nii, S., C. Morgan, and H. M. Rose. 1968. Electronmicroscopy of herpes simplex virus. II. Sequence ofdevelopment. J. Virol. 2:517-536.

18. Nii, S., C. Morgan, H. M. Rose, and K. C. Hsu. 1968.Electron microscopy of herpes simplex virus. IV. Stud-ies with ferritin-conjugated antibodies. J. Virol. 2:1172-1184.

19. Olitsky, P. K., and R. W. Schlesinger. 1941. Effect oflocal edema and inflammation in the skin of the mouseon the progression of herpes virus. Science. 93:574-575.

20. Paine, T. F., Jr. 1964. Latent herpes simplex infection inman. Bacteriol. Rev. 28:472-479.

21. Porter, D. D., H. G. Porter, and B. B. Deerhake. 1969.Immunofluorescence assay for antigen and antibody inlactic dehydrogenase virus infection of mice. J. Immu-nol. 102:431-436.

22. Rabin, E. R., A. B. Jenson, and J. L. Melnick. 1968.Herpes simplex virus in mice: electron microscopyof neural spread. Science 162:126-127.

23. Roizman, B. 1969. The-herpesviruses - a biochemicaldefinition of the group. Cur. Top. Microbiol. Immunol.49:1-79.

24. Schmidt, N. J. 1969. Tissue culture technics for diag-

VOL. 7, 1973 287


.org/D

ownloaded from

http://iai.asm.org/

288 COOK AND STEVENS

nostic virology, p. 113-114. In E. H. Lennette andN. J. Schmidt (ed.), Diagnostic procedures for viraland rickettsial infections. American Public HealthAssociation, Inc., New York.

25. Severin, M. J., and R. J. White. 1968. The neural trans-mission of herpes simplex virus in mice. Light andelectron microscopic findings. Amer. J. Pathol. 53:1009-1020.

26. Stevens, J. G., and M. L. Cook. 1971. Latent herpessimplex virus in spinal ganglia of mice. Science 173:843-845.

27. Stevens, J. G., G. J. Kado-Boll, and C. B. Haven. 1969.Changes in nuclear basic proteins during pseudo-rabies virus infection. J. Virol. 3:490-497.

28. Weiss, P. 1969. Neuronal dynamics. p. 255-299. In F. 0.

INFECT. IMMUNITY

Schmitt, T. Melnechuk, G. C. Quanton, and G. Adel-man (ed.), Neurosciences research symposium sum-maries, vol. 3. The M.I.T. Press, Cambridge.

29. Wildy, P. 1967. The progression of herpes simplexvirus to the central nervous system of the mouse. J.Hyg. 65:173-192.

30. Yamamoto, T., S. Otani, and H. Shiraki. 1965. A studyof the evolution of viral infection in experimentalherpes simplex encephalitis and rabies by means offluorescent antibody. Acta Neuropathol. 5:288-306.

31. Zelena, J. 1969. Bidirectional shift of mitochondria inaxons after injury, p. 73-94. In S. H. Barondes (ed.),Cellular dynamics of the neuron, symposium of theinternational society for cell biology, vol. 8. AcademicPress Inc., New York.


.org/D

ownloaded from

http://iai.asm.org/

Documents

Feb. Vol. for Pathogenesis ofHerpetic Neuritis ... · in the peripheral nervous system constitutes the present report (an abstract ofthis material has been presented, Bacteriol, Proc