j clin distribution zoster virus and simplexinfections by the varicella zoster virus (vzv) andherpes...

6
J Clin Pathol 1996;49:243-248 Distribution of varicella zoster virus and herpes simplex virus in disseminated fatal infections A F Nikkels, P Delvenne, C Sadzot-Delvaux, P Quatresooz, G E Pierard Abstract Aims-To study the cutaneous and visceral distribution of herpes simplex virus (HSV) and varicelia zoster virus (VZV) in fatal infections. Methods-Standard histology, immuno- histochemistry (monoclonal antibodies VL8 and VL2 and polyclonal antibody IE63 directed against VZV; monoclonal anti- bodies IBD4 and HH2 and polyclonal antibodies directed against HSVI and HSVII) and in situ hybridisation (anti- HSV and anti-VZV probes) were applied to formalin fixed, paraffin wax sections. Results-On histological examination, Herpesviridae infection was evident in various organs including the lungs, liver and skin. In addition, immunohisto- chemistry and in situ hybridisation re- vealed the presence of HSV and VZV an- tigens and nucleic acids in several cell types and tissues showing no cyto- pathological alterations suggestive of Her- pesviridae infection. The organs with histological evidence of infection also con- tained VZV or HSV antigens and their genes. Conclusions-These findings suggest that organ failure in disseminated VZV and HSV infections is primarily caused by HSV or VZV induced cell damage and lysis. They also indicate that immunohisto- chemistry and in situ hybridisation can provide an accurate, type-specific diag- nosis on formalin fixed, paraffin wax em- bedded tissue even when classic histological and cytological characteristics are lacking. (J7 Clin Pathol 1996;49:243-248) Department of Dermatopathology, CHU Sart Tilman, B-4000 Liege, Belgium A F Nikkels G E Pierard Department of Pathology P Delvenne G Lipcsei P Quatresooz Department of Fundamental Virology C Sadzot-Delvaux S Debrus J Piette B Rentier Correspondence to: Dr G E Pierard. Accepted for publication 18 October 1995 Keywords: varicella zoster virus, herpes simplex virus, immunosuppression, immunohistochemistry, in situ hy- bridisation, lethal infection. Infections by the varicella zoster virus (VZV) and herpes simplex virus (HSV) types I and II are potentially life-threatening, particularly in the immunocompromised host. Recent reports show a clear upward shift in the age of VZV infection and the associated mortality rate.13 The clinical diagnosis of varicella, herpes zoster and herpes simplex infections may be difficult, especially when the cutaneous erup- tion is atypical as in immunocompromised patients." Extensive visceral dissemination can occur several days before the skin lesions appear. Typical histocytological diagnostic clues for HSV and VZV infections include the S Debrus, J Piette, B Rentier, G Lipcsei, presence of syncytial keratinocytes, Cowdry type A intranuclear inclusions, "ground-glass" nuclei, and/or limited haemorrhagic necrotic areas. However, the typical cytopathological changes are not always present and the VZV pneumonitis, which is the most frequent com- plication of varicella, may have non-specific histological aspects.9 Conversely, intranuclear inclusions are not always indicative of viral infection. The mechanisms of organ dysfunction that may lead to a fatal outcome are not entirely known. Pathogenic hypotheses are based mainly on histological observations of rare nec- ropsy reports. The direct effect of the viruses was only suspected by the presence of Cowdry type A intranuclear inclusions. Currently, type specific identification of Herpesviridae infections on formalin fixed, paraffin wax embedded tissue is possible using polyclonal or monoclonal antibodies directed against envelope glycoproteins or nucleocapsid constituents of HSV and VZV. 103 These results can be confirmed by in situ hybridisation with specific HSV and VZV nucleic acid probes.'4 15 During viral latency, HSV and VZV nucleic acids and some ct or immediate early (IE) gene transcripts, but not early (E) and late (L) proteins, are present in morphologically normal cells in the dorsal root ganglia.' '7 During pro- ductive VZV infection, some viral genome se- quences and viral glycoproteins gE and gB have been detected in dermal cells without cytological alterations.'5 Further char- acterisation of viral expression in organs other than skin should highlight the distribution and the pathogenesis of disseminated infections by VZV and HSV. This study, using histology, immuno- histochemistry and in situ hybridisation, shows that the distribution of VZV and HSV in various tissues and organs, collected at necropsy, of patients with visceral dissemination of HSV and VZV infection correlates with the major clinical signs. We also show that VZV and HSV can be detected in tissues lacking histo- pathological alterations during disseminated in- fections. Case reports CASE 1 A 73 year old man presented with acute atypical abdominal pain. The past medical history in- cluded moderate idiopathic lymphopenia, par- tial prostatectomy for benign hyperplasia, type II diabetes, cholecystectomy, ancient myo- cardial infarction, corticosteroid dependent chronic obstructive lung disease (methyl- 243 on February 21, 2021 by guest. Protected by copyright. http://jcp.bmj.com/ J Clin Pathol: first published as 10.1136/jcp.49.3.243 on 1 March 1996. Downloaded from

Upload: others

Post on 07-Oct-2020

4 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: J Clin Distribution zoster virus and simplexInfections by the varicella zoster virus (VZV) andherpes simplexvirus (HSV) types I andII are potentially life-threatening, particularly

J Clin Pathol 1996;49:243-248

Distribution of varicella zoster virus and herpessimplex virus in disseminated fatal infections

A F Nikkels, P Delvenne, C Sadzot-Delvaux,P Quatresooz, G E Pierard

AbstractAims-To study the cutaneous and visceraldistribution ofherpes simplex virus (HSV)and varicelia zoster virus (VZV) in fatalinfections.Methods-Standard histology, immuno-histochemistry (monoclonal antibodiesVL8 and VL2 and polyclonal antibody IE63directed against VZV; monoclonal anti-bodies IBD4 and HH2 and polyclonalantibodies directed against HSVI andHSVII) and in situ hybridisation (anti-HSV and anti-VZV probes) were appliedto formalin fixed, paraffin wax sections.Results-On histological examination,Herpesviridae infection was evident invarious organs including the lungs, liverand skin. In addition, immunohisto-chemistry and in situ hybridisation re-vealed the presence of HSV and VZV an-tigens and nucleic acids in several celltypes and tissues showing no cyto-pathological alterations suggestive ofHer-pesviridae infection. The organs withhistological evidence ofinfection also con-tained VZV or HSV antigens and theirgenes.Conclusions-These findings suggest thatorgan failure in disseminated VZV andHSV infections is primarily causedbyHSVor VZV induced cell damage and lysis.They also indicate that immunohisto-chemistry and in situ hybridisation canprovide an accurate, type-specific diag-nosis on formalin fixed, paraffin wax em-bedded tissue even when classichistological and cytological characteristicsare lacking.(J7 Clin Pathol 1996;49:243-248)

Department ofDermatopathology,CHU Sart Tilman,B-4000 Liege, BelgiumA F NikkelsG E Pierard

Department ofPathologyP DelvenneG LipcseiP Quatresooz

Department ofFundamental VirologyC Sadzot-DelvauxS DebrusJ PietteB Rentier

Correspondence to:Dr G E Pierard.Accepted for publication18 October 1995

Keywords: varicella zoster virus, herpes simplex virus,immunosuppression, immunohistochemistry, in situ hy-bridisation, lethal infection.

Infections by the varicella zoster virus (VZV)and herpes simplex virus (HSV) types I and IIare potentially life-threatening, particularly inthe immunocompromised host. Recent reportsshow a clear upward shift in the age of VZVinfection and the associated mortality rate.13The clinical diagnosis of varicella, herpes

zoster and herpes simplex infections may bedifficult, especially when the cutaneous erup-tion is atypical as in immunocompromisedpatients." Extensive visceral disseminationcan occur several days before the skin lesionsappear. Typical histocytological diagnosticclues for HSV and VZV infections include the

S Debrus, J Piette, B Rentier, G Lipcsei,

presence of syncytial keratinocytes, Cowdrytype A intranuclear inclusions, "ground-glass"nuclei, and/or limited haemorrhagic necroticareas. However, the typical cytopathologicalchanges are not always present and the VZVpneumonitis, which is the most frequent com-plication of varicella, may have non-specifichistological aspects.9 Conversely, intranuclearinclusions are not always indicative of viralinfection.The mechanisms of organ dysfunction that

may lead to a fatal outcome are not entirelyknown. Pathogenic hypotheses are basedmainly on histological observations of rare nec-ropsy reports. The direct effect of the viruseswas only suspected by the presence of Cowdrytype A intranuclear inclusions.

Currently, type specific identification ofHerpesviridae infections on formalin fixed,paraffin wax embedded tissue is possible usingpolyclonal or monoclonal antibodies directedagainst envelope glycoproteins or nucleocapsidconstituents ofHSV andVZV. 103 These resultscan be confirmed by in situ hybridisation withspecific HSV and VZV nucleic acid probes.'4 15During viral latency, HSV and VZV nucleic

acids and some ct or immediate early (IE)gene transcripts, but not early (E) and late (L)proteins, are present in morphologically normalcells in the dorsal root ganglia.' '7 During pro-ductive VZV infection, some viral genome se-quences and viral glycoproteins gE and gBhave been detected in dermal cells withoutcytological alterations.'5 Further char-acterisation of viral expression in organs otherthan skin should highlight the distribution andthe pathogenesis of disseminated infections byVZV and HSV.

This study, using histology, immuno-histochemistry and in situ hybridisation, showsthat the distribution ofVZV and HSV in varioustissues and organs, collected at necropsy, ofpatients with visceral dissemination of HSVand VZV infection correlates with the majorclinical signs. We also show that VZV and HSVcan be detected in tissues lacking histo-pathological alterations during disseminated in-fections.

Case reportsCASE 1A 73 year old man presented with acute atypicalabdominal pain. The past medical history in-cluded moderate idiopathic lymphopenia, par-tial prostatectomy for benign hyperplasia, typeII diabetes, cholecystectomy, ancient myo-cardial infarction, corticosteroid dependentchronic obstructive lung disease (methyl-

243

on February 21, 2021 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.49.3.243 on 1 March 1996. D

ownloaded from

Page 2: J Clin Distribution zoster virus and simplexInfections by the varicella zoster virus (VZV) andherpes simplexvirus (HSV) types I andII are potentially life-threatening, particularly

Nikkels, Delvenne, Sadzot-Delvaux, Debrus, Piette, Rentier, et al

prednisolone, 4 mg/day for over 10 years),and partial gastrectomy after perforation ofcardio-oesophagal peptic ulcers. A varicella-likeeruption developed on his head and soon dis-seminated over his entire skin surface. Thepatient was febrile (39-40°C). Histological andimmunohistological evaluations ofa skin biopsyspecimen led to the diagnosis of varicella. Acy-clovir therapy (7 5 mg/kg every eight hours)was initiated. Two days later, the patient ex-perienced severe respiratory distress in as-sociation with disseminated intravascularcoagulation, massive hepatolysis, and acuterenal insufficiency. Anti-VZV IgM titres roseafter four days, while IgG titres remained neg-ative. The patient died from multisystem fail-ure after six days. Necropsy was carried outfour hours after death.

CASE 2A 40 year old woman presented with pain andzosteriform eruption present for three days inthe L2-4 dermatomes. The relevant past med-ical history included centrocytic/centroblasticlymphoma, which had been treated three yearsbefore by corticosteroids, chemotherapy(cyclophosphamide, adriamycin, vincristine,prednisolone) and splenectomy. For the pre-vious year, treatment had been limited tomethylprednisolone at a dose of 8 mg daily.Herpes zoster was diagnosed clinically andintravenous acyclovir therapy (10 mg/kg everyeight hours) was initiated. Extradermatomalsatellite lesions were absent. Liver and renalfunction tests were not altered. Enzyme linkedimmunosorbent assay (ELISA) yielded evid-ence of past VZV and HSV infections. Twodays later the patient rapidly developed hy-perthermia and respiratory distress. Chest x rayshowed extensive bilateral interstitial infiltra-tion suggestive of viral pneumonitis. Despiteassisted ventilation in the intensive care unit,her respiratory dysfunction worsened and shedied. Necropsy was carried out 12 hours afterdeath.

CASE 3A 29 year old man was admitted to hospitalbecause of hyperthermia and mental confusiondeveloping one week after a journey to In-donesia. The diagnosis of malaria (Plasmodiumfalciparum) was reached. The rapid de-terioration of the neurological status was sug-gestive of cerebral malaria. A brain scan wasnormal and electroencephalography showeddiffuse encephalopathy. Intravenous quinidine(10mg/kg/day) was initiated. Two days later,the patient rapidly developed severe respiratorydistress (P02 54-8 mmHg; base excess7 1 mmol/l), which was treated by assisted vent-ilation and high dose corticosteroids. Severaldays later his serum hepatic enzymes activitiesincreased (lactate dehydrogenase 32 500 IUll;aspartate aminotransferase 10 950 IU/l; alanineaminotransferase 7280 IU/1), indicating ful-minant hepatitis. ELISA revealed a past epi-sode of VZV infection (IgG 2560, IgMnegative). Tests for cytomegalovirus, Epstein-

Barr virus, HSV, adenoviruses, influenza, para-influenza 1, 2, 3, and hepatitis A, B and Cwere repeatedly negative. No mucocutaneouslesions were observed. A severe haemorrhagicsyndrome developed with worsening hepa-tolysis. A liver biopsy specimen showed ex-tensive haemorrhagic necrosis with no evidenceof viral infection. A bone marrow examinationshowed pancytopenia. Gastrointestinal bleed-ing, intravascular coagulation and acute renalfailure developed rapidly. The patient died dur-ing liver transplantation. A necropsy was car-ried out the same day.

MethodsHISTOLOGICAL ASSESSMENTNecropsy material was fixed in formalin (10%neutral buffered formalin) and embedded inparaffin wax for routine processing and stainingwith haematoxylin and eosin. The histologicaldiagnosis of Herpesviridae infection was as-sessed according classic criteria, including thepresence of Cowdry type A intranuclear in-clusions, "ground-glass" nuclei, large eosino-philic nuclei, syncytial giant cells, and areasof limited haemorrhagic necrosis."819 Sections5 ptm thick were prepared on silanised slidesfor immunohistochemistry and in situ hybrid-isation.

IMMUNOHISTOCHEMISTRYThe monoclonal and polyclonal antibodiesused are listed in table 1. The VL8 and VL2mouse monoclonal antibodies react, re-spectively, with the VZV glycoproteins gE andgB.'213 The rabbit polyclonal antibody anti-IE63 is directed against the VZV immediateearly protein (IE63). Rabbit hyperimmuneserum was obtained after inoculation ofa fusionprotein (molecular weight 45 kDa). Specificityof the polyclonal antibody was controlled bywestern blotting, immunofluorescence andimmunohistochemistry.20 The monoclonalantibody 1BD4 (Seralab, Crawley Down, UK)recognises an 89 kDa cytoplasmic protein ofHSVI and the monoclonal antibody HH2 (Ser-alabs) is specific for HSVII. The rabbit poly-clonal antibodies anti-HSVI and anti-HSVII(Dako, Glostrup, Denmark) react with HSVIand HSVII antigens. The anti-HSV and anti-VZV antibodies do not cross-react.

Immunohistochemistry was carried out atroom temperature. An avidin-biotin based de-tection system was used for all tissues exceptfor the liver where the alkaline phosphatase-antialkaline phosphatase (APAAP) techniquewas chosen because of the presence of en-dogenous biotin. Tissue sections (5 gm thick)on silanised slides were deparaffinised andplaced in Tris buffered saline (TBS, Tris0 15M, NaCl 0 05M, pH 7 6) baths. A non-immune blocking serum (swine or rabbit, JRHBiosciences, Lenexa, Kansas, USA) containing3% bovine serum albumin (BSA, Behring,Marburg, Germany) was applied for 30 min-utes. After incubation with the primary anti-body, biotinylated secondary antibodies (Swineanti-rabbit diluted 1 in 300, rabbit anti-swine

244

on February 21, 2021 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.49.3.243 on 1 March 1996. D

ownloaded from

Page 3: J Clin Distribution zoster virus and simplexInfections by the varicella zoster virus (VZV) andherpes simplexvirus (HSV) types I andII are potentially life-threatening, particularly

Disseminated Herpesviridae infections

Table 1 Antibody panel

IncubationClone/ascites Antigen Type Dilution (minutes) Source

vzvVL8 gE Mouse MAb 1 in 20 30 CHU Sart Tilman"VL2 gB Mouse MAb 1 in 10 30 CHU Sart Tilman"anti-IE63 IE63 Rabbit PAb 1 in 300 60 CHU Sart Tilman20

HSVIBD4 HSVI Mouse MAb 1 in 20 90 SeralabsHH2 HSVII Mouse MAb 1 in 20 90 Seralabsanti-HSVI HSVI Rabbit PAb Prediluted 60 Dakoanti-HSVII HSVII Rabbit PAb Prediluted 60 Dako

MAb =monoclonal antibody; PAb =polyclonal antibody. Source: CHU Sart Tilman, Department of Virology, Liege; Seralabs,Crawley Down, UK; Dako, Glostrup, UK.

diluted in 400; Dako) were added and theslides incubated for 30 minutes. The slideswere rinsed in TBS for 20 minutes and coveredby avidin-biotin complex coupled to alkalinephosphatase (ABCAP, Dako) for 30 minutes.New Fuchsin (Dako) was used as chromogen.Slides were counterstained with Mayer's haem-alum and mounted in glycergel (Dako).

IN SITU HYBRIDISATIONIn situ hybridisation was performed with di-goxigenin labelled probes and biotinylatedprobes, respectively, for the detection of VZVand HSVDNA. The 16 -6 kilobase EcoRI-A and14 5 kilobase EcoRI-B restriction endonucleasefragments of VZV DNA corresponding, re-

spectively, to gE and gpB,2' subcloned intopUC 9,22 were labelled in vitro by randompriming23 with digoxigenin d-UTP (Boehr-inger, Mannheim, Germany) to be used as

probes for in situ hybridisation. The bio-tinylated anti-HSV probe (Enzo Diagnostics,New York, USA) is derived from a mixtureof two clones of HSV-DNA sequences in theBamHI site of pBR322. The insert sizes are

16-0 and 8-0 kilobases.Tissue sections on silanised slides were de-

paraffinised and digested with proteinase K(100 ng/ml, Boehringer) for 15 minutes. Afterdenaturing the probe mixtures and tissuesamples together at 1 00°C for five minutes, theprobes were hybridised for 12 hours in a humid

chamber at 37°C. After washing, slides were

incubated with an alkaline phosphatase con-

jugated anti-digoxigenin antibody or Strept-avidin for 30 minutes. Nitroblue tetrazoliumand bromo-chloro-indolylphosphate were usedas chromogen substrates and nuclear fast redwas used for counterstaining. Slides were de-hydrated and mounted in DePex (Gurr, Poole,Dorset, UK).The following controls for immunohisto-

chemistry and in situ hybridisation were used.Biopsy specimens of clinically evident herpeszoster served as positive controls for VZV anti-bodies and nucleic acid probes. Biopsy speci-mens of herpes labialis and herpes genitaliswere used, respectively, as controls for HSVIand II antibodies and HSV probes. Negativecontrols included normal skin obtained duringsurgery and at necropsy. Additional negativecontrols omitted the primary antibody or theDNA probe.

ResultsGENERAL FINDINGSRoutine histologyTable 2 summarises the results of routinehistological examination. Some intranuclearinclusions, "ground-glass" nuclei, large eos-

inophilic nuclei, and giant cells were identifiedin the liver, lungs and skin of case 1, in theskin and lungs of case 2, and in the lungs and

Table 2 Clinical, histological and virological correlations in organ dysfunction leading to fatal outcome

PredominantMajor clinical signs histological findings IHC ISH

Case 1: VZVHepatic insufficiency Patchy hepatocellular necrosis, + +

inclusionsRenal insufficiency Tubular hydropic + +

degeneration, thromboticmicroangiopathy

Pulmonary distress Patchy haemorrhagic necrosis, + +inclusions

Abdominal pain Ulcerations of the digestive + +tract

Intravascular coagulation Thrombotic microangiopathy + (dendrocytes, + (dendrocytes,endothelial cells) endothelial cells)

Varicella Spongiotic, vesicular + +dermatitis, inclusions

Case 2: VZVHerpes Zoster Spongiotic, vesicular + +

dermatitis, inclusionsPulmonary distress Patchy haemorrhagic necrosis, + +

inclusionsCase 3: HSVI

Hepatic insufficiency Extensive hepatolysis, +inclusions

Pulmonary distress Patchy haemorrhagic necrosis, + +inclusionsCortical haemorrhagic necrosis + +of the adrenal glands

IHC= immunohistochemistry; ISH =in situ hybridisation.

245

on February 21, 2021 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.49.3.243 on 1 March 1996. D

ownloaded from

Page 4: J Clin Distribution zoster virus and simplexInfections by the varicella zoster virus (VZV) andherpes simplexvirus (HSV) types I andII are potentially life-threatening, particularly

Nikkels, Delvenne, Sadzot-Delvaux, Debrus, Piette, Rentier, et al

Table 3 Detection of VZV and HSV antigens and nucleic acids by immunohistochemistry and in situ hybridisation in cell types lackingcytopathological features indicative of infection

Organ

Subject Skin Digestive tract Lungs Thyroid Lymph nodes Spleen Liver Pancreas

Case 1 Dendrocytes, Smooth muscle Pneumocytes, Endothelial cells, Mononuclear Reticulated cells, Periductalendothelial cells cells, glandular smooth muscle smooth muscle cells endothelial cells connective tissue

cells cells, fibroblast-like cells cellscells

Case 2 Pneumocytes, Scattered Isolatedsmooth muscle hepatocytes glandularcells, fibroblast-like cellscells

Case 3 Smooth muscle Pneumocytes, Reticulated cells, Hepatocytescells, glandular endothelial cells, endothelial cellscells inflammatory cells

liver of case 3. The inclusion bodieslocalised in areas ofhaemorrhagic necrosisor in the immediate vicinity. Only the amentioned organs met the histological crof viral infection. Immunohistochemistryin situ hybridisation showed that the infewas much more widespread than susp(

Figure 1 VZV glycoprotein gE present in cells scattered in the spleen (VL8 monocliantibody).

Figure 2 VZVglycoprotein gE present in a centrolobular area of the liver and inconnective tissue cells surrounding a bile duct (VL8 monoclonal antibody).

wereand/iboveiteriar andbction

on histological assessment only (table 3). Theother organs listed in tables 2 and 3 did notconform to the histological criteria for a diag-nosis of viral infection.

ectea Localisation of VZV in cellsImmunostaining with the VL8 and VL2 mono-clonal antibodies essentially highlighted thecell membranes. The staining intensity wasstronger for VL8 than for VL2. Weak cyto-plasmic staining was observed in some infectedcells. Diffuse extracellular positivity occurredin areas of extensive necrosis. The anti-IE63polyclonal antibody predominantly stained thenuclei and, with a weaker intensity, the peri-nuclear zone and the cytoplasm.The EcoRI-A and EcoRI-B fragments ofVZV

DNA, used as probes, produced a stainingpattern that was generally restricted to thenuclei. Weak perinuclear or cytoplasmic stain-ing was sometimes present. The informationgained from immunohistochemistry and in situhybridisation for VZV infections correlatedwell, although the intensity of labelling variedin the various cell populations.

Localisation ofHSV in cells'onal The polyclonal and monoclonal antibodies dir-

ected against HSVI and II revealed mem-branous and cytoplasmic staining in thedifferent infected cell lines. Cross-reactivity wasobserved with the polyclonal antibody directedagainst HSVI and II and the type-specific iden-tification of HSV was better established usingthe 1BD4 and HH2 monoclonal antibodies.The HH2 monoclonal antibody never pro-duced a signal in the test specimens, althoughthe HSVII controls were positive. The poly-clonal antibodies gave stronger staining signalsthan the 1BD4 monoclonal antibody. In situhybridisation always revealed a predominantlynuclear signal and rarely perinuclear positivity.There was a clear correlation between the res-ults obtained by immunohistochemistry and insitu hybridisation.

CASE REPORTSTable 2 summarises the major clinical ob-servations in relation to the histological findingsand virological results on immunohisto-chemistry and in situ hybridisation (figs 1-4).

246

on February 21, 2021 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.49.3.243 on 1 March 1996. D

ownloaded from

Page 5: J Clin Distribution zoster virus and simplexInfections by the varicella zoster virus (VZV) andherpes simplexvirus (HSV) types I andII are potentially life-threatening, particularly

Disseminated Herpesviridae infections

Figure 3 HSVI immunoreactivity in the lung (IBD4 monoclonal antibody).

XIJ

t .. ;6

.''8,i.I i;V

K

44

-.4 %,+

Figure 4 HSVI immunoreactivity in the lamina propria of the stomach (anti-Ipolyclonal antibody).

The diagnoses at necropsy in these thriwere, respectively: case 1, disseminataneous and visceral varicella with muliorgan failure and fatal outcome; case

dissemination ofVZV after herpes zosttion; and case 3, primary extensiveHSVI dissemination responsible fotisystem organ failure and fatal outcon

Table 3 shows the various organs ancin which VZV and HSVI were deteimmunohistochemistry and in situisation in areas where typical cytopathfeatures of Herpesviridae infection were(figs 2 and 4). The other organs not mein tables 2 and 3 did not contain any e

of viral infection on histology, immurchemistry or in situ hybridisation. Inand 2, anti-HSV antibodies and HSVacid probes yielded negative results insues. Anti-VZV antibodies and DNAwere consistently negative in case 3.normnal skin and normal necropsy materconsistently negative.

Discussion^tX̂ Fatal outcome following infection with VZV or

t 4 HSV is rare, and few necropsy studies of suchpatients have been published.48 8226 In themain, the patients were severely immuno-compromised, although only minor immuno-suppression has been noted in other cases.27The clinical presentation of fatal HSV andVZV infections is often atypical. The diagnosisremains difficult, particularly in the absence oftypical mucocutaneous lesions. In over 80%

: of cases of disseminated HSV infection thediagnosis is only established at necropsy.5

*1tt.' Histological examination is frequentlyr- t^4 ^$4 needed to confirm the diagnosis. However, the

classic features of Herpesviridae infection areit . not always easy to recognise, and may even

be absent.9 When Herpesviridae infection is~..~̂ suspected, type-specific identification can be

achieved by nucleic acid hybridisation and an-tigen detection.'01113 The most sensitivemethod to detect HSV antigens in tissue sec-tions relies on the use ofthe polyclonal antibodyto HSVI and HSVII. However, the monoclonalantibodies 1BD4 and HH2 are required fortype-specific identification.This study indicates that HSV or VZV an-

tigens and nucleic acids can be present anddetectable in some cell types or tissues in theabsence of specific Herpesviridae cyto-pathological changes. This suggests that evenwhen histological examination remains in-conclusive, immunohistochemistry and in situhybridisation can confirm the diagnosis.

In our three patients, intranuclear inclusionswere found only in small amounts at the peri-phery of necrotic areas. In one of the casesreported by Cheatham et all8 intranuclear in-clusions were numerous even in cells withoutclear cytolytic features. This patient, however,died 19 days after the cutaneous onset of va-ricella. It is likely that VZV infection developsvery slowly in some cell types and takes time to

iSVI produce the characteristic nuclear inclusions.The results also demonstrate an in vivo cell-

specific permissiveness to VZV and HSV in-fection. In keratinocytes, hepatocytes and pneu-

ee cases mocytes, VZV and HSV rapidly produce ated cut- productive infection that causes host cell lysistisystem and destruction. In other cell types, such as2, fatal dermal dendritic cells, endothelial cells, smoother erup- muscle cells, and fibroblast-like cells, viral an-visceral tigens and DNA were detected before the in-r mul- tranuclear inclusions developed. In these cells,ne. no cytopathological effect was observed ini tissues haematoxylin and eosin stained sections. Thiscted by kind of infection does not meet the criteria ofhybrid- Herpesviridae latency because VZV late an-tological tigens are expressed, indicating a complete viraleabsent replicative cycle.ntioned In some cell types, VZV and HSV are re-vidence sponsible for productive infection with cyto-nohisto- pathological effects. The same viruses can,cases 1 however, also induce a chronic infection withnucleic few cytopathological changes. This phe-all tis- nomenon is observed in patients with AIDS inprobes whom HSV and VZV induce acute vesicular le-Control sions or chronic wart-like cutaneous lesions.2829^ial were Such distinct evolutions probably depend on the

cytokine environment and on the activity of the

247

II0

I'l ). d V.

40

on February 21, 2021 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.49.3.243 on 1 March 1996. D

ownloaded from

Page 6: J Clin Distribution zoster virus and simplexInfections by the varicella zoster virus (VZV) andherpes simplexvirus (HSV) types I andII are potentially life-threatening, particularly

Nikkels, Delvenne, Sadzot-Delvaux, Debrus, Piette, Rentier, et al

local immune system. They may also be due toa disturbed viral replicative cycle.30The present study also shows that the extent

of VZV dissemination varies from one patientto another. This variability is probably relatedto the cell mediated immunity and antibodystatus ofthe patients. HSV was found in similarvisceral cell types as those involved in VZVinfection. HSV infection, however, occurredwith less prominence in fibroblast-like cells,smooth muscle cells and endothelial cells.Compared with VZV infection, HSV antigensandDNA were not detected in dermal dendriticcells, monocytes/macrophages or endothelialcells.'3 Whether HSV has a tropism for variouscell types similar to that of VZV remains to beestablished by further studies.Another striking finding is the patchy dis-

tribution ofHSV and VZV infections in organs

such as the liver, lungs and spleen, mimickingthe distribution observed in cutaneous in-fections or in cell cultures. The mechanismslimiting the extent of each infectious focus isunknown.The detection of VZV and HSV in various

cell types by immunohistochemistry and in situhybridisation seemed to be highly correlated,and both methods were sensitive and specific.In some instances, however, immunohisto-chemistry is preferable to in situ hybridisation.This is particularly the case in necrotic lesions,in which immunohistochemical antigen de-tection revealed VZV gE, gB and IE63 or HSVantigens while nucleic acid hybridisation was

negative. It is possible that the resistance ofviral antigens to proteases during necrosis isgreater than that of nucleic acids to nucleases.An identical phenomenon is observed in HSVand VZV cutaneous granulomas, which are

likely to result from delayed-type hyper-sensitivity reactions against viral proteinconstituents.3'

Cases 1 and 3 underline the potential severityof primary VZV or HSV infections in sero-

negative, adult patients. The negative HSV sero-

logy in case 3 is probably because fulminantevolution outran the development of the B cellresponse.The major clinical alterations in these

patients correlated well with the histological,immunohistochemical and in situ hybridisationfindings. These results suggest that replicationof HSV and VZV has a direct, causative rolein multisystem organ failure, underlining theimportance of early diagnosis, immune de-ficiency disclosure and antiviral treatment.

In conclusion, this study confirms that type-specific diagnosis of HSV and VZV infectionsis possible on formalin fixed, paraffin wax em-

bedded tissue. As viral components can alsobe detected in cells that do not exhibit cy-topathological features, immunohistochemistryand in situ hybridisation are of potential valueto confirm the diagnosis of HSV and VZVinfections, especially when they cannot be diag-nosed on routine histological examination.The results also suggest that HSV and VZVhave a direct role in the pathogenesis of multi-system organ failure.

1 Gray GC, Palinkas LA, Kelley PW. Increasing incidence ofvaricella hospitalisations in the United States Army andNavy personnel: are today's teenagers becoming moresusceptible? Should recruits be vaccinated? Pediatrics 1990;86:867-73.

2 Sloan DSG, Burlinson A. Shift in age in chickenpox. Lancet1992;340:974.

3 Miller E, Vurdien J, Farrington P. Shift in age in chickenpox.Lancet 1993;341:308-9.

4 Dehertogh DA, Brettman LR. Hemorrhagic cystitis due toherpes simplex virus as a marker of disseminated herpesinfection. Am J Med 1988;84:632-5.

5 Shlien RD, Meyers S, Lee JA, Discher R, Janowitz HD.Fulminant herpes simplex hepatitis in a patient with ul-cerative colitis. Gut 1988;29:257-61.

6 Schiller GL, Nimer SD, Gajewski JL, Golde DW. Ab-dominal presentation of varicella-zoster infection in re-cipients of allogenic bone marrow transplantation. BoneMarrow Transplant 1991;7:489-91.

7 Perez-Oteyza J, Pascual C, Garcia-Larana J, Odriozola J,Rocamora A, Navarro JL. Abdominal presentation ofvaricella zoster infection after bone marrow trans-plantation. Bone Marrow Transplant 1992;9:217.

8 Stemmer SM, Kinsman K, Tellschow S, Jones RB. Fatalnoncutaneous visceral infection with varicella-zoster virusin a patient with lymphoma after autologous bone marrowtransplantation. Clin Infect Dis 1993;16:497-9.

9 Miliauskas JR, Webber BL. Disseminated varicella at aut-opsy in children with cancer. Cancer 1984;53: 1518-25.

10 Martin JR, Holt RK, Langston C, Gilden DH, RichardsonEP, Manz HJ, et al. Type-specific identification of herpessimplex and varicella-zoster virus antigen in autopsy tis-sues. Hum Pathol 1991;22:75-80.

11 Muraki R, Baba T, Iwasaki T, Sata T, Kurata T. Im-munohistochemical study of skin lesions in herpes zoster.Virchows Arch A Pathol Anat Histopathol 1992;420:71-6.

12 Nikkels AF, Sadzot-Delvaux C, CloesJM, Rentier B, PierardGE. Granulomatous reactions following herpes zoster con-tain varicella zoster glycoprotein gpI. J Invest Dermatol1992;98:522.

13 Nikkels AF, Debrus S, Sadzot-Delvaux C, Piette J, DelvenneP, Rentier B, et al. Comparative immunohistochemicalstudy of herpes simplex and varicella-zoster infections.Virchows Arch A Pathol Anat Histopathol 1993;422:121-6.

14 Nikkels AF, Delvenne P, Debrus S, Rentier B, Pierard GE.VZV glycoproteins gpl and gpII are present in dermalcells without their corresponding genome. J Cutan Pathol1992;19:544.

15 Nikkels AF, Delvenne P, Debrus S, Sadzot-Delvaux C,Piette J, Rentier B, et al. Distribution of varicella-zostervirus gpI and gpII and corresponding genome sequencesin the skin. Jf Med Virol 1995;46:91-6.

16 Baringer JR. Recovery of herpes simplex virus from humansacral ganglions. New Engl J Med 1974;291:828-30.

17 Mahalingam RM, Wellish M, Wolf W, Dueland AN, CohrsR, Vafai A, et al. Latent varicella zoster viral DNA inhuman trigeminal and thoracic ganglia. N Engl 7 Med1990;323:627-31.

18 Cheatham WJ, Weller TH, Dolan TF, Dower JC. Varicella:report of two fatal cases with necropsy, virus isolation,and serological studies. Am J7 Pathol 1956;32:1015-35.

19 McSorley J, Shapiro L, Brownstein MH, Hsu KC. Herpessimplex and varicella zoster: comparative histopathologyof 77 cases. IntJtDermatol 1974;13:69-75.

20 Debrus S, Sadzot-Delvaux C, Nikkels AF, Piette J, RentierB. Varicella-zoster virus gene 63 encodes an immediateearly protein abundantly expressed during latency. J Virol1995;69:3240-5.

21 Davison AJ, Scott JE. The complete DNA sequence ofvaricella-zoster virus. J Gen Iirol 1986;67:1759-816.

22 Merville-Louis MP, Sadzot-Delvaux C, Delree P, Piette J,Rentier B, Moonen G. Molecular cloning of varicella-zoster virus DNA and its detection in situ in infectednerve cells. Arch Int Physiol Biochem 1987;95:B87.

23 Feinberg AP, Vogelstein B. A technique for radiolabellingDNA restriction endonuclear fragments to high specificactivity. Ann Biochem 1983;132:6-13.

24 Sherman RA, Silva J, Gandour-Edwards R. Fatal varicella inan adult: case report and review of the gastrointestinalcomplications of chickenpox. Rev Infect Dis 199 1;13:424-7.

25 Ross JS, Fanning WL, Beautyman W, Craighead JE. Fatalmassive hepatic necrosis from varicella-zoster hepatitis.Am J Gastroenterot 1980;84:423-7.

26 Silk HJ, Guay-Woodford L, Perez-Atayde AR, Geha RS,Broff MD. Fatal varicella in steroid-dependent asthma. JTAllergy Clin Immunol 1988;81:47-51.

27 Close GC, Houston IB. Fatal haemorrhagic chickenpox ina child on long-term steroids. Lancet 1981;i:480.

28 Disler RS, Dover JS. Chronic localized herpes zoster inthe acquired immunodeficiency syndrome. Arch Dermatol1990;126: 1105-6.

29 Hoppenjans WB, Bibler MR, Orme RL, Solinger AM.Prolonged cutaneous herpes zoster in acquired im-munodeficiency syndrome. Arch Dermatol 1990;126: 1048-50.

30 Nikkels AF, Debrus S, Sadzot-Delvaux C, Piette J, RentierB, Pierard GE. Localization of varicella-zoster virus nuc-leic acids and proteins in the human skin. Neurology (inpress).

31 Nikkels AF, Debrus S, Delvenne P, Sadzot-Delvaux C,Rentier B, Pierard GE. Viral glycoproteins in herpesviridaegranulomas. Am Jf Dermatopathol 1994;16:588-94.

248

on February 21, 2021 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.49.3.243 on 1 March 1996. D

ownloaded from