Skip to main content
eScholarship
Open Access Publications from the University of California

Herpes simplex virus (HSV)-associated erythema multiforme (HAEM): A viral disease with an autoimmune component.

  • Author(s): Aurelian, Laure
  • Ono, F
  • Burnett, J
  • et al.
Main Content

Herpes simplex virus (HSV)-associated erythema multiforme (HAEM): A viral disease with an autoimmune component.
Aurelian, L1, Ono, F1, and J Burnett2.
Dermatology Online Journal 9 (1): 1

1. Virology-Immunology Laboratories, Department of Pharmacology and Experimental Therapeutics; and 2. Department of Dermatology2, University of Maryland, School of Medicine, Baltimore, Maryland 21201

Abstract

Erythema multiforme (EM) is a clinical conundrum the name of which reflects the broad morphological spectrum of the lesions. Molecular and immunologic evidence that herpes simplex virus (HSV) causes a subset of EM lesions [herpes-associated EM (HAEM)] is reviewed, and new data are presented which suggest that autoreactive T-cells triggered by virus infection play an important role in HAEM pathogenesis. Disease development begins with viral DNA fragmentation and the transport of the DNA fragments to distant skin sites by peripheral blood mononuclear cells (PBMCs). HSV genes within DNA fragments deposited on the skin [notably DNA polymerase (Pol)] are expressed, leading to recruitment of HSV-specific CD4+ Th1 cells that respond to viral antigens with production of interferon-γ (IFN-γ). This step initiates an inflammatory cascade that includes expression of IFN-γ induced genes, increased sequestration of circulating leukocytes, monocytes and natural killer (NK) cells, and the recruitment of autoreactive T-cells generated by molecular mimicry or the release of cellular antigens from lysed cells. The PBMCs that pick up the HSV DNA [viz. macrophages or CD34+ Langerhans cells (LC) precursors], their ability to process it, the viral proteins expressed in the skin and the presence of epitopes shared with cellular proteins may determine whether a specific HSV episode is followed by HAEM development. Drug-associated EM (DIEM) is a mechanistically distinct EM subset that involves expression of tumor necrosis factor α (TNF-α) in lesional skin. It is our thesis that the polymerase chain reaction (PCR) assay for HSV DNA detection in lesional skin and staining with antibodies to IFN-γ and TNF-α, are important criteria for the diagnosis of skin eruptions and improved patient management.



Introduction

Erythema multiforme (EM) was first described by Ferdinand von Hebra over a century ago [1]. It is a clinical conundrum, the name of which means multiple forms of redness and reflects the broad morphological spectrum of the lesions. It is characterized by a polymorphous eruption composed of symmetrically distributed macules, papules, bullae, and typical target lesions with a central vesicle or bulla. The lesions have a propensity for the distant extremities and the oral mucosae. Desquamation of less than 10% of the body surface can occur. Spontaneous resolution occurs with scaling and hyperpigmentation within 3–5 weeks. Fatalities are rare. The incidence of EM is estimated at 0.01–1%/year [2 ,3 ,4]. Histopathological features include accumulation of mononuclear cells around upper dermal blood vessels, endothelial swelling, and epidermal damage with hydropic degeneration of basal cells and focal keratinocyte necrosis. Exocytosis of lymphocytes and mononuclear cells is characteristic. Epidermal damage is often prominent in target lesions [5,6 ,7 ,8].

Some clinicians consider EM a single disease with different degrees of clinical severity that includes Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) [9]. However, most agree that all three are distinct entities [9,10 ,11,12 ]. EM minor is a mild recurrent condition that clears without sequelae and is associated with infection, including infection with Mycoplasma pneumoniae, vaccinia or varicella-zoster virus (VZV) [13 ,14 ], Patient history, clinical observations and prospective studies indicate that most cases of EM minor follow infection with herpes simplex virus (HSV). Recurrent herpes associated EM (HAEM) can be precipitated by sun exposure (like recurrent HSV), with a mean of 2.6 episodes per year in prepubertal children [15]. Histopathological features of HAEM lesions are consistent with delayed type hypersensitivity (DTH) reactions. The preponderance of cells infiltrating the dermis is T-cells, primarily CD4+ [7,11 ,12 ,16 ]. EM major, a broader category, is caused by drug administration. It occurs as an isolated finding and is accompanied by constitutional symptoms. In stark contrast to HAEM, the abnormal histopathologic features include a dermal infiltration of preponderantly non-T and CD8+ T-cells [5]. Acrosyringeal concentration of necrotic keratinocytes is a potential clue to drug etiology [17 ].

The recent international Severe Cutaneous Adverse Reactions (SCAR) study is an excellent attempt to relate the type of skin eruption with its cause. It concludes that EM patients are younger, more often male, they have a 10-fold higher rate of recurrence and less frequently have body temperature above 38.5° C and involvement of two or more mucous membranes. Clinical presentation consists of acrally distributed target lesions that fit the original description of EM. However, our past series of patient investigations have identified many exceptions to the concept of HAEM as an eruption of only acral targetoid lesions. Among the unusual presentations, we have seen EM (i) follow both recurrent and primary cutaneous HSV infections; (ii) occur after many, but not all, HSV recurrences experienced by a given patient; (iii) appear in the same patient producing targetoid lesions in one instance but plaques in a subsequent event; and (iv) present as frequent recurrences with oral mucosal targetoid lesions in a patient repeatedly denying HSV infection but having a positive polymerase chain reaction (PCR) for HSV in the involved lesion [18].

In this report, we discuss our thesis that a diagnosis of HAEM should be applied to a subgroup of EM patients in whom HSV was identified as a probable cause by virological, molecular and/or immunological criteria. PCR can be used to identify the presence of HSV DNA in HAEM lesions and/or the tissues can be studied for HSV gene expression by reverse transcriptase PCR (RT-PCR) and/or immunohistochemistry with antibodies to specific viral genes. Detection of interferon-γ (IFN-γ) in HAEM lesions can also be used as evidence of virus involvement [19]. Another patient subset that we define as drug-induced EM (DIEM), is characterized by a symmetrical eruption of the skin and, in most cases, oral mucosae and is distinct from SJS and TEN. This rash has vesicles, bullae, erythematous plaques, and, occasionally, a few targetoid lesions. Most often it is central rather than acral. Pruritus may be common. Histologic examination shows a preponderance of monocytes with T-cells that are primarily CD8+, and expression of tumor necrosis factor a (TNF-α ) [19] (Table 1). Here, we briefly review molecular and immunological evidence that HSV causes HAEM and present new data that suggest that HAEM pathogenesis includes a virus-triggered autoimmune component.


1. HAEM lesions contain HSV DNA fragments.


Figure 1
Fig. 1. Skin localization of HSV DNA. (A) In situ PCR with Pol primers of formalin fixed HAEM lesional skin reveals positive signals in basal and spinous keratinocytes and the hair follicle. Arrow identifies cluster of positive cells.

Although suspected as the cause of HAEM since the 19th century, evidence that HSV is involved in HAEM etiology is recent. Historical difficulties stem from the almost-universal failure to isolate infectious virus from HAEM lesional skin or detect viral DNA in skin tissues by In situ hybridization, the detection limit of which is ~ 1 HSV genome copy/cell [20,21,22,23]. During the last decade, however, numerous studies reported the presence of HSV DNA in HAEM lesions, by using the PCR assay, the detection limit of which is on the average 10-5 HSV genome copies per cell equivalent [22 ,23 ,24 ,25 ,26 ,27 ,28 ,29 ]. These findings suggest that either the entire HSV genome is present in rare cells or larger numbers of cells contain viral DNA fragments. Arguing against the presence of the entire HSV genome, PCR with equally sensitive primers found that a high proportion of HAEM tissues (75%) was positive for the viral DNA polymerase (Pol) gene while only rare tissues were positive for additional HSV genes or genes that span the entire HSV genome [23 ,28 ]. In situ PCR indicated that Pol DNA is located in keratinocytes in the basal and spinous cell layers of the lower third of the epidermis. The number of positive cells differed in various specimens and was between 10 and 95 per field. In some tissues the signals had a focal distribution consisting of random clusters of 2–15 keratinocytes, which may reflect clonal expansion of DNA positive cells or transfection of adjacent cells by the viral DNA (Fig. 1).


Figure 2
Fig. 2. HAEM lesions contain HSV DNA fragments. Schematic representation of HSV genome indicating the location of the UL and US components bracketed by inverted repeats (ab,b'a', a'c',ca) the sites of the origins of replication (oriL and oriS) and 7 genes that encode proteins required for viral DNA replication (UL5, UL8, UL9, UL29, UL30, UL42 and UL52) (black boxes). Genes targeted for PCR analysis of HAEM lesional skin encompass most of the viral genome and include UL5, UL9, UL19, UL23, UL29, UL30, UL39, UL42 and UL54 that encode proteins listed below the UL designation. They include OBP (origin binding protein), TK (thymidine kinase), ICP8, Pol (DNA polymerase), RR1 (large subunit of ribonucleotide reductase), Pol Access (polymerase accessory protein) and immediate early (regulatory) protein ICP27. Results of PCR are shown for 7 HAEM patients positive for viral DNA fragments (patients 1081, 3731, 6464, 2222, 7128-1, 012024, TB). DNA fragments present in the skin from the 7/10 positive patients (listed) are shown below the diagram. Bars encompass all genes positive by PCR. DNA fragments in patient 012024 are not contiguous, suggesting that the lesions contain at least 2 distinct DNA fragments. Lesional skin from patients 49531, 9051 and 10063 are negative with all the studied primers.

More recently, we used PCR with primers for nine genes located along the HSV genome in order to examine the presence of viral DNA in lesional skin from five patients with HSV lesions without HAEM and 10 patients with HAEM. Lesional skin from all the HSV patients [5/5 (100%)] were positive for viral DNA with all the studied primers. By contrast, Pol was the only viral DNA that was reproducibly detected in most HAEM skin tissues (7/10 or 70%). All the HAEM tissues were negative for UL5 (component of helicase-primase complex) and UL9 (origin-binding protein, OBP), both of which are involved in viral DNA replication. In most HAEM tissues, positivity initiated between UL19 and UL29 and terminated between UL30 and UL42. In at least one tissue, DNA positivity was not contiguous. The data are summarized in Figure 2 for the 7 HAEM patients positive for HSV DNA. They do not reflect viral DNA replication at the origin (oriL), because replication requires the protein products of 7 HSV genes [UL5, UL8, UL9, UL29, UL30, UL42 and UL52] [30] and most of these are not present in HAEM lesions (Fig.2). In view of accumulating evidence that skin retains viral or bacterial DNA fragments (31,32), the most likely interpretation of the data is that HAEM lesions contain HSV DNA fragments. Lesional skin from HAEM patients 49531, 9051 and 10063 were negative for HSV DNA with all the studied primers.E The exact interpretation of the failure to detect HSV DNA in these tissues is unclear.E It may reflect unrecognized technical difficulties, the presence of viral DNA fragments other than those sampled by the studied primers, inappropriate clinical diagnosis or tissue sampling, or time relative to lesion onset.E Indeed, although HSV DNA clearance from HAEM lesional skin is relatively slow, the duration of HSV positivity differs in various patients and even in various recurrent episodes experienced by the same patient (Table 2).


2. HSV gene expression is involved in HAEM lesion development.


Figure 3
Fig. 3. Skin localization of HSV gene expression. Immunofluorescent staining with Pol antibody of HAEM lesional skin from the patient shown in Fig. 1. Gene expression is in basal and spinous keratinocytes.

The following clinical findings suggest that HSV causes HAEM. First, HAEM always follows, and never precedes, HSV lesions. Second, skin-delivered viral antigens cause HAEM [36], and third, continuous prophylactic treatment with acyclovir prevents both HSV and HAEM recurrences, with drug failure attributed to patient noncompliance or insufficient tissue concentrations [35 ,37 ,38 ]. Significantly, HAEM is not always preceded by clinically overt HSV lesions although lesional skin contains viral DNA, suggesting that subclinically shed virus can also trigger HAEM [33 ,34 ,35 ]. Viral DNA clearance from HAEM lesions is relatively slow, such that healed HAEM skin could still be positive for HSV DNA at 1–5 months after lesion resolution. Viral genes are expressed (RNA and protein) in HAEM lesions (Fig. 3), but by contrast to viral DNA, gene expression is found only in acute HAEM lesions (Table 2). Moreover, HAEM episodes with relatively short-lived HSV gene expression resolve faster than those with longer expression (10–11 and 15–18 days, respectively), providing molecular evidence that HSV is involved in HAEM causation [14 ,23 ,35 ].

  Table 1. Characteristic features of HAEM
and drug-induced EM (DIEM)
  HAEMa DIEM
Causative agentHSV1/2Drugs
Disease courseAcute, self-limited, recurrent (7-21 days after HSV lesion)Acute, self-limited not recurrent, does not follow HSV lesions
ProdromeAbsent/moderatePresent
Predilection sites Skin lesionAcral extremities Target lesions, commonAcral extremities, Face Target lesions rare, blisters
Mucosal involvementAbsent/minimalProminent
Constitutional symptomsAbsent/moderatePresent/severe
ComplicationsNoneInfreqeuent (pneumonia, hemorhage, GI, renal failure)
MortalityNone5-15%
  HistopathologybFocal necrotic KC; moderate/ pronounced edema; mononuclear infiltrate with predominant CD4+ T cells Extensive KC necrosis; acrosyringeal concentration of necrotic KC; less pronounced edema; mononuclear infiltrate and CD8+ T cells
Laboratory diagnosisLesional skin positive for HSV DNA (PCR) and IFN-g (immunohistochemistry)Lesional skin negative for HSV DNA (PCR); positive for TNF-a (immunohistochemistry)
aMycoplasma pneumoniae can cause a similar clinical picture but it is negative for HSV DNA by PCR. Bullous EM following VZV infection/vaccination was reported.
bHistologically, acrosyringeal concentration of necrotic keratinocytes (KC) was reported in DIEM, but not HAEM (17).  

We considered the possibility that HAEM development is determined by axonal transport of HSV DNA. In HAEM, this would entail centripetal spread up the dorsal root ganglia followed by centrifugal spread out the peripheral nerve elsewhere to the skin. Arguing against this interpretation, however, HSV DNA was not seen in skin from the proximal area of the same dermatome as HAEM lesional skin [23 ]. HAEM development is also unrelated to Pol catalytic activity, because formation of a heterodimeric complex of Pol with UL42 protein is required for catalytic activity [30 ], and UL42 is expressed in only 57% of Pol-positive tissues [19 ]. HSV proteins could provide aberrant regulatory signals that alter the cell cycle and/or trigger apoptosis, as suggested by the finding that the UL42 protein interacts with cdc2 cyclin-dependent kinase, which is involved in the G2-M transition [39 ]. Indeed, apoptosis was seen in skin from 33% of HAEM patients, with apoptotic cells sparsely distributed in the epidermis [40] in a pattern reminiscent of that seen for HSV gene expression [19 ,23 ]. Although these findings may be suggestive, the major role of HSV proteins in HAEM pathogenesis appears to be the induction of a DTH response (see sections 3 and 4).

An intriguing question is why most HSV patients do not develop HAEM. Some studies have implicated HLA alleles B15(B62), B35, A33, DR53, DQB1*0301 and DQw3 in HAEM development [46 ,47 ]; other HLA alleles (viz. A1) were associated with the propensity to develop recurrent HSV lesions [48]. Moreover, the incidence of DQB1*0302 was increased in HAEM patients with mild mucous membranes involvement; whereas, an association with DQB1*0402 was seen in the disease form showing severe mucous membrane involvement [49]. However, longitudinal studies indicate that HAEM does not follow all the recurrent HSV episodes experienced by a given patient, and the factors that determine which HSV episode will result in HAEM causation are still unknown. Most likely, HAEM development is determined by the efficacy of HSV DNA dissemination to distant skin sites and its fragmentation during transport.


3. HAEM pathogenesis includes an HSV-specific immune component.

It has long been suspected that the DTH-like presentation of HAEM reflects an immune response induced by HSV antigens. However, the involvement of virus-specific immunity in HAEM pathogenesis has been difficult to demonstrate because the titers and activities of HSV-specific antibody, as well as the function and CD phenotype distribution of HSV responding T-cells, are similar in HAEM and HSV patients [14 ,41 ]. Because the T-cell receptor (TCR) repertoire used by infiltrating T-cells at DTH sites is restricted in various infections [42 ,43 ] and autoimmune diseases [44 ], we compared the usage of HSV-specific TCR in HAEM and HSV patients.

Approximately 95% of circulating human T-cells and most T-cells in human skin have the α β receptor [45], which is encoded by discontinuous variable (V), diversity (D), joining (J), and constant (C) gene segments. The Cα region consists of only one gene having the identical sequence from individual to individual, but the Vβ consists of approximately 70 individual genes belonging to 20 families, each with 1–7 members. Accordingly, we examined Vβ expression in HSV-stimulated PBMCs by RT-PCR using Cα as an internal control. The results for a specific Vβ were averaged in all patients per group and expressed as mean percent total integration units plus or minus standard deviation. A twofold or higher change in the percentage of a Vβ family over several time points was considered significant [43 ,44 ]. We found that virtually all Vβ families were expressed in PHA-stimulated PBMCs from patients with acute or healed (1–3 months after lesion resolution) HAEM or HSV lesions. All families were also seen in HSV-stimulated PBMCs from HSV patients, but only one-half of the families were seen in HSV-stimulated PBMCs from patients with acute HAEM [14 ]. Some families (viz. Vβ 2) were significantly increased in HAEM as compared to HSV patients (p<0.01). Restricted TCR usage in HAEM patients probably reflects the limited antigenic diversity of responding T-cells and is consistent with the small number of HSV proteins expressed in HAEM lesions. Significantly, there was a good correlation between HAEM lesion duration and the restricted TCR repertoire usage (Table 2), indicating that HSV-specific T-cell responses are involved in HAEM pathogenesis [16].

  Table 2. Clinical presentation and laboratory results
DateClinical Presentationa SpecimenbSkin BiopsyHSV DNA/ expressioncAltered TCR repertoire/skin Vb 2d
7/92eHSV then HAEM0-0

1-0

1-1

0-1
HSV lesion (day 2)
EM lesion (day 1)
EM lesion (day 3)
Healed HSV lesion
(day 30)
+/+

+/+

+/+

-/-
+/+

+/+

+/ND

ND/ND
3/94HSV and HAEM1-2

1-3

1-4

1-5
Perilesional skin
Normal skin hip
Normal skin shoulder
EM lesion (day 5)
-/-

-/-

-/-

+/+
 

 



+/+
4/94Lesion-free
(30d)
1-6Healed lesional skin+/--/-
7/95HAEM1-7-1

1-7-5
EM lesion
(day 1)
EM lesion
(day 5)
+/+

+/-
+/+

-/-
8/95HAEM1-8-3
1-8-10
ND EM lesion
(day 10)
ND
+/+
+/ND
+/+
9/95Lesion-free
(21d)
Lesion-free
(42d)
1-8-41

1-8-62
ND+/-

ND
-/-

-/ND
10/95HAEM1-9
1-10
EM lesion
(day 2)
Perilesional skin
+/+

-/-
+/+

-/-
11/95Lesion-free
(30d)
1-11Healed lesional skin +/-ND/+
11/96Lesion-free (13m)1-12Healed lesional skin-/--/-
6/97HAEM1-13

1-14

1-15
EM lesion (day 1)
EM lesion (day 10)
Healed lesion
(day 13)
+/+

+/+

+/-
+/+

+/+

-/-
2/93HSV2-1

2-2
HSV lesion,
(day 2)
HSV lesion
(day 10)
+/+

+/-
ND

-/-
2/93*HAEM2-3EM lesion
(12 hrs)
+/++/+
4/93Lesion-free (33d)2-4Healed lesional skin+/--/-
7/94HAEM2-5-6

2-5-11
EM lesion (day 6)
EM lesion (day 11)
+/-

-/-
-/-

-/-
10/96HAEM2-6-3

2-6-9
EM lesion (day 3)
EM lesion (day 9)
+/+

+/+
+/+

+/+
3/98Lesion-free (15m)2-7Healed lesional skin-/-ND
5/95HAEM 3-1EM lesion (day 1) +/++/+
11/95Lesion free (5m) 3-2Healed lesional skin-/--/-
6/96HAEM3-3EM lesion (day 5)+/++/-
4/98HAEM3-4EM lesion (day 1) +/++/+
a Parentheses represent time after lesion onset (d, days; m, months; hrs, hours) b Specimens 1-1 to 1-15 were obtained from patient 1; 2-1 to 2-7 from patient 2 and 3-1 to 3-4 from patient 3.
* HAEM onset was 7 days after preceding HSV episode (2/93). Specimens were obtained 12 hrs after onset of HAEM lesion.
c Expression of Pol RNA (by RT-PCR) and protein (by immunohistochemistry with Pol antibody)
d PBMC were cultured with HSV antigen and TCR repertoire alterations determined. Skin biopsies were stained with Vb 2 antibody. However, the antigenic specificity of the T cells in the skin is unclear
ND - not done


4. HSV-specific T-cells in HAEM lesions are CD4+ T helper 1 (Th1) and produce IFN-γ .

Having seen that HSV-specific T-cells from HAEM patients primarily express Vβ 2, we wanted to test the hypothesis that HSV-specific T-cells are recruited to HAEM lesions, giving rise to their characteristic DTH-like appearance. Serial sections of lesional skin were studied for expression of HSV proteins and T-cell infiltration by staining with antibodies to Pol, CD4, Vβ 2 and Vβ 3 (used as control). HAEM tissues positive for the Pol protein stained with antibodies to CD4 and Vβ 2, but not Vβ 3, with staining located in the epidermis and at the epidermis-dermis junction. Healed HAEM lesional skin was negative for both Pol antigen and T-cells. We conclude that the infiltrating T-cells are involved in disease manifestation because there is a good correlation between Pol expression, T-cell infiltration, and the duration of clinical symptoms [16 ,19 ,35] (Table 2).

In situ RT-PCR and immunohistochemistry studies indicate that both HAEM and HSV lesional skin tissues are positive for IFN-γ, a cytokine produced by CD4+ T helper type 1 (Th1) cells that are potential effectors of tissue damage [50 ]. IFN-γ is located within mononuclear cells and inflammatory infiltrates at the epidermis-dermis junction in lesional skin, but it is not expressed in healed tissues that are negative for HSV proteins and T-cell infiltration. There is a strong correlation between IFN-γ presence and HSV-protein expression in the skin tissues (p<0.0001 by Pearson correlation coefficient), indicating that HSV-specific Th1 cells are involved in HAEM pathogenesis [19]. HAEM skin tissues are also positive for transforming growth factor β (TGF-β ) [19] and IFN-γ -induced proteins that function as lymphocyte attractants [56], suggesting that IFN-γ activates tissue macrophages and keratinocytes to release a wide spectrum of cytokines and inflammatory molecules that amplify cutaneous inflammatory events and increase accumulation of circulating leukocytes in the epidermis [52,53 ,54,55]. By upregulating the expression of HLA class I–II antigens, IFN-γ may also enhance the antigen-presenting cell capability of keratinocytes [57 ], thereby amplifying the immune-inflammatory process. Amplification could also be a result of the function of IFN-γ in the differentiation of inflammatory Th1 cells while inhibiting the generation of regulatory, Th2 cells. Finally, IFN-γ may increase the susceptibility of Pol-positive keratinocytes to lysis by macrophages and cytotoxic T-cells (CTLs), as evidenced by the observations that (i) HSV-infected human epidermal keratinocytes treated with IFN-γ express HLA-DR and are targets for such lysis in vitro (58,59), and (ii) CTLs recognize intracellular (nonstructural) HSV proteins [60 ] and could, therefore, lyse Pol-positive keratinocytes. Indeed, EM tissues express perforin, a cytolytic protein produced by CTLs [61 ].

By contrast to HAEM, DIEM lesional skin is positive for TNF-α [19 ], a pro-inflammatory cytokine that is produced by monocytes and macrophages [51 ], and induces migration of neutrophils into the skin, in preference to lymphocytes [52 ]. This finding is consistent with previous reports that the DIEM inflammatory response is primarily mediated by macrophages [5 ] and indicates that HAEM and DIEM are mechanistically distinct syndromes.


5. The majority of T-cells in HAEM lesional tissues recognize cellular antigens.

Our recent studies sought to obtain a better understanding of T-cell infiltration of HAEM lesional skin. Biopsies (6 mm) were obtained from 4 HAEM patients, one of which (GA) was studied within 30 hours of lesion onset, and the others (RM, TW and RL) on days 4–6 after lesion onset. Biopsies were also obtained from two HSV patients (CS and JB) on days 3 and 4 after lesion onset, respectively. The dermis and epidermis were physically separated after overnight digestion in dispase, and the epidermis was digested with trypsin-EDTA (20 min at 37° C), dissected with forceps and washed in modified Hank's solution with 50% fetal calf serum and 0.1% DNase. T-cells were 3%–8% of the total epidermal cells, as determined by immunohistochemistry with CD3 antibody on cytospin slides. Keratinocytes and T-cells were separated by density gradient centrifugation over metrizamide, and the lymphocytes were expanded by stimulation with 1 x 106/well of irradiated HSV-infected autologous PBMCs (generated, as described [62]) in medium supplemented with 10% heat-inactivated, pooled human AB serum and human IL-2 (40U/ml). In some experiments, lymphocytes were cultured with autologous PBMCs and a viral antigen mixture consisting of 400 µg/ml of an UV-irradiated extract from 20 hours HSV-1/2 infected human epidermoid (HEp-2) cells [63] and 10 µg/ml of purified Pol protein. The Pol protein was prepared from Sf9 (insect) cells infected with the ACNPV-UL30 baculovirus Pol vector (72 hrs; multiplicity of infection = 5) and purified by chromatography on phosphocellulose followed by hydroxyapatite [64]. T-cells were stimulated three times at weekly intervals to establish T-cell lines. Single cell clones were established from these lines by limiting dilution in Terasaki wells, in medium containing 10% human serum, 200 ng/ml PHA-P, 40U/ml human IL-2, 25ng/ml human IFN-γ and 25ng/ml human GM-CSF and with 104 autologous feeder cells per well. The plates were fed at 4–5 days intervals by adding medium with IL-2 until day 14 when T-cell growth was apparent. Cells were removed from the positive wells and expanded by culture in 96-well round-bottom plates with 105 autologous feeders per well in medium containing PHA and IL-2. After 48 hours the medium was replaced with medium containing IL-2, and the cultures were refed on days 6 and 10. On day 14, the cultures were split into two replicates—one for screening and the other for expansion. Expansion was by cultivation in 24-well and then 6-well plates in medium containing 40 U/ml of IL-2 and restimulation with PHA at 2–3 week intervals.

  Table 3. HSV-specificity of T cell clones from skin lesions of HAEM patient GA.
ClonebCDcTCRcProliferation to antigenaIFNg c
   cellularHSVPol 
GA clone A CD4Vb 21531
± 240
3419
± 49
3171
± 888
+
GA clone B CD4ND1711
±421
3879
±211
1910
±111
+
GA clone 1CD4ND3484
±41
4188
± v113
3700
±801
ND
GA clone 18 CD4ND7180
±1550
10716
±356
ND-
a Cells (0.5-1x103 cells) were cultured with irradiated autologous PBMC (105) and HSV, uninfected HEp-2 cell extract (cellular) or purified Pol antigen (10&micro:/ml) for 5 days and assayed for [3H]-TdR incorporation after an 18 hrs pulse.
bIn addition to the 4 HSV-specific T cell clones in this Table, 21 clones from this patient proliferated equally well or preferentially with cellular as compared to HSV antigen (Table 4).
cCells were stained with CD3, CD4, CD8 and Vb antibodies and assayed by FACS.
dSupernatants of HSV stimulated T cell clones were assayed for IFN-γ production by ELISA, as previously described [67]
ND = not done

The T-cell lines established from these patients were stained with antibodies to CD3, CD4, CD8, or isotype control mouse immunoglobulins fixed with PBS-azide containing 2% D-glucose and 2.6% formaldehyde and analyzed by fluorescence-activated cell sorting (FACS). Five thousand cells were analyzed for each sample. In all lines, there was a higher proportion of CD4+ than CD8+ cells (60%–68% and 13%–33% CD4+ and CD8+ cells, respectively). To test for HSV specificity, cells (as few as 0.5-1x103) were cultured (5 days) with the irradiated autologous PBMCs 105 and 10µ g/ml HSV antigen (extract of 20 hours HSV-1/2 infected Hep-2 cells [63 ]) or cellular antigens [uninfected Hep-2 cell extract [63 ]] and pulsed with tritiated thymidine ([3H]-TdR) for 18 hours [63]. Lines from HAEM patient GA and from HSV patients CS and JB had higher proliferative activity with HSV than cellular antigens, indicating that they are HSV-specific. Lines from HAEM patients RM, TW and RL proliferated equally well to the viral and cellular antigens, indicating that they do not recognize HSV antigens. Four of the 25 T-cell clones established from patient GA (16%) were HSV-specific (evidenced higher levels of proliferation to HSV than cellular antigens), although the intensity of the virus-specific response was relatively low (Table 3). They were CD4+ and at least one of them (clone A) was Vβ 2+. Supernatants from two clones (A and B) cultured with HSV antigen were positive for IFN-γ, as determined by ELISA. Clone A proliferated equally well with HSV antigen and purified Pol protein, suggesting that it is Pol specific (Table 3). The other clones from patient GA and all the clones from HAEM patients RM, TW, and RL proliferated equally well with HSV and cellular antigens, indicating that they recognize cellular proteins (Table 4). All the clones from HSV patients CS and JB had a relatively high-intensity HSV-specific response (Table 4).

 Table 4. HSV specificity of T cell clones from HAEM, HSV and DIEM patients
  Patient/DxCloneProliferation with antigen 
  cellularHSVPol
GA/HAEMaclone C1541
   ±187
1422
   ±57
ND
GA/HAEMaclone 123552
   ±126
2921
   ±160
ND
GA/HAEMaclone 169941
   ±364
10436
   ±142
ND
RM/HAEMbclone 1612080
   ±606
11468
   ±270
ND
RM/HAEMbclone 88555
   ±701
8319
   ±1003
ND
TW/HAEMcclone2838
   ±617
2633
   ±161
ND
TW/HAEMcclone A1427
   ±230
1369
   ±228
ND
CS/HSVdclone 1343
   ±10
4239
   ±71
438
   ±12
CS/HSVdclone 2554
   ±20
5450
   ± 89
499
   ±62
CS/HSVdclone 3399
   ±15
8120
   ±99
402
   ±120
JB/HSVdclone 21677
   ±117
3250
   ±100
ND
a Cells (0.5-1x103 cells) were cultured with irradiated autologous PBMC (105) and HSV, uninfected Hep-2 cell or purified Pol antigens (10m g/ml) for 5 days and assayed for [3H]-TdR incorporation after and 18 hrs pulse (63). Twenty one clones established from HAEM patient GA (30 hrs after lesion onset) were not HSV specific.
b Eighteen clones established from HAEM patient RM (day 4 after onset) were not HSV-specific
c Twelve clones established from HAEM patient TW (day 5 after lesion onset) were not HSV specific
d Three clones established from HSV patient CS (day 3 after onset) and 2 clones established from HSV patient JB (day 4 after onset) were assayed as above and were HSV-specific. Clone 1 from patient CS proliferated better in response to Pol than cellular antigen, but the response was of low intensity presumably reflecting suboptimal antigen dose  

Collectively, the data suggest that HSV-specific T-cells in HAEM lesions represent a low-intensity transient response (not seen in patients studied at 4–6 days after lesion onset). This finding is consistent with the limited expression of HSV genes at this site (both in terms of antigenic complexity and duration), and the restricted HSV-specific TCR repertoire usage in HAEM patients. The virus-specific response in HAEM lesional skin is followed by an inflammatory amplification loop of higher intensity and longer duration that involves recruitment of T-cells that respond to cellular antigens (autoreactive). Supporting this interpretation are the observations both that virus-induced inflammatory responses trigger the release of self-antigens that stimulate autoreactive T-cells [65], and also that viral determinants that mimic host antigens stimulate self-reactive T-cell clones to attack host tissues [66 ,67,68 ,69 ]. Similar conclusions were reached for HSV-induced stromal keratitis in the mouse, where an epitope from a HSV coat protein is recognized by autoreactive T-cells [68]. Indeed, molecular mimicry may be an important mechanism to translate a low-level virus infection into an autoimmune response.

Autoreactive T-cells are likely recruited to HAEM lesional skin by cytokines and chemokines that are produced by keratinocytes at the site of inflammation and DTH. We conclude that the production of such cytokines and chemokines is stimulated by IFN-γ generated during the virus-specific response, because (i) IFN-γ is expressed in both HAEM and HSV lesional skin; (ii) IFN-γ stimulates keratinocytes to express HLA-DR antigens, the adhesion molecule ICAM-1 (CD54), and pro-inflammatory cytokines and chemokines; [52 ,53 ,54 ,55 ], and (iii) HAEM lesional skin is positive for the lymphocyte attractant chemokines Mig (macrophage IFN-γ inducible gene), IP10 (IFN-γ inducible protein 10) and RANTES [56 ] as well as TGF-β [19], a pro-inflammatory cytokine that recruits and activates immature leukocyte populations (70,71). Significantly, TGF-β induces cell-cycle arrest involving the cyclin-dependent kinase inhibitor p21waf [72], and p21waf is also expressed in HAEM lesions. Neither TGF-β nor p21waf are expressed in healed lesional skin, and their expression is related to HSV infection, because they are also present in HSV but not in DIEM lesional skin [19].


6. HSV DNA transport to the HAEM lesional site.

PBMCs from HAEM patients contain Pol DNA, suggesting that HSV DNA is transported by the vascular route [41 ]. Expression of the adhesion molecule ICAM-1 by microvascular endothelial cells is significantly increased by co-cultivation with HSV-1-infected PBMCs, suggesting that ICAM-1 may be involved in preferential adhesion and extravasation of PBMCs that carry viral DNA. Viral gene expression is required for ICAM-1 upregulation, since upregulation is not mediated by UV-inactivated HSV [29]. Endothelial cells co-cultured with HSV-infected PBMCs also express higher levels of HLA class I, thus setting the stage for the dermal inflammatory response characteristic of HAEM [29]. Presumably, HSV DNA fragments in circulating or decaying PBMCs are released at the time of PBMC deposition on the skin [74]. However, ICAM-1 upregulation is not restricted to HAEM, since it was also implicated in PBMC recruitment to DIEM skin lesions [75].

The PBMCs responsible for HSV transport are still unknown. We proposed that HSV is picked up by monocytes and macrophages [74 ], an interpretation that is supported by recent in vitro findings that monocytes evidence the highest rate of infection with HSV (80%) [29 ]. However, HSV might also be transported by CD34+ hematopoetic Langerhans cells (LC) progenitors. These bone-marrow-derived leukocytes undergo phenotypic changes after antigenic challenge, changes that allow them to migrate to regional lymphoid organs, where they activate resting T-cells. Approximately one-half of the CD34+ cells in peripheral blood are positive for the skin homing receptor CLA, and they may deposit the viral material on immigration back to the skin where they differentiate into LC [76]. The development of LC from CD34+ precursors and their localization in the epidermis is facilitated by TGF-β [76 ], which is upregulated in HAEM lesions [19]. Implicit in this interpretation is the conclusion that patients who develop HAEM differ from other HSV patients in the proportion of CD34+ LC progenitors in peripheral blood. An alternative—but not mutually exclusive—interpretation is that CD34+ cells from HAEM patients are less efficient in their ability to degrade the HSV DNA they picked up from the site of the HSV lesion, thereby generating relatively stable DNA fragments that are deposited at distant skin sites. Because viral DNA is likely to be picked up at random by various PBMC subpopulations, this interpretation predicts that HAEM lesion development after some—but not other—HSV episodes experienced by the same patient may be determined by the specific PBMC subpopulation (i.e., macrophages or CD34+ LC precursors) that picks up the HSV material. The mechanisms of (i) HSV DNA processing; (ii) transfer of viral DNA fragments from macrophages or LC to keratinocytes; and (iii) the generation of autoreactive T-cells, are still unclear.


Conclusions

Convincing molecular and immunologic data support previous clinical and histological evidence that HSV causes HAEM. The most significant aspects of disease pathogenesis as understood today are summarized in Table 5. Disease is initiated in the absence of virus replication by the expression of HSV genes contained within DNA fragments that are transported and generated by PBMCs, notably macrophages and CD34+ LC precursors. HSV DNA fragmentation could also occur on extravasation of decaying PBMCs. Inflammatory responses are initiated by viral gene expression in the skin and the recruitment of HSV specific CD4+ Th1 cells. IFN-γ generated by this response upregulates cytokines and chemokines that amplify cutaneous inflammatory events both with increased sequestration of circulating leukocytes, monocytes, and NK cells and with the recruitment of autoreactive T-cells to the epidermis. The mechanism of autoreactive T-cells generation is still unclear and may involve molecular mimicry or release of cellular antigens from lysed cells that express HSV proteins. The outcome is epidermal damage resulting from lysis of surrounding keratinocytes, release of various cytotoxic factors, and keratinocyte growth arrest and apoptosis. Inasmuch as HSV DNA fragmentation is likely to be a random event, different viral genes are presumably deposited on the skin during distinct HAEM episodes, a possibility that may explain the broad morphological spectrum of the lesions and their severity, even in the same patient. By contrast to HAEM, T-cells are a relatively small component of the inflammatory cell population in DIEM lesions, and DIEM lesional skin is positive for TNF-α , an inflammatory cytokine produced by monocytes rather than activated T-cells. TGF-β and p21waf are not expressed in DIEM lesions. Together with PCR for HSV DNA, immunohistochemistry for IFN-γ and TNF-α , can serve as laboratory-based diagnostic tests to differentiate HAEM from DIEM.

Table 5. Important Steps in HAEM pathogenesis
  1. Primary or recurrent HSV infection which may be subclinical or accompanied by visible vesicular lesions
  2. Macrophages and/or CD34+ hematopoetic progenitors engulf HSV and DNA is fragmented resulting in generation of DNA fragments that encompass HSV genes including Pol
  3. Peripheral extravasation of PBMC carrying HSV DNA fragments possibly related to ICAM-1 expression on endothelial cells
  4. Deposition of HSV DNA at distant skin sites
  5. Expression of HSV DNA in keratinocytes. Duration of gene expression determines the length of HAEM eruption
  6. Infiltration of HSV-specific CD4+ Th1 cells in the dermis/epidermis of HAEM lesions Activated T cells have a restricted TCR repertoire and include increased proportion of Vb 2 cells.
  7. HSV-specific T cells respond to HSV antigens and generate IFN-g
  8. Generation of cytokine and chemokines amplification cascade, including TGF-b , Mig, IP10 and RANTES and recruitment of NK cells, monocytes and leukocytes
  9. Recruitment of autoreactive T cells to the lesion site resulting in an autoimmune amplification loop
  10. Epidermal cell damage results from attacks by cytotoxic T cells, NK cells and monocytes and/or chermokines in varying combinations
  11. TGF-b and p21waf are expressed in keratinocytes at the site and adjacent to the epidermal damage, possibly contributing to apoptotic cell death

ACKNOWLEDGMENTSL Studies from our laboratory were supported by grant AR42647 from the National Institute of Arthritis and Musculoskeletal and Skin Diseases, NIH.

References

1. Hebra, F. von, On Diseases of the Skin including Exanthemata, vol. 1. translated by C.H. Fagge, London, New Sydenham Society pp 285-289, 1866

2. Huff JC, Weston WL, Tonnesen MG: Erythema multiforme: a critical review of characteristics, diagnostic criteria, and causes. J Am Acad Dermatol 1983;8:763-775.

3. Leigh IM, Mowbray JF, Levene GM, Sutherland S: Recurrent and continuous erythema multiforme - a clinical and immunologic study. Clin Exp Dermatol 1984;10:58-67.

4. Dowd RM, Champion RH. Erythema multiforme. In: Champion RH, Burton JL, Burns DA, Breatnach SM eds. Textbook of Dermatology 6th edition, . Blackwell Science Pub. Oxford UK. pp: 2081-2084, 1998.

5. Margolis RJ, Tonnesen MG, Harrist TJ, Bhan AK, Wintroub BV, Mihm MC, Soter NA: Lymphocyte subsets Langerhans cells and indeterminate cells in erythema multiforme. J Invest Dermatol 1983;81:403-406.

6. Zaim MT, Giorno RC, Golitz LE, Kunke KS, Huff JC: An immunopathologic study of herpes-associated erythema multiforme. J Cutan Pathol 1987;14:257-262.

7. Malmstrom M, Ruokonen H, Konttinen YT, Bergroth V, et al. Herpes simplex virus antigens and inflammatory cells in oral lesions in recurrent erythema multiforme. Acta Derm Venereol 1990;70:405-410.

8. Ford, MJ., Smith KL, Croker BP, Hacker SM, Flowers FP. Large granular lymphocytes within the epidermis of erythema multiforme. J Am Acad Dermatol 1992;27:460-462.

9. Katz, J., Livneh, A., Shemer, J., Danon, Y.L., Peretz, B. herpes simplex-associated erythema multiforme (HAEM): a clinical therapeutic dilemma. Pediatr Dent 1999;21:359-362.

10. Assier H, Bastuji-Garin S, Revuz J, Roujeau J-C. Erythema multiforme with mucous membrane involvement and Stevens-Johnson syndrome are clinically different disorders with distinct causes. Arch Dermatol 1995;131:539-542.

11. Paquet P, Pierard GE. Erythema multiforme and toxic epidermal necrolysis: a comparative study. Am J Dermatopathol 1997;19:127-132.

12. Auquier-Durant A, Mockenhaupt M, Naldi L, Correia O, Scroder W, Roujeau JC. Correlations between clinical patterns and causes of erythema multiforme majus, Stevens-Johnson Syndrome and Toxic Epidermal necrolysis. Arch Dermatol 2002;138:1019-1024.

13. Villiger RM, vonVigier RO, Ramelli GP, Hassink RI, Bianchetti MG. Precipitants in 42 cases of erythema multiforme. Eur J Pediatr 1999;158:929-932.

14. Onishi I, Kishimoto S. Erythema multiforme after resolution of herpes zoster by acyclovir. Eur J Dermatol 2002;12:370-2.

15. Weston WL, Morelli JG. Herpes simplex virus-associated erythema multiforme in prepubertal children. Arch Pediatr Adolesc Med 1997;151:1014-1016.

16. Kokuba S, Imafuku S, Huang S, Aurelian L, Burnett J. Expression of a herpes simplex virus (HSV) gene and the qualitative nature of the HSV-specific T cell response are associated with the development of erythema multiforme lesions. J Brit Dermatol 1998;138: 952-964.

17. Zohdi-Mofid M, Horn TD. Acrosyringeal concentration of necrotic keratinocytes in erythema multiforme: a clue to drug etiology. Clinicopathologic review of 29 cases. J Cutan Pathol 1997;24:235-240.

18. Kokuba H, Kauffman CL, Burnett JW, and Aurelian L. Clinical and virologic comparison of three patients with erythema multiforme. Acta Dermato-Venereologica. 1999;79:247-248.

19. Kokuba H., Aurelian, L., Burnett, J.W. Herpes simplex virus associated erythema multiforme (HAEM) is mechanistically distinct from drug-induced EM: IFN-γ is expressed in HAEM lesions and TNF-α in drug-induced EM lesions. J Invest Dermatol 1999;113:808-815.

20. Smith EM, McLaren LC: Attempt to recover simplex virus from skin sites of recurrent infection. Int J Dermatol 1977;16:748-757.

21. Orton PW, Huff JC, Tonnesen MG, Weston WL: Detection of herpes simplex virus antigen in skin lesions of erythema multiforme. Ann Intern Med 1984;101:48-50.

22. Miura S, Smith CC, Burnett JW, Aurelian L: Detection of viral DNA within skin of healed recurrent herpes simplex infection and erythema multiforme lesions. J Invest Dermatol 1992;98:68-72.

23. Imafuku S, Kokuba H, Aurelian L, Burnett J. Expression of herpes simplex virus DNA fragments located in epidermal keratinocytes and germinative cells is associated with the development of erythema multiforme lesions. J Invest. Dermatol 1997;109:550-556.

24. Brice SL, Krzemien D, Weston WL, Huff JC: Detection of herpes simplex virus DNA in cutaneous lesions of erythema multiforme. J Invest Dermatol 1989;93:183-187.

25. Cao M, Xiao X, Egbert B, Darragh TM, Yen TSB: Rapid detection of cutaneous herpes simplex virus infection with polymerase-chain reaction. J Invest Dermatol 1989;92:391-392.

26. Darragh TM, Egbert M, Berger TG, Yen TSB: Identification of herpes simplex virus DNA in lesions of erythema multiforme by the polymerase chain reaction. J Am Acad Dermatol 1991;24:23-26.

27. Aslanzadeh J, Helm KF, Espy MJ, Muller SA, Smith TF: Detection of HSV-specific DNA in biopsy tissue of patients with erythema multiforme by polymerase chain reaction. Br J Dermatol 1992;126:19-23.

28. Yokoi K. Erythema multiforme and HSV. The Japanese J Dermatol 1995;105:1661-1664.

29. Larcher C, Gasser A, Hattmannstorfer R, Obexer P, Furhapter C, Fritsch P, Sepp N. Interaction of HSV-1 infected peripheral blood mononuclear cells with cultured dermal microvascular endothelial cells: a potential model for the pathogenesis of HSV-1 induced erythema multiforme. J Invest Dermatol 2001;116:150-156.

30. Reddig PJ, Grinstead LA, Monahan SJ, Johnson PA, Parris DS. The essential in vivo function of the herpes simplex virus UL42 protein correlates with its ability to stimulate the viral DNA polymerase in vitro. Virology 1994;200:447-456.

31. Pancake BA, Wassef EH, Zucker-Franklin D. Demonstration of antibodies to HTLV-I tax in patients with cutaneous T cell lymphoma, mycosis fungoides, who are seronegative for antibodies to the structural proteins of the virus. Blood 1996;88:3004-3009.

32. Persing DH, Rutledge BJ, Rys PN, Podzorski DS, et al. Target imbalance: disparity of Borrelia burgdorferi genetic material in synovial fluid from Lyme arthritis patients. J Inf Dis 1994;169:668-672.

33. Aurelian L, Kessler II: Subclinical herpes virus infections of the genital tract are commonly associated with viral shedding. Cervix 1985;3:235-248.

34. Wald A, Zeh J, Selke S, Ashley RL, Corey L: Virologic characteristics of subclinical and symptomatic genital herpes infections. New Engl. J. Med 1995;333:770-775.

35. Kokuba, H., Imafuku, S., Burnett, J.W., Aurelian, L. Longitudinal study of a patient with herpes simplex virus associated erythema multiforme: viral gene expression and T cell repertoire usage. Dermatology 1999;198:233-242.

36. Shelley WB: Herpes simplex virus as a cause of erythema multiforme. JAMA 1967;201:153-156.

37. Tatnall FM, Schofield JK, Leigh IM. A double-blind placebo-controlled trial of continusous acyclovir therapy in recurrent erythema multiforme. Br J Dermatol 1995;132:267-270.

38. Kerob, D. Recurrent erythema multiforme unresponsive to acyclovir prohylaxis and responsive to valacyclovir continuous therapy. Arch Dermatol 1998;134:877.

39. Advani SJ, Weichselbaum RR, Roizman B. cdc2 cyclin-dependent kinase binds and phosphorylates herpes simplex virus 1 Ul42 DNA synthesis processivity factor. J Virol 2001;75:10326-10333.

40. Inachi S, Mizutani H, Shimizu M. Epidermal apoptotic cell death in erythema multiforme and Stevens-Johnson Syndrome. Arch Dermatol 1997;133:845-849.

41. Brice SL, Stockert SS, Bunker JD, Bloomfield D, Huff JC, Norris DA, Weston WL. The herpes-specific immune response of individuals with herpes-associated erythema multiforme copmpared with that of individuals with recurrent herpes labialis. Dermatol Res 1993;285:193-196.

42. Wang X-H, Ohmen JD, Uyemura K, Rea TO, Kronenberg M, Modlin RL. Selection of T lymphocytes bearing limited T-cell receptor _ chains in the response to a human pathogen. Proc Natl Acad Sci 1993;90:188-192.

43. Pantaleo G, Demarest JF, Schacker T, Vaccarezza M, et al. The qualitative nature of the primary immune response to HIV infection is a prognosticator of disease progression independent of the initial level of plasma viremia. Proc. Nat. Acad. Sci 1997;84:254-258.

44. Davies TF, Martin A, Concepcion ES, Graves P, Cohen L, Ben Nun A. Evidence of limited variability of antigen receptors on intrathyroidal T cells in autoimmune thyroid disease. N Engl J Med 1991;325:238-244.

45. Ohmen JD, Moy RL, Zovich D, Lieberman A, et al.: Selective accumulation of T cells according to T cell receptor V_ gene usage in skin cancer. J Invest Dermatol 1994;103:751-757.

46. Schofield, JK, Tatnall FM, Brown McCloskey JD, Navarrete C, I. Leigh IM. Recurrent erythema multiforme: tissue typing in a large series of patients. Br J Dermatol 1994;131:532-535.

47. Khalil I., Lepage Douay VC, Morin L, al-Daccak R, Wallach D, et al. HLA DQB1*0301 allele is involved in the susceptibility to erythema multiforme. J Invest Dermatol 1991;97:697-700.

48. Legendre C, Russell AS, Jeannet M. HLA antigens in patients with recrudescent herpes simplex infections. Tissue Antigens 1982;19:85-89.

49. Malo A, Kampgen E, Wank ER. Recurrent herpes simplex virus-induced erythema multiforme: different HLA DQB1 alleles associate with severe mucous membrane versus skin attcks. Scand J Immunol 1998;47:408-411.

50. Staats HF, Lausch RN. Cytokine expression in vivo during murine herpetic stromal keratitis. J Immunol 1993;151:277-283.

51. Fegali CA, Wright TM. Cytokines in acute and chronic inflammation. Front. Biosc 1997;2:12-26.

52. Colditz IG, Watson DL. The effect of cytokines and chemotactic agonsist on the migration of T lymphocytes into skin. Immunology 1992;76:272-278.

53. Barker JNWN, Mitra RS, Griffiths CEM, Dixit VM, Nickoloff BJ. Keratinocytes as initiators of inflammation. Lancet 1991;337:211-214.

54. Arany I, Brysk MM, Brysk H, Tyring SK. Regulation of inducible nitric oxide synthase mRNA levels by differentiation and cytokines in human keratinocytes. Biochem. Biophys. Res. Comm 1996;220:618-622.

55. Nacy CA, Maltzer MS. T cell-mediated activation of macrophages. Curr. Opin. Immunol 1995;3:330-335.

56. Spandau U, Brocker EB, Kampgen E, Gillitzer R. CC and CXC chemokines are differentially expressed in erythema multiforme in vivo. Arch Dermatol 2002;138:1027-1033.

57. Basham TY, Nickoloff BJ, Merigan TC, Morhenn VB. Recombinant gamma interferon HLA-DR on cultured human keratinocytes. J Invest Dermatol 1984;83:88-92.

58. Kalish RS. Non-specifically activated human peripheral blood mononuclear cells are cytotoxic for human keratinocytes in vitro. J Immunol 1989;142:74-80.

59. Miklsoka Z, Cunningham AL. Herpes simplex virus type 1 glycoproteins gB, gC and gD are major targets for CD4 T lymphocyte cytotoxicity in HLA-DR expressing humnan epidermal keratinocytes. J Gen Virol 1998;79:353-361.

60. Banks TA, Allen EM, Dasgupta S, Sandri-Goldin R, Rouse BT. Herpes simplex virus type 1-specific cytotoxic T lymphocytes recognize immediate-early protein ICP27. J Virol 1991;65:3185-3191.

61. Sayama N, Watanabe Y, Tohyama M, Miki Y. Localization of perforin in viral vesicles and erythema multiforme. Dermatology 1994;188:305-309.

62. Tigges MA, Koelle D, Hartog K, Sekulovich RE, et al. Human CD8+ Herpes simplex virus-specific cytotoxic T lymphocyte clones recognize diverse virion protein antigens. J Virol 1992;66:1622-1634.

63. Miura S, Kulka M, Smith CC, Imafuku S, et al. Cutaneous UV radiation inhibits herpes simplex virus-induced lymphoproliferation in latently infected subjects. Clin Imm Immunopathol 1994;72:62-69.

64. Hernandez TR, Lehman IR. Functional interaction between the herpes simplex-1 DNA polymerase and UL42 protein. J Biol Chem 1990;265:11227-11232.

65. Rouse, BT. Virus-induced immunopathology. Adv Virus Res 1996;47: 353-376.

66. Hemmer B, Fleckenstein BT, Vergelli M, Jung G, McFarland H, et al. Identification of high potency microbial and self ligands for a human autoreactive class II-restricted T cell clone. J Exp Med 1997;185:1651-1659.

67. Gyotoku, T, Ono, F Aurelian, L. Development of HSV-specific CD4+ Th1 responses and CD8+ cytotoxic T lymphocytes with antiviral activity by vaccination with the HSV-2 mutant ICP10_PK. Vaccine 2002;20:2796-2807.

68. Zhao Z-S, Granucci F, Yeh L, Schaffer PA, Cantor H. Molecular mimicry by herpes simplex virus type 1: autoimmune disease after viral infection. Science 1998;279:1344-1347.

69. Koelle DM, Chen HB, McClurkan CM, Petersdorf EW. Herpes simplex virus type 2-specific CD8 cytotoxic T lymphocyte cross-reactivity against prevalent HLA class I alleles. Blood 2002;99:3844-3847.

70. Sporn MB, Roberst AB. Transforming growth factor-β: recent progress and new challenges. J Cell Biol 1992;119:1017-1021.

71. McCartney-Francis NL, Wahls SM. Transforming growth factor β: a matter of life and death. J. Leukocyte Biol 1994;55:401-409.

72. Ravitz MJ, Wenner CE. Cyclin dependent regulation during G1 phase and cell cycle regulation by TGF-β. Adv Cancer Res 1997;71:165-207.

73. Strobl H, Riedl E, Schneicker C, Bello-Fernandez C, et al. TGF-β1 promotes in vitro development of dendritic cells from CD34+ hemopoietic progenitors. J Immunol 1996;157:1499-1507.

74. Aurelian L, Kokuba H., Burnett JW. Understanding the pathogenesis of HSV-associated erythema multiforme. Dermatology 1998;197:219-222.

75. Bruynzeel I, VanDer Raaij EMH, Boorsma DM, DeHaan P, Willemze R. Increased adherence to keratinocytes of peripheral blood mononuclear leucocytes of a patient with drug-induced erythema multiforme. Br J Dermatol 1993;129:45-49.

76. Strunk D, Egger C, Leitner G, Hanau D, Stingl G. A skin homing molecule defines the Langerhans cell progenitor in human peripheral blood. J Exp Med 1997;186: 1131-1136.

© 2003 Dermatology Online Journal