Dermatopathology

Acute Graft-versus-Host Disease

George F. Murphy
Brigham and Women's Hospital and Harvard Medical School
Boston, MA


Clinical History
A 28-year-old male was who had recently undergone induction therapy for acute myelocytic leukemia developed malaise and a vague maculopapular exanthem seven days after allogeneic hematopoietic stem cell transplantation. The rash was described as mildly pruritic. He had a history of atopic disease and had once developed a similar rash after taking an antibiotic. There was no clinical evidence of hepatic or gastrointestinal dysfunction. Because of the clinical circumstances that resulted in a broad differential, a biopsy was performed.


Slide 1
Click to view with ImageScope
Click to view with a Web-Based Viewer



Figure 1
Figure 2

Figure 3
Figure 4


Key Histologic Findings
  • Sparse superficial dermal lymphoid infiltrate about post-capillary venules, along the dermal-epidermal junction, and focally within the epidermal layer

  • Preferential association of lymphocytes with basal keratinocytes at tips of rete ridges and in lower infundibula of hair follicles

  • Lymphocytes surrounding keratinocytes, some with degenerative changes

  • Mild epidermal maturation disarray (due to chemotherapy)

Differential and Final Diagnosis
The differential diagnosis of interface alterations in an immunocompromised patient is quite broad. In this case, it is further complicated by the subtlety of the histologic findings. The differential includes interface drug eruptions, including erythema multiforme-like lesions; 'toxic' exanthems due to antigen-triggered cytokine release; cytotoxic response to viral infection; the exanthem of 'lymphocyte recovery'; and early acute graft-versus-host disease. Based on the restriction of interface changes to the rete tips and lower infundibula of hair follicles, the relatively 'pure' yet scant lymphoid inflammatory component, and evidence supporting early 'satellitosis' whereby lymphocytes surround target keratinocytes restricted to specific epithelial microdomains, a presumptive diagnosis of early acute graft-versus-host disease (GVHD) was issued. Differential diagnostic considerations were also provided in a comment, with a recommendation for repeat biopsy if lesions continued to evolve. A second biopsy 5 days later showed evolution of changes and confirmed the previous findings at a time when the patient had begun to experience diarrhea and elevated liver function tests--in aggregate further supporting the diagnosis of GVHD.

Acute Graft-versus-Host Disease: General Overview
Graft-versus-host disease (GVHD) involving the skin was first described in 1955 as a "secondary disease" (to distinguish it from the primary toxicity of serum sickness) in a murine model of hematopoietic stem cell transplantation [1]. The term acute GVHD has conventionally been used to describe a distinctive syndrome of hepatitis, gastroenteritis, and dermatitis developing within 100 days of an allogeneic hematopoietic stem cell transplant. The term chronic GVHD has indicated a syndrome involving dermatitis often with sclerosis and systemic manifestations suggestive of sclerodermoid autoimmunity that evolve after day 100. However, in an era of increased numbers of haploidentical, cord blood, nonmyeloablative, and unrelated transplants; "stem cell boosts"; and donor lymphocyte infusions, the temporal parameters for distinguishing acute versus chronic GVHD become potentially less helpful. For example, the histopathological alterations, as assessed by skin biopsy, need not adhere to a specific time course, with the occasional patient showing contemporaneous changes of both acute and chronic GVHD.

The clinical and histopathologic changes of cutaneous GVHD may occur in a number of clinical settings in addition to traditional bone marrow transplantation, including syngeneic transplants [2, 3], following blood transfusions [4, 5], and following solid organ transplants [6]. In certain patients, acute cutaneous GVHD may be difficult to distinguish from certain viral exanthemas as well as from some toxic and allergic drug eruptions [7]. Moreover, GVHD in the autologous setting may be difficult to distinguish from that in the allogeneic setting [8]. Thus, the possibility of this life-threatening disorder should be considered in the context of clinical setting more common than bone marrow transplantation (e.g. in patients who have received blood transfusions).

Skin biopsies, especially when performed serially to assess disease evolution and progression, may be useful in distinguishing GVHD from other disorders in the post-transplant period. However, the skin biopsy is neither entirely sensitive nor specific for GVHD detection. Thus, close correlation of skin biopsy results with other pathological and clinical parameters is necessary to enhance diagnostic accuracy. When an exanthem appears early (e.g. within one week) after infusion of a cellular product (e.g. a stem cell or donor lymphocyte infusion), a single skin biopsy may be of limited utility in diagnosing GVHD [9] due to the subtlety of the findings. Agents, such as intravenous contrast media and exposure to ultraviolet (UV) light, that in themselves may cause skin rashes, also have the capability to trigger or worsen GVHD [10], underscoring the importance of a detailed clinical history upon pathological examination of a biopsy. Controversy exists as to whether certain infectious processes, such as cytomegalovirus, can provoke or exacerbate GVHD, as opposed to simply arising secondary to the immune dysregulation that typically accompanies this disease [11, 12].

Clinical Considerations
The first manifestation of acute GVHD in the skin is a maculopapular rash. This rash may be punctate, correlating with early and preferential involvement of hair follicles and thus resembling folliculitis. It also may be more diffuse, resembling sunburn. Pruritis and/or burning may be prominent components. Rarely (<7% of cases), there is epidermal necrosis with sloughing and formation of bullae, resulting in a clinical presentation resembling toxic epidermal necrolysis [13]. This more severe type of cutaneous acute GVHD appears to be decreasing in incidence, possibly as the result of improved immunosuppressive regimens for prevention, alterations in conditioning regimens, improved HLA matching, or potentially as a consequence of earlier diagnosis and more rapid intervention with immunosuppressive agents. Recently a syndrome has been described in which patients receiving transplants, particularly from unrelated donors, exhibit hyperacute GVHD manifested by generalized erythroderma with desquamation and diarrhea developing before any signs of myeloid engraftment [14].

The clinical manifestations of chronic cutaneous GVHD are often more protean than its acute counterpart, although the skin is the most commonly involved organ in this manifestation of the disease. Chronic GVHD may or may not be preceded by clinically-detectable acute GVHD and may begin at sites of trauma, UV irradiation, or localized infection, such as with varicella zoster. There are two major variants of cutaneous chronic GVHD: (1) sclerodermoid and (2) lichenoid, a form of interface cytotoxic injury with histologic feature that closely resemble lichen planus. In the sclerodermoid form, affected individuals develop indurated plaques with hypo- and hyperpigmentation. Initially localized, the skin thickening may become progressively more diffuse, as in systemic scleroderma (progressive systemic sclerosis). Those cutaneous manifestations of chronic GVHD that are predominantly lichenoid show no typical pattern of distribution. Individual papules and small plaques are erythematous to violaceous and exhibit a variably shiny to scaling, flat-topped surface unlike that produced by the epidermal injury in acute GVHD. With chronicity, hypo- and hyperpigmented areas may develop as a result of repeated basal cell layer injury, and over time larger plaques may arise due to coalescence of individual papules. Poikilodermatous changes may develop in some patients. The lichenoid form may occur earlier after transplantation than the sclerodermoid form, and occasional patients exhibit both lichenoid and sclerodermoid features of chronic GVHD at the same time. Moreover, the finding of papillary dermal fibrosis in lichenoid GVHD has been espoused by some to be a harbinger for the eventual development of sclerodermoid disease.

Mucosal alterations may be encountered in both acute and chronic GVHD, although lesions tend to be more widespread in the chronic form, with involvement of the mouth, conjunctiva, and genitalia. Oral involvement is often characterized by focal to diffuse mucosal atrophy and erythema, especially affecting the tongue, pallet, lips, and buccal surfaces [15]. Ocular lesions producing keratitis sicca may result in symptoms of burning, itching, pain, and photophobia [16, 17]. Vaginal pathology, when severe, is manifested in the formation of strictures [18], and penile lesions may eventuate in phimosis [19] or Peyronie's disease [20].

Diagnostic Pathology of Acute GVHD
The first histological alterations that occur in acute cutaneous GVHD involve adhesion and transvascular diapedesis of lymphocytes through postcapillary venules situated within the uppermost dermis (superficial vascular plexus). Such changes are often observed in biopsies 2 to 3 weeks after transplantation or in association with the first clinical signs and symptoms of cutaneous involvement. Mast-cell degranulation about the affected venules appears to be related to activation of microvascular endothelial cells responsible for initial recruitment of effector cells. It is of interest, therefore, that pruritis, a common sign of mast-cell degranulation, may also accompany early clinical lesions of acute GVHD. The result of adhesive interaction between microvascular endothelial cells and effector cells is migration of the latter into the perivascular interstitium of the superficial (papillary) dermis. Remarkably few cells seem to be involved in this phase, and their migratory fate appears to be influenced by secondary induction of chemokines and adhesion molecules in target tissue, such as by cells in the overlying epidermal layer.

The diagnostic histology of this endothelial phase of acute GVHD is nonspecific, and serial biopsies are often required to establish a definitive diagnosis. Maculopapular viral exanthemas are probably the most common cause of a superficial perivascular lymphocytic infiltrate, and this possibility must always be considered when the alternative possibility of acute GVHD is entertained. Although viral exanthemas may be hemorrhagic and therefore associated with superficial perivascular extravasation of erythrocytes, altered coagulation status may also produce this finding in the posttransplant period. Drug eruptions must also be considered in the early endothelial phase of acute GVHD. Generally, drug eruptions will involve both superficial and deep dermal vessels, and the inflammatory infiltrate will include eosinophils as well as lymphocytes. However, in the setting of the immunosuppression that accompanies the posttransplant period, these characteristic features may be lacking. Moreover, although most lesions of acute GVHD, in our experience, do not harbor eosinophils, occasional lesions do. Cutaneous eruption of lymphocyte recovery, which is seen after chemotherapy and a period of marrow aplasia, shows an upper dermal, usually perivascular infiltrate with variable exocytosis of lymphocytes and spongiosis. Apoptotic keratinocytes are seen rarely. Thus, histological and clinical evolution over time (see below), as assessed in serial biopsies. is often the best diagnostic indicator of the presence of evolving acute GVHD.

Lymphocytes that initially accumulate about superficial dermal venules migrate from the perivascular interstitium into the overlying epidermal layer (epidermotropism). Some of these cells may seem to align along the dermal-epidermal junction, while others are present at all levels of the epidermis. The finding of lymphocytes in the epidermis is diagnostically important in early acute GVHD, for it is the harbinger of target-cell injury. Unlike other forms of dermatitis showing lymphoid epidermotropism, such as various forms of eczematous dermatitis, acute GVHD generally fails to show significant intercellular edema (spongiosis) within the epidermal layer. However, we have seen examples of true spongiotic dermatitis that, upon serial sampling, have demonstrated transition to lesions more typical of acute GVHD. Whether this represents the blending of two independent immune responses in the post-transplant period, or the potential for rare examples of early GVHD to show prominent spongiosis, remains unclear.

The most characteristic histologic finding in acute GVHD is the finding of "satellitosis," indicating target-cell injury within the epidermis or follicular epithelium. Satellitosis consists of multiple lymphocytes intimately surrounding one or several keratinocytes that show signs of eosinophilic degeneration and death. Such keratinocytes usually contain condensed, hyperchromatic, and sometimes fragmented nuclei. The cytoplasm is dark red due to increased eosin uptake (acidophilia). Recent evidence indicates that such dying target keratinocytes are actually undergoing apoptosis [21], although original reports referred to such cells as "dyskeratotic" [22, 23]. In general, the degree of epidermal injury in acute GVHD seems out of proportion to the sparse numbers of lymphocytes present within the superficial dermal and epidermal layers. It is important to realize that a small number of degenerating keratinocytes independent of lymphocytic apposition may be the sequelae of the pretransplant conditioning regimen alone. Accordingly, care should be taken to attribute epidermal necrosis to acute GVHD only when such changes are associated with unequivocal clustering of epidermotropic lymphocytes. A helpful feature in early acute GVHD is the preferential lymphoid infiltration and apoptosis affecting keratinocytes within the tips of epidermal rete ridges. With disease progression, vacuolization of contiguous basal cells may accompany satellitosis, and the former may be so pronounced in some patients as to result in diffuse epidermal sloughing in a manner akin to toxic epidermal necrolysis.

In addition to infiltration of the epidermis, migration by lymphocytes into the upper third of the hair follicle (follicular infundibulum) is also commonly observed in acute GVHD. This phenomenon has been termed "cytotoxic folliculitis" and is a helpful indicator of early GVHD even in biopsies that fail to show significant lymphocytic infiltration of the interfollicular epidermis. When GVHD is extensive and severe, separation of follicular epithelium from surrounding adventitia may occur as a result of contiguous apoptosis within the basal cell layer.

Acute GVHD also may involve other squamous epithelial surfaces that differ in architecture and cytology from most cutaneous regions. These include acral skin and oral mucosa. Biopsies of acral acute GVHD frequently reveal subtle alterations, consisting of superficial perivascular lymphoid infiltration and focal epidermotropism and keratinocyte apoptosis that selectively involves the rete tip epithelium that is often well-developed at this site. It must be remembered that stratum corneum in these regions is normally thick and compacted, and thus should not be interpreted as a component of the pathology. Squamous mucosal acute GVHD is characterized by a band-like infiltrate of lymphocytes (and usually some plasma cells) directly beneath the epithelial layer. Infiltration of the epithelium by lymphocytes may be associated with apoptosis and satellitosis, although this feature is variable. Occasionally the inflammation extends to involve the epithelium of minor salivary glands, which shows prominent lymphoid infiltration and, with chronicity, periductal fibrosis.

Sensitivity, Specificity, and Predictive Value of Histopathology
In murine models of acute GVHD, the number of apoptotic target epidermal cells [21] have been found to correlate with the number of effector T cells in the donor marrow inoculum, the presence and extent of hepatic and intestinal disease [22], and the overall severity of clinical disease [23]. It therefore is reasonable to ask whether the severity of disease, as assessed histopathologically in a skin biopsy, permits similar predictions in humans. By inference, the Lerner system of histologic grading [24] suggests that the extent of epidermal injury may be of biological significance in predicting clinical disease severity and outcome. Massi and coworkers have suggested that one potential drawback of this approach, however, is failure to take into account inflammatory infiltrate as a criterion in early (stage I and II) lesions [25]. This issue has been addressed in one study where 69 cyclosporin-treated, allogeneic BMT patients were evaluated to determine early clinical, laboratory, or histopathologic indicators for the development of progressive, fatal acute GVHD [26]. Whereas substantial increases in total bilirubin, stool output, extent of rash and overall clinical GVHD stage proved useful in predicting severe, potentially fatal forms of the disease, the number of lymphocytes entering into the epidermal layer and number of dyskeratotic (apoptotic) keratinocytes did not. In other studies, skin biopsy findings were found to correlate poorly with outcome in patients treated based on clinical severity of skin rashes suggestive of acute GVHD [9]. Accordingly, although we provide qualitative descriptions for suspected GVHD skin biopsy specimens in the form of a note indicating the nature and extent of effector cell infiltration and epidermal injury, we do not provide histologic grades as predictors of disease severity. It must also be emphasized that in any given patient, the skin biopsy may be of limited sensitivity for early detection of disease, particularly with respect to other target sites (e.g. gut). In addition, because the histopathology of early acute GVHD has considerable overlap with other conditions that occur in the posttransplant period (e.g. viral exanthemas, drug eruptions, eruptions associated with immunologic reconstitution), specificity is enhanced dramatically when the biopsy is correlated with other signs of disease.

Adjunctive Studies for Diagnosis of Cutaneous Acute GVHD
While histopathology remains the mainstay of diagnosis for both acute and chronic GVHD, problems with specificity and sensitivity have resulted in interest in development of special ancillary techniques that may assist in more accurate recognition of this often protean and diagnostically elusive condition.

Histochemical approaches
The potential ability to differentiate between acute GVHD and a cytotoxic hypersensitivity condition that may closely simulate it, erythema multiforme, has been examined using histochemical stains to detect bile pigment associated with hyperbilirubinemia within the epidermal layer [27]. No lesions of erythema multiforme stained for intraepidermal bile pigment (n=50), whereas 6% of the acute GVHD cases showed bile pigment within the epidermal layer (3 of 50 cases evaluated). Thus it appears that in a very small minority of acute GVHD cases, intraepidermal bile pigment serves as a built-in marker for the hyperbilirubinemia of GVHD-associated liver dysfunction, and therefore may assist in distinguishing between erythema multiforme and acute GVHD. The practical significance of this observation, however, is limited by the small number of cases that actually harbor intraepidermal bile pigment, and the questionable assumption that clinical parameters of liver involvement may not be available to the pathologist.

Immunohistochemical approaches
Interface drug eruptions, where cytotoxic lymphocytes attack the epidermal basal cell layer, may closely mimic acute GVHD. In one study that sought to compare acute interface drug eruptions with acute GVHD using immunohistochemistry, both conditions were found to have a predominance of CD8+ T cells in the infiltrate, reduction in the number of CD1a+ Langerhans cells, and increased epidermal expression of HLA-DR and ICAM-1 [28]. Thus, by these parameters, both acute interface drug eruptions and acute GVHD could not be reliably distinguished. In contrast, a similar approach taken to differentiate epidermal-type chronic GVHD (lichen planus type) from true lichen planus showed more promising results [29] based on a limited numbers of patients. Whereas lichen planus was characterized by infiltration of CD4+ T cells and increased numbers of Langerhans cells, lichenoid chronic GVHD demonstrated predominantly CD8+ T cells, associated LAK cell markers CD16 and CD28, and diminished numbers of Langerhans cells.

Molecular approaches
In one study, the unanticipated development of acute GVHD after massive blood transfusion was reportedly confirmed by examination of T lymphocytes infiltrating a skin biopsy by PCR directed against a Y chromosome-specific sex-determining region Y (SRY) gene [30]. The diagnosis was established by HLA-DNA typing with PCR-sequence-specific oligonucleotide that revealed the presence of complex HLA-DR chimerism in the peripheral lymphocytes collected after the onset of GVHD. Thus, the use of SRY-directed PCR could hold promise as a rapid technique for the early diagnosis of GVHD in female patients. However, it remains unclear as to whether mere infiltration of skin by chimeric donor cells implies effector function, since passive infiltration of donor cells at sites of inflammation due to non-GVHD related stimuli (e.g. drug-induced and viral exanthemas) could produce similar findings.

In vitro assays
Recently, skin organ culture has been rediscovered for the study of numerous histologic and immunopathological changes relevant to acute GVHD, such as induction of class II and related adhesion pathways [31, 32]. As a predictive assay, human skin has been used to anticipate the likelihood of acute GVHD developing in HLA-matched sibling bone marrow transplantation [33]. This model involves sensitizing donor lymphocytes in vitro in a primary mixed lymphocyte reaction and then evaluating the secondary response in patient skin biopsies by grading the resultant GVHD-like alterations histopathologically. Sviland and Dickinson have suggested that this model permits an 82% correlation of histopatholgic changes of GVHD in the explants with clinical outcome [34]. This approach may be refined further by using the TUNEL approach to mark apoptotic target cells in the explants [35], thereby enhancing prediction of situations likely to give rise to more severe grades of GVHD. These approaches remain in developmental stages, however. Moreover, a potential limitation lies in the fact that degenerative alterations that occur within the epidermal layer upon skin explantation in vitro may bear marked similarity to early evolutionary stages of acute GVHD. An exciting potential alternative to this approach has recently been discovered where viable and structurally normal human skin xenografts on genetically immunosuppressed (SCID) mice can be induced to express the phenotype of cytotoxic dermatitis with features of epidermal-type GVHD [36].

Pathobiology
Acute and chronic GVHD may appear to occur on a continuum in a single individual. However, they probably represent two different, albeit related, disease processes. Acute GVHD is a cytotoxic attack of donor lymphocytes on host tissues, most dramatically on cells associated with epithelial compartments (gut, liver, skin), which are seen as foreign. Chronic GVHD, on the other hand, not only involves cytotoxicity, but also a derangement of host immune function, stimulated by donor lymphocytes and allowing the development of autoimmunity in a permissive genetic background. In situations of GVHD where there is no tissue incompatibility (syngeneic GVHD), this immune dysfunction becomes paramount.

We have conceptually divided acute GVHD into three distinct phases: allostimulation, homing, and targeting. A fourth phase, resolution/chronicity, also exists, and will be discussed separately under the pathobiology of chronic GVHD. Throughout the three phases of acute GVHD, and ostensibly driving these events to a crescendo by the point of targeting, is an ever-increasing background of cytokine stimulation. As will be seen, these circulating cytokines appear to direct and enhance events related to cell stimulation, organ-directed migration, and cytotoxicity, and thus represent an integral part of the disease process.



During the allostimulatory phase of GVHD, the relative contributions of CD4 and CD8 T cells in the genesis of GVHD across miHA barriers has been one of our major interests [37, 38, 39]. Using H2-matched donor-recipient murine models, we have found that CD8 T cells, and in a more limited number of experimental situations, CD4 T cells, may mediate lethal GVHD across certain miHA barriers [38, 39, 40]. In the latter situation, allostimulation involves function of the CD4 molecule as a co-receptor for interaction of the T-cell receptor (TCR)-CD3 complex with specific antigen bound by MHC class II molecules on antigen presenting cells (APCs). CD4, a structural member of the immunoglobulin superfamily, contains four globular domain-like regions (D1-D4), and three complementary determining regions (CDR) in each domain. These CDRs are potential sites for protein-protein interactions, and synthetic peptides derived from CDR sequences may possess binding properties similar to those of the parent protein. Thus, recent work in this and other laboratories has also focused in selective inhibition of allostimulation via discovery and application of synthetic peptides that mimic CDR3-region mediated binding [41]. More recently, by examining the TCR Vb repertoire of effector T cells using CDR3-size spectratyping in murine models of miHA-induced GVHD, we also have identified expansion of specific V b families, correlated them with miHA differences between donor and host, and evaluated them over the course of disease progression [42, 43, 44]. This approach provides an important opportunity to potentially modulate disease activity by manipulating specific V b families that undergo clonal or oligoclonal expansion during early disease. Indeed, we have found that BMT with appropriate positive selection for relevant TCR Vb CD4+ and CD8+ T cell subsets that are expanded during allostimulation results in fatal GVHD induction, whereas mice transplanted with non-expanded T cells from the same donors survive with minimal morbidity [45].

During the homing phase, localization-specific effector Vb T cell families to microvascular beds within target organs is mediated by expression of a coordinated array of cytokine-inducible adhesion molecules by endothelial cells, facilitating progressive binding and eventual diapedesis of effector leukocytes. Cytokines known to play a key role in GVHD [45, 46] are capable of inducing endothelial molecules that facilitate leukocyte adhesion. We have also found that one of these, TNFa, is stored in sizable quantities in mast cells that express Ig eRI receptors and surround postcapillary venules in human and murine dermis [47, 48, 49]. Moreover, mast cell degranulation and related local TNFα release precedes effector T cell influx in experimental GVHD [50]. Animals treated with peptide analogs that inhibit IgE-Fc eRIa interactions as well as mast cell deficient murine transplant recipients show amelioration of GVHD [50, 51]. Molecules relevant to T cell recruitment to extracutaneous target tissues in GVHD include LPAM (Lymphocyte Peyer's Patch Adhesion Molecule) or α4β7 integrin, an important homing integrin on alloreactive gut-homing T cells. Indeed, we have found that donor T cells genetically deficient in α4β7 develop less intestinal GVHD, while cutaneous GVHD and GVL responses are not diminished [52].

While these and many other recent studies have partially elucidated the complexity of the effector stages of acute GVHD, considerably less is known of the target phase. Original descriptions in the rhesus monkey model of GVHD detailed epithelial cell death characterized by apposition of lymphocytes that aggregated at the perimeter of affected cellular targets. This phenomenon, termed "satellitosis", implied a requirement for direct interaction of infiltrating effector lymphocytes in the genesis of tissue pathology. In mouse models, we initially described injured cutaneous epithelial cells as "dyskeratotic" due to their intensely eosinophilic cytoplasm resembling aggregated keratin [53]. Quantitation of dyskeratotic target cells (the "dyskeratotic index") proved to correlate with number of T cells in donor inocula, disease severity, and treatment responses. The term dyskeratotic conveyed little, however, concerning the pathogenesis of target cell injury. In 1996, we learned that the mechanism of target cell injury in experimental GVHD was apoptosis, a finding soon validated in humans [54]. This suggested a fundamental role for expression and ligation of death receptors (e.g. Fas, TNFR-1, TRAIL) in GVHD target cell injury. While apoptotic cells were often directly associated with infiltrating T cells, they also occurred before T cell infiltration, suggesting a possible contributory role for soluble mediators in their generation. Moreover, apoptotic cells were not diffusely distributed in the basal cell layer, as traditional dogma suggested. Rather, they proved to be spatially restricted to basal cell layer domains at tips of human epidermal rete ridges, rete ridge-like prominences (RLPs) of murine tongue, and bulge regions of hair follicles known to express cytokeratin 15 (K15; 55) [55].

Summary
GVHD in its acute and chronic forms is all too often clinically protean, histologically deceptive, and pathogenically mysterious. As we learn more about its diagnosis and pathobiology, the complexity of this disorder seems to paradoxically increase. Yet enormous strides have been accomplished at a molecular level in an effort to tease out pathways that may themselves be targets for therapeutic intervention. For example, synthetic peptides crafted by computer-modeling to bind to and block specific molecular interactions involving the T cell receptor have shown promise in experimental systems [41]. Inhibition of mast cell activity, cytokines, and the adhesion molecules they induce are known to influence effector cell homing [51, 56], and thus represent additional strategies for disease amelioration. Approaches to alter molecular interactions that result in apoptotic targeting while preserving the ability of donor T cells to kill residual leukemic cells are under active exploration, and early results are encouraging [57]. All of these initiatives depend upon more sensitive and specific modalities to diagnose disease and assess its activity, particularly in major target organs such as the skin. Recognition of the three-step pathogenesis of GVHD (allostimulation, homing, and targeting) and the mechanisms that are responsible for these progressive phases is integral to identification of novel ways of interrupting the relevant disease pathways at molecular levels. Because acute and chronic GVHD are possible paradigms for a number of potentially related and more common skin diseases (e.g. erythema multiforme and Stevens-Johnson Syndrome, lichen planus, scleroderma), the understanding gained from the study of GVHD at clinical and experimental levels is likely to have implications that reach far beyond the population of patients receiving allogeneic bone marrow transplantation.

Acknowledgements
Dr. John L. Wagner was of particular assistance with the clinical details. Support for the assembly and reporting of much of the pathobiology-related data was assisted by federal funding in the form of R01, P01, and SDRC grants from the National Institutes of Health:

References
  1. Barnes DWH, Loutit JF. Spleen protection: The cellular hypothesis. In Bacq ZM, Radiobiology Symposium. Butterworth: London, 1955: 134-135.

  2. Gluckman E, Devergie A, Sohier J, Saurat J, Saurat JH. Graft-versus-host disease in recipients of syngeneic bone marrow. Lancet 1: 253-254, 1980.

  3. Tokime K, Isoda K, Yamanaka K, Mizutani H. A case of acute graft versus host disease following autologous peripheral blood stem cell transplantation. J Dermatol 27: 446-449, 2000.

  4. Greenbaum BH. Transfusion-associated graft-versus-host disease: Historical perspectives, incidence, and current use of irradiated blood products. J Clin Oncol 9: 1889-1902, 1991.

  5. Hull RJ, Bray RA, Hillyer C, Swerlick RA. Transfusion-associated chronic cutaneous graft-versus disease. J Amer Acad Dermatol 33: 327-332, 1995.

  6. Schmuth M, Vogel W, Weinlich G, Margreiter R, Fritsch P, Sepp N. Cutaneous lesions as the presenting sign of acute graft-versus-host disease following liver transplantation. Brit J Hematol 141: 779-780, 1999.

  7. Takatsuka H, Takemoto Y, Yamada S, Mori A, Wada H, Fujimori Y, Okamoto T, Kanamaru A, Kakishita E. Similarity between eruptions induced by sulfhydryl drugs and acute cutaneous graft-versus-host disease after bone marrow transplantation. Hematol 7: 55-57, 2002.

  8. Esteban JM, Somolo G. Skin biopsy in allogeneic and autologous bone marrow transplant patients: a histologic and immunohistochemical study and review of the literature. Mod Pathol 8: 59-64, 1995.

  9. Zhou Y, Barnett MJ, Rivers JK. Clinical significance of skin biopsies in the diagnosis and management of graft-versus-host disease in early post allogeneic bone marrow transplantation. Arch Dermatol 136: 717-721, 2000.

  10. Vavricka SR, halter J, Furrer K, Wolfensberger U, Schanz U. Contrast media triggering cutaneous graft-versus-host disease. Bone Marrow Trans 29: 899-901, 2002.

  11. Boström L, Ringdén O, Sundberg B, Ljungman P, Linde A, Nilsson B. Pretransplantation herpes virus serology and chronic graft-versus-host disease. Bone Marrow Trans 4: 547-552, 1989.

  12. Ljungman P, Niederwieser D, Pepe MS, Longton G, Storb R, Meyers JD. Cytomegalovirus infection after bone marrow transplantation for aplastic anemia. Bone Marrow Trans 6: 295-300, 1990.

  13. Villada G, Roujeau J, Cordonnier C, Bagot M, Kuentz M, Wechsler J, Vernant JP. Toxic epidermal necrolysis after bone marrow transplantation; study of nine cases. J Amer Acad Dermat 23: 870-875, 1990.

  14. Sullivan KM, Deeg HJ, Sanders J, Klosterman A, Amos D, Shulman H, Sale G, Martin P, Witherspoon R, Appelbaum F, et al. Hyperacute graft-versus-host disease in patients not given immunosuppression after allogeneic marrow transplantation. Blood 67: 1172-1175, 1986.

  15. Schubert MM, Sullivan KM, Morton TH, Izutsu KT, Peterson DE, Flournoy N, Truelove EL, Sale GE, Buckner CD, Storb R, et al. Oral Manifestations of chronic graft-versus-host disease. Arch Int Med 144: 1591-1595, 1984.

  16. Gratwohl AA, Moutsopoulous HM, Chused TM, Akizuki M, Wolf RO, Sweet JB, Deisseroth AB. Sjögren-type syndrome after allogeneic bone marrow transplantation. Ann Int Med 87: 703-706, 1977.

  17. Tichelli A, Duell T, Weiss M, Socie G, Ljungman P, Cohen A, vanLint M, Gratwohl A, Kolb HJ. Late onset keratoconjunctivitis sicca syndrome after bone marrow transplantation: Incidence and risk factors. Bone Marrow Trans 17: 1105-1111, 1996.

  18. Corson SL, Sullivan K, Batzer F, August C, Storb R, Thomas ED. Gynecologic manifestations of chronic graft-versus-host disease. Obst Gynecol 60: 448-492, 1992.

  19. Karni M, Kanda Y, Sasaki M, Takeda N, Tanaka Y, Saito T, Ogawa S, Honda H, Ohba S, Mitani K, Hirai H, Yazaki Y. Phimosis as a manifestation of chronic graft-versus-host disease after allogeneic bone marrow transplantation. Bone Marrow Trans 21: 721-728, 1998.

  20. Grigg AP, Underhill C, Russell J, Sale G. Peyronie's disease as a complication of chronic graft versus host disease. Hematology 7: 165-168, 2002.

  21. Gilliam A, Whitaker-Menezes, Korngold R, Murphy GF. Apoptosis is the predominant form of epithelial target cell injury in acute experimental graft-versus-host disease. J Invest Dermatol 1996; 107:377-383.

  22. Murphy GF, Whitaker D, Sprent J, Korngold R. Characterization of target injury of murine acute graft-versus-host disease directed to multiple minor histocompatibility antigens elicited by either CD4+ or CD8+ effector cells. Amer J Pathol 1991, 138:983-990.

  23. Ferrara J, Guillen FJ, Sleckman B, Burakoff SJ, Murphy GF. Cutaneous acute graft-versus-host disease to minor histocompatibility antigens in a murine model: histologic analysis and correlation to clinical disease. J Invest Dermatol. 1986; 86:371-375.

  24. Lerner KG, Kao GF, Storb R, Buckner CD, Clift RA, Thomas ED. Histopathology of graft-vs.-host reaction (GvHR) in human recipients of marrow from HL-A-matched sibling donors. Transplant Proc 1974;6:367-71

  25. Massi D, Franchi A, Pimpinelli N, Laszlo D, Bosi A, Santucci M. A reappraisal of the histopathologic criteria for the diagnosis of cutaneous allogeneic acute graft-vs-host disease. Am J Clin Pathol 112:791-800, 1999.

  26. Darmstadt GL, Donnenberg AD, Vogelsang GB, Farmer ER, Horn TD. Clinical, laboratory, and histopathologic indicators of the development of progressive acute graft-versus-host disease. J Invest Dermatol 99:397-402, 1992.

  27. Dilday BR, Smoller BR. Intracytoplasmic bile pigment in skin biopsy specimens for graft-versus-host disease versus erythema multiforme. Mod pathol 11:1005-1009, 1998.

  28. Osawa J, Kitamura K, Saito S, Ikezawa Z, Nakajima H. Immunohistochemical study of graft-versus-host reaction (GVHR)-type drug eruptions. J Dermatol 21:25-30, 1994.

  29. Hitchins L, Fucich LF, Freeman SM, Millikan LE, Marrogi AJ. Immunophenotyping as a diagnostic tool to differentiate lichen planus from chronic graft-versus-host disease: diagnostic observations on two patients. J Invest Med 45:463-468, 1997.

  30. Hayakawa S, Chishima F, Sakata H, Fujii K, Ohtani K, Kurashina K, Hayakawa J, Suzuki K, Nakabayashi H, Esumi M. A rapid molecular diagnosis of posttransfusion graft-versus-host disease by polymerase chain reaction. Transfusion 33:413-417, 1993.

  31. Messadi DV, Pober JS, Fiers W, Gimbrone MA, Murphy GF. Induction of an activation antigen on post-capillary venular endothelium in human skin organ culture. J Immunol. 1987; 139:1557-1562.

  32. Messadi DV, Pober JS, Murphy GF. Effects of recombinant gamma-interferon on HLA-DR and DQ expression by skin cells in short-term organ culture. Lab Invest. 1988; 58:61-67.

  33. Sviland L, Dickinson AM. A human skin explant model for predicting graft-versus-host disease following marrow transplantation. J Clin Pathol 52:910-913, 1999.

  34. Sviland L, Hromadnikova I, Sedlacek P, Cermakova M, Holler E, Eissner G, Schulz U, Kolb HJ, Jackson G, Wang XN, Dickinson AM. Histological correlation between different centers using the skin explant model to predict graft-versus-host disease following bone marrow transplantation. Human Immunol 62:1277-1281, 2001.

  35. Jarvis M, Schulz U, Dickinson AM, Sviland L, Jackson G, Konur A, Wang XN, Hromadnikova I, Kolb HJ, Eissner G, Holler E. The detection of apoptosis in a human in vitro skin explant assay for graft versus host reactions. J Clin Pathol 55:127-132, 2002.

  36. Christofidou-Solomidou M, Albelda SM, Bennett FC, Murphy GF. Experimental production and modulation of human cytotoxic dermatitis in human-murine chimeras Amer J Pathol 1997; 150:631-639.

  37. Sprent J, Schaefer M, Lo D, Korngold R. Properties of purified T cell subsets. II. In vivo responses to class I vs. class II H-2 differences. J Exp Med 163:998-1011, 1986.

  38. Korngold R, Sprent J. Variable capacity of L3T4+ T cells to cause lethal graft-vs-host disease across minor histocompatibility barriers in mice. J Exp Med 165:1552-1564, 1987.

  39. Murphy GF, Whitaker D, Sprent J, Korngold R. Characterization of target injury of murine acute graft-versus-host disease directed to multiple minor histocompatibility antigens elicited by either CD4+ or CD8+ effector cells. Am J Pathol 138:983-990, 1991.

  40. Korngold R: Immunobiology of graft-versus-host disease. Amer J Ped Hem/Onc 15:18-27, 1993.

  41. Townsend RM, Briggs C, Marini JC, Murphy GF, Korngold R. Inhibitory effect of a CD4-CDR3 peptide analog on graft-vs-host disease across a MHC-haploidentical barrier. Blood 1996; 88:3038-3047.

  42. Friedman TM, Statton D, Jones SC, Berger MA, Murphy GF, Korngold R. Vb spectratype analysis reveals heterogeneity of CD4+ T cell responses to minor histocompatibility antigens invoved in graft-versus-host disease: correlations with epithelial tissues infiltrate. Biol Blood Marrow Trans 2001; 7:2-13.

  43. Jones SC, Friedman TM, Murphy GF, Korngold RM. Specific donor Vb-associated CD4+ T-cell responses correlate with severe acute graft-versus-host disease directed to multiple minor histocompatibility antigens. Biol Blood Marrow Trans 2004; 10:91-105.

  44. Friedman TM, Jones SC, Statton D, Murphy GF, and Korngold R. Evolution of responding CD4+ and CD8+ T cell repertoires during the development of graft-versus-host disease directed to minor histocompatibility antigens. Biol Blood Marrow Trans 2004; 10:224-235.

  45. Schmaltz C, Alpdognan O, Muriglan SJ, Kappel BJ, Rotolo JA, Ricchetti ET, Murphy GF, Crawford JM, Van den Brink MRM. Donor T cell-derived TNF is required for graft-versus-host disease, graft-versus-tumor activity, and complete T cell chimerism after bone marrow transplantation. Blood 2003, 101:2440-2445.

  46. Korngold R, Marini JC, de Baca ME, Murphy GF, Giles-Komar, J. Role of tumor necrosis factor-α in graft-versus-host disease and graft-versus-leukemia responses. Biol Blood Marrow Trans 2003, 9:292-303.

  47. Klein LM, Lavker RM, Matis WL, Murphy GF. Degranulation of human mast cells induces an endothelial antigen central to leukocyte adhesion. Proc Natl Acad Sci USA 1989; 86:8972-8976.

  48. Walsh LJ, Trinchieri G, Waldorf HA, Whitaker D, Murphy GF. Human dermal mast cells contain and release tumor necrosis factor- a which induces endothelial leukocyte adhesion molecule-1. Proc Natl Acad Sci USA 1991; 88:4220-4224.

  49. Ioffreda MD, Murphy GF. Mast cell activation and leukocyte recruitment responses into skin sites: Role of cell adhesion molecules. In: Bochner BS, ed. Adhesion Molecules in Allergic Disease Marcel Dekker, New York 1997, pp. 257-278.

  50. Murphy GF, Sueki H, Teuscher C, Whitaker D, Korngold R. Role of mast cells in early epithelial target cell injury in experimental acute graft-vs-host disease. J Invest Dermatol 1994; 102:451-461.

  51. Korngold R, Jameson B, McDonnell JM, Leighton C, Sutton BJ, Gould HJ, Murphy GF. Peptide analogs which inhibit IgE-Fc eRIa interactions ameliorate the development of lethal graft-versus-host disease. Biol Blood Marrow Trans 1997; 3:187-193.

  52. Petrovic A, Alpdogan O, Willis L, Eng JM, Greenberg AS, Kappel BJ, Liu C, Murphy G, Heller G, Van den Brink MRM. LPAM (α4β7 integrin) is an important homing integrin on alloreactive T cells in the development of intestinal graft-versus-host disease. Blood (in press).

  53. Ferrara J, Guillen FJ, Sleckman B, Burakoff SJ, Murphy GF. Cutaneous acute graft-versus-host disease to minor histocompatibility antigens in a murine model: histologic analysis and correlation to clinical disease. J Invest Dermatol 1986; 86:371-375.

  54. Gilliam A, Whitaker-Menezes, Korngold R, Murphy GF. Apoptosis is the predominant form of epithelial target cell injury in acute experimental graft-versus-host disease. J Invest Dermatol 1996; 107:377-383.

  55. Whitaker-Menezes D, Jones SC, Friedman TM, Korngold R, Murphy GF. An epithelial target site in experimental graft-versus-host disease and cytokine-mediated cytotoxicity is defined by cytokeratin 15 expression. Biol Blood Marrow Trans 2003, 9:559-570.

  56. Yan, H-C, Juhasz I, Pilewski J, Murphy GF, Herlyn M, Albelda SM. Human/SCID mouse chimeras: An experimental in vivo model system to study the regulation of human endothelial cell-leukocyte adhesion molecules. J Clinical Invest. 1993; 91:986-996.

  57. Schmaltz C, Alpdogan O, Kappel BJ, Muriglan SJ, Rotolo JA, Ongchin J, Willis LM, Greenberg AS, Eng JM, Crawford JM, Murphy GF, Yagita , Wlaczak H, Peschon JJ, van den Brink MRM. T cells require TRAIL for optimal graft-versus-host activity. Nature Med 8:1433-1437, 2002.