Oral graft-versus-host disease

Abstract

OBJECTIVE:

Graft-versus-host disease (GVHD) is a leading cause of morbidity and mortality in patients receiving hematopoietic cell transplant. It is estimated that 40–70% of engrafted patients surviving the initial transplant eventually develop chronic GVHD (cGVHD), which can persist for months to years and require long-term management from multiple disciplines. This review describes the oral component of this transplant complication.

DESIGN:

The search related to GVHD patho-biology, salivary gland disease after hematopoietic cell transplant and treatments for oral GVHD encompassed literature from 1966 through 2008. Searches were limited to the MEDLINE/PubMed database and English language literature in peer-reviewed journals.

RESULTS:

Our understanding of the patho-biology of oral cGVHD is based on studies of other affected tissues. It is difficult to determine the prevalence and incidence of salivary gland disease after transplant because there is no universally accepted case definition. In general, clinical trials for treatment of oral cGVHD have been too small to make strong recommendations for use in clinical practice.

CONCLUSIONS:

Larger well-designed clinical studies are needed to understand the patho-biology of oral cGVHD and determine best treatments for this disease.

Introduction

The increasing success of allogeneic hematopoietic stem cell transplantation (thereafter referred to as HSCT) for treatment of neoplastic and non-neoplastic hematologic diseases has contributed to its steady increase in use (Flowers et al, 2002; Laughlin et al, 2004; Thomas et al, 2004; Appelbaum, 2007; O'Keefe et al, 2007). Worldwide over 40 000 individuals receive HSCT annually, out of which 15 000 are performed with cells from an allogeneic donor (Flowers et al, 2002). After the discovery that grafted allogeneic hematopoietic cells could eliminate residual tumor cells of the recipient via an immune mediated mechanism (graft-versus-tumor effect), HSCT became an experimental treatment for other malignancies. On-going clinical trials are testing HSCT as an adjunctive therapy for solid tumors such as renal cell carcinoma (Ueno and Childs, 2007), and also as a new treatment for severe autoimmune diseases (Pavletic and Illei, 2005). These trends may significantly increase the annual number of transplanted patients in the near future.

Graft-versus-host disease (GVHD) is a principal impediment for broader use of HSCT and is a leading cause of morbidity and mortality in HSCT recipients. It is more likely to occur if the host receives a graft from an unrelated donor, or if the host or donor is older. Both acute and chronic phases of this complication develop, often involving multiple organ systems. Acute GVHD has relatively uniform clinical picture, classically manifested by erythematous rash, diarrhea and/or liver involvement, generally occurs early post-transplant (most commonly within the first 3–4 months) and is the major cause of early lethality. Chronic GVHD (cGVHD) is a distinct syndrome that can affect virtually every major organ system but most commonly involves skin, oral, vaginal and conjunctival mucosa, salivary and lacrimal glands and the liver. It is estimated that 40–70% of engrafted patients surviving the initial transplant eventually develop cGVHD, which can persist for months to years and require long-term management from multiple disciplines (Fraser et al, 2006). This adds significantly to the burden of disease in patients, who often have had prolonged illness before undergoing transplant. cGVHD is the leading cause of death in long-term survivors of transplant, even though it is associated with a decreased risk of relapse in those receiving HSCT for leukemia (Fraser et al, 2006).

Given that the use of HSCT will increase with time, that there will be an undersupply of related donors, and that related donors and transplant recipients will age as the population ages, the incidence of cGVHD will probably increase in coming years. As the oral component of this transplant complication is a source of significant morbidity, the oral health research community is uniquely positioned to help define the pathogenesis and management of this important disease. This review concentrates on the known patho-physiology of oral GVHD, its histopathology and clinical features, GVHD salivary gland disease and clinical trials for treatment of oral GVHD.

Patho-physiology

Acute and cGVHD were traditionally distinguished by time of onset following transplant. However, newer consensus criteria define each by their characteristic clinical and pathologic features rather than chronologically (Filipovich et al, 2005). Acute GVHD often occurs within 100 days of HSCT, but can persist beyond that time or recur (Table 1). Classic features of acute GVHD include erythematous maculo-papular rash, diarrhea and elevation of liver enzymes. The clinical presentation of cGVHD is variable and may include lichen planus-like changes of skin and mucosa, sclerosis of the skin, sicca syndrome secondary to lacrimal and/or salivary gland damage, liver involvement with cholestatis, and decrease in pulmonary function secondary to bronchiolitis obliterans.

Table 1.

Predominant features of acute and chronic graft-versus-host disease (Filipovich et al, 2005)

Site	More common in Acute	Chronic
Skin	Erythema	Poikiloderma
	Maculopapular rash	Lichen planus-like features
	Pruritus	Sclerotic features; Morphea-like features, Lichen sclerosus-like features
		Often areas of depigmentation, hypopigmentation or hyperpigmentation
Nails		Dystrophy; longitudinal ridging, splitting, or brittle nails
		Onycholysis; pterygium unguis; nail loss
Hair		Alopecia (after recovery from induction chemotherapy or radiotherapy); scaling, papulosquamous lesions
Oral	Mucositis	Lichen-planus like features
	Erythema	Xerostomia
	Pain	Hyperkeratotic plaques
		Mucocele
		Restriction of mouth opening from sclerosis
		Mucosal atrophy
		Pseudomembranes and ulcers
Eyes		New onset dry, gritty, or painful eyes
		Cicatricial conjunctivitis
		Keratoconjunctivitis sicca
Gastro-intestinal	Anorexia; nausea; vomiting	Esophageal web
	Diarrhea; weight loss; failure to thrive	Strictures or stenosis in the upper to mid third of the esophagus
Genitalia		Lichen-planus like features with possible scarring
Liver	Elevation of total bilirubin, alkaline phosphatase, alanine aminotransferase or aspartate aminotransferase to >2 times the upper limit of normal with no other cause	Same features as acute
Lung		Bronchiolitis obliterans
Muscles, joints		Fasciitis; joint stiffness or contractures secondary to sclerosis, sometimes myositis or polymyositis
Other		Thrombocytopenia; eosinophilia; lymphopenia; hypo- or hyper-gammopathy; autoantibodies

Billingham first proposed basic conditions necessary for development of GVHD (Billingham, 1966): (i) immunocompetent cells in the graft; (ii) the inability of the host cells to reject the graft (immune compromise); and (iii) the ability of the infused cells to recognize the host as foreign (antigenic mismatch). Each element plays a critical role in the development of both acute and cGVHD.

Acute GVHD

The acute GVHD occurs in three closely overlapping and interrelated stages: conditioning, activation and expansion of alloreactive cells and the effector phase (Ferrara and Reddy, 2006; Shlomchik, 2007; Sun et al, 2007).

Conditioning

Antigenic differences between the donor and the host will lead to graft rejection unless the donor cells in the graft are given advantage over host cells. Such advantage is achieved through depletion of the host immune cells by irradiation and/or chemotherapy with immunosuppressive conditioning. More aggressive myeloablative conditioning is commonly associated with higher risk and increased severity of GVHD in contrast to gentler, non-myeloablative regimens (Couriel et al, 2004; Saliba et al, 2007). Conditioning is accomplished with chemotherapeutic agents and/or total body irradiation, with the goal to deplete the host immune system and allow successful engraftment of donor stem cells. However, increasing degrees of conditioning are associated with progressive cellular damage, particularly in the gastrointestinal (GI) tract and skin, leading to increased permeability of epithelial barriers, leakage of microbial products such as lipopolysaccharide, and release of inflammatory cytokines (Cooke et al, 2001). This ultimately leads to activation of the host antigen presenting cells (APCs), characterized by increased expression of co-stimulatory molecules and migration to the secondary lymphoid organs. The role of microbial products in this stage of GVHD initiation is underscored by the fact that germ-free animals have a much lower incidence of GVHD (Heidt and Vossen, 1992). Similarly, prior gut decontamination was effective in decreasing the risk of GVHD in clinical trials (Beelen et al, 1999).

Activation and expansion of alloreactive cells

As T-cell numbers are primarily limited by the availability of homeostatic cytokines, most notably interleukin (IL)-7, conditioning creates ‘space’ for newly infused donor T cells (Alpdogan et al, 2003). These immuno-competent cells are critical to the development of GVHD, as evidenced by the finding that GVHD can be abrogated effectively or eliminated through T-cell depletion of the graft. However, grafted T cells are vital for transplant success, as they provide adaptive immunity to infections and control the malignancy (the primary reason for allogeneic transplantation). Past clinical studies established that patients who receive T-cell depleted grafts were at a much greater risk for graft rejection, serious infections and relapse of the malignancy (Marmont et al, 1991).

Following migration to secondary lymphoid organs, host APCs encounter donor lymphocytes (Ferrara and Reddy, 2006). Naïve donor T lymphocytes recognize host antigens on the surface of APCs, activate, proliferate and express effector cytokines. The number of T cells that are ultimately activated depends on the prior frequency of the T cells specific for host antigens (precursor frequency). In major histocompatibility antigen (MHC) mismatched transplants, large numbers of donor T cells recognize host HLA antigens as foreign, leading to increased expansion and ultimately more severe GVHD. The degree of antigenic mismatch is one of the most critical factors in development of GVHD, and is the primary limiting factor preventing wider use of allogeneic HSCT (Sasazuki et al, 1998). The incidence and severity of GVHD increase with the degree of MHC mismatch with the best results obtained when matched siblings are used as donors. In MHC-matched transplants (such as those from matched siblings), MHC molecules are identical between the donor and the host, and antigen recognition is dependent on the differences in minor histocompatibility antigens. These are peptides derived from any number of polymorphic genes throughout the genome that differ between the donor and recipient. Peptides generated within the APC through proteosomal processing of the endogenous cell proteins are presented to donor CD8 cells with MHC-I molecules. In addition, antigens that are taken up by APC from the surrounding milieu are processed in the phagolysosomal pathway and are presented to donor CD4 cells in complex with MHC-II molecules (Goulmy, 1996).

Effector phase

Following activation, donor T cells acquire an effector phenotype, exit secondary lymphoid organs, enter the blood stream and migrate to the host target tissues (Ferrara and Reddy, 2006).

Activation of CD8 cells is accompanied by increased production of cytotoxic effector molecules such as perforin and granzymes and secretion of cytokines such as interferon-gamma (IFN-γ). Effector function of CD4 cells is primarily mediated through secretion of cytokines. Depending on the conditions of initial activation and subsequent cytokine profile, CD4 cells can be roughly subdivided into Th1, Th2, and more recently Th17 subtypes characterized by secretion of ‘signature cytokines’ of IFN-γ, IL-4, and IL-17, respectively (Harrington et al, 2005). In addition to acquisition of effector functions, donor T cells acquire specific migratory properties depending on the origin of the APCs that mediated the initial activation. Thus, donor T cells activated by lamina propria APCs in the mesenteric lymph nodes migrate preferentially into the gut and cause GI GVHD, while those activated in the skin draining lymph nodes will migrate to skin. This specificity is determined by the unique pattern of chemokine receptor expression on T cells that direct migration to the tissues expressing corresponding chemokines (Wysocki et al, 2005). Once the effector T cells reach the target organs, they cause tissue damage via direct cytotoxicity against epithelia and release of cytokines such as IFN-γ. This subsequently activates resident macrophages to release proinflammatory molecules such as tumor necrosis factor-alpha (TNF-α), IL-1 and IL-6 leading to further attraction of proinflammatory cells and amplification of the inflammatory cascade (Reddy and Ferrara, 2003).

Chronic GVHD

Graft-versus-host disease occurring or lasting beyond day 100 post-transplant has been defined classically as chronic. Although practical for the epidemiologic purposes, such definition is inadequate for the following reasons:

1. Symptoms and signs typical of cGVHD frequently develop before day 100.

2. Signs classically associated with acute GVHD, e.g. diarrhea or diffuse erythematous rash, may occur after day 100, particularly in the setting of donor lymphocyte infusion (DLI).

Therefore, the latest NIH consensus criteria recommend classification based on characteristic symptoms and signs rather than a rigid temporal definition (Table 1), (Filipovich et al, 2005).

Chronic GVHD is the leading long-term complication of allogeneic stem cell transplantation. While there have been significant advances in prevention and treatment of acute GVHD, the pathogenesis of cGVHD is poorly understood. Several factors could be responsible for this:

1. Most animal studies use major mismatched strain combinations and focus on acute GVHD with short-term mortality and weight loss as the primary outcomes.

2. Although some animal models of cGVHD have been proposed, none adequately reproduces the complexity of human condition.

3. Chronic GVHD develops relatively late after transplant, necessitating prolonged follow-up and data collection. This makes prospective studies difficult and expensive to perform.

4. Almost any organ system can be affected by cGVHD, leading to a wide variety of symptoms and signs.

5. Until recently, a universally accepted system of diagnosis, staging and response criteria in cGVHD has not been available, preventing interpretation of clinical studies of different centers.

6. Most human studies focused on describing easily accessible cell populations (i.e. those from the peripheral blood) which may not accurately reflect events occurring in target organs.

In the following section, we will summarize and discuss the evidence available related to patho-biology from animal and human studies to date.

Risk factors

Prior diagnosis of acute GVHD is the risk factor associated most consistently with subsequent cGVHD (Ferrara and Reddy, 2006). If acute GVHD is not fatal, it will either resolve or develop into cGVHD (progressive onset). cGVHD may also develop in patients without clinical acute GVHD (de novo onset) or following complete resolution of acute GVHD (quiescent onset). Other risk factors for cGVHD are increasing donor and recipient age, increasing CD3 (T cell) dose in the graft, female donor and male recipient combination, unrelated donors, mismatched HLA donors, diagnosis of chronic myelogenous leukemia or myelodysplastic syndrome, total body irradiation, and the use of mobilized peripheral blood stem cell transplant (Przepiorka et al, 2001; Pavletic et al, 2005).

Clinical similarity to autoimmunity

In contrast to acute GVHD, cGVHD affects multiple target organs and produces a constellation of clinical manifestations (Filipovich et al, 2005; Ferrara and Reddy, 2006). In many cases, these manifestations closely resemble both clinically and histologically those of common autoimmune disorders including lichen planus, Sjögren's syndrome, scleroderma, systemic lupus erythematosus, dermatomyositis and primary biliary cirrhosis (Filipovich et al, 2005; Baird and Pavletic, 2006). This prompted the suggestion that cGVHD is the disorder of disregulated immunity similar to that in autoimmunity. Unlike classical autoimmune disorders in which tolerance to native antigens is lost and the immune system recognizes and mounts a detrimental immune response to host antigens, donor T cells recognize antigens of the host and perpetuate the disease process in cGVHD. To better understand the potential mechanisms of development and progression of cGVHD, it is important to review basic aspects of mechanisms that ensure tolerance in the healthy individual and contrast them with the setting of allo-HSCT.

Central vs peripheral tolerance

The concept of tolerance is fundamental to the normal immune response (Mak, 2006). The development of T cells occurs in the thymus through the closely linked processes of positive and negative selection. Positive selection ensures that the developing T cell is able to recognize the host MHC molecules, either MHC-I or MHC-II for CD8 and CD4 cells, respectively. As T-cell receptor rearrangement is a random process, it is important to ensure that naïve T cells do not initiate responses to the host antigens. Host proteins are expressed in the thymus on the epithelial and dendritic cells (DCs) and the T cells whose receptors recognize host MHC with high affinity are clonally deleted (negative selection). Only cells which recognize the host MHC with low affinity are allowed to mature and exit to the periphery. This process is fundamental to development of central tolerance. However, despite efficiency of clonal deletion in elimination of the overwhelming majority of autoreactive T cells, some escape the thymus and are thought to be controlled in the periphery, through such mechanisms as anergy (loss of ability to proliferate upon encounter with harmless antigens) and active suppression by regulatory cells (Schwartz, 2003; Kyewski and Klein, 2006).

Regulatory T cells

The concept of regulatory cells has received extraordinary attention in the recent years following the report describing profound autoimmunity developing in animals following neonatal thymectomy. It was later shown that this could be prevented by transfusion of a particular T-cell type expressing high levels of IL-2 receptor α-chain (CD25). These cells were termed natural regulatory T cells and were shown to develop in the thymus along with conventional T cells. Natural regulatory T cells characteristically express transcription factor FoxP3 and recognize a large number of host antigens. Following exit from the thymus, they co-localize with effector T cells in secondary lymphoid organs and target tissues and control immune responses through mechanisms that may involve secretion of immunosuppressive cytokines such as transforming growth factor-beta (TGF-β) and IL-10, contact dependent inhibition, apoptosis of effector T cells, and competition for IL-2. In addition to ‘natural’ regulatory T cells that develop and acquire their function in the thymus, peripheral regulatory T cells developing from conventional T cells under certain conditions (e.g. high TGF-β or IL-10 levels) have also been described (Sakaguchi, 2004; Grazia Roncarolo et al, 2006; Roncarolo and Battaglia, 2007).

T-cell development following allogeneic HSCT

Having reviewed the events of normal T-cell development, we will examine factors that influence the T-cell repertoire in recipients of allo-HSCT. Following successful engraftment, donor stem cells begin producing myeloid and lymphoid progenitors that repopulate the host (Ferrara and Reddy, 2006). Unlike B cells which develop in the bone marrow, T cells require the presence of a functional thymus for development. Most allo-HCST is performed in older adults when thymic function is impaired and the development of naïve T cells is greatly compromised. In addition, the thymus and its epithelial components critical for proper T-cell development are thought to be targets in acute GVHD. Therefore, most of the T cells that eventually repopulate the host have developed in the donor thymus. As both negative selection and generation of regulatory T cells are dependent on the nature of the antigen presentation in the thymus, it is likely that failure of central tolerance contributes to the development of the autoimmune-like manifestations of cGVHD. Empirical evidence in support of this comes from several sources:

1. The ability of the adult thymus to support renewed thymopoiesis is age dependent, with older recipients having less increase in thymic size and naïve T-cell generation post-transplant.

2. The incidence of cGVHD increases with increasing age of the recipient. In contrast, young children with functional thymus have a lower incidence of cGVHD.

Therefore, the ultimate T-cell repertoire in the older transplant recipients is primarily determined by the initial T-cell composition of the graft. If the T-cell repertoire contains cells recognizing host antigens but not the natural regulatory T cells that can counteract them, this situation could promote the development of persistent allo-reactivity and cGVHD. Similarly, even if thymopoiesis is preserved post-transplant, the damage inflicted upon the host thymus during acute GVHD may be sufficient to impair mechanisms of central tolerance and thymic generation of regulatory T cells (Fukushi et al, 1990; Hakim and Gress, 2002).

Dendritic cells

Following the initial period of mixed chimerism post-transplant, donor cells replace all hematopoietic compartments (Ferrara and Reddy, 2006). Donor DC precursors emerge from the bone marrow and travel to the peripheral tissues where they mature into fully functional DCs. As both DCs and T cells are now of donor origin, how does the induction of allo-reactivity occur? In the MHC-matched setting, donor CD8 cells should not be able to respond to host antigens as minor histocompatibility antigens recognized by CD8 cells in the context of MHC-I molecules originate from proteins produced within the donor DC itself. In contrast, host molecules can be taken up from the environment by donor MHC, processed in the MHC class II pathway and presented to donor CD4 cells. However, the involvement of CD8 cells in cGVHD is well documented. One possibility is that cells induced at the time of acute GVHD by host APCs persist as memory cells and mediate cGVHD without involvement of donor DCs. Another possibility is explained by the phenomenon of cross-presentation – redirection of exogenous antigens from MHC-II (vacuolar) into the MHC-I (cytoplasmic) pathway with subsequent presentation to CD8 cells (Shlomchik, 2003).

T cells

Several studies have examined the relative roles of various T-cell subsets in development of cGVHD. Depending on the murine model used, both CD4 and CD8 cells have been shown to be critical for the induction of cGVHD. In the human setting, several studies have examined the dynamics of T-cell populations. Yamashita et al (2004) evaluated T-cell subsets in the peripheral blood of patients with cGVHD and found that patients with severe disease had increased numbers of effector-memory CD4 T cells. In a subsequent study, they found that the number of effector-memory cells decreases in those patients who respond to extracorporeal therapy, a treatment for GVHD (Yamashita et al, 2006).

In recent animal studies, Chen et al (2007) demonstrated that Th1 and Th17 cells formed early post-transplant persist and mediate cGVHD immunopathology. They also suggest that the relative lack of regulatory T cells is crucial for the development of this process (Chen et al, 2007). Few human studies have examined the role of regulatory T cells in cGVHD. Clark et al found increased frequency of CD4⁺CD25⁺ T cells in the peripheral blood of cGVHD patients and suggested that no numerical deficiency of regulatory cells is responsible for development of cGVHD. However, although CD25 is a relatively reliable marker of regulatory T cells in mice, it is expressed by activated cells and may represent an effector population (Clark et al, 2004). In contrast, a later study found decrease in the regulatory T cells in the cGVHD population (Zorn et al, 2005). The only study published to date that examined levels of regulatory T cells in tissues as well as the peripheral blood found that the ratio of regulatory T cells (defined by FoxP3 expression) relative to CD8 cells similar in cGVHD and normal controls, but decreased compared to patients post-transplant with GI infections (Rieger et al, 2006).

B cells

Good evidence of B-cell involvement in cGVHD comes from the trials using rituximab, a monoclonal antibody (anti-CD20 antibody) that binds subsets of B cells. Following initial reports (Ratanatharathorn et al, 2000, 2003), recent phase II trials demonstrated that rituximab was moderately efficacious in reducing cGVHD, including in steroid-refractory patients (Cutler et al, 2006; Okamoto et al, 2006; Zaja et al, 2007).

Initial evidence for a potential role of autoantibodies in the pathogenesis of cGVHD in humans was provided by a study of antibodies against Y chromosome encoded antigen, H-Y. The presence of H-Y antibodies correlated with cGVHD and maintenance of disease remission (Miklos et al, 2005). Although numerous other autoantibodies have been described in cGVHD, including those against double stranded DNA, other nuclear components, mitochondria and rheumatoid factor, their significance has not been established (Wechalekar et al, 2005; Patriarca et al, 2006). Recently, activating antibodies to platelet derived growth factor (PDGF) receptor have been described both in scleroderma and in chronic sclerodermatous GVHD and may be an important contributor to the development of fibrosis in these conditions (Svegliati et al, 2007).

BAFF (BLyS) is a TNF family cytokine produced by stromal and other cells of non-hematopoietic origin and active B cells. BAFF promotes survival of marginal zone B cells producing low affinity antibodies – the subset that has been linked to autoantibody production. Increased BAFF levels have been shown in several classic autoimmune diseases, in particular systemic lupus erythematosus (Mackay et al, 2007). Recently, elevated BAFF levels have been demonstrated in patients with active cGVHD (Sarantopoulos et al, 2007).

In conclusion, the pathogenesis of cGVHD is complex and poorly understood, but is likely to involve dysfunction of tolerance determining mechanisms similar to classic autoimmune diseases. Further hypothesis driven prospective studies in humans focusing on target tissue events in conjunction with clinical trials are most likely to build our understanding of this challenging condition.

Oral findings in the timeframe of acute GVHD

Mucosal erythema, ulcerations, and painful desquamative oral lesions occur often in patients undergoing conditioning for HSCT. In one study of 150 patients transplanted for acute myelocytic leukemia, stomatitis developed in 26% during the engraftment period (Neudorf et al, 2004). However, a true clinical case definition of oral acute GVHD is lacking, as several factors, including conditioning chemotherapy, concurrent radiation, neutropenia, herpes simplex infection (HSV), treatment with cytokines such as granulocyte-macrophage colony-stimulating factor (GM-CSF), and infusion of donor hematopoietic cells contribute to oral lesion development during the first 28 days following transplant. Oral mucositis induced by chemotherapy in a non-transplanted cancer patient presents clinically as erythema progressing to ulceration of non-keratinized oral tissues. Ulcers often occur within 2 weeks of chemotherapy initiation, healing 2–4 weeks after chemotherapy ceases (Lalla et al, 2008). Palifermin (recombinant human-keratinocyte growth factor), a drug that reduces chemotherapy and radiation induced mucositis, reduces the severity and duration of oral mucositis in patients undergoing autologous HSCT and allo-HSCT patients conditioned with cyclophosphamide and fractionated total-body irradiation (Cy/TBI), suggesting conditioning contributes to early oral lesions (Spielberger et al, 2004; Blazar et al, 2006). Therefore, some investigators feel that oral acute GVHD is not a distinct entity.

While infectious agents can contribute to oral lesions in transplanted patients, most patients receive prophylactic anti-fungal and anti-viral agents. Anti-viral prophylaxis with agents such as acyclovir, given to prevent HSV and cytomegalovirus (CMV) reactivation after transplant, reduces viral shedding and herpetic lesions in most transplanted individuals. Therefore, the contribution of HSV and other herpetic viruses to oral lesions in the first 100 days post-transplant is most likely small (Epstein et al, 1996). However, reactivation of herpetic viruses is still possible in patients receiving this prophylaxis. In a randomized, double-blind study of 618 subjects comparing oral acyclovir and oral valacyclovir to prevent CMV reactivation, seven taking acyclovir and five taking valacyclovir developed herpetic lesions. A clinical diagnosis of herpes zoster was made in four (two per group) (Ljungman et al, 2002).

Oral presentation of cGVHD

Classic features of oral cGVHD include lichenoid changes, ulcerations and mucosal atrophy (Figure 1, panel a), salivary gland dysfunction, and restricted oral opening (Schubert and Sullivan, 1990; Woo et al, 1997; Filipovich et al, 2005). Separately, these presentations mirror several oral autoimmune conditions. The mucosal lesions are very similar to those found in oral lichen planus (OLP, Figure 1, panel b), the salivary gland infiltrates mimic those found in Sjögren's syndrome, and the fibrosis and restricted oral range of motion suggest scleroderma. Clinical findings proposed to be most diagnostic of oral cGVHD are lichenoid changes, white patches, hyperkeratotic plaques, or restricted oral range of motion in patients with sclerotic features of skin GVHD (Filipovich et al, 2005).

Figure 1.

Similarities in clinical picture of oral cGVHD and oral lichen planus. Ulcerative and erythematous lesions of the buccal mucosa in oral cGVHD (a) and lichen planus (b)

In accordance with the new NIH staging criteria, the extent of mucosal lesions can be documented using a scoring system that considers the severity and extent of the most common findings. The score can be totaled to generate a cumulative score of severity for use in clinical studies (Pavletic et al, 2006).

Oral mucosal lesions

Mucosal lesions can be a source of significant pain, and when extensive, can limit nutritional intake and reduce the patient's ability to maintain oral hygiene. Almost any oral site may be involved, with buccal mucosa and tongue among most commonly affected (Treister et al, 2008). Clinically, these lesions exhibit erythema and white striae (hyperkeratotic plaques) and have been described as reticular, erythematous, atrophic and ulcerative (Woo et al, 1997; Filipovich et al, 2005; Imanguli et al, 2006; Sari et al, 2007). Extensive ulcerative lesions may be covered with a pseudomembrane.

Several clinical studies have reported that oral lesions are common in patients with cGVHD, estimated to occur in 45–83% of patients (Treister et al, 2005; Schubert and Correa, 2008). In one longitudinal study, the oral mucosa was the second most commonly affected site in those who developed cGVHD (Flowers et al, 2002). In another study of 101 patients, 51% of surviving patients developed oral cGVHD sometime during the 3 years (Mohty et al, 2002). A similar 3-year cumulative incidence of oral lesions was found in the cohort of 126 patients followed by Flowers et al (2002), with just under 50% of patients developing oral cGVHD sometime during the 3 years.

Mucoceles

Superficial mucoceles are subepithelial extravasations of sialomucin that occur at the epithelial-connective tissue interface and appear to be directly related to minor salivary glands. Clinically, the presentation is a fluid-filled, smooth elevation of the epithelium surrounding the duct of the minor salivary gland. Mucoceles develop when the duct is physically occluded, forcing the saliva into the surrounding tissues. The current belief is that salivary gland inflammation in GVHD, coupled with decreased salivary fluid secretion and viscous saliva, blocks excretory ducts (Garcia et al, 2002; Filipovich et al, 2005).

Fibrosis/limited opening

Continued inflammation and scarring of the buccal mucosa can restrict oral opening to an extent that limits a patient's ability to perform oral hygiene, and increase morbidity by decreasing food intake. This can eventually lead to weight loss (Schubert and Correa, 2008).

Oral squamous cell carcinoma

A rare, but serious complication associated with chronic oral mucosal GVHD is oral squamous cell carcinoma. Squamous cell carcinoma, often in the oral cavity, is one of the most commonly diagnosed secondary cancers in transplanted patients. This has been observed in multiple, multi-national cohorts of persons surviving BMT, (Kolb et al, 1999; Curtis et al, 2005; Shimada et al, 2005). In multivariate analysis, extensive cGVHD and older age at the time of transplantation were associated with a higher risk for squamous cell cancers (Shimada et al, 2005), but it is not clear whether GVHD itself or its treatment is most strongly associated with tumor development.

Histopathology

Most oral mucosal tissues as well as minor salivary glands can be affected by GVHD. Labial (minor) salivary gland involvement occurs more frequently than oral mucosa, possibly due to the higher tissue expression of histocompatibility antigens by salivary tissues and the accessibility of the glands to pathogenic lymphocytes (Soares et al, 2005). Salivary gland histopathology is described in detail below.

It should be noted that large studies examining the histopathologic features of acute GVHD have not been performed in humans, as biopsies are often contraindicated immediately post-HSCT due to the high infectious risk and risk of other complications (Woo et al, 1997). Therefore, most histopathologic knowledge of acute GVHD comes from skin biopsies. Studies that examined the development of acute GVHD skin lesions over time identified the first changes as basal cell layer cytoplasmic vacuolar alterations with perivascular inflammation of the vessels in the superficial plexus (Farmer, 1985). The lymphocytic infiltrate subsequently moves toward the epithelium and is associated with spongiosis. Dyskeratotic cells then appear, subepidermal clefts develop and the epithelium may separate from the connective tissue (Farmer, 1985; Woo et al, 1997).

Biopsy to confirm the clinical diagnosis of active cGVHD is recommended when alternative diagnoses are possible, when no diagnostic clinical features of cGVHD are present, when prior changes might be altering clinical judgment or when only internal organs exhibit clinical signs of active cGVHD. Along with those indications, whenever drug toxicity is suspected and infections with atypical clinical features are present, biopsy is essential to establish the correct diagnosis (Nash et al, 2006).

Histologic interpretation of cGVHD is not always straightforward and can be altered by a large number of factors. Inflammatory activity can be suppressed by immunosuppressive medications (Woo et al, 1997; Nash et al, 2006), drug reactions can mimic GVHD, the tissue sample may not be representative of the involved site, and prior tissue damage can be difficult to separate from new GVHD activity.

Minimal histologic criteria for oral mucosal cGVHD are localized or generalized epithelial changes consisting of lichenoid interface inflammation, exocytosis, and apoptosis (Figure 2) (Shulman et al, 2006). Histopathologic features include hydropic degeneration of the basal cells, interspersed areas of hyperparakeratosis or atrophy, subepithelial clefting, and apoptotic spinal layer cells with pyknotic nuclei (Soares et al, 2005). The connective tissue is characterized by variable amounts of perivascular inflammation and lymphocytic infiltration that is altered by the use of immunosuppressive medications, with occasional separation of the epithelium from the connective tissue (Woo et al, 1997; Soares et al, 2005; Lew and Smith, 2007). Immunohistochemical studies have demonstrated that the infiltrate is predominately lymphocytes and macrophages (Figure 3). The reported ratio of CD8 T cells to CD4 T cells, however, has differed among studies, with some finding a slight predominance of CD8 cells (Soares et al, 2005) and others a predominance of CD4 cells (Nakamura et al, 1996). The increased number of macrophages suggests an important role for this cell type in oral cGVHD (Soares et al, 2005).

Figure 2.

Histopathologic changes of oral cGVHD. Note “dyskeratotic’ (apoptotic) keratinocyte and adjacent lymphocyte (small arrow) and subepithelial mononuclear infiltrate (large arrow).

Figure 3.

Immunohistology of oral cGVHD. Immunofluorescent staining and confocal imaging for CD3 (red, a), CD8 (cyan, b), and CD45RO, a marker for effector-memory T cells (green, c). Nuclei are stained with DAPI, blue. Note predominance of CD8 cells in the infiltrate (overlap, d). Line shows the approximate position of the basement membrane.

Salivary gland changes

Initial observations

In 1977, a paper was published describing clinical features of a 25-year-old male who received a BMT for acute myelogenous leukemia (Lawley et al, 1977).

Approximately 8 months following transplant, he presented with features of cutaneous scleroderma, xerophthalmia, and xerostomia. Results of the Schirmer's tear test, which quantifies tear production, and the minor salivary gland biopsy suggested that the patient had Sjögren's syndrome. That same year, another case series described four patients with complaints of xerophthalmia and xerostomia who had received allogeneic BMT and subsequently developed cGVHD. Three patients had keratoconjunctivitis sicca (KCS) similar to that found in patients with Sjögren's syndrome, and all four had significant minor salivary gland lymphocytic infiltration. Other features of various autoimmune diseases were present, including sclerodermatous skin and discoid lupus erythematosus-like lesions (Gratwhol et al, 1977).

Prevalence and risk factors for GHVD salivary gland disease and Sjogren's-like syndrome

Since the original descriptions, multiple reports of a Sjögren-type syndrome in patients with GVHD have been published (Izutsu et al, 1983a,b; Janin-Mercier et al, 1987; Lindahl et al, 1988; Rouquette-Gally et al, 1988; Nakhleh et al, 1989). However, as different case definitions have been used to define GVHD salivary gland disease, it is difficult to determine the prevalence and time course of salivary gland pathology after transplant (Table 2). In several studies, salivary gland infiltration was more common if patients had oral mucosal GVHD. Other factors, such as conditioning regimens, transplant cell source, GVHD prophylaxis protocols and patient age, may influence the development of oral and ocular lymphocytic infiltration post-transplant. In particular, data suggest that external beam total body irradiation during conditioning is a risk factor for GVHD salivary gland disease (Heimdahl et al, 1985; Dahllof et al, 1988; Brattstrom et al, 1991; Dens et al, 1996). Other studies suggest that autoimmune markers such as serum anti-nuclear antibodies are associated with GVHD salivary gland disease (Gratwhol et al, 1977; Rouquette-Gally et al, 1988), and salivary flow rate assessments across time suggest that there is an initial depression of flow after conditioning that may recover with time (Dens et al, 1996).

Table 2.

Histopathologic studies of minor glands and clinical findings of GVHD salivary gland disease

Author, year	Patients	Cells	Time post-transplant	Salivary pathology criteria	Biopsy controls	Subjective findings	Salivary flow findings	Scintigraphy	Summary findings
Lawley et al, 1977	1 male with GVHD	BM		Greenspan and Daniels	None	Patient had oral dryness	↓ Stimulated parotid flow	No	Patient also had features of scleroderma
Gratwhol et al, 1977	4 with GVHD	BM	14–19 months	Talal	2 autologous BMT patients	1 of 6 had dry mouth; 3 of 6 had dry eyes	Two with GVHD had ↓ parotid flows	1 of 5 was abnormal	Lymphocytic infiltrates and acinar atrophy in 4/4 withGVHDand 0/2 controls
Shulman et al, 1980	20 patients with GVHD	BM	13–365 days	Qualitative assessment	32 controls without GVHD	Not reported systematically	None	None	One of the first papers describing the spec trum of systemic clinico-pathologic features of GVHD
Sale et al, 1981	37 patients with GVHD	BM	75–2172 days	Grade I and Grade II	47 grafted patients without GVHD	Not reported	None	None	Grade I findings were sensitive for GVHD (>80%), but only 50–56% specific. Grade II findings less sensitive but more specific for GVHD. No association of histo logic findings and irradiation
Izutsu et al, 1983a,b	11 with chronic GVHD	BM	76–704 days	Qualitative assessment	6 allogeneic patients without GVHD and 4 syngeneic patients	Not reported systematically	Slight ↓ compared to healthy controls	No	5/10 GVHD glands abnormal vs 1/8 controls. GVHD saliva had ↓ salivary IgA, and ↑ salivary Na+, IgG and albumin
Janin-Mercier et al, 1987	60 allogeneic and 8 syngeneic	BM	100 days; some followed up to 3 years	Qualitative assessment	8 syngeneic patients	Not reported systematically	No	Yes; graded 0–IV on maximal rate of emptying;	Minor glands = Class II in 21/60 allogeneic patients and 0/8 syngeneic. Scintigraphy abnormal in 46 GVHD patients and 8/8 syngeneic patients
Nakhleh et al, 1989	34	BM	24–1653 days	Qualitative assessment	None	No	No	No	GVHD in the salivary glands predicts GVHD in other organs
Horn et al, 1995	59	BM	Mean = 136 days	Horn criteria	None	No	No	No	No association with oral histologic changes and BMT outcome
Hiroki et al, 1996	11 with chronic GVHD	BM	Not given	Scored 0, 1+, and 2+	Compared with 24 Sjogren's patients	8/11 reported oral dryness	↓ Stimulated whole flow 4/11 GVHD patients	Abnormal in 3/11 GVHD patients	HLA-DR, ICAM-1, and E-selection expressed on ductal and endothe lial cells in areas of lymphocytic infiltration. CD4/CD8 ratio = 0.9 in GVHD glands
Nakamura et al, 1996	18 patients with GVHD	BM	Not given	Scored 0, 1+, and 2+	19 BMT patients without GVHD	+ oral dryness in 11/18 with GVHD and 10/19 without	↓ Whole stim. flow in 8/18 GVHD patients and 7/19 non-GVHD patients	Abnormal in 3/11 GVHD and 0/4 non-GVHD patients	16/18 GVHD patients had 1+ or 2+ salivary biopsy vs 5/19 non-GVHD (differs at P < 0.001). No difference in acinar atrophy or destruction
Alborghetti et al, 2005	14 patients with oral GVHD	BM = 7; PBSC = 7	100–1006 days	Horn criteria	9 BMT patients without GVHD	2/14 GVHD + oral dryness	14/14 by clinical exam	No	Acinar volume ↓ in GVHD group; fibrosis ↑ in GVHD. Biopsies evaluated without clinical diagnosis

Keratoconjunctivitis sicca can be diagnosed without invasive methods; therefore, eye examinations are performed more often than salivary gland biopsies in patients post-transplant. The prevalence of KCS after BMT was 19% in a multicenter retrospective cohort study of 248 persons, and was more common in those with cGVHD. Other risk factors for KCS were female sex, age > 20 years, conditioning with single dose irradiation and methotrexate for prevention of GVHD (Tichelli et al, 1996).

Two recent studies suggest that salivary glands may be less affected by modern conditioning, which may be non-myeloablative and not use irradiation. In one study, salivary gland function was judged as normal in 47 of 49 pediatric patients who were treated with HSCT for a variety of conditions. Notably, these patients were young, many had not received total body irradiation as part of conditioning, and the source of the transplanted cells was stem cells rather than bone marrow cells (Tichelli et al, 1996; Treister et al, 2005). In another prospective cohort study of 20 adult patients receiving HSCT, ⁹⁹mTc-pertechenate scintigraphy was employed to evaluate early changes in major salivary gland function following transplant (Coracin et al, 2006). Scans before HSCT as well as those at Days +30, +60 and, +100 post-transplant did not demonstrate functional changes over time. However, concurrent ⁶⁷Ga scanning, an imaging method that detects inflammation, revealed inflammatory infiltration following HSCT, primarily in submandibular glands, up to Day +100 after transplant. Together, these findings suggest that salivary glands are still impacted by modern transplant techniques, but the long-term consequences that modern HSCT has on salivary gland function have not been defined.

Salivary compositional changes after HSCT

As salivary glands constitute part of the mucosal immune system, a few studies have examined changes in salivary constituents after transplant. Izutsu et al (1983b) reported that patients (n = 12) with cGVHD of salivary glands had decreased salivary immunoglobulin A (IgA) and inorganic phosphate, with increased [Na⁺], [Cl⁻], albumin and immunoglobulin G (IgG) as compared to transplant patients without GVHD (n = 10) or healthy controls (n = 8). These observations were confirmed by others (Nagler and Nagler, 2004). In another study of 61 subjects, elevated salivary [Na⁺] from the minor salivary glands was highly predictive of cGVHD in patients, especially those who had not had total body irradiation (Izutsu et al, 1983a). Other reported salivary changes in patients with cGVHD include elevated salivary epidermal growth factor (Nagler et al, 1996).

To achieve complete immune reconstitution after transplant, grafted cells must migrate to mucosal tissues, receive appropriate signals from antigen-presenting cells, and produce secretory immunoglobulins. In successfully grafted patients, secretory IgA can be detected in saliva within 6–12 months of transplant, and secre-tory IgA recovery precedes serum IgA by 2–3 months (Steinbrenner et al, 2005). Depressed levels of serum IgA (Fujimaki et al, 2001), salivary IgA (Izutsu et al, 1985) and IgA-bearing plasma cells (Loughran et al, 1990) within minor salivary glands are positively associated with the presence or development of cGVHD, and may explain the increased susceptibility of these patients to sinopulmonary infections.

A recent study demonstrated that transplant has a significant impact on the salivary proteome. Serially collected saliva samples from 41 patients undergoing allogeneic-HSCT were evaluated at three time points for 6 months after transplant. Multiple changes in salivary proteome were detected by SELDI-TOF mass spectrometry across time, with significant elevations of lactoferrin and secretory leukocyte protease inhibitor that persisted at least 6 months post-transplant (Imanguli et al, 2007).

Histopathology of salivary glands in GVHD

Three features characterize the histopathologic changes in GVHD salivary gland disease: lymphocytic infiltration of the salivary gland ducts, individual ductal epithelial cell necrosis (apoptosis), and destruction of acinar (at least 10%) tissues with periductal fibrosis (Soares et al, 2005; Shulman et al, 2006). Children may also exhibit oncocytic ductal metaplasia. A challenge for those evaluating biopsy specimens is distinguishing past disease from active disease. Dense fibrosis and acinar destruction most probably reflect past disease, while fibroplasia, acinar and periductal inflammation and ductal damage most probably reflect current GVHD activity. The lymphocyte infiltrate primarily contains T lymphocytes, with a slight predominance of CD8 vs CD4 cells and macrophages (Hiroki et al, 1996; Soares et al, 2005). Expression of HLA-DR is found on ductal epithelial cells associated with lymphocytic infiltration (Lindahl et al, 1988; Hiroki et al, 1996).

The hallmark of Sjögren's syndrome is a lymphoplasmocytic periductal infiltrate within the major or minor salivary glands, and three validated scoring methods are used to assess minor gland biopsies obtained for disease diagnosis. The presence of a lymphoplasmocytic cluster of at least 50 cells is termed a ‘focus’. Pathologists often use the focus score (Greenspan et al, 1974), a number between 0 and 12 that is generated by counting the foci in a 4-mm² representative section of salivary tissue, or the Chisholm–Mason scale (Chisholm and Mason, 1968) to semi-quantify the mononuclear infiltrate. A focus score of 1 (which equals a Chisholm–Mason score of 3) is consistent with Sjögren's syndrome in the modified American–European criteria for Sjögren's syndrome (Vitali et al, 2002). The other scale often used to evaluate biopsies is the Tarpley scale (Tarpley et al, 1974), which assesses both salivary gland fibrosis and lymphocytic infiltration.

Most GVHD literature does not specify what, if any, criteria were used to evaluate and diagnose GVHD involving the salivary glands. Literature often states lymphocytic foci or infiltrates were noted in the glands, but focus scores usually were not reported (Table 2). While two histologic grading systems (Sale et al, 1981; Horn et al, 1995) have been proposed for cGVHD of minor salivary gland evolution that incorporate amount of lymphocytic infiltration with destruction of glandular acini and fibrosis, these scales are not used extensively. One small study compared the utility of the Horn scale to distinguish patients with and without GVHD salivary gland disease (Alborghetti et al, 2005), but the different proposed scales have not been validated using both normal minor gland tissues and tissues from transplanted patients.

Treatment

Multiple factors must be considered when treating a patient with oral GVHD. Even though oral GVHD may have serious health consequences, GVHD of the liver or lungs is life-threatening. Therefore, therapeutic decisions must include consideration of the patients' medical regimens for non-oral GVHD conditions and, as such, cooperation between treating physicians and dentists is mandated. Furthermore, individual response variation for oral GVHD is expected. The clinician may need to try various therapeutics and possibly therapeutic combinations to manage successfully oral GVHD symptoms. The primary treatment goals are to diminish patient pain, maintain the patient's ability to eat, increase the patient's quality of life and prevent destruction of oral tissues and dentition.

Pharmacotherapy for oral GVHD may be systemic, topical, or injectable (Imanguli et al, 2006; Couriel et al, 2006). In general, these agents have not been tested in well designed, randomized controlled trials, and trials of therapies for systemic GVHD may not include oral assessments. Therefore, the evidence for treatment efficacy should be considered preliminary.

Systemic pharmacotherapy for oral mucosal lesions

Systemic immunosuppressive therapy is used for extensive cGVHD involving multiple organs or body sites. The primary limitation of systemic immunosuppression is the increased risk for opportunistic infections, a leading cause of mortality in HSCT patients. Furthermore, there is the possibility of reduced graft-versus-tumor effect (Couriel et al, 2006; Imanguli et al, 2006). The two systemic immunosuppressive drugs used most commonly are cyclosporine and systemic corticosteroids, either alone or in combination. Cyclosporine, a calcineurin inhibitor, suppresses T-cell proliferation and prevents transcription of genes for IL-2, IL-2 receptor, and IFN-γ. Therefore, patients taking cyclosporine are at increased risk for oral mucosal infections such as oral candidiasis and herpetic infections. Systemic corticosteroids also predispose patients to fungal and viral infection. Therefore, clinicians managing severe mucosal lesions in GVHD patients must consider infectious agents as possible etiologic causes or contributors when patients are maintained using long-term systemic immunosuppression.

Other systemic agents used for systemic GVHD include tacrolimus (FK-506) and sirolimus (rapamycin), which inhibit T-cell proliferation similarly to cyclosporine, and pentostatin and mycophenolate mofetil (MMF), which suppress both T and B lymphocytes (Couriel et al, 2006; Imanguli et al, 2006). GI and hematologic toxicity can limit MMF use, and the infectious susceptibilities of patients taking this drug are significant (Krejci et al, 2005).

Hydroxychloroquine, an antimalarial drug with anti-inflammatory properties, clofazimine and thalidomide, which have anti-inflammatory properties including decreasing TNF-α activity, have also been tested as treatments for GVHD (Gilman et al, 2000).

Several trials evaluated thalidomide in cGVHD with modest overall responses. Systemic thalidomide is associated with multiple side effects which limit its tolerable dose and reduce its therapeutic benefit (Parker et al, 1995; Biagi et al, 2001; Imanguli et al, 2006).

Emerging systemic therapies being tested in patients with recalcitrant disease include monoclonal antibodies to block mediators of inflammation, such as the anti-TNF-α antibodies infliximab and etanercept, and the anti-IL-2 receptor antibody daclizumab (Imanguli et al, 2006). Rituximab, an anti-CD20 monoclonal antibody, has demonstrated promise for treatment of GVHD (Ratanatharathorn et al, 2003; Zaja et al, 2007). The drug was effective in over 50% of patients with refractory cGVHD and may have a beneficial impact upon survival. Cutler et al (2006) reported a clinical response rate of 70% in the treatment of steroid refractory GVHD (Kim, 2007).

A non-pharmacologic treatment for systemic GVHD is extracorporeal photophoresis (ECP), a process that separates the patient's mononuclear cells through apheresis and exposes them to ultraviolet light A (UVA). The cells are subsequently re-infused in the patient. Though not completely elucidated, the process is believed to induce apoptosis of alloreactive T lymphocytes, normalize the CD4/CD8 ratio and induce regulatory T cells. Preliminary data suggest that the process is efficacious for oral cGVHD, but the procedure is limited by its long duration (4 h) and the availability of ECP facilities (Imanguli et al, 2006).

Topical and local therapy

Topical and local therapy for oral GVHD offer several advantages, including fewer systemic side effects and drug interactions, the ability to intensify therapy to one specific area while preventing systemic host immuno-suppression, and maintenance of graft-versus-tumor effects. Despite these possible advantages, there are few controlled trials which have examined the efficacy of topical treatments for oral GVHD or have compared topical and systemic approaches for management of oral GVHD (Imanguli et al, 2006).

Corticosteroids

Topical corticosteroids, commonly used for many oral mucosal inflammatory conditions, are the most popular local therapy for oral cGVHD. Agents tested in studies include topical budesonide rinse (Elad et al, 2003) and topical dexamethasone rinse (0.1 mg cc⁻¹) (Wolff et al, 2004). Other topical agents that might be used include flucinonide, clobetasol, beclomethasone, and triamcin-alone (Couriel et al, 2006; Imanguli et al, 2006; Kim, 2007).

Cyclosporine and tacrolimus

Topical cyclosporine has been tested in small open label studies for oral GVHD (Epstein and Reece, 1994) and topical tacrolimus is presently used in dermatology for atopic dermatitis and cutaneous cGVHD (Choi and Nghiem, 2001). Recently, there have been reports indicating possible efficacy of topical tacrolimus ointment in oral cGVHD. Preliminary findings suggest that it has therapeutic benefit with limited side-effects (Eckardt et al, 2004; Albert et al, 2007).

Local phototherapy

Exposure to UVA after oral administration of 8-methoxypsoralen (PUVA) causes cross-linkage of DNA, leading to cellular apoptosis. Rapidly proliferating cells such as activated T lymphocytes are particularly sensitive to effects of PUVA. (Yoo et al, 1996; Imanguli et al, 2006). Small studies tested PUVA for treatment of oral cGVHD, with preliminary reports of success. However, PUVA therapy may increase the risk for squamous cell and basal cell carcinoma in transplanted patients (Imanguli et al, 2006).

Treatment GVHD salivary gland disease

Guidelines for the treatment of patient with decreased salivary flow are designed to increase patient comfort and decrease caries risk (Atkinson et al, 2005). Agents to increase patient comfort include cholinergic agonists pilocarpine or cevimeline HCl that increase resting salivary flow rates, ‘artificial’ salivas or coating agents that moisten the mouth, and sugarless candies, sugarless mints or sugarless gums that stimulate salivary flow through mechanical and gustatory stimuli. Patients with reduced salivary flow have an increased risk for caries and oral fungal infections. Topical fluorides should be prescribed as needed for patients with active carious lesions. Other options to reduce caries include concentrated fluoride varnish (such as 5% sodium fluoride containing fluoride at 25 000 ppm) and monitoring the patient every 3 months to restore new lesions as needed.

Treatment studies of cholinergic agonists for HSCT patients with significant salivary complaints have been very small, consisting of one double-blind, placebo-controlled study and a few open label trials (Nagler and Nagler, 1999, 2001; Carpenter et al, 2006; Agha-Hosseini et al, 2007). The only controlled study found pilocarpine four times per day minimally increased salivary flow at one time point in the study and decreased xerostomia complaints, though it is unclear how perceptions of dryness were collected (Agha-Hosseini et al, 2007).

Study limitations

Most therapeutic trials for oral GVHD were not randomized, had no placebo group and were too small to control for confounding systemic therapies. Clinical outcomes were not graded using standardized scales, making interpretation of the data difficult. Large trials of systemic therapies often evaluate the mouth using the oral mucositis WHO grading scale, which was not developed to assess oral GVHD. Though two standardized scales for grading oral GVHD exist, (Piboonniyom et al, 2005; Pavletic et al, 2006), they need to be validated in prospective studies.

Limitations of GVHD salivary gland disease studies

Most studies of salivary gland pathology after transplantation are very small and should be interpreted with caution. At this time, there is no accepted case definition of GVHD salivary gland disease to use in clinical studies. Some investigations only use patient perceptions of oral dryness, which may not indicate salivary gland pathology. Studies comparing flow rates of transplanted patients with and without salivary gland disease should be large enough to account for the wide variance in salivary flow rates of healthy populations, and control for medication usage (Ship et al, 1991; Yeh et al, 1998).

Histologic interpretation of minor salivary glands has it own set of challenges. At least five different grading systems (Greenspan, Chisholm–Mason, Sale, Horn, and Talal, Table 2) were used to diagnosis GVHD salivary gland disease in various studies, and those developed for GVHD (Sale, Horn, and Talal) have not been extensively validated in studies that include healthy un-transplanted controls. This is essential, as modest lymphocytic infiltration with an occasional focus of lymphocytes (Radfar et al, 2002) is found in normal tissues. Therefore, one could expect that salivary glands would be re-populated after transplant with some lymphocytes. In addition, salivary gland fibrosis, a reported feature of GVHD salivary gland disease, occurs with age in normal glands (Syrjanen, 1984). More comprehensive, larger studies of the histopathologic changes associated with GVHD salivary gland disease are needed. Evaluators scoring minor gland biopsy specimens should not know the group assignment of specimens they are evaluating.

Conclusion

Chronic GVHD remains the most significant long-term challenge in the allogeneic HSCT. Despite high significance and prevalence, patho-physiology of cGVHD remains elusive and progress in the development of effective therapeutic and preventive strategies has been slow. Intensive basic, translational and patient-oriented research in this important area is critically needed to make allo-HSCT a more predictable and successful treatment modality. In addition, given the close clinical and histopathologic similarities between cGVHD and many autoimmune disorders, we believe that cGVHD serves as a unique human model of autoimmunity. Therefore, progress in cGVHD research may benefit not only those post-allo-HSCT but also a much larger community of patients.