|Home | About | Journals | Submit | Contact Us | Français|
Viral infections are often associated with salivary gland pathology. Here we review the pathogenesis of HIV-associated salivary gland disease (HIV-SGD), a hallmark of diffuse infiltrative lymphocytosis syndrome. We investigate the presence and contributions of viral diseases to the pathogenesis of salivary gland diseases, particularly HIV-SGD. We have detected BK viral shedding in the saliva of HIV-SGD patients consistent with viral infection and replication, suggesting a role for oral transmission. For further investigation of BKV pathogenesis in salivary glands, an in vitro model of BKV infection is described. Submandibular (HSG) and parotid (HSY) gland salivary cell lines were capable of permissive BKV infection, as determined by BKV gene expression and replication. Analysis of these data collectively suggests the potential for a BKV oral route of transmission and salivary gland pathogenesis within HIV-SGD.
HIV-SGD is AIDS defining in pediatric HIV infection and has increased in the adult HIV population. The term ‘HIV-SGD’ was first coined by Schiødt et al. (1989) to describe an enlargement of major salivary glands and/or a complaint of xerostomia in the absence of xerogenic agents, medications, and diseases known to cause xerostomia in HIV patients (Schiødt et al., 1989, 1992). In HIV-SGD, quantitative changes occur in the saliva—such as lower secretory rates of sodium, calcium chloride, cystatin, lysozyme, and total anti-oxidant capacity—which affect the homeostasis of the oral cavity and account for significant morbidity during the progression of HIV disease (Lin et al., 2003). Typically, HIV-SGD presents a unilateral or bilateral diffuse soft swelling resulting in facial disfigurement, and may be associated with pain. Histologically, HIV-SGD is characterized by hyperplastic, intraparotid lymph nodes and/or lymphatic infiltrates within the salivary gland tissue. This lymphocytic infiltrate consists predominantly of CD8 T-cells that are also CD29+, indicative of a memory cell phenotype. Recent histochemical analysis has revealed the location of the infiltrate as periductal with acinar atrophy, ductal dilation, and mild to moderate fibrosis with collagen deposition (McArthur et al., 2003; Rivera et al., 2003). HIV-SGD also affects the minor salivary glands, with labial salivary glands demonstrating features of sialadenitis.
HIV-SGD has been universally established as among the most important AIDS-associated oral lesions. Oral lesions are important clinical indicators of HIV/AIDS by suggesting HIV infection in the undiagnosed individual, indicating clinical disease progression, and predicting development of AIDS (Patton and van der Horst, 1999). Specific common oral lesions are strongly associated with immune suppression, as measured by CD4 cell counts, and are modestly associated with high viral burden, thus serving as potential clinical markers of HIV viremia and the consequent destruction of the immune system with progressive HIV disease. Since the introduction of highly active anti-retroviral therapy (HAART) in the mid-1990s, the prevalence of types of oral cavity infections of HIV-infected patients has changed over time, with hairy leukoplakia (HLP), oral candidiasis (OC), and oral Kaposi’s sarcoma (KS) decreasing and the incidence of HIV-SGD increasing (Patton et al., 2000; Greenspan and Greenspan, 2002). For example, in an HIV-infected North Carolina population, HIV-SGD has gone from 1.8% in an early cohort to 5.0% in a late cohort (Patton et al., 2000).
This increase in HIV-SGD after the induction of a HAART drug regimen is reminiscent of previously described cases of immune reconstitution disease (IRD). IRD among HIV-infected patients is an adverse consequence of restoration of immune responses during the initial months of anti-retroviral treatment (ART) (Lawn et al., 2005). Rizos et al. (2003) described a patient on triple anti-retroviral therapy with excellent immunological and virological responses and parotid gland enlargement that may have reflected immune reconstitution. IRD typically occurs as a result of a pre-existing opportunistic infection or a subclinical infection becoming “unmasked” upon the introduction of HAART treatment, when HIV viral load decreases and CD4 count increases. Therefore, HIV-SGD presenting as an IRD implies that an infectious agent may play a role in its etiology.
In a prospective study, the association of a Sjögren’s-like syndrome (SLS) with the progression of HIV-related disease in a cohort of HIV patients treated with HAART was evaluated (Mastroianni, 2004). The study identified four cases of SLS in a cohort of 150 HIV-positive patients treated with HAART, whereas none of the HIV-positive patients not treated with HAART fulfilled the diagnostic criteria for SLS. In conclusion, the study suggested that HAART treatment may play an important role in the etiology of SLS, and that SLS may be a new and important complication of long-term HAART regimens (Mastroianni, 2004). Interestingly, 75% of these patients did not express the autoantibodies typical of SS, and thus probably had HIV-SGD. This study highlights the clinically identical nature of the two diseases and provides further evidence that HIV-SGD increases with HAART therapy. As more people receive HAART therapy, the potential for increasing incidence of HIV-SGD grows.
In developing countries, the incidence of HIV-SGD has been reported to be as high as 48% of HIV-1-infected patients (McArthur et al., 2000). Parotid gland lesions are generally much larger and more disfiguring in pediatric patients than in adults, and in most children xerostomia is not usually associated with parotid swelling, although in adult patients it is a fairly common complaint. In both adults and children, the lesion often occurs in conjunction with generalized lymphadenopathy. Reasons for the differences between the adult and pediatric manifestations are not yet understood (Leggott, 1992). In children with HIV infection, it has been shown that the presence of salivary gland enlargement is associated with less rapid progression to death than is found in those with OC or herpes simplex. The time to death averages 5.4 years, compared with 3.4 years in those with OC (Katz et al., 1993). This may be due to a protective effect of the CD8 T-cells that predominate in the lymphocytic infiltrate in the salivary glands in that condition and are part of diffuse infiltrative lymphocytic syndrome (DILS) (Itescu et al., 1989, 1990). This is perhaps a marker of the ability of the host to respond to viral infection. The HIV-SGD disease process is of particular interest, because in 1-2% of patients, malignant lymphomas have been described in association with these glandular lesions, making this disease a pre-malignant lesion (Ioachim and Ryan, 1988; DiGiuseppe et al., 1996).
HIV-SGD reflects a localized manifestation of DILS that occurs in the parotid gland of HIV-positive persons. DILS is characterized by salivary and lacrimal glandular swelling and sicca symptoms of various degrees of intensity, accompanied by persistent circulating and visceral CD8+ lymphocytic infiltration of parotid glands, lacrimal glands, kidneys, muscles, nerves, lungs, and lymph nodes, in association with HIV infection (Itescu et al., 1989). DILS has also been designated as a cystic lymphoid hyperplasia of AIDS and is a close phenotypic mimic of Sjögren’s syndrome in terms of sicca symptoms, salivary glandular enlargement, histology, and predisposition to develop non-Hodgkin’s lymphoma. The difference between DILS and Sjögren’s syndrome (Table 1), however, lies in the frequent occurrence of extraglandular sites of lymphocytic infiltration in DILS, the nature of the infiltrating lymphocytes—that is, CD8 in DILS vs. CD4 in Sjögren’s syndrome—and the scarcity of serum autoantibodies in DILS (Itescu et al., 1993). Other conditions that may mimic DILS include salivary gland tumors, especially papillary cystadenoma lymphomatosum, viral conditions such as parotitis, bacterial infections such as tuberculosis, or systemic diseases, for instance, sarcoidosis. Sialadenosis, related to endocrine and nutritional conditions or neurogenic medications, may also be included in the differential diagnosis (Ramos-Gomez, 2002). As in HIV-SGD, patients with DILS appear to be at a significantly increased risk of malignant lymphomas (Mandel et al., 1998).
In HIV-SGD, epidemiologic data point to an antigen-driven process. A prospective study that recognized that the epidemiology, clinical presentation, and certain extraglandular manifestations of HIV-SGD/DILS have changed concomitant with the introduction of HAART suggests that DILS is an antigen (viral)-driven response, and that the primary treatment for it is anti-HIV therapy (Basu et al., 2006). Further evidence alluding to an infectious etiologic agent is in the form of data showing different rates of HIV-SGD in children (20-47%) vs. adults (3-7.8%) (DiGiuseppe et al., 1996). We postulate that this may indicate primary viral infection in children vs. residual immunity in adults. Further, the infiltrating lymphoid cells present during HIV-SGD point to an antigen-driven MHC-determined host immune response (Itescu et al., 1989). That the lymphocytosis within HIV-SGD is phenotypically an antigenically driven response may also implicate an infectious agent present during HIV immunosuppression.
It has been postulated that SS autoimmune disease is primed by viral infection. There is evidence that the type 1 interferon pathway is highly stimulated locally and systemically in SS (Ronnblom et al., 2003). It has been suggested that viruses may initially infect the gland, leading to production of type 1 interferon by plasma dendritic cells. Interferon would then activate the adaptive immune system. Activation of the adaptive immune response would then result in auto-antibody immune response, even in the absence of virus.
There is a long association between viral infections and diseases of the salivary gland. In general, viral infections of the salivary gland may result in reduced saliva production and/or swelling. Mumps is the most common salivary gland infection caused by paramyxovirus (Fields et al., 1996). Paramyxovirus is transmitted by droplet infections carried in the saliva. There is a primary viral infection of the oral cavity and a multiplication of the virus in the upper respiratory tract and the regional lymph nodes. Paramyxoviral infection causes bilateral swelling of the parotids, although the other salivary glands may also be affected. However, subsequent to vaccine development, the incidence of mumps has been drastically reduced.
Several viruses, including DNA, RNA, and retroviruses, have been considered important co-factors in the development of SS. The two DNA viruses that have been studied in association with pSS are cytomegalovirus (CMV) and Epstein-Barr virus (EBV) (Saito et al., 1989; Mariette et al., 1991; Maitland et al., 1995). Overall, the data regarding CMV and EBV as causative agents for SS are contradictory, and, because pSS does not occur in most cases of viral reactivation in vivo, the link between reactivation and autoimmunity induction remains to be established. Two RNA viruses have been detected within the salivary glands of SS patients: hepatitis C and Coxsackie virus, both of which require more investigation before they are classified as being the etiologic agent of SS (Triantafyllopoulou et al., 2004; Carrozzo, 2008). Interestingly, four different retroviruses have been associated with pSS: human T-lymphotropic virus type 1 (HTLV-1), human immunodeficiency virus (HIV), intracisternal A-type particles, and human retrovirus 5 (Garry et al., 1990; Shattles et al., 1992; Griffiths et al., 1997; Yamano et al., 1997). However, only HTLV-1 and HIV produce a Sjögren’s-like syndrome with significant pathologic and clinical differences from pSS.
In HIV-positive patients, HIV p24 protein was detected in the salivary glands by an immunohistochemistry technique (Rivera et al., 2003). Rivera’s group was also able to identify Epstein-Barr virus (EBV) protein, but not cytomegalovirus (CMV), by immunohistochemistry in the salivary glands of HIV patients (Rivera et al., 2003). Yen et al. (2004), however, concluded that CMV and KSHV were not associated with Benign Lymphoepithelial Cysts (BLC) of the parotid gland, although EBV was found more frequently in the disease than in normal parotid controls (Yen et al., 2004). Human herpes virus 6 (HHV-6) and HHV-7 were also demonstrated in the salivary glands and in the saliva in HIV-positive persons (Di Luca et al., 1995). However, their presence has not been attributed to etiological importance for salivary gland lesions. The lymphocytic infiltration observed in HIV-SGD/DILS may be an antigen-driven event, due to the presence of infection of the salivary gland. Papovaviruses have also long been associated with salivary gland disease development in experimental animals. Murine infection with polyomaviruses is associated with parotid gland tumorigenesis and has been implicated in human carcinogenesis (Imperiale, 2000). As suggested by the body of work reviewed above, viral infection is tightly associated with salivary gland pathology (Table 2). With the exception of mumps virus and PIV, the majority of these human studies detected only viral nucleic acid, but did not determine whether salivary gland tissues are actually permissive for these agents.
The potential for a viral etiologic agent in HIV-SGD is highly feasible. We have detected BKV in the saliva of patients with HIV-associated salivary gland pathology and in healthy individuals (Jeffers et al., 2009). Primary infections with BKV are typically subclinical or linked to mild respiratory illness (Fields et al., 1996), followed by viral dissemination to the sites of lifelong persistent infection. The major site of persistence for BKV is the kidney and urinary tracts, with epithelial cells of the kidney, ureter, and bladder as the predominant cell types that are persistently infected (Fields et al., 1996; Imperiale, 2000). Interestingly, with quantitative real-time PCR analysis, BKV DNA was detected in saliva samples from HIV-positive patients and healthy control individuals (Fig.). Like most opportunistic pathogens, BKV DNA was detected in the saliva at higher levels in immunocompromised patients compared with healthy individuals. BKV DNA levels were highest in persons with HIV-SGD.
To begin to decipher BKV pathogenesis within the salivary gland cell, it was essential to develop an in vitro model system. BKV infection was characterized in submandibular (HSG) and parotid (HSY) gland cell lines. These cells were able to support viral entry, transcription, translation, and virion production, and BKV infection could be inhibited by saturation of the capsid protein with its ganglioside receptor (Jeffers et al., 2009). BKV entry into salivary gland cells was inhibited by the blocking of VP1 via ganglioside saturation, as had previously been shown in Vero cells (Sinibaldi et al., 1990). As expected, viral transcription was more significantly diminished in the presence of 100 μg/mL ganglioside compared with 60 μg/mL in the salivary gland cells. Similar trends were observed in the Vero cell line upon ganglioside treatment. Likewise, a consistent decrease in viral protein expression was detected with ganglioside treatment in both salivary gland and Vero cells. The detection of labeled BK virions within the cytoplasm of salivary gland cells confirmed BKV entry (Fig., C). In salivary gland cells, labeled virus was still detected at 48 hpi, while labeled virus was no longer detected in Vero cells at the same time-point, perhaps due to viral uncoating and entry into the nucleus. Viral gene products were first observed at 24 hpi for both T Ag and VP1 with real time RT-PCR, with viral transcription peaking at about 3 dpi for both submandibular and parotid gland cell infection. Protein expression was first detected at 4 dpi by immunoblot and immunofluorescence. While BKV transcripts were detected in abundance, the expression of viral-encoded proteins was more modest. The difference in mRNA vs. protein levels may reflect the decreased stability of the BKV T Ag protein. Productive infection was further substantiated by the detection of virions by EM at 5 days post-infection, as well as the presence of DNAse-resistant virus in the supernatant. These virions increased over time and were used to infect naïve Vero cells. Interestingly, the kinetics of infection appeared to be slower in salivary gland cells than in kidney cells (Jeffers et al., 2009).
We thus demonstrated that BKV was detected in oral fluids, and that BK infection and replication occur in vitro in salivary gland cells. Analysis of these collective data suggests the potential for the BKV oral route of transmission and oral pathogenesis.
This investigation was supported by USPHS Research Grants K23 DE 00460-01 and R03 DE14444, from the National Institute of Dental and Craniofacial Research, National Institutes of Health, Bethesda, MD 20892. UNC CFAR 9P30AI50410.