Abstract
Human T cell leukemia virus type I (HTLV-I) sequences were sought in labial salivary glands of patients with HTLV-I—associated myelopathy or tropical spastic paraparesis and of seropositive neurologically healthy carriers. HTLV-I proviral DNA was found by polymerase chain reaction amplification in DNA extracted from lip biopsies of every patient. Viral RNA was found by in situ hybridization in the acini epithelium, as well as in lymphocytic infiltrates. This observation suggests that HTLV-I expression in labial salivary glands could participate in the inflammatory lesions observed in these patients. Some seronegative patients with Sjögren's syndrome or dryness syndrome were also positive for viral transactivator tax DNA (41% in Martinique and 16% in non-HTLV-I—endemic region). Despite histologic signs of lymphocytic infiltration, no viral expression was found in the labial salivary glands of these patients.
Human T cell leukemia virus type I (HTLV-I)—infected individuals may develop various diseases, including adult T cell leukemia (ATL), tropical spastic paraparesis or HTLV-I—associated myelopathy (TSP/HAM), polymyositis, polyarthritis, uveitis, lymphocytic alveolitis, and Sjögren's syndrome (SS) (reviewed in [1]). Except for ATL, the pathologies associated with HTLV-I infection are inflammatory conditions. The mechanism by which HTLV-I causes these pathologies is not well understood.
SS is a chronic disease characterized by the destruction of salivary and lachrymal glands, leading to dryness of mouth and eyes [2]. Histologic observations of salivary glands from SS patients show important lymphocytic infiltration. Vernant et al. [3] were the first to propose an association between HTLV-I and SS. In fact, a large proportion of HAM/TSP patients have clinically and histologically confirmed SS. In Japan, a high rate of HTLV-I seroprevalence has been observed among SS patients [4]. The pathogenesis of SS in regions where HTLV-I is not endemic remains to be clarified, but viral infection has been suggested as a possible etiologic factor. Several studies have shown the surprising presence of HTLV-I Tax transactivator gene in lip biopsies of HTLV-I—seronegative patients with SS [5, 6]. Furthermore, transgenic mice expressing the HTLV-I tax gene presented an exocrinopathy resembling SS [7].
In HTLV-I—infected individuals, the infiltration of salivary glands by infected or autoreactive activated T lymphocytes might contribute to the development of exocrinopathy. Alternatively, HTLV-I—specific cytotoxic CD8 T cells could be responsible for the destruction of infected epithelial cells. The aim of this study was to search for the presence of HTLV-I expression in labial salivary glands (LSGs) of patients with TSP/HAM, patients with SS associated with HTLV-I, healthy HTLV-I carriers, and HTLV-I—seronegative patients with SS from non-HTLV-I—endemic regions. Polymerase chain reaction (PCR) was used to detect proviral DNA, and reverse transcription (RT) PCR or in situ hybridization (ISH) was used to detect viral RNA.
Materials and Methods
Patients and tissue specimens
Twelve HAM/TSP patients, 6 HTLV-I—seropositive healthy carriers, and 12 HTLV-I—seronegative patients with SS or dry mouth and eyes condition, all from Martinique, were referred for study at La Meynard University Hospital (Fort de France, Martinique). Twelve patients from a region without endemic HTLV-I who presented with SS or dry mouth and eyes condition were selected at the Foch University Hospital (Suresnes, France). Lip biopsies were either fixed in 4% paraformaldehyde and embedded in paraffin or frozen in liquid nitrogen—cooled isopentane. Mononuclear cell infiltration of the LSG was quantified, and marked fibrous thickening of the acini was estimated. As a control, peripheral blood mononuclear cells (PBMC) from 22 healthy blood donors (Centre de Transfusion Sanguine, St. Joseph Hospital, Paris) were collected.
PCR and RT-PCR
DNA was extracted from paraffin sections as previously described [8]. RNA was extracted from frozen unfixed sections using TRI-Reagent (Molecular Research Center, Cincinnati). cDNA was synthesized from 10 μg of RNA with avian myeloblastosis virus reverse transcriptase and hexamers as described [8]. Amplifications were performed with Taq DNA polymerase in 100 μL of standard buffer, with 200–500 ng of DNA or cDNA. HTLV-I proviral DNA sequences were amplified with 35 cycles (tax: 340 nucleotides, positions 7432–7772, in Seiki ATK1 sequence [9]; env: 276 nucleotides, positions 5116–5392; gag: 205 nucleotides, positions 1616–1821, in Seiki ATK1 sequence). For RT-PCR, an HTLV-I tax—spliced mRNA sequence (507 nucleotides, positions 5098–7723) was amplified with 35 cycles. Amplified products were detected by Southern blot hybridization with specific oligonucleotide internal probes labeled with 32P in 3′ end by terminal transferase.
HTLV-I tax cloning and sequencing
PCR-amplified products from lip biopsies were treated with the Klenow fragment of DNA polymerase I, phosphorylated, ligated to the SmaI cut and dephosphorylated M13mp18 vector, and electroporated in competent XL1-Blue Escherichia coli. Positive clones were detected by hybridization with specific probes, and single-strand DNA was prepared. Sequencing was accomplished by amplification with a sequencing kit (Taq DyeDeoxy Terminator Cycle; Applied Biosystems, Cambridge, UK) and electrophoresis on an automated sequencer (Applied Biosystems).
ISU
ISH was performed as described with a 35S-labeled antisense riboprobe corresponding to the tax region [8]. HTLV-I—infected MT2 cells were used as the positive control. Sense riboprobe complementary to the same HTLV-I region was used as the negative control.
Results
Histologic classification of patients
Lip biopsies were obtained from 12 HAM/TSP patients and 6 neurologically healthy HTLV-I—seropositive individuals from Martinique. Most HAM/TSP patients and some neurologically healthy carriers presented mild clinical symptoms of mouth and eye dryness. LSG were analyzed and graded according to classic criteria for SS (table 1). Histologic observation showed in most HAM/TSP patients the presence of several inflammatory infiltrates with >50 mononucleated cells per lobule (figure 1). Among seronegative patients from Martinique, 6 were diagnosed with primary SS and 6 with dry mouth and eyes condition without all criteria of primary SS. Among 12 patients from non—HTLV-I—endemic region with dry mouth and eyes condition, 2 were diagnosed with primary SS, 9 with nonspecific dryness syndrome, and 1 with sarcoidosis (table 1).
Summary of results obtained by tax gene polymerase chain reaction (PCR) and reverse transcription-PCR amplification on DNA from labial salivary glands (LSGs) from patients with HTLV-I—associated pathologies and seronegative patients from a non—HTLV-I—endemic region.
Histologic examination of labial salivary gland (LSG) sections from different patients. A, Control: HTLV-I—seronegative individual from Martinique with no clinical symptoms (magnification ×344). B, HTLV-I—seronegative patient with Sjögren's syndrome from non—HTLV-I—endemic region (magnification ×344). C, HTLV-I—seropositive patient from Martinique with TSP/HAM (magnification ×344). D, HTLV-I—seropositive patient from Martinique with no TSP/HAM but with dryness syndrome (magnification ×344). E, HTLV-I—seropositive patient from Martinique with TSP/HAM (magnification ×860). F, HTLV-I—seronegative patient with Sjögren's syndrome from non—HTLV-I—endemic region (magnification ×860).
Detection of HTLV-I DNA and RNA in salivary glands
HTLV-I proviral sequences were amplified in DNA extracted from LSG. As shown in table 1, all seropositive patients were positive for tax, gag, and env genes, indicating that HTLV-I proviral DNA was present in LSG of these patients. Five seronegative patients with dry mouth and eyes syndrome out of 12 from Martinique (41%) were also PCR-positive for tax but not for gag or env genes. LSG from seronegative patients with dry mouth and eyes condition from a non-HTLV-I—endemic area were also studied. Two (16%) of 12 samples were PCR-positive for the tax gene and negative for the gag and env genes. PBMC from seropositive patients were PCR-positive for all HTLV-I primers tested. In contrast, PBMC from patients from non—HTLV-I—endemic regions were always negative for tax PCR, even when LSGs were positive. DNA extracted from the PBMC of 22 healthy control blood donors were negative.
To study viral expression, we looked for spliced mRNA sequences of the tax gene in the RNA extracted from frozen LSG (RNA could not be extracted from paraffin-embedded tissues in a form suitable for PCR amplification). As shown in table 1, tax-spliced mRNA could be amplified in seropositive patients' LSGs, indicating that the virus was expressed. In contrast, seronegative patients were negative for viral expression, even when positive for tax DNA. As a control for cDNA synthesis, the housekeeping gene glyceraldehyde phosphate dehydrogenase mRNA was amplified in every sample.
Sequencing of PCR products
To verify the sequences amplified in patients with tax primers, the 340-nucleotide-long PCR product was cloned, and five different clones per patient were sequenced (positions 7432–7772 in Seiki ATK1 sequence [9]). Except for one or two silent mutations in some patients, the amplified sequences were similar to the cosmopolitan HTLV-I sequence, even in the patients from non—HTLV-I—endemic regions. No variation was found between individual clones from the same patient, indicating that the observed mutations were not introduced during the PCR by Taq polymerase. This result confirms that the amplified sequences were not due to plasmid contamination.
ISH
ISH was performed on many sections of different frozen and paraffin-embedded LSG samples from individuals for whom the expression of the tax gene had been detected by RT-PCR. No background was observed on frozen tissue when ISH was performed with the tax antisense riboprobe, even after 1 month of exposure. Specific HTLV-I tax RNA signal was detected in frozen sections following autoradiographic exposure times of 2–4 weeks. As shown in figure 2, tax gene expression was observed in TSP/HAM patients as well as in neurologically asymptomatic patients with clinical and histologic signs of dry mouth and eyes syndrome. This expression was observed reproducibly with different sections from the same patient. Control ISH with tax sense riboprobe was always negative. Seronegative patients from regions with or without endemic HTLV-I were negative for tax expression by both RT-PCR and ISH, even when positive for tax DNA by PCR. As shown in figure 2, positive signal was observed both in luminal epithelial cells and in the acini structure (figure 2D), as well as in inflammatory infiltrates (figure 2A, 2B).
In situ hybridization with HTLV-I tax riboprobe on labial salivary gland (LSG) sections from different patients. Arrows = tax gene expression. A, Sample from HTLV-I—seropositive patient from Martinique with TSP/HAM, hybridized with an antisense probe (magnification ×860). B, Sample from HTLV-I—seropositive patient from Martinique with no TSP/HAM but with dryness syndrome, hybridized with antisense probe (magnification ×860). C, Sample from HTLV-I—seropositive patient from Martinique with TSP/HAM, hybridized with sense probe as control (magnification ×860). D, Sample from HTLV-I—seropositive patient from Martinique with TSP/HAM, hybridized with antisense probe (magnification ×2165). In each case, exposure time was 4 weeks.
Discussion
In the present work we show by PCR amplification that the HTLV-I provirus is present in the labial salivary glands of TSP/HAM patients and of neurologically healthy HTLV-I carriers with dryness syndrome. Using RT-PCR amplification, we show that viral RNA is expressed in LSG. Virus expression was detected by ISH in both acini cells and inflammatory infiltrates, suggesting that tissue damage could be due either to a direct viral effect or to the inflammatory cells present in the focal lesions. Our data suggest that HTLV-I may infect glandular epithelial cells and induce dryness syndrome. Pathologic changes in exocrine glands linked to HTLV-I Tax expression have already been shown in tax-transgenic mice that presented with exocrinopathy affecting lachrymal and salivary glands and resembling SS [7].
In the peripheral blood, HTLV-I infects primarily CD4 T cells. The provirus load is high in infected individuals, but the virus replicates weakly and is expressed in only 1 in 10,000 circulating lymphocytes. This may indicate either that the circulating lymphocytes are not the only virus reservoir or that viral integration leads to proliferation of infected lymphocytes. Because of the lymphotropism of HTLV-I, a reservoir in which the latently integrated virus could replicate has been sought in lymphoid organs. The virus has been detected in lymph nodes of ATL patients, circulating dendritic cells, and lymphocyte progenitors in bone marrow [1]. Outside the immune system, low levels of HTLV-I antigen and mRNA have been found in central nervous system and muscle lesions of TSP/HAM and HTLV-I—induced polymyositis [8, 10, 11]. Excepted for a few infected astrocytes in a case of HAM/TSP [12], infected cells in the central nervous system appear to be infiltrating CD4 lymphocytes. Our observation suggests that salivary glands could represent a reservoir in which the virus replicates. Recently, persistent HTLV-I infection of breast luminal epithelial cells has been demonstrated, suggesting that chronically infected epithelial cells could serve as a long-term reservoir of HTLV-I infection in patients [13].
Our observation that the LSGs of 41% of seronegative patients from Martinique were positive for tax proviral DNA is surprising. This issue is still controversial [14]. Our result could reflect the greater sensitivity of tax primers in PCR amplification compared with env and gag primers. This greater sensitivity is probably due to the small size of the tax fragments that are usually amplified (<150 nucleotides long) [5, 6]. To avoid this problem, we amplified a 340-nucleotide-long tax fragment (larger than env and gag, 276 and 205 nucleotides, respectively).
Nevertheless, the higher frequency of “tax only” PCR-positive patients observed in Martinique (41%) compared with a non-HTLV-I—endemic region (16%) may indicate an implication of Tax in this condition, even in seronegative patients in an HTLV-I—endemic region. These patients might harbor a defective virus unable to induce an antibody response or might have been infected recently and have not yet seroconverted. Alternatively, their provirus load might have been too low to induce an immune response, or they could harbor a defective provirus unable to induce an immune response. This observation is in accord with a recent study of HTLV-I prevalence in the United States, indicating that a substantial fraction of HTLV-I—seronegative individuals were positive by PCR [15].
Regarding the seronegative patients from a region without endemic HTLV-I, our finding confirms previous observations in which HTLV-I tax sequences have been detected in the salivary gland epithelium of some seronegative SS patients in France as well as in a region of Japan without endemic HTLV-I [5, 6]. The tax sequence amplified in these patients could correspond to a defective provirus or could belong to another exogenous or endogenous unknown retrovirus. Recently, a novel exogenous retroviral sequence was identified in a salivary gland of a patient with SS [16]. This sequence corresponding to a pol gene and close to human endogenous HervK retrovirus has been related to B and D oncoviruses and designated human retrovirus 5. However, in a recent paper, the association of this new virus and SS has not been confirmed [17].
In conclusion, our results suggest that the presence of HTLV-I tax sequences in the LSGs of HTLV-I—positive patients could participate in the inflammatory lesions observed. A precise characterization of infected cells and a larger epidemiologic study comparing different stages of the disease will be required to fully evaluate the implications of this observation for the pathogenesis of SS.
Acknowledgments
We thank Séverine Boullier for providing healthy blood donors and Julie Cosserat for patient selection.
References
Grant support: Agence Nationale de Recherche sur le SIDA, Centre National de la Recherche Scientifique, European Community Human Capital and Mobility Program (contract CHRX-CT94-0670), and Pasteur Institute Foundation.
