Keywords

Hashimoto’s thyroiditis (HT) is a very common autoimmune disease of the thyroid. In addition to genetic background, several viruses, including herpesviruses, have been suggested to play a role as possible environmental triggers of disease, but conclusive data are still lacking.

HHV-6 infection is common and has a worldwide distribution (Caselli and Di Luca 2007). Viral strains cluster in two variants that were recently classified as different viral species, on the basis of characteristic biological differences: HHV-6A, with still unknown disease association, and HHV-6B, the etiologic agent of roseola (exanthem subitum), a childhood benign febrile disease. HHV-6 species in vitro replicate most efficiently in primary T-cells and in selected T-cell lines. However, the in vivo tropism of HHV-6 is considerably broader, including macrophages, endothelial cells, salivary glands, and brain (Caruso et al. 2002; Di Luca et al. 1994; Thomas et al. 2008). After primary infection, HHV-6 establishes a latent infection and resides mainly in peripheral blood mononuclear cells (PBMCs) and in macrophages (Di Luca et al. 1994; Kondo et al. 1991). During latency, HHV-6 expresses specific viral transcripts. In particular, expression of U94, in the absence of other viral lytic transcripts, is considered a molecular marker of viral latency (Caselli et al. 2006; Rotola et al. 1998).

HHV-6 has been tentatively associated to several chronic autoimmune inflammatory processes (Scotet et al. 1999), including Sjogren syndrome (Fox et al. 1989; Ranger-Rogez et al. 1994), multiple sclerosis (Alvarez-Lafuente et al. 2010; Challoner et al. 1995; Rotola et al. 2004), rheumatoid arthritis and systemic lupus erythematosus (Alvarez-Lafuente et al. 2005; Krueger et al. 1991). In addition, recent case reports suggested that HHV-6 infection might be related to the onset of autoimmune disorders, including purpura fulminans, severe autoimmune acquired protein S deficiency (Boccara et al. 2009), autoimmune connective tissue diseases (Broccolo et al. 2009), and severe autoimmune hepatitis (Potenza et al. 2008). Our analysis of fine needle thyroid aspirates (FNAs) and blood from HT patients and controls showed that HHV-6A prevalence and load are highly increased in thyroid tissue of HT patients (Caselli et al. 2012b). Furthermore, HT-derived thyrocytes harbor active HHV-6A, whereas the virus is strictly latent in the few virus-positive controls. We also reported that HHV-6A infects thyroid cells, inducing de novo expression of HLA-II surface antigens. Consequently, HT patients have increased CD4+ and CD8+ T-cell responses to HHV-6 U94 protein and infected thyrocytes become a target for innate Natural Killer (NK) cell killing. NK cells comprise about 10–15 % of all circulating lymphocytes and are able to lyse target cells that have lost the protective signal mediated by human leukocyte antigen (HLA) class I surface molecules (Storkus et al. 1987) as in viral infections. In particular, non-productive HHV-6A and HHV-6B infection is known to lead to the up-regulation of HLA (Human Leukocyte Antigens)-A, -B, -C molecules on dendritic cells (Bertelsen et al. 2010; Gustafsson et al. 2013), via autocrine IFN (Interferon)-α signaling, as well as the up-regulation of HLA-DR and CD86 molecules. This modification may result in the inability of NK cells to recognize target infected cells, as they still present HLA expression. Moreover, HHV-6A infection suppress DC stimulation of allogenic T cell proliferation. The ability to block innate and adaptive immune responses might be a successful strategy by which HHV-6A avoids the induction of appropriate host defense mechanisms, and thus facilitating persistent infection. Human NK cells can be divided into two subsets based on their cell-surface density of CD56 molecule in CD56bright and CD56dim, each with distinct phenotypic properties. There is evidence to suggest that these NK-cell subsets have unique functional attributes and, therefore, distinct roles in the human immune response. The CD56dim NK cell subset is more naturally cytotoxic while CD56bright NK-cell subset has the ability to produce abundant cytokines following activation and has low natural cytotoxicity (Cooper and Caligiuri 2001). NK cell activities have been evaluated during HHV-6 infection. NK cell activation was high in the acute phase of HHV-6 infection and declined gradually during convalescence. These results suggest that NK cells play a major role in resolving acute phase infection while specific lymphocyte activity develops later (Kumagai et al. 2006). In this study we analyzed changes in the activity and subtype of NK cells in peripheral blood cells from HT patients.

Clinical samples derived from 8 HT patients and 8 patients with benign follicular epithelial lesions (controls). The 8 HT patients included 2 males and 6 females, with a mean age of 57 ± 15 years (range 37–78 years), with anti-thyroperoxidase antibodies (TPO Ab) > 35 IU/ml (mean value = 835 IU/ml, range 343–3000 IU/ml), and anti-thyroglobulin antibodies (Tg Ab) > 115 IU/ml (mean value = 205 IU/ml, range 120–366 IU/ml). The 8 control patients included 4 males and 4 females with a mean age of 64 ± 18 years (range 30–91 years) (there was no statistically significant difference between the two groups), and showed TPO Ab < 35 IU/ml (mean value = 10 IU/ml, range 8–13 IU/ml), and Tg Ab < 115 IU/ml (mean value = 16 IU/ml, range 11–21 IU/ml). None of the patients enrolled in the study presented other autoimmune diseases.

Patients were characterized for HHV-6 viral load in their peripheral blood mononuclear cells (PBMCs) and thyroid FNAs, obtained as part of routine clinical work from patients undergoing FNAs for diagnostic purposes, and were used after receiving approval from the Local Ethical Committee of the University of Ferrara and S. Anna Hospital of Ferrara. The patients provided written informed consent for both FNA procedure (which is part of the clinical practice) and for biomolecular analyses, to which purpose the samples were anonymized. PBMCs were isolated by Ficoll-Hypaque gradients. DNA was isolated from FNAs and PBMCs as described (Caselli et al. 2012a). HHV-6 DNA presence and load were analyzed by PCR and real time quantitative PCR (qPCR) specific for the U94 and U42 genes (Caruso et al. 2009); samples were considered positive when 1 ug of cell DNA harbored more than 100 copies of viral DNA (Caselli et al. 2012a). Amplification of the house-keeping human RNase P gene was used as a control. All clinical samples were analyzed in a randomized and blinded fashion. NK cell activity and number was measured by flow cytometer.

NK cells were characterized with a specific anti-CD panel (CD3-PerCp-Cy5.5, CD56-FITC, CD107a-PE, CD16-PE) (e-Bioscience). For the CD107a degranulation assay, that shows NK cell activation status, cells were stained with CD107a-PE (e-Bioscience) after 1 h of incubation at 37 °C and 3 h of treatment with Golgi Stop solution (Becton Dickinson) (Rizzo et al. 2012). Ten thousand events were acquired. Cell viability was assessed by propidium iodide staining. Anti-isotype controls (Exbio) were performed.

HT and control subjects were also characterized for their antibody response against HHV-6 (whole virus) or its U94/Rep protein, by testing plasma samples by specific ELISA assays (Caselli et al. 2002). As a control, the plasma samples derived from 12 healthy donors were also assayed.

The results showed that HHV-6 was more prevalent in HT FNAs (8/8, 100 %) than in FNAs derived from controls (2/8, 25 %) (p < 0.001) (Table 1). Furthermore, viral load was higher in HT specimens (mean 1.2 × 104 copies/μg of cellular DNA, range 8 × 102–4.7 × 104 copies/μg DNA) than in the few controls which resulted positive for HHV-6 (mean 3.9 × 102 copies/μg DNA, range 2.2–5.7 × 102 copies/μg DNA) (p < 0.01). Similar results were obtained in PBMCs. In particular, HHV-6 was detected in 8/8 HT PBMCs (100 %) and only in 3/8 PBMCs derived from controls (37 %) (p < 0.01) (Table 1). Furthermore, viral load was higher in HT specimens (mean 1.8 × 104 copies/μg of cellular DNA, range 1.8 × 102–3.9 × 104 copies/μg DNA) than in the few controls which resulted positive for HHV-6 (mean 3.7 × 102 copies/μg DNA, range 2.8–4.9 × 102 copies/μg DNA) (p < 0.01). Where possible, virus species characterization, performed as previously described (Caselli et al. 2012a), showed the presence of HHV-6A in the thyroid tissue and of HHV-6B in PBMCs (data not shown), confirming the previously observed different tropism of the two viruses (Caselli et al. 2012a).

Table 1 Presence of HHV-6 in specimens obtained from HT patients and controls
Full size table

The anti-HHV-6 antibody response, evaluated by ELISA using a whole virus lysate (obtained by treatment of purified virions with 0.25 % Triton followed by brief sonication) as the antigen, showed no significant differences in antibody prevalence or titer between HT and control subjects or healthy donors, as well as the anti-tetanus toxin/toxoid (TT) IgG response, used as a control (measured by an ELISA kit, Alpha Diagnostic) (Fig. 1). On the contrary, the antibody response specifically directed against the HHV-6 U94/Rep protein was more prevalent in HT patients (8/8) than in controls (6/8), and especially the titer was significantly higher in HT vs control subjects (1:1624 vs 1:543) (p < 0.01), who showed prevalence and titer values similar to those of healthy donors (10/12; titer 1:442) (Fig. 1). These results confirmed that HT patients not only have a specific anti-U94/Rep cellular immune response (Caselli et al. 2012a), but also develop specific antibodies against this virus protein.

Fig. 1
figure 1

Anti-HHV6 antibody response in HT and control subjects. Humoral response against (a) whole HHV-6 virus, (b) Tetanus Toxoid (TT), and (c) HHV-6 U94/Rep protein were evaluated by ELISA in Hashimoto’s thyroiditis (HT), control (CTR) and healthy donors (HD) groups. Results are expressed as (a) mean absorbance at OD405nm ± standard deviation, (b) mean Ab titer (U/ml) ± standard deviation, and (c) mean Ab titer (dilution−1) ± standard deviation. *p ≤ 0.01, obtained with two tailed Student t test

Full size image

The analysis of CD3−CD56bright NK cell percentages reported a higher amount of these cells in the samples from HT patients compared with controls (p = 0.02; two tailed Student t test) (Fig. 2a). The activation status of CD3−CD56bright NK cells was higher in HT patients compared with controls (p = 0.01; two tailed Student t test) (Fig. 2b). On the contrary, CD3−CD56dim NK cells did not present differences in the two groups of subjects (Table 2). Since NK cells are also subdivided into different populations based on the relative expression of CD16, we analyzed the levels of this surface markers. CD16 is a Fc receptor that, upon recognition of antibody-coated cells, delivers a potent signal to NK cells, which eliminate targets through direct killing and cytokine production. When we considered CD16 expression (Fig. 2c, Table 2), we observed that almost all CD3−CD56dim NK cells are CD16+, as previously reported (Poli et al. 2009). On the contrary, there was a slight difference in the percentage of CD3−CD56brightCD16− NK cells between the two groups (Fig. 2c, Table 2).

Fig. 2
figure 2

Cell percentage of (a) CD3−CD56bright NK cells and (b) CD3−CD56brightCD107a+ in HT and control (CTR) subjects. Data are reported as Mean ± standard deviation. *p value obtained with two tailed Student t test. (c) Representative dot plots for CD56 and CD16 staining in HT and control (CTR) subjects. Cell percentages are reported

Full size image
Table 2 Percentages of NK cells in HT and control subjects
Full size table

When we analyzed the possible association between cell activation of CD3−CD56bright NK cells in HT patients and TPO and Tg Ab levels, we observed a slight correlation between these parameters and CD3−CD56bright NK cell CD107a expression (Fig. 3a, b).

Fig. 3
figure 3

Correlation between cell activation (CD107a expression) of CD3−CD56dim NK cells from HT patients and disease status, reported as (a) TPO and (b) Tg Ab levels. *p value obtained with Correlation Z test

Full size image

These results indicate that NK cells might have an important role for the control of disease activity and viral infection. In fact, we observed an increased NK cell activity in HT patients characterized by HHV-6 infection in FNAs. Previous researches documented the implication of NK cells in the control of both viral infections (Kumagai et al. 2006; Rizzo et al. 2012; Wu et al. 2015) and autoimmune thyroid disease exacerbation (Hidaka et al. 1992). The increase in CD3−CD56bright NK cells, that are characterized by a cytokine-secreting phenotype, during HHV-6 infection could modify the cytokine environment in HT patients with a possible implication in the disease. In particular, we found an increase in CD3−CD56brightCD16− NK cells, that are known to abundantly produce IFN-γ (Vitale et al. 2004). It is known that cytokines are involved in the pathogenesis of thyroid diseases working in both the immune system and directly targeting the thyroid follicular cells. They are involved in the induction and effector phase of the immune response and inflammation, playing a key role in the pathogenesis of autoimmune thyroid disease. Finally, cytokines can directly damage thyroid cells, leading to functional disorders and may also stimulate the production of nitric oxide and prostaglandin, thus increasing the inflammatory response in HT patients (Mikoś et al. 2014). Moreover, our findings on the increase in CD56bright NK cells in HT patients are in agreement with a previous study that documented the increase in CD56bright NK cells and inflammatory cytokines in the cerebrospinal fluid and serum of a 15-month-old girl with acute necrotizing encephalopathy (ANE) associated with HHV-6 (Kubo et al. 2006).

We are aware that it is difficult to prove etiologic links between viral infections and diseases, especially in the case of a ubiquitous agent such as HHV −6. Moreover, the number of subjects enrolled in this study is limited and a larger cohort is necessary to confirm these results. Nevertheless, our findings indicate that HHV-6 might contribute to HT development, increasing NK cell secretion of inflammatory cytokines sustaining the persistence of an inflammatory status in HT patients.