Cytokine production in patients with recurrent acute tonsillitis: analysis of tonsil samples and blood

The aim of this study was to examine T cell function in tonsils of patients with recurrent acute tonsillitis (RAT) or peritonsillar abscess (PTA) by analyzing the cytokine production following T cell receptor (TCR) and co-receptor stimulation with a combination of anti-CD3 and anti-CD28 antibodies. The release of IFN-γ, TNF-α, IL-2, IL-4, IL-6, IL-10 and IL-17A from isolated, stimulated T cells of 27 palatine tonsils (10 RAT, 7 PTA, 10 tonsils without inflammation) was measured via a bead-based flow cytometric analysis. The results were compared with the cytokine release of isolated peripheral T cells in a subset of the same patients (6 PTA, 4 patients without signs of inflammation in the blood). TCR stimulation increased the concentration of released cytokines in tonsil and blood as well as in different forms of inflammation and tissue with no inflammation. Stimulation increased the pro-inflammatory cytokines TNF-α, IFN-γ, and IL-2 more than the anti-inflammatory cytokines IL-4 and IL-10 in tonsil and blood samples in RAT, PTA, and samples without inflammation. Blood of patients with PTA showed a higher pro-inflammatory cytokine level compared to the samples of patients without inflammation. T cells in tonsils are fully responsive and competent for antigen-induced cytokine production in RAT and PTA. One should be aware that tonsillectomy, if indicated, might remove a functioning immune organ. Tonsillotomy might be an alternative even in adults to maintain immunological function.

T cell purification from tonsils. T cells were isolated following a protocol optimized for tonsillar T cells that has been published before 14 . Briefly, tonsil tissue is minced, homogenized, and pressed through a cell strainer. Mononuclear leucocytes were enriched by density gradient centrifugation on Ficoll Histopaque-1077 (Sigma-Aldrich, Steinheim, Germany) and T cells are isolated out of the resulting mononuclear fraction via automated magnetic sorting on the AutoMACs (Miltenyi Biotec, Bergisch Gladbach, Germany) using combined positive selection on CD4 and CD8 magnetic beads (Miltenyi Biotec, Bergisch Gladbach, Germany). The purity of final T cell preparations ranged 90-95%, as ascertained by surface CD3 staining via flow cytometry 13 . T cells were eluted into the cell culture medium (RPMI 1,640 medium supplemented with penicillin and streptomycin and 10% heat-inactivated fetal calf serum (FCS) and kept in a humidified atmosphere at 37 °C and 5% CO 2 for 24 h before stimulation. Peripheral T cells from a donor and patient blood were isolated and cultured in the same way out of PBMCs, except that a step of erythrocyte lysis was included before magnetic sorting 15 . Antibodies and T cell stimulation. T cell stimulation was performed after isolation via two different protocols, one each for low-grade and high-grade TCR stimulation for 48 h. For low-grade TCR stimulation, T cells were stimulated with soluble anti-CD3 and anti-CD28. To this end, cells were challenged with 1.7 μg each of biotinylated anti-CD3ε (clone OKT3), and anti-human CD28 purified (clone CD28.2), both from eBioscience www.nature.com/scientificreports/ Inc. (San Diego, USA). High-grade stimulation was accomplished using surface-immobilized anti-CD3/CD28 coated on beads. To this end, T cells were challenged with 10 µl of anti-biotin MAC SiBead Particles loaded with biotinylated anti-CD2, anti-CD3, and anti-CD28 per 10 6 cells (T Cell Activation/Expansion Kit, Miltenyi Biotec, Bergisch Gladbach, Germany), following the kit's instructions. As a positive control, T cells were treated with a mix of 0.3 µM phorbol 12-myristate 13-acetate (PMA) and 1 µg/ml ionomycin, which circumvents the T cell receptor and receptor-proximal steps to activate the T cell signaling machinery. T cells were purified from patient tonsils by an optimized and streamlined protocol 14 and have been previously characterized with regard to T cell subclasses (CD4/CD8) and activation markers 13 . Cytokine measurements. Levels of TNF-α, IFN-y, IL-2, IL-4, IL-6, IL-10 and IL-17A in the supernatants of stimulated T cells were analyzed via a magnetic bead-based array using the BD Cytometric Bead Array (CBA)-Human Th1/Th2/Th17 Kit, following the manufacturer´s instructions. All flow cytometry data were acquired using a FACS Canto or FACS Calibur (BD Pharmingen™, Franklin Lakes, USA). Data processing was done with BD Cytometric Bead Array (CBA) FCAP Array Software.

Statistical analysis.
Results are presented as means ± standard deviation (SD), if not otherwise indicated.
For comparison between two groups, a unpaired t-tests and for multiple comparison, a one-way-ANOVA with post-hoc Bonferroni correction were performed by using IBM SPSS statistics software version 23.0.0.0, SPSS Inc., Chicago, IL. The level of significance was set to p < 0.05.

Results
Stimulation of tonsillar T cells with soluble anti-CD3/CD28, bead-immobilized anti-CD3/CD28, and PMA/ionomycin increased the concentration of released cytokines ( Table 2 and Supplementary Fig. 1). TCR stimulation led to a stronger increase of genuine pro-inflammatory cytokines TNF-α, IFN-γ, and IL-6 than anti-inflammatory cytokines, prominently IL-4 and IL-10 ( Fig. 1, Table 2). Blood of patients with PTA had a higher concentration of pro-inflammatory cytokines in comparison to patients with no sign of inflammation in the blood. Importantly, the secretion of IL-2, the autocrine mitogen responsible for clonal T cell expansion, was more markedly upregulated by immobilized versus soluble anti-CD3/CD28 (Table 2), consistent with many previous studies. Also, T cells from non-inflamed tonsil tissue featured the strongest IL-2 response, in contrast to all other measured cytokines, which were either equally or more prolifically generated by T cells from inflamed tonsils. Tonsil tissue of RAT reacted less after stimulation (Fig. 1).
In general, we observed a clear preference of tonsillar versus peripheral T cells for distinct cytokines. Thus, tonsillar T cells of RAT produced higher levels of IL-10 and IL-17 than peripheral T cells (  Table 2). Notably, IL-6 production was significantly higher in T cells from RAT in comparison to not inflamed tonsil tissue (RAT-T vs. NO INFL-T p = 0.02). However, IL-6 levels generated in tonsillar T cells were low compared to the levels produced by peripheral T cells, suggesting that IL-6 may not be a physiologically

Discussion
The current study documented cytokine release after stimulation of tonsillar T cells from two inflammatory scenarios of tonsillar pathology, i.e. RAT and PTA. The results underscore the notion that the palatine tonsils are a functional immune organ even in adults and adults with chronic inflammatory tonsillar disease.
The strength of this study was the analysis of T cells originating from different forms of inflammation in tonsil side-by-side to non-inflamed tonsil specimens and peripheral lymphocytes from the same patients (the latter limited to a subgroup of donors). At the same time, the small group numbers, inherent to this type of study, was certainly a limitation preventing more robust statistics.
We also took care to compare distinct modes of T cell stimulation, as this process is finely tuned in lymphocytes to discriminate productive antigen-challenge from non-productive cases. Thus, antibody-based clustering of the TCR via CD3 and its co-receptor CD28 is known to trigger productive and full-blown T cell activation (culminating in IL-2 secretion and clonal expansion) only when antibodies are administered immobilized on  www.nature.com/scientificreports/ surfaces 16 . Thus, our stimulation protocol included low-grade TCR stimulation (soluble anti-CD3/CD28) and high-grade stimulation conditions (Abs immobilized on beads), plus a positive control with PMA/ionomycin. In a previous study, we observed a trend to a higher percentage of T helper cells in the blood of patients with PTA in comparison to RAT. Furthermore, tonsil tissue of RAT contained more PD1(+) CD4(+) T cells reflecting T cell exhaustion due to chronic infection 13 .
TCR stimulation increased the pro-inflammatory cytokines TNF-α, IFN-γ, and IL-2 much higher than of the anti-inflammatory cytokines IL-4 and IL-10 in tonsil and blood, suggesting that a Th1-response predominated over Th2-response. Komorowska et al. reported similar results. In their work, which analyzed cytokine production by lymphoid cells from adenoids and tonsils, the Th1 response dominated over the humoral immune Th2 reaction in palatine tonsils as compared to adenoids in 25 children between 4 and 15 years 17 . In an RNA-PCR-analysis of adults, it was also shown that the Th1 based immune response predominated in patients with chronic inflammatory tonsillitis, whereas the Th2 immune reactions were more pronounced in adenoids of adenoid hyperplasia 19 . In our own present work, Th1 immune response dominated in RAT clearly over Th2 immune response.
Along these lines, pro-inflammatory cytokines dominate in children with chronic tonsillitis not just in lymphoid tissue, but also in saliva 18 . Specifically, levels of TNF-α, IL-1, IL-6 are higher in the saliva of children with chronic tonsillitis, although statistically significant differences were only noted for the pro-inflammatory cytokine IL-6 18 . In our study, IL-6 was also significantly higher in tonsil tissue of RAT than tonsil tissue of PTA and non-inflamed tonsil tissue. However, it was striking that IL-6 production by tonsillar T cells of all scenarios was negligibly low compared to the levels produced by peripheral T cells. Importantly, this feature was specific for IL-6, as other cytokines where secreted to comparable orders of magnitude in lymphocytes from tonsils and blood. We conclude that tonsillar T cells feature a skewed TCR response with low IL-6 production, indicating that IL-6 may not be a particularly relevant cytokine in the context of the tonsillar immune response.
Our findings also document robust secretion of the anti-inflammatory cytokines IL-10 and IL-17 by tonsillar T cells, in accordance with previous reports by others 20,21 . This anti-inflammatory response may be owed to tonsillar follicular T helper cells (Tfh), as Tfh cells are abundant in human tonsils and prolific in the production of IL10 21 . We did not use surface markers to discriminate Thf cells within our T cell preparations, but the findings reported warrant analyzing the contribution of Tfh cells to the tonsillar immune response to infections in future studies.
The present findings are consistent with the concept that human tonsils are immunologically functional entities. Immunomodulatory therapies aiming at the manipulation of the Th1/Th2 balance response are thus a plausible and conceivable new therapeutic strategy in tonsillar inflammation 22 . Immunomodulators with inhibition of Th1 response could contribute to an improvement of RAT. Drugs for inhibition of IL-2 and TNF-α are used for autoimmune diseases like rheumatoid arthritis, Crohn's disease, colitis ulcerosa, or psoriasis 23 . Such therapies are currently not in use for RAT. This may be owed to the fact that tonsillotomy or tonsillectomy is expected to have fewer systemic side effects compared to immune modulation. However, in the case of contraindications to  www.nature.com/scientificreports/ surgery or in the case of simultaneous autoimmune diseases, immunomodulators could conceivably contribute to an improvement of tonsil inflammatory diseases.
In summary, we suggest that in future immunomodulatory therapies or tonsillotomy, i.e. only a partial resection of the tonsils should be discussed and considered more seriously as an alternative to tonsillectomy in the case of RAT, as tonsillotomy should preserve, at least partially, tonsil immunity. In the case of PTA, abscess incision or cranial tonsillotomy 24 as an alternative to abscess tonsillectomy could similarly be a promising strategy from an immunological perspective. www.nature.com/scientificreports/ conclusion T-cells in palatine tonsils from adult patients with recurrent acute tonsillitis and peritonsillar abscess are functional and responsive and feature a characteristic, specific cytokine response with low IL-6 and features of Th1. We propose that tonsillectomy should be performed only in severe disease with high impairments. Tonsillotomy could be an alternative therapeutic option for preserving tonsillar immunity. Immune-based therapies aiming at the modulation of the Th1/Th2 balance represent a promising new therapeutic concept for exceptional cases of tonsillar inflammation.