IL-35 is a heterodimer of EBV-induced gene 3 and of the p35 subunit of IL-12, and recently identified as an inhibitory cytokine produced by natural Treg in mice, but not in humans. Here we demonstrate that DC activated by human rhinoviruses (R-DC) induce IL-35 production and release, as well as a suppressor function in CD4+ and CD8+ T cells derived from human peripheral blood but not in naïve T cells from cord blood. The induction of IL-35-producing T cells by R-DC was FOXP3-independent, but blocking of B7-H1 (CD274) and sialoadhesin (CD169) on R-DC with mAb against both receptors prevented the induction of IL-35. Thus, the combinatorial signal delivered by R-DC to T cells via B7-H1 and sialoadhesin is crucial for the induction of human IL-35+ Treg. These results demonstrate a novel pathway and its components for the induction of immune-inhibitory T cells.
One of the main functions of the immune system is to control infections 1. The contact with a pathogen requires a strong and efficient response of the immune system to prevent harm for the organism. Yet, potent immune responses may be accompanied by severe side-effects, with immune-pathology as a final result. Thus, anti-pathogen responses need to be controlled adequately. There is increasing evidence that suppressor cells or Treg are critically involved in this process. In fact, recent studies even suggest that pathogens actively provoke the generation of Treg, thereby harnessing these regulatory cells to evade the immune system.
Two major subsets of Treg have been proposed – natural and inducible – that differ in terms of their development, specificity, and mechanism of action. Natural occurring Treg consist of CD4+ T cells, generated in the thymus 2 and are characterized by the constitutive expression of CD25 and the transcription factor FOXP3. Natural Treg inhibit effector T-cell responses via so far unclear mechanisms that involve cell–cell contact. More recently, Collison et al. demonstrated that IL-35 contributes to the inhibitory function of murine natural Treg 3, 4. IL-35 is a novel heterodimeric cytokine consisting of EBV-induced gene 3 (EBI3) and the p35 subunit of IL-12 5. However, human CD4+CD25+FOXP3+ Treg do not constitutively express IL-35 and induction of FOXP3 upregulates neither EBI3 nor p35 mRNA 6, 7. Inducible Treg develop from mature T-cell populations under certain conditions, e.g. upon stimulation with tolerogenic DC or by IL-10 treatment 8, 9. Inducible Treg primarily act via soluble mediators and typically produce high levels of immune-suppressive cytokines IL-10 and/or TGF-β. The suppressive function of human inducible Treg seems to be FOXP3-independent 10, 11.
Human rhinoviruses (HRV), the major cause of common cold in humans, can blunt adaptive immune responses through the induction of a novel DC activation program. We have recently reported that cocultivation of DC with HRV-14 (R-DC, DC activated by human rhinovirus) induces the expression of inhibitory receptors B7-H1 (PD-1L, CD274) and sialoadhesin (Sn, Siglec-1, CD169), a sialic acid binding lectin, without affecting the expression of stimulatory receptors such as CD80 or CD86. The consequence of this altered accessory repertoire on R-DC is that cocultured T cells acquire a deep anergic state 12, 13.
Here we demonstrate that R-DC induce suppressor function in cocultured T cells. The inhibitory effect was found to be caused by the culture supernatant (SN) of CD4+ and CD8+ peripheral blood T cells, activated with R-DC, but not with naïve T cells from cord blood (CB). We found that R-DC-induced Treg produced and released IL-35, which is responsible for the inhibitory effect of the Treg SN. Most importantly, blocking of B7-H1 and sialoadhesin on R-DC with specific mAb against both receptors prevented the induction of IL-35. Thus, inhibitory signals delivered from R-DC to T cells via B7-H1 and sialoadhesin were essential to the induction of human IL-35-producing Treg, defining a novel route of T-cell instruction.
DC treated with HRV induce FOXP3-independent Treg
We have recently demonstrated that HRV is able to subvert the T-cell stimulatory function of human DC. Cocultured T cells acquire a deep anergic state 12. This study was aimed to investigate whether T cells stimulated with R-DC gain a regulatory function.
The results presented in Fig. 1A demonstrate that addition of T cells prestimulated with R-DC strongly inhibited T-cell proliferation, induced by untreated DC. Such an effect did not occur when T cells were primed with untreated DC. Prior fixation of R-DC-induced Treg reverted the inhibitory function (Fig. 1B). Addition of the inhibitor WIN 52035-2 at the time of transfer, which specifically blocks HRV binding to its cellular receptor ICAM-1 14, did not remove the suppressive effect, which indicates that viral transfer is not involved in the inhibitory response (Fig. 1C). Depletion of R-DC from the coculture with T cells did not alter the results: the purified prestimulated T cells still showed eminent inhibitory effects when added to an MLR (data not shown). We conclude that T cells prestimulated with R-DC are responsible for the inhibitory effect observed.
FOXP3 is a forkhead family transcription factor important in the development of Treg; therefore we evaluated its levels in R-DC-induced Treg. Analysis of FOXP3 expression in CD25- T cells revealed that induction of FOXP3 does not differ between DC and R-DC stimulated T-cells, indicating that the suppressive capacity of R-DC stimulated T cells is not directly correlated with an increased induction of FOXP3. Also FOXP3 levels drop again after 48 and 96 h in T cells, stimulated with R-DC or DC (Fig. 1D). Activation-induced transient FOXP3 expression in human T cells has been frequently observed before 10, 15–17 and is not necessarily correlated with regulatory activity 7, 10, 18.
R-DC-induced peripheral blood Treg release an inhibitory factor
In order to analyze whether the inhibitory effect was mediated through (a) soluble factor(s), we added the SN of the R-DC-induced Treg to T-cell/DC cocultures. The SN of R-DC alone, as we have already described, does not lead to diminished T-cell proliferation 12. However, the inhibitory effect was found in the SN of R-DC-induced Treg (Fig. 2A). Both purified CD4+ and CD8+ peripheral blood T cells were cocultured with R-DC and each of their SNs contained this suppressive factor (Fig. 2B and C), because the SN again showed a strong T-cell inhibitory capacity. This factor was not released by naïve T cells, isolated from human CB, as the SN of naïve T cells cocultured with R-DC was not inhibitory (Fig. 2D). Additionally, naïve T cells cocultured with R-DC did not show a reduced proliferation (Supporting Information Fig. 1A and B). This contrasts strongly to the finding that in the coculture of peripheral blood T cells and R-DC, T-cell proliferation is impaired 12. Thus, R-DC-mediated inhibition is specific for CD4+ and CD8+ effector T cells, but not for naïve T cells.
R-DC-induced human Treg act independent of IL-10, TGF-β and IFN-α
Inducible Treg can develop from mature T-cell populations under certain conditions, e.g. upon stimulation with tolerogenic DC. They act via release of soluble factors such as IL-10, a well established inhibitory molecule. In order to elucidate if the inhibitory effect was mediated through IL-10 or other factors, we added the SN of our R-DC-induced Treg to an MLR and investigated whether the inhibitory effect was reversible with neutralizing Ab to IL-10, TGF-β, or IFN-α 11, 19. The levels of the respective factors in the T-cell/R-DC SN were determined previously 12. The inhibitory quality of the SN of R-DC-induced Treg was not reversible with mAb against IL-10, IFN-α, and TGF-β (Fig. 3A). The inhibitory effect of IL-10, TGF-β or IFN-α on a T-cell/DC coculture and the reversibility of this effect with neutralizing Ab is depicted in Supporting Information Fig. 2. Furthermore, size fractionation of the T-cell/R-DC SN revealed that the inhibitory factor is found in the >50 kDa fraction and not in the <50 kDa molecular weight range (Fig. 3B).
R-DC induce IL-35 expression in human T cells
The observation that the inhibitory factor is expected to be >50 kDa leads us to investigate IL-35, a heterodimeric cytokine consisting of EBI3 and the p35 subunit of IL-12 with inhibitory function and a molecular size of 78 kDa 5. We found that T cells cultured with R-DC showed elevated levels of EBI3 and p35 mRNA, but no changes in the p28 levels, which forms IL-27 together with EBI3 (Fig. 4A). Furthermore, intracellular stainings showed that EBI3 was also upregulated at the protein level in peripheral blood T cells, stimulated with R-DC in comparison to T cells cocultured with DC (Fig. 4B, left column). In naïve T cells stimulated with R-DC we did not observe an upregulation of EBI3 (Fig. 4B, right column). P35 is constitutively expressed in DC or R-DC stimulated peripheral blood T cells or naïve T cells (Fig. 4B). This is in accordance to previous findings, which show that p35 is constitutively expressed in various types of human T cells 6. T cells previously activated with R-DC additionally stimulated with PMA/Ionomycin showed a slight increase in p35 and EBI3 (Supporting Information Fig. 4). Resting peripheral blood T cells or T cells prestimulated with DC did not express IL-35 subunits upon PMA/Ionomycin stimulation (data not shown). In order to find out whether R-DC induced inhibitory T cells release IL-35, we co-immunopreciptated the cytokine out of SNs of T cells and R-DC or DC cocultures. As shown in Fig. 4C, R-DC-treated T cells release eminently more IL-35 as the DC stimulated T cells. Also the anti-p35-mAb-coated beads used for immunoprecipitation of IL-35 out of the T-cell/R-DC SN show clear reactivity with the EBI3 Ab when analyzed via flow cytometry and weak reactivity is observed with the respective beads precipitating out of the T-cell/DC SN (Fig. 4D). Only weak reactivity of the beads was observed with anti-p40 mAb (IL-12) and no reactivity was observed with anti-IL-27 mAb (Fig. 4E).
IL-35 contributes to the inhibitory function of R-DC-induced Treg
As R-DC-treated T cells display a regulatory phenotype and release IL-35, the following experiments were designed to examine whether the observed effects were mediated by this cytokine. We added the inhibitory SN of the R-DC-induced Treg to an allogeneic MLR together with a polyclonal Ab to EBI3 or a mAb against p35. We could show that the inhibitory effect of the SN from T cells was abolished and proliferation restored. Figure 5A and B illustrate that Ab directed against both subunits were able to neutralize the inhibitory capacity of the T-cell/R-DC SN, whereas Ab against IL-12p40 or IL-27 did not alter the inhibitory function of the SN (Fig. 5C and D). In addition, purified CD4+ and CD8+ T cells also express EBI3 and gain regulatory function upon stimulation with R-DC and the inhibitory effect of the SN can be reverted by Ab against IL-35 (EBI3 and p35; Supporting Information Fig. 5). Next we used the p35-depleted SN (from Fig. 4), which was no longer inhibitory in an MLR as depicted in Fig. 5E, whereas the T-cell/R-DC SN, precipitated with a control Ab or mock treated, was still inhibitory. Thus the inhibitory effect of R-DC-induced Treg is mediated by IL-35. IL-12p40- or IL-27-depleted SN of a T-cell/R-DC coculture was still inhibitory in an MLR (Fig. 5 F and G) and Supporting Information Fig. 6 shows that IL-12 can be precipitated with the utilized anti-p40 mAb.
B7-H1 and sialoadhesin on R-DC induce the production of IL-35
We have recently found that R-DC work via B7-H1 and sialoadhesin, because blocking of the accessory molecules B7-H1 and sialoadhesin on R-DC with specific mAb against both receptors reverted the inhibitory phenotype of R-DC 12. Now neutralizing Ab to B7-H1 and sialoadhesin were added to the T-cell/R-DC coculture. The production of EBI3 and therefore the production of IL-35 could be effectively blocked by a combination of the two mAb as presented in Fig. 6A. P35 expression did not change considerably with addition of the neutralizing Ab (Fig. 6A right column). The neutralizing Ab were added to a T-cell/R-DC coculture and the cell culture SN of these cells was able to inhibit T-cell proliferation, the Ab alone partially reverted the inhibitory effect. By using a combination of both Ab, proliferation could be restored (Fig. 6B). Thus, the cell surface structures sialoadhesin and B7-H1 are involved in the induction of the IL-35+ Treg.
We demonstrate in this study that IL-35 production and release is induced in human peripheral blood CD4+ and CD8+ T cells by B7-H1 and sialoadhesin co-stimulation, provided by DC. Such IL-35+ T cells are potent Treg, which, in contrast to IL-10-driven type-1 Treg (Tr1), do not suppress T-cell responses via IL-10 and/or TGF-β 11. Several pieces of evidence support the conclusion that the R-DC-induced Treg act via IL-35. Neutralization with anti-EBI3 and anti-p35 Ab and depletion of IL-35 removed the inhibitory effect of the SN of Treg and naïve T cells from CB, which do not produce IL-35 upon stimulation with R-DC, lack suppressor function. Thus, induction of IL-35 represents a novel activation program in human T cells responding to viral infection.
EBI3 is a member of the IL-12 family. It was first identified in B lymphocytes based on its induction following EBV infection. It codes for a 34 kDa-secreted glycoprotein homologous to the p40 subunit of IL-12. Recent studies have shown that EBI3 can dimerize with IL-12 p35 and EBI3/p35 was called IL-35. The in vivo association between EBI3 and p35 was originally evidenced in human placental extracts 20. Data presented in Fig. 4 and 5 demonstrate that IL-35 and not IL-27 or even IL-12 is responsible for the inhibitory effect of the SN.
More recent studies demonstrated that IL-35 is constitutively expressed by mouse CD4+CD25+FOXP3+ Treg 3, 5. Transcripts coding for EBI3 and p35 were observed to be constitutively coexpressed by mouse Treg and EBI3/p35 heterodimer was coprecipitated from the cell culture SN of these cells. In addition, in vitro and in vivo studies suggested that the expression of IL-35 by mouse Treg contributed to their suppressive function 21. However, human CD4+CD25+FOXP3+ Treg do not constitutively express IL-35 and induction of FOXP3 upregulates neither EBI3 nor p35 mRNA in human T cells 6, 7. Yet, recombinant mouse IL-35 was shown to inhibit the proliferation of mouse effector T cells in vitro. In another recent study, a single chain mouse IL-35-Fc fusion protein was demonstrated to enhance the proliferation of mouse Treg, while inhibiting the development of Th17 cells 5.
The data of this study demonstrate for the first time that IL-35 is a potent regulatory cytokine, also in the human immune system, and that a combinatorial signal delivered from DC to T cells via B7-H1 and sialoadhesin is crucial to the induction of human IL-35+ Treg.
We observe transient FOXP3 expression in T cells stimulated by R-DC as well as DC. Such temporal activation-induced FOXP3 expression in human T cells has been described before and is not obligatory correlated with a regulatory function, whereas natural CD4+CD25+ Treg show constitutive FOXP3 expression 10, 22.
B7-H1, a member of the B7-molecule family, is a well-defined accessory molecule with inhibitory effects through its receptor PD-1 on the T-cell side 23, 24. We, and other groups, have recently demonstrated that B7-H1 is essentially involved in the induction and maintenance of T-cell anergy 25. There is abundant evidence that different viruses abuse B7-H1 to turn-off effector T-cell responses 26–28. The findings of this study imply that B7-H1-mediated inhibition of T-cell responses is, at least partly, due to its capacity to contribute to the induction of IL-35 production. Yet, B7-H1 alone was not sufficient to induce IL-35, but required co-signaling via sialoadhesin. Sialoadhesin, a member of sialic acid binding lectin family of I-type lectins, preferentially binds to sialylated carbohydrate structures (e.g. NeuAcα2,3-Gal) 29 and CD43 has been recently described as ligand for sialoadhesin on T cells 30. Sialoadhesin is a frequently used marker for macrophages because it is typically not expressed on monocytes, lymphocytes, and DC. Yet, type-I IFN have lately been reported to up-regulate sialoadhesin on monocytes 30–33, but also on DC (our unpublished data). Thus, sensing of viral infections by DC leads to the up-regulation of the inhibitory receptor pair B7-H1 and sialoadhesin, which is critical for the induction of IL-35+ Treg.
We have discovered this novel pathway of immune-regulation by analyzing the impact of HRV on DC. HRV are specialized pathogens and only infect humans with all the well-known symptoms of a cold. HRV infection is probably the most frequent human infectious disease, which indicates that the host/HRV relationship is highly evolved. HRV utilizes a variety of tricks to blunt our immune-system and induction of IL-35+ Treg may represent a further prominent immune-evasion mechanism 13. Since induction of B7-H1 and sialoadhesin expression on DC seem to be induced by many other viruses as well, it is intriguing to suggest that the induction of IL-35+ Treg is a general theme in viral infections.
Materials and methods
Media, reagents, and chemicals
Cells were maintained in RPMI 1640 (Gibco, Paisley, Scotland), supplemented with 2 mM L-glutamine, 100 U/mL penicillin, 100 μg/mL streptomycin, and 10% FBS (Sigma-Aldrich, St. Louis, MO, USA). Recombinant human GM-CSF and IL-4 were kindly provided by Novartis Research Institute (Vienna, Austria). The HRV-blocking reagent WIN 52035-2 14 was a kind gift from the Sterling-Winthrop Research Institute (Rensselaer, NY, USA) and was used at a final concentration of 5 μg/mL. IL-10,TGF-β and IL-12 were purchased from R&D Systems (Minneapolis, MN, USA). IFN-α (isoform 2c) was purchased from Boehringer Ingelheim (Vienna, Austria). Monensin, PMA, and Ionomycin were obtained from Sigma-Aldrich. Human IL35:Fc was obtained from Alexis Biochemicals (San Diego, CA, USA).
The following murine mAb were generated in our laboratory: negative control Ab VIAP (against calf intestine alkaline phosphatase), 5-272 (B7-H1), 7-239 (CD169, sialoadhesin), VIT6b (CD1a). The polyclonal murine Ab against EBI3 and GNAI2 (guanine nucleotide binding protein, α-inhibiting activity polypeptide 2; used as isotype control for EBI3) were purchased from Abnova (Taipei, Taiwan). Results presented in Supporting Information Fig. 3 show that the polyclonal anti-EBI3 Ab is specific for EBI3. The monoclonal Ab against p35 (clone27537), IL-12p40 (mAb609), IL-12p70 (mAb611), IL-27 (mAb25261), and TGF-β (mAb240) were purchased from R&D Systems. The neutralizing polyclonal anti–IL-10 Ab (PAL-hIL10) was obtained from Strathmann Biotech (Hannover, Germany). MAb EB-I against IFN-α was kindly provided by G. Adolf (Boehringer Ingelheim).
PBMC were isolated from buffy coats obtained from the Red Cross in Austria. Heparinized whole blood of healthy donors was separated by standard density gradient centrifugation with Ficoll-PaqueTM Plus (GE Healthcare Chalfont St. Giles, UK). Subsequently, T cells (total-CD3+ T cells used unless stated otherwise), CD4+, CD8+ and CD25– T cells, and monocytes were separated by magnetic sorting using the MACS technique (Miltenyi Biotec, Bergisch Gladbach, Germany) as described previously 34. Naïve T cells were isolated from CB. CB samples from healthy donors were collected during healthy full-term deliveries. Approval was obtained from the Medical University of Vienna institutional review board for these studies. CB-T cells used in this study were CD45RA+ (92±3%) and CD45RO−. DC were generated by culturing purified blood monocytes for 7 days with a combination of GM-CSF (50 ng/mL) and IL-4 (100 U/mL). Preparation and purification of rhinoviruses were performed as described 34. DC were treated with HRV14 for 1 day (R-DC) at a titer of 1 TCID50 (50% tissue culture infectious dose) per cell.
T-cell–DC coculture/SN generation
To examine the suppressor activity of the SN of R-DC-induced Treg, T cells were added to R-DC or DC in a 10:1 or 5:1/T-cell:DC ratio. These SN were harvested after 1–3 days of coculture and 100 μL/well were added to different MLR. Centricon YM-50 filters (Millipore, Bedford, MA, USA) were used for size fractionation of the SN. The fraction containing molecules >50 kDa was compared to the fraction containing molecules <50 kDa in an allogeneic MLR. The T cells of the coculture were also investigated by intracellular staining or analyzed via real-time PCR.
T-cell proliferation assays
For the MLR, allogeneic, purified T cells (1×105) were incubated with graded numbers of DC. Experiments were performed in 96-well round bottom cell culture plates in RPMI 1640 medium supplemented with 10% FBS. Proliferation of T cells was monitored by measuring (methyl-3H)TdR (ICN Pharmaceuticals, Irvine, CA, USA) incorporation on day 5 of culture. Cells were harvested 18 h later, and radioactivity was determined on a microplate scintillation counter (Packard Instruments, Meriden, CT, USA). Assays were performed in triplicates. For Fig. 1 preactivated T cells were harvested, irradiated (30 Gy, 137Cs source) and tested for their suppressive function, for Supporting Information Fig. 1 and 4A preactivated T cells were not irradiated. For that purpose, increasing numbers of preactivated T cells were added to an allogeneic MLR with a fixed number of T cells (1×105) and DC (1×104). In neutralization assays Ab were added at final concentration of 10 μg/mL and IL-10, IFN-α, TGF-β were used at 5 ng/mL.
Flow cytometric analysis
For intracellular staining monensin (5 μM) (and for Supporting Information Fig. 4 also PMA/Ionomycin (both 100 nM)) was added to the cells for 12 h. Cells were harvested, fixed with FIX-solution (An der Grub, Kaumberg, Austria) for 20 min, washed twice with PBS, and permeabilized for 20 min with PERM-solution (An der Grub) in the presence of the primary Ab. Oregon Green-conjugated goat anti-mouse Ig Ab from Molecular Probes (Carlsbad, CA) was used as second step reagent. Flow cytometric analysis was performed using a FACScalibur flow cytometer (Becton Dickinson, Franklin Lakes, NJ, USA).
Immunoprecipitation, cytokine depletion, and Western blotting
For immunoprecipitation mAb p35 or mAb VIAP (isotype control) was loaded onto 7×107 sheep anti-mouse IgG coupled Dynabeads (Dynal, Oslo, Norway) with 2.8 μm diameter as described in detail elsewhere 35, 36. After washing twice with PBS, the beads were incubated with cell culture SN for 12 h at 4°C on a rotator. The SN of the beads was considered depleted of p35, p40, or IL-27 and tested in an MLR. The beads themselves were washed twice and a part of the beads (1×106) was analyzed via flow cytometry using a FACScalibur flow cytometer. Therefore beads were incubated for 30 min. at 4°C with unconjugated Ab against EBI3, IL-12p40, IL-27, or isotype control. After washing, Oregon Green-conjugated goat anti–mouse-Ig from Invitrogen (Carlsbad, CA) was used as a second-step reagent. Flow cytometric analysis was performed using a FACScalibur flow cytometer (BD Biosciences, San Diego, CA).
Concerning the rest of the beads bound protein was eluted with reducing sample buffer (Biorad, Richmond, CA, USA) by boiling for 5 min and monitored by Western blot analysis. Western blotting was performed under standard conditions using mAb at 1 μg/mL. Bound mAb were detected using HRP-conjugated goat Ab to mouse Ig (DAKO, Glostrup, Denmark; 1/10000). Signals were detected on Kodak Biomax XAR films (Sigma-Aldrich) and quantified using the ImageJ 1.32 software (National Institutes of Health, Bethesda, MD, USA).
Total cellular RNA was isolated using TRI reagent (Sigma-Aldrich), chloroform extraction, and subsequent isopropanol precipitation according to the manufacturer's protocol. cDNA was generated using the Revert Aid MuLV-RT kit (Fermentas, Burlington, Canada) using Oligo (dT) 18 primers according to the manufacturer's protocol. cDNA was stored at −20°C until use. Quantitative real-time PCR was performed by the Mx3005P QPCR system (Stratagene, Cedar Creek, TX, USA) using Sybr Green detection. In all assays, cDNA was amplified using a standard program (2 min at 50°C, 10 min at 95°C, 40 cycles of 15 s at 95°C/15 s at 60°C/30 s at 72°C). G3PDH was used as a housekeeping gene. The following primers were used: hFOXP3 forward: 59′-GAA ACA GCA CAT TCC CAG AGT TC-3′ and reverse: 5′-ATG GCC CAG CGG ATG AG-3′, hEBI3 forward: 5′-TCT GAG ATC TCT GCC CGC CCT GCA GTG GAA GG-3′ and reverse: 5′-CTT GAG ATC TGC CCA GGC TCA TTG TGG CAG TG-3′; hIL-12 p35 forward: 5′-TTT GCG GCC GCA CCT CCC CGT GGC CAC TCC AG-3′, and reverse: 5′-TTT GCG GCC GCA TTC AGA TAG CTC ATC ACT CT-3′ and hp28 forward: 5′-GCG GAA TCT CAC CTG CCA-3′ and reverse: 5′-GGA AAC ATC AGG GAG CTG CTC-3′, G3PDH forward: 5′-CGACCACTTT GTCAAGCTCA-3′ and G3PDH reverse: 5′-AGGGGAGATT CAGTGTGGTG-3′.
The following cell lines were used in this study: the EBV-transformed lymphoblastoid B cell line (EBV CL) OTMA was generated in our laboratory 37. The Daudi cell line was obtained from American Type Culture Collection (ATCC, Manassas, VA, USA).
Statistical analysis was performed using a two-tailed Student's t test using unpaired nonparametric test (Mann–Whitney). Significance is represented as p<0.05 (*), p<0.01 (**) and p<0.001 (***), n.s. not significant.
The authors thank Petra Cejka, Saro Künig, and Claus Wenhardt for expert technical assistance. This work was supported by a grant of the Austrian Science Fund (FWF, APP20266FW to JS).
Conflict of interest: The authors declare no financial or commercial conflict of interest.