Receptor Expression, Blood Immune Cells, but Not Keratinocytes Or Melanocytes, Have an Impaired Respons

ORIGINAL ARTICLE Despite IFN-l receptor expression, blood immune cells, but not keratinocytes or melanocytes, have an impaired response to type III interferons: implications for therapeutic applications of these cytokines

K Witte1, G Gruetz2, H-D Volk2, AC Looman3, K Asadullah4,5, W Sterry5,6, R Sabat1 and K Wolk1 1Interdisciplinary Group of Molecular Immunopathology, Dermatology/Medical Immunology, University Hospital Charite´, Berlin, Germany; 2Institute of Medical Immunology, University Hospital Charite´, Berlin, Germany; 3Gene Analysis Service, Berlin, Germany; 4Target Discovery, Bayer Schering Pharma Inc., Berlin, Germany; 5German Skin Research Center, Berlin, Germany and 6Department of Dermatology, University Hospital Charite´, Berlin, Germany

Interferon (IFN)-l1, -2 and -3 (also designated as interleukin (IL)-29, IL-28a and IL-28b) represent a new subfamily within the class II cytokine family. They show type I IFN-like antiviral and cytostatic activities in affected cells forming the basis for IFN-l1 therapy currently under development for hepatitis C infection. However, many aspects of IFN-ls are still unknown. This study aimed at identifying the target cells of IFN-ls within the immune system and the skin. Among skin cell populations, keratinocytes and melanocytes, but not fibroblasts, endothelial cells or subcutaneous adipocytes turned out to be targets. In contrast to these target cells, blood immune cell populations did not clearly respond to even high concentrations of these cytokines, despite an IFN-l receptor expression. Interestingly, immune cells expressed high levels of a short IFN-l receptor splice variant (sIFN-lR1/sIL-28R1). Its characterization revealed a secreted, glycosylated protein that binds IFN-l1 with a moderate affinity (KD 73 nM) and was able to inhibit IFN-l1 effects. Our study suggests that IFN-l therapy should be suited for patients with verrucae, melanomas and non-melanoma skin cancers, apart from patients with viral hepatitis, and would not be accompanied by immune-mediated complications known from type I IFN application. Genes and Immunity (2009) 10, 702–714; doi:10.1038/gene.2009.72; published online 1 October 2009

Keywords: interleukin-28; interleukin-29; interferon-b; soluble receptor; IL-22BP; hepatocyte

Introduction their encoding genes, their spatial protein structure and their receptor usage, but their activities mimic many Human interferon (IFN)-l1, -2 -and -3 (also designated as functional aspects of type I IFNs.1,2,6–8 interleukin (IL)-29, IL-28a and IL-28b, or type III IFNs) All three cytokines share the same receptor complex are highly related cytokines that were independently that consists of IFN-lR1 (also designated as IL-28R1) and discovered in 2003 by two research groups.1,2 The amino- IL-10R2,1,2 with IL-10R2 also being part of the receptor acid sequences of IFN-l2 and IFN-l3 are virtually complexes for IL-10, IL-22 and IL-26.4,5 As for the type I identical (94%), those of IFN-l2 and IFN-l1 share 66%. IFN system, engagement of the IFN-l receptor complex These mediators are part of a larger family of structurally leads to the activation of signal transducers and related cytokines called the class II cytokine family, activators of transcription (STAT)1 and STAT2, which, which also contains the IL-10-related cytokines together with IFN regulatory factor-9, form a transcrip- (IL-10, IL-19, IL-20, IL-22, IL-24 and IL-26), type I IFNs tional complex called IFN-stimulated gene factor-3.1,7,9 and IFN-g.3–5 Within the class II cytokine family, IFN-ls This factor induces the so-called IFN-stimulated genes form a theoretical bridge between IL-10-related cytokines (ISGs). Moreover, an engagement of the IFN-l receptor on the one side and type I IFNs on the other, because activates STAT3 and STAT5,1 which also represents they are closer to the first in terms of the structure of a main feature of signal transduction induced by IL-10- related cytokines.5 Up to now, research has mainly focused on the role of Correspondence: Dr K Wolk, Interdisciplinary Group of Molecular IFN-ls in viral infections, which is also an important Immunopathology, Dermatology/Medical Immunology, Campus— feature of the type I IFN action. IFN-ls seem to be Charite´ Mitte, University Hospital Charite´, Charite´platz 1, produced by any nucleated cell type after viral infection D-10117 Berlin, Germany. or viral infection-mimicking agents such as polyinosinic- E-mail: [email protected] 10 Received 30 June 2009; revised 20 August 2009; accepted 31 August polycytidylic acid (poly(I:C)). Aparticularlyhighexpres- 2009; published online 1 October 2009 sion was observed in infected plasmacytoid dendritic cells Target cells of type III interferons K Witte et al 703 (DCs),11 the distribution of which may influence the Results quantity of IFN-l production in a respective tissue. l Moreover, IFN-ls exert direct antiviral activities by Expression of IFN- R1 mRNA in human tissues and in cell inducing antiviral proteins,1,2,6,12–14 and upregulating populations of the immune system and the skin the cellular major histocompatibility complex class I The IFN-ls are newly discovered cytokines, the target (MHC I).1,15,16 Apart from their role during viral cells of which have not been completely uncovered. We infection, IFN-ls, similar to type I IFNs, may be therefore aimed at further characterizing the IFN-l important in fighting against bacterial infections. In fact, targets in the human system. We first analyzed 28 these cytokines are produced by monocytic cells after human tissues for the expression of the IFN-l receptor. activation with bacterial lipopolysaccharides (LPS), and As IL-10R2, one of the two components of this receptor especially during LPS-induced maturation of monocytic complex, which is also shared by other cytokines, was DCs,11,17 and they enhance the expression of bacterial already known to be ubiquitously expressed,23 we component sensing Toll-like receptors.17 Finally, an focused on the expression of the IFN-lR1 receptor antiproliferative effect on some tumor cell lines was component. As demonstrated by real-time PCR on demonstrated for IFN-ls,7,12,18,19 an effect that is also reverse-transcribed cDNA (qPCR), IFN-lR1 could be shared by type I IFNs. detected in all analyzed tissues, although, unlike The overlapping activities exerted by type I and type IL-10R2, the expression level greatly varied among III IFNs would suggest a redundancy between both IFN tissues (Figure 1a). Interestingly, the liver and lungs systems. However, there is growing evidence that in showed the highest levels, and the main organs of the contrast to the lacking cell specificity of type I IFNs, IFN- immune system (spleen, thymus, peripheral blood ls do not target every type of cell.15,19–22 For instance, mononuclear cells (PBMCs)), as well as the skin, were whereas IFN-ls exerted clear effects on different epithe- also among those organs expressing above-average lial cell populations and, at least in humans, on IFN-lR1 levels. In contrast, the brain showed very low hepatocytes, no influence could be detected on endothe- levels (Figure 1a), which is in line with previous results lial cells of the brain and kidney.2,6,12,17,22 A more selective on limited brain cell/tissue receptor expression and target cell range would be of great importance for the IFN-l sensitivity.2,19 possible therapeutic applications of IFN-ls. Although it Because organs are constructed of different cell types, would reduce the number of potential medical indica- we further analyzed the expression of IFN-lR1 in cell tions, the limitation of side effects would represent a big populations of two organs that we are especially advantage over type I IFNs. To better understand which interested in, the immune system and the skin. cells are responsive to the action of IFN-ls, this work Primary human monocytes, B, T and NK cells, as the aimed at identifying target cells of type III IFNs. components of blood immune cells, and primary human

0.1

0.01

expression 0.001

normalized mRNA 0.0001 skin liver lung heart brain testis colon uterus kidney spleen PBMCs thymus trachea prostate placenta stomach fetal liver pancreas fetal brain cerebellum spinal cord bone marrow thyroid gland adrenal gland salivary gland small intestine mammary gland skeletal muscles

1 1 1

0.1 0.1 0.1

0.01 0.01 0.01 expression expression 0.001 expression 0.001 0.001 normalized mRNA normalized mRNA 0.0001 normalized mRNA 0.0001 0.0001 Mo Bc Tc NKc KcEc Fb Mc Ac Hc HepG2 Figure 1 Expression of interferon (IFN)-lR1 mRNA in various human tissues and selected cells. IFN-lR1 expression was analyzed by qPCR and data are given as normalized to hypoxanthine phosphoribosyl-transferase 1 (HPRT) expression. (a) Expression in human tissues. Data from the measurement of samples pooled from different donors are shown. (b) Expression in primary monocytes (Mo), B (Bc), T (Tc) and NK cells (NKc) isolated from peripheral blood mononuclear cells (PBMCs), in primary keratinocytes (Kc), dermal microvascular endothelial cells (Ec), dermal fibroblasts (Fb), melanocytes (Mc) and subcutaneous adipocytes (Ac), as well as in primary hepatocytes (Hc) and in the hepatocyte-derived HepG2 cell line. Data from three (PBMC subpopulations; mean±s.e.m.) or two (liver and skin cells; mean±range) independent experiments are shown.

Genes and Immunity Target cells of type III interferons K Witte et al 704 keratinocytes, dermal microvascular endothelial cells, poly(I:C), a TLR3 ligand. After extensive washing, cells dermal fibroblasts, melanocytes and subcutaneous adi- were stimulated with IFN-l2, IFN-l1andIFN-b1fora pocytes, as the components of skin and subcutis, were further 24 h and assessed for MHC I expression. However, investigated by qPCR. none of the pretreatment protocols could induce a Furthermore, primary human hepatocytes and HepG2 convincing responsiveness of the different PBMC sub- hepatoma cells were included in this study as positive populations to IFN-ls. Figure 3c shows respective data control cells. As demonstrated in Figure 1b, B cells and obtained with monocytes. Interestingly, the specific keratinocytes showed a very high IFN-lR1 expression, prestimulations actually led to a decrease in IFN-lR1 followed by hepatocytes/HepG2, T cells and melano- expression in the different PBMC subpopulations cytes. In contrast, NK cells and monocytes expressed (Figure 3b). rather moderate levels of IFN-lR1, and minimal or even no expression at all was found in microvascular Protein expression of IFN-lR1 and IL-10R2 in blood immune endothelial cells, fibroblasts and adipocytes. cells and keratinocytes The lack or minimal sensitivity of tested immune cells to Sensitivity of selected cell populations toward type III IFNs IFN-ls raised the question of the cause of this phenom- The different cellular IFN-lR1 expression levels sug- enon. A possible explanation is a lacking translation of gested different sensitivities of the analyzed cell popula- IFN-lR1 mRNA into protein in these cells. We therefore tions toward IFN-ls. To prove this assumption, we analyzed the IFN-lR1 protein levels in PBMCs from investigated signal transduction pathways in selected different donors compared with keratinocytes and cell populations. To begin with, total PBMCs, keratino- HepG2 cells using the western blot technique. As shown cytes, melanocytes, dermal endothelial cells and HepG2 in Figure 4, in all cell populations including PBMCs, a cells were stimulated with IFN-l2 and IFN-l1 for 0 clear main band of a size of approximately 70 kDa and, (controls), 10 and 20 min, and were analyzed for with the exception of keratinocytes, some ancillary bands cellular phospho-STAT1, -STAT3 and -STAT5 content probably corresponding to differently glycosylated forms by western blot analysis. As shown in Figure 2a, of membrane-associated IFN-lR1 could be detected. keratinocytes, melanocytes and HepG2 cells responded Whereas the main band was most pronounced in HepG2, to these cytokines by mainly activating STAT1. With the its intensity was comparable or even more pronounced exception of melanocytes, IFN-l1 was much more potent in PBMCs compared with keratinocytes. As expected, than IFN-l2, and signals increased with time. In contrast, all analyzed cells also expressed IL-10R2 (55–60 kDa; stimulation of PBMCs with IFN-l2 and IFN-l1 resulted Figure 4). These results exclude that immune cells fail to in only minimal, if any, STAT1 activation (and no STAT3 express IFN-lR1 protein. or STAT5 activation), whereas IFN-g and IFN-b1 strongly induced STAT1 and STAT3 activation in these cells. In Characterization of a short IFN-lR1 variant as a potential line with the minimal IFN-lR1 expression, neither IFN- mediator of cellular insensitivity toward IFN-ls l2 nor IFN-l1 induced STAT activation in endothelial Another cause of the minimal sensitivity of tested cells (data not shown). We then investigated STAT immune cells toward IFN-ls could be the possible activation by IFN-ls separately in monocytes, B, T and expression of an inhibitory IFN-lR1 variant. In fact, NK cells using intracellular staining of phospho-STAT1 Sheppard et al.2 and Dumoutier et al.24 have described the and -STAT3 and flow cytometry. In these experiments, existence of different IFN-lR1 mRNA splice variants. we increased the concentration of IFN-ls to 100 ng mlÀ1. One of these variants lacks the sequence that is currently We found that IFN-l2 and IFN-l1 did not activate STAT1 designated by the NCBI as the first part of exon VII, or STAT3 at all in monocytes, T or NK cells. In B cells, a resulting in a 29 amino-acid deletion within the reproducible, although minimal, activation of STAT1 and intracellular domain.2 Another splice variant lacks the STAT3 was seen (Figure 2b). In further experiments, an sequence that corresponds to the transmembrane increase in IFN-l1 concentration to even 1 mgmlÀ1 could domain-containing exon VI, resulting in a frameshift not induce STAT1 or STAT3 activation in monocytes or T with an alternative, premature stop codon.24 So far, cells, as shown by the dose–response analysis presented neither the existence as proteins nor the cellular sources, in Figure 2c. expression levels or the function of the alternative IFN- To include a second parameter indicating a possible lR1 forms is known. We therefore established a PCR sensitivity of cells toward IFN-ls, we analyzed the IFN- assay to simultaneously amplify all these IFN-lR1 forms l-dependent increase of MHC I surface expression. In (spanning exons V–VII), and analyzed cDNAs from line with the results obtained from the signal trans- PBMCs, keratinocytes and HepG2 cells in this regard. As duction analysis, IFN-l2 and IFN-l1 showed a dose- demonstrated in Figure 5a, the PCR products obtained dependent upregulation of MHC I surface expression on with all cells each represented a long and a short IFN- HepG2 cells, keratinocytes and melanocytes, whereas 10 lR1 form. Interestingly, keratinocytes and HepG2 cells or 100 ng mlÀ1 of these cytokines did not provoke this expressed predominantly the longer form, whereas effect in the blood immune cell populations (Figure 3a). PBMCs expressed considerably less of the long IFN- As immune cells could be responsive toward IFN-ls lR1 form but more of the shorter form. By sequencing, only after previous stimulation, we then prestimulated we identified the long form as the complete IFN-lR1 PBMCs for 24 h with the following cell-specific stimuli: cDNA fragment, representing the membrane-associated LPS for the stimulation of monocytes, fixed Staphylococcus variant (memIFN-lR1). The shorter form (sIFN-lR1) aureus cells for B cells, immobilized anti-CD3/anti-CD28 turned out to lack the exon VI sequence and, therefore, monoclonal antibodies (mAbs) for the stimulation of corresponds to the mRNA splice variant predicted by T cells and IL-2/IL-12 for the preferential stimulation of Dumoutier et al.24 The deduced protein sequence (244 NK cells. Additional PBMCs were prestimulated with amino acids) lacks the transmembrane and intracellular

Genes and Immunity Target cells of type III interferons K Witte et al 705 PBMCs Keratinocytes Melanocytes HepG2

P-STAT1

P-STAT3

P-STAT5

STAT3

0 10 20 10 20 10 20 10 20 0 10 20 10 20 0 10 20 10 20 0 10 20 10 20 C IFN-λ2 IFN-λ1 IFN-γ IFN-β1 C IFN-λ2 IFN-λ1 C IFN-λ2IFN-λ1 C IFN-λ2IFN-λ1

250 Monocytes B-cells T-cells NK-cells * * * 200 * * * P-STAT1 150 * * * * 100

% of control control 50 IFN-λ2 0 IFN-λ1 250 γ * IFN- 200 * * β * IFN- 150 * * P-STAT3 * * 100 % of control 50 0

Monocytes T-cells

60 P-STAT1 P-STAT3 60 P-STAT1 P-STAT3 IFN-β 40 40 IFN-λ1 MFI MFI 20 20

0 0 0 0.01 0.1 1 0 0.01 0.1 1 00.01 0.1 1 0 0.01 0.1 1 [μg/ml] [μg/ml] [μg/ml] [µg/ml] Figure 2 Sensitivity of immune cells, keratinocytes, melanocytes and HepG2 cells toward interferon (IFN)-ls as assessed by induction of signal transduction. (a) Peripheral blood mononuclear cells (PBMCs), keratinocytes, melanocytes and HepG2 cells were stimulated or not (control) for 10 and 20 min with 40 ng mlÀ1 IFN-l2 or IFN-l1. For PBMCs, stimulation with 10 ng mlÀ1 IFN-g and 500 U mlÀ1 IFN-b1 was additionally included as a positive control. Tyrosine phosphorylation of STAT molecules was assessed by western blot analysis. One representative of two experiments of each cell type is shown. (b) PBMCs were stimulated or not (control) for 20 min with 100 ng mlÀ1 IFN-l2, 100 ng mlÀ1 IFN-l1, 10 ng mlÀ1 IFN-g and 500 U mlÀ1 IFN-b1. Levels of tyrosine phosphorylation of STAT1 and STAT3 in different PBMC subpopulations were assessed by respective intracellular staining and flow cytometry. Data from five independent experiments are given as percentage of untreated controls calculated on the basis of measured mean fluorescence intensities (MFIs; mean±s.e.m.). Significance of differences to respective untreated controls was analyzed using the Wilcoxon matched-pairs signed-rank test (*Po0.05). (c) PBMCs were stimulated for 20 min with 0, 10, 100 and 1000 ng mlÀ1 IFN-l1 or IFN-b1. Monocyte and T-cell tyrosine phosphorylation of STAT1 and STAT3 was assessed as in (b). Data from three independent experiments are given as MFIs calculated as differences between treated and untreated cells (mean±s.e.m.). domain encoding parts and corresponds to the extra- sequence (exons II–VII) and subsequent sequencing (data cellular domain of memIFN-lR1 with the exception of not shown). No further variant was found in any of these the five C-terminal amino acids that are replaced by 21 reactions. amino acids generated by a frame shift (Figures 5b and c). In the next step, we quantified the expression of both Therefore, sIFN-lR1 may represent a secreted, single- IFN-lR1 forms using specific qPCR assays and calcu- chain receptor that should have IFN-l-binding sites lated their expression ratios in separated monocytes, B, identical to memIFN-lR1. The identity of these IFN-lR1 T and NK cells, as well as in keratinocytes, hepatocytes forms was additionally confirmed at the mRNA level by and HepG2 cells. Consistent with our results from the PCR assays covering longer fragments of the IFN-lR1 conventional PCR, the expression ratio between the long

Genes and Immunity Target cells of type III interferons K Witte et al 706 Monocytes B-cells T-cells NK-cells 300 * 250 * * 200 * MHC I * * 150 * * 100 % of control 50 0

Keratinocytes Melanocytes HepG2 control 300 IFN-λ2 10 ng/ml 250 * IFN-λ2 100 ng/ml * * * * 200 * * * λ * * IFN- 1 10 ng/ml MHC I 150 * * IFN-λ1 100 ng/ml 100

% of control γ 50 IFN- 10 ng/ml β 0 IFN- 1 500 U/ml

IFN-λR1 MHC I 1 300 control 250 0.1 control IFN-λ2 200 stimulated IFN-λ1 0.01 150 β 100 IFN- expression 0.001 pretreatment % of control w/o 50 normalized mRNA 0.0001 0 MoBc Tc NKc LPS poly(I:C) Figure 3 Sensitivity of immune cells, keratinocytes, melanocytes and HepG2 cells toward interferon (IFN)-ls as assessed by the upregulation of major histocompatibility complex class I (MHC I) expression. (a) Peripheral blood mononuclear cells (PBMCs), keratinocytes, melanocytes and HepG2 cells were stimulated or not (control) with 10 and 100 ng mlÀ1 IFN-l2 and IFN-l1 for 24 h. For PBMCs, stimulation with 10 ng mlÀ1 IFN-g and 500 U mlÀ1 IFN-b1 was additionally included as positive controls. Cellular MHC I expression was assessed by respective cell-surface staining and flow cytometry. Data from five independent experiments are given as percentage of untreated control calculated on the basis of mean fluorescence intensities (MFIs; mean±s.e.m.). Significance of differences to respective untreated controls was investigated by use of the Wilcoxon matched-pairs signed-rank test (*Po0.05). (b) PBMCs were stimulated with 100 ng mlÀ1 bacterial LPS (stimulation of monocytes), with heat-killed, formalin-fixed S. aureus cells (stimulation of B cells), IL-2/IL-12 (stimulation of NK cells) and anti-CD3 mAb/anti-CD28 mAb (stimulation of T cells) for 6 h. IFN-lR1 expression was analyzed by qPCR and representative data from one out of two experiments are given as normalized to hypoxanthine phosphoribosyl-transferase 1 (HPRT) expression. (c) PBMCs were prestimulated with bacterial LPS, poly(I:C) or were left without prestimulation for 24 h. After washing, prestimulated samples were restimulated or not (control) with 100 ng mlÀ1 IFN-l2, 100 ng mlÀ1 IFN-l1 and 500 U mlÀ1 IFN-b1 for 24 h. Monocyte MHC I expression was assessed by respective cell-surface staining and flow cytometry. Data from three independent experiments are given as percentage of untreated control without prestimulation, calculated on the basis of MFIs (mean±s.e.m.).

and the short form was cell specific (Figure 5d), whereas keratinocytes, hepatocytes and HepG2 cells showed mean expression ratios of higher than seven (7.3, 10.4 and 15.2, respectively); the ratio in the different immune PBMCs1PBMCs PBMCs2 Keratinocytes3 HepG2 80 cell populations was as low as 1.6–3.6. This finding 60 provoked the hypothesis that for the sensitivity of IFN-λR1 l 50 cells to IFN- s not only the expression level of the signaling competent IFN-lR1 receptor but also the 80 memIFN-lR1/sIFN-lR1 expression ratio may be of IL-10R2 60 importance. 50 To investigate whether this hypothesis is true, in the 40 GAPDH next part of our study, we biochemically and functionally characterized sIFN-lR1. For that purpose, this splice

Figure 4 Peripheral blood mononuclear cells (PBMCs), keratino- variant fused to a His6 and a V5 tag was transiently cytes and HepG2 cells express interferon (IFN)-lR1 and IL-10R2 overexpressed in HepG2 cells by transfection. As a protein. Protein expressions of IFN-l receptor components and of control, a construct containing His6- and V5-tagged IL-22 the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were analyzed in the cell lysates of PBMCs from three binding protein (IL-22BP), the secreted single-chain donors as well as of keratinocytes and HepG2 cells using the receptor for IL-22, was transfected in parallel. Western western blot technique. blot analysis using anti-V5 tag polyclonal antibodies

Genes and Immunity Target cells of type III interferons K Witte et al 707

Figure 5 Detection of a short interferon (IFN)-lR1 splice variant with different relative expression levels in immune cells, keratinocytes and liver cells. (a) A PCR assay was performed using primers spanning IFN-lR1 exons V–VII. PCR products (941 and 810 bp) obtained with HepG2 cells, keratinocytes and peripheral blood mononuclear cells (PBMCs), and separated by gel electrophoresis, are shown. (b) On the basis of PCR fragment sequences generated in (a), deduced mRNA structures and underlying exons of the two detected splice forms are presented schematically. Whereas the long form corresponds to the full length, cell membrane-associated receptor chain (memIFN-lR1), the shorter form turned out to be a splice variant that lacks exon VI, resulting in a truncated receptor chain without transmembrane or intracellular domain-coding sequences (sIFN-lR1). For memIFN-lR1, the predicted protein domains are indicated as well: SP, signal peptide; ED, extracellular domain; TD, transmembrane domain; ID, intracellular domain. (c) Deduced amino-acid sequences and protein domain structures of memIFN-lR1 and sIFN-lR1 are shown. Predicted protein domains are indicated in blue (SP), green (ED), black (TD), red (ID) or magenta (altered amino-acid sequence because of changed reading frame shift). Potential N- and O-glycosylation sites are in bold and boxed. (d) The expression of both memIFN-lR1 and sIFN-lR1 was quantified in separated monocytes (Mo), B (Bc), T (Tc) and NK (NKc) cells, as well as in keratinocytes (Kc), hepatocytes (Hc) and HepG2 cells using variant-specific qPCR assays, and expression of sIFN-lR1 relative to that of memIFN-lR1 was calculated. Data from two (HepG2; mean±range) or three (all other cell types; mean±s.e.m.) independent experiments are given.

(pAbs) showed that the sIFN-lR1 variant, similar to as a control, IL-22BP were run at different concentrations IL-22BP, could be detected in both the cell lysate and the over a sensor chip surface coated with IFN-l1 in the first cell culture supernatant of transfected HepG2 cells flow cell and IL-22 in the second flow cell. The increase (Figure 6a). These data show that sIFN-lR1 is in fact a in resonance units (RU) was recorded as a measure of the soluble receptor as it is at least partly secreted. The density of receptor molecules bound to the chip in a occurrence of several bands in these analyses addition- defined period of time. As shown in Figure 6c, sIFN-lR1 ally suggested that sIFN-lR1 is a glycosylated protein. Its dose dependently bound to immobilized IFN-l1, but not glycosylation pattern in the cell lysate, similar to that of to immobilized IL-22. It is noteworthy that binding to IL-22BP, differed from that found in the cell culture IFN-l1 seemed to be less pronounced than that of supernatant, suggesting that these proteins are not IL-22BP to IL-22. secreted until full glycosylation is achieved. Enzymatic Soluble receptors may act as negative regulators of N-deglycosylation led to clear bands of about 32 kDa for cytokine actions by preventing their binding to the both the IFN-lR1 variant and IL-22BP, proving respective membrane-bound receptor. Most of such soluble recep- glycosylation (Figure 6b). tors, including IFNaRba, are splice variants correspond- We then investigated whether the recombinant-pro- ing to the extracellular moieties of respective membrane- duced sIFN-lR1 was able to bind IFN-l1. By means of associated receptor variants.25 To get further insight into the surface plasmon resonance technique, sIFN-lR1 and, the biological role of sIFN-lR1, we aimed at investigating

Genes and Immunity Target cells of type III interferons K Witte et al 708 native N-deglycosylated 210 210

100 100 80 80 60 60 50 50 40 40 30 30

20 20

cell lysate SN cell lysate SN R1 R1 R1 R1 λ λ λ λ λ control control control control IL-22BP IL-22BP IL-22BP IL-22BP sIFN- sIFN- sIFN- sIFN-

λ 150 sIFN- R1 400 IL-22BP

300 100 RU] RU] Δ Δ 200 50 Binding [

Binding [ 100

0 0 040 80 120 160 08020 40 60 [μg/ml] [µg/ml] IFN-λ1 immobilized IL-22 immobilized Figure 6 Overexpressed sIFN-lR1 is a glycosylated, secreted molecule that binds interferon (IFN)-l1. HepG2 cells were transfected with

plasmids encoding either sIFN-lR1 or, as a control protein, IL-22BP, both as V5-His6-tagged constructs. (a) Overexpressed proteins were analyzed in cell culture supernatants and cell lysate using V5-tag on the basis of the western blot technique. (b) Cell lysates and cell culture supernatants of transfected cells were treated with N-deglycosylating enzyme and analyzed as in (a). (c) sIFN-lR1 and, as a control, IL-22BP

were purified from the cell culture supernatant of respective overexpressing HepG2 cells on the basis of their His6 tag and, using the surface plasmon resonance technique, were run at different concentrations over IFN-l1/IL-22-coated sensor chip surfaces, and the dose-dependent increase in RU (DRU) as a measure of the binding was recorded. Data from four (interaction between sIFN-lR1 and IFN-l1, mean±s.e.m.), two (interaction between IL-22BP and IL-22; mean±range) and one experiment (interaction between sIFN-lR1 and IL-22 and between IL-22BP and IFN-l1) are shown.

(1) its affinity to IFN-l1 and (2) its ability to inhibit the candidate responsible for the observed lack/minimal cellular IFN-l1 action. As the answers to these questions response of investigated immune cells to IFN-ls. require a high amount of recombinant sIFN-lR1, for these experiments, we used a commercially available Fc- coupled molecule construct (sIFN-lR1-Fc). The affinity between sIFN-lR1-Fc (immobilized) and IFN-l1 was Discussion measured by means of the surface plasmon resonance Type I IFNs are mediators that are very important for technique. Calculated on the basis of their interaction on first-line antiviral and antitumor effector mechanisms, binding equilibrium, the KD value was approximately and they also show immunomodulatory activities. Their 73 nM (Figure 7a). importance is especially reflected by their widespread To investigate whether sIFN-lR1 would inhibit the clinical use in viral infections such as hepatitis C and B IFN-l1 action on cells, we used HepG2 cells. These cells viruses, in cancers such as melanoma and hairy cell were stimulated with 10 ng mlÀ1 IFN-l1 or 500 U mlÀ1 leukemia and in multiple sclerosis.26–29 However, a major IFN-b1, which had been preincubated with 1 and limitation of type I IFN therapy is the frequent 10 mgmlÀ1 of sIFN-lR1-Fc. The cytokine-induced MHC occurrence of side effects: in particular influenza-like I increase was assessed by flow cytometric analysis as symptoms, fatigue, suppression of bone marrow, induc- before. As demonstrated in Figure 7b, sIFN-lR1-Fc dose tion or exacerbation of autoimmune disorders, neuro- dependently prevented the IFN-l1-induced upregulation toxic reactions and emotional depression. Moreover, of MHC I. The prevention obtained with 10 mgmlÀ1 was people may become resistant to type I IFNs.30 Therefore, approximately 70%. No influence of sIFN-lR1-Fc on the the therapeutic potential of the only recently discovered IFN-b1-induced MHC I upregulation was observed type III IFNs IFN-l1, 2 and 3, which seem to mimic the (Figure 7b). direct antiviral and antitumor activities of type I IFNs, These data suggest that sIFN-lR1, at least at high local seems to be very promising. The antiviral effects of IFN- concentration, inhibits IFN-l1 action and may be a ls were so convincing that an application of IFN-l1to

Genes and Immunity Target cells of type III interferons K Witte et al 709 and IFN-l2 were similarly potent in melanocytes. In contrast to keratinocytes and melanocytes, IFN-lR1 cytokine receptor K [nM] D expression in dermal endothelial cells and fibroblasts and in subdermal adipocytes was very low or even IFN-λ1 sIFN-λR1-Fc 73 ±4.1 absent, and endothelial cells did not respond to IFN-ls (fibroblasts and adipocytes were not tested). Interestingly, with the exception of a very weak IL-22 IL-22BP-Fc ~1(23) response of B cells, human blood immune cells did not respond to type III IFNs. In these cells, cytokines neither IFN-α1 sIFN-αRB 100 (46) activated typical signal transduction elements such as IFN-α2 STAT 1 or 3 nor caused MHC class I upregulation even at sIFN-αRB 0.6 –5 (45-49) very high doses. In the literature, the responsiveness of

IFN-β sIFN-αRB > 0.4 (45, 46, 48, 49) immune cells was very controversially discussed: for example, Ank et al.20 demonstrated that no relevant ISG induction could be achieved by these cytokines in mouse macrophages or myeloid DCs. Moreover, Lasfar et al.15 MHC I 300 demonstrated the lack of STAT activation by type III IFNs * in mouse splenocytes and bone-marrow-derived macrophages. Sommereyns et al.,22 by using plasmid electro- 200 * control transfer-mediated expression of IFN-l3 and measuring λ IFN- 1 the IFN-l in vivo response by monitoring organ ISG 100 IFN-β % of control expression, found no clear response in the spleen. In contrast to these investigations, some studies did 0 propose direct effects of IFN-ls on monocytes,32,33 sIFN-λR1-Fc: 1 µg/ml +++ DCs34 and T cells.35–40 We currently cannot explain the 10 µg/ml +++ discrepancy in these observations in terms of whether IFN-ls influence, for example, monocytes or T cells. Figure 7 Interferon (IFN)-lR1 shows a moderate affinity to IFN-l1 As we demonstrated, it seems that the preactivation and is able to inhibit IFN-l1 action. (a) Using the surface plasmon resonance technique, IFN-l1 at different concentrations was run with cell-specific stimuli or virus-mimicking agents over sensor chip surfaces coated with sIFN-lR1-Fc (second flow could not confer any clear responsiveness to these cell)/IL-22BP-Fc (first (reference) flow cell), and steady-state affinity cytokines on peripheral immune cells. However, these was evaluated. The KD value is presented from two independent data do not totally exclude any sensitivity to IFN-ls, experiments as the mean±range. As a comparison, published data for example, occurring in small subpopulations or regarding the affinity between IL-22BP-Fc and IL-22 and between achieved on special activation/differentiation states of sIFN-aRB and selected type I IFNs, as well as references of underlying studies, are indicated. (b) HepG2 cells were stimulated immune cells. or not (control) with 10 ng mlÀ1 IFN-l1 and 500 U mlÀ1 IFN-b1, Interestingly, the lack of or minimal immune cell which had been preincubated for 2 h with 0, 1 and 10 mgmlÀ1 sIFN- responsiveness to IFN-ls observed in our study was not lR1-Fc. Major histocompatibility complex class I (MHC I) cell- simply because of the lack or minimal expression of the surface expression was measured by flow cytometry. Data from six respective receptor in the immune cells. In fact, of all cell independent experiments are given as a percentage of the respective populations tested in our study, B cells even showed the group without sIFN-lR1-Fc, calculated on the basis of mean l fluorescence intensities (MFIs; mean±s.e.m.). Significance of highest IFN- R1 mRNA expression and the expression in differences to the respective untreated control was analyzed using T cells was at least comparable with those of melanocytes the Wilcoxon matched-pairs signed-rank test (*Po0.05). and hepatocytes/HepG2 cells. We also clearly detected IFN-lR1 protein in PBMCs. A similarly lacking response of cells that express significant levels of IFN-lR1 mRNA was previously recognized by Meager et al.19 on analysis treat chronic hepatitis C virus infection is currently being of a range of human tumor cell lines. developed by ZymoGenetics Inc. (Seattle, WA, USA). Many different mechanisms could be responsible for However, many aspects of the biology of IFN-ls, the fact that, despite a clear receptor expression, immune including their cellular target specificity, are not fully cells do not respond to IFN-ls. These mechanisms understood. That means that we can neither estimate the include the possible lack of surface expression of undesired side effects of IFN-l treatment nor do we have IFN-lR1 or IL-10R2; the existence of alternative receptor a reliable overview of additional, potential fields of a glycosylation patterns determining the functionality therapeutic application. of the receptor; the existence of intracellular factors Aiming to identify target cell populations especially that negatively regulate IFN-lR1 signal transduction and within the human immune system and the skin for these the presence of a blocking decoy or soluble receptor for mediators, in this study we observed that keratinocytes IFN-ls. We, along with others, demonstrated that and melanocytes expressed significant levels of IFN-lR1 immune cells clearly express IL-10R2 mRNA23,41,42 and and clearly responded to IFN-l stimulation. Although protein (Figure 4), and that these cells react to IL-10,41 this sensitivity of keratinocytes has already been de- which proves the functionality of IL-10R2 on the cell scribed by our group,17 our study now revealed for the surface. first time that IFN-ls influence human melanocytes. In In our study, we focused on the investigation of the line with previously published data,19,31 IFN-l1 was possible expression of nonfunctional or blocking much more potent than IFN-l2 in keratinocytes, as was IFN-lR1 variants. No expression of mRNA encoding the case in hepatocytes. Interestingly, however, IFN-l1 a cell-associated IFN-lR1 variant with a truncated

Genes and Immunity Target cells of type III interferons K Witte et al 710 intracellular domain that could function as a decoy Materials and methods receptor was found by us in any of the cells tested. However, we actually found an increased relative Cell culture expression of a very short variant in immune cells. The Peripheral blood mononuclear cells were separated from recombinant expression of the respective protein re- the blood of healthy donors by Ficoll-Paque density- vealed a glycosylated, secreted molecule. We proceeded gradient centrifugation (Pharmacia, Freiburg, Germany). on the assumption that the affinity of sIFN-lR1 to IFN-l1 For some experiments, monocytes, B-, T- and NK cells may be very similar to that of sIFN-lR1-Fc and that of were isolated from PBMCs using the MACS system memIFN-lR1, because, with the exception of the five (Miltenyi Biotec, Bergisch-Gladbach, Germany), and C-terminal amino acids, the sequence of sIFN-lR1 either positive (monocytes, B cells, T cells used without comprises that encoding the extracellular domain of culture in Figures 1b and 5d) or negative (NK cells used memIFN-lR1. We demonstrated that this protein was without culture in Figures 1b and 5d, and all subpopula- able to bind IFN-l1 with a moderate affinity. In fact, the tions used for stimulation experiments in Figure 3b)

KD value of 73 nM indicated an approximately 100 times selection was used. Cells were either prepared directly lower binding than that for the interaction of IL-22 with after separation for IFN-lR1 mRNA expression analysis its potently inhibiting soluble receptor IL-22BP, as or, for stimulation experiments, were cultured under low previously determined by our laboratory.43 Moreover, endotoxin conditions as described previously.49 Primary the sIFN-lR1-Fc-IFN-l1 binding had a similar (IFN-a1) human hepatocytes were obtained from Lonza (Verviers, or lower (IFN-a2, IFN-b) affinity compared with the Belgium) and cultured with HGM medium (Lonza) on binding of the extracellular domain of the R1-type IFN thin-layer Biocoat Matrigel (BD Biosciences, Heidelberg, receptor subunit IFN-aRB to the respective type I IFN44–48 Germany) according to the manufacturer’s instructions. (Figure 7a). Our yet preliminary data also indicated that HepG2 cells were purchased from the European Collec- the affinity of sIFN-lR1-Fc to IFN-l1 was similar to the tion of Cell Cultures (Salisbury, UK) and maintained affinity to IFN-l2. according to the supplier’s instructions. Primary human Importantly, sIFN-lR1-Fc was able to inhibit cellular epidermal keratinocytes and dermal fibroblasts were IFN-l responses effectively, suggesting that it may exert obtained from Lonza and cultured in KGM and FGM-2 an effect, at least under some circumstances, as a medium, respectively, from the same supplier. Adipo- negative regulator of the action of these cytokines. cytes were generated from primary human subcutaneous Therefore, although a threshold memIFN-lR1 level preadipocytes (Lonza) by culturing them for 10 days seems to be the prerequisite for cellular IFN-l responses,9 with PGM-2 differentiation medium (Lonza). Primary the ratio of the local expressions of memIFN-lR1 and human melanocytes were purchased from Invitrogen sIFN-lR1 should also be of importance. The elevated (Karlsruhe, Germany) and cultured in the respective expression of an sIFN-lR1 splice variant may therefore medium (Invitrogen) according to the supplier’s instruc- represent one, but not necessarily the exclusive mechan- tions. Primary human dermal microvascular endothelial ism of the minimal IFN-l sensitivity of immune cells. cells were purchased from Invitrogen and ProVitro However, there remain some questions to be answered (Berlin, Germany) and cultured according to the suppli- in further studies. For instance, are there other mechan- ers’ instructions. All these cells were used for direct isms that contribute to the minimal IFN-l sensitivity of analysis or stimulation experiments after a 24–48 h immune cells? Are the sIFN-lR1 levels secreted in vivo preculture period. For western blot analysis of STAT by immune cells high enough to also influence the phosphorylation, PBMCs, after a preculture period of responsiveness of neighboring cells? Are there mechan- 1.5 h, as well as keratinocytes, melanocytes, endothelial isms to concentrate sIFN-lR1 around the cells that cells and HepG2 cells, were each stimulated with produce it? 40 ng mlÀ1 IFN-l2 and 40 ng mlÀ1 IFN-l1 for 10 and In summary, our study supports that IFN-ls do not 20 min, or were left unstimulated (controls); for PBMCs, influence every type of cell and that memIFN-lR1 additional stimulation with 10 ng mlÀ1 IFN-g or with expression is not sufficient to ensure cellular sensitivity 500 U mlÀ1 IFN-b1 was performed. For flow cytometric toward these cytokines. In particular, the data show that analysis of STAT phosphorylation in PBMC subpopula- immune cells are not major targets of IFN-ls, and that tions, we stimulated PBMCs, after being precultured for sIFN-lR1 may be a factor contributing to the cellular 1.5 h, if not indicated otherwise (as in Figure 2c), with insensitivity toward these cytokines. In contrast to 100 ng mlÀ1 IFN-l1, 100 ng mlÀ1 IFN-l2, 10 ng mlÀ1 IFN-g immune cells, a couple of skin cell populations, namely, or 500 U mlÀ1 IFN-b1 for 20 min, or were left unstimu- keratinocytes and melanocytes, are major targets of lated (controls). For the flow cytometric quantification of these cytokines. We conclude that type III IFNs may MHC I expression, PBMCs, keratinocytes, melanocytes be promising candidates for therapeutic application and HepG2 cells were stimulated or not (control) not only in viral hepatitis but also in epidermal viral each with 10 and 100 ng mlÀ1 of IFN-l2 and IFN-l1 for infections such as human papillomavirus-associated 24 h; PBMCs were additionally stimulated with verrucae, in epidermal carcinoma such as basal cell and 10 ng mlÀ1 IFN-g and 500 U mlÀ1 IFN-b1. In a separate squamous cell carcinoma, as well as in melanoma. experimental setting, PBMCs, before measuring their Because cells of the immune system and of the central MHC I response to IFN-l2, IFN-l1 and IFN-b1 with nervous system are not major targets of type III IFNs, respect to different subpopulations, were prestimulated a therapeutic application of these cytokines should not with 100 ng mlÀ1 LPS from Escherichia coli 0127:B8 be accompanied by flulike symptoms, induction or (Sigma-Aldrich Chemie, Deisenhofen, Germany) (stimu- exacerbation of autoimmune disorders or neurotoxic lation of monocytes); with 0.001% (w/v) heat-killed, reactions. Further studies are needed to prove the formalin-fixed S. aureus cells (PANSORBIN; Calbiochem- efficacy of type III IFNs in treating these skin disorders. Novabiochem, Bad Soden, Germany) (stimulation of B

Genes and Immunity Target cells of type III interferons K Witte et al 711 cells); 10 ng mlÀ1 IL-2/10 ng mlÀ1 IL-12 (stimulation of approved by sequencing by means of the ABI Prism NK cells); anti-CD3 mAb (R&D Systems, Wiesbaden- BigDye Terminator cycle sequencing kit and the ABI 310 Nordenstadt, Germany) and anti-CD28 mAb (Orthoclone capillary electrophoresis genetic analyzer (both from OKT3; Jannsen Cilag, Sulzbach, Germany) coated on Applied Biosystems, Darmstadt, Germany). Restriction culture vessel (each at 1 mgcmÀ2) (stimulation of T cells); endonucleases were purchased from New England Biolabs, 10 mgmlÀ1 poly(I:C) (Sigma-Aldrich) as well as with except AaRI, which was purchased from Fermentas medium, for 24 h, and were then extensively washed. In (St Leon-Rot, Germany). addition, isolated monocytes, B, T and NK cells were stimulated with LPS, S. aureus, IL-2/IL-12 and anti- Purification of recombinant sIFN-lR1 and IL-22BP CD3/anti-CD28 mAb as described above for 6 h, and For overexpression of recombinant sIFN-lR1 and IL- were lysed for IFN-lR1 mRNA analysis. 22BP and subsequent purification, we adapted HepG2 For recombinant expression of sIFN-lR1 and IL-22BP cells to serum-free EMEM medium (Sigma-Aldrich), in HepG2 cells, we performed cellular transfection with supplemented with SynQ (Cell Culture Services, corresponding plasmids (see below) using Lipofectamine Hamburg, Germany) and transfected afterward as 2000 (Invitrogen) according to the manufacturer’s in- described above, whereby the medium was changed structions after subsequent culturing in Eagle’s Minimal 6 h after transfection to a new medium supplemented Essential Medium (EMEM; Sigma-Aldrich) supplemen- with gentamycine, amphotericin B (Lonza) and protease ted with 10% fetal bovine serum and 2 mM glutamine inhibitors (P1860; Sigma-Aldrich). Culture supernatants (both from Biochrom, Berlin, Germany), gentamycine were collected 4 days after transfection and dialyzed sulfate and amphotericin B (both from Lonza) for 72 h. against a specific native binding buffer (25 mM TRIS-HCl Afterward, cell culture supernatants were collected and (pH 8.0), 500 mM NaCl, 2 mM Imidazole; Amocol, Teltow, cells were lysed for protein analyses. Germany). Overexpressed sIFN-lR1 and IL-22BP were To investigate the influence of sIFN-lR1 on IFN-l1- purified by His-tag using cobalt affinity chromatography induced MHC I expression, we stimulated HepG2 cells columns (Amocol) under native conditions according to for 24 h with 10 ng mlÀ1 IFN-l1, 500 U mlÀ1 IFN-b1ora the manufacturer’s instructions, whereby all buffers respective volume of culture medium, each of which had used were supplemented with protease inhibitors been preincubated with 1 or 10 mgmlÀ1 of sIFN-lR1-Fc (Complete EDTA-free; Roche Diagnostics, Mannheim, (R&D Systems) for 2 h. The sIFN-lR1-Fc construct Germany). Concentrations of protein solutions were contains two bound copies of the complete extracellular obtained by repeated ultracentrifugation using Amicon domain of IFN-lR1 linked by a peptide of six amino Ultra-4 devices (Millipore, Schwalbach/Ts, Germany), acids (IEGRMD) to a fragment (P100 to L330) of human whereby the buffer was changed to phosphate-buffered immunoglobulin (Ig)G1. All cytokines and antibodies saline. The quality of purified protein was checked by mentioned above were purchased from R&D Systems. western blot analysis as described below and protein All experiments in a given test series on immune cells concentration was assessed using BCA protein quantifi- and keratinocytes were conducted with cells from cation assay (Perbio, Bonn, Germany). different donors. Western blot analysis Plasmid construction and transfection For these analyses, cell lysis, measurement of protein Plasmids encoding either sIFN-lR1 (NM_173065) or IL- concentration, protein electrophoresis and western blot- 22BP (NM_181309) (open reading frame (ORF) 1) under ting were performed as described previously.50 For signal the control of the human EF-1a promoter, which in transduction analyses, blotted samples were incubated addition contain the sequence of a His6 and V5 protein with pAbs against phospho-STAT1 (Tyr701), phospho- tag (ORF 2), were purchased from RZPD (Berlin, STAT3 (Tyr705), phospho-STAT5 (Tyr694) or total STAT3 Germany). To fuse the two ORFs, the stop codon of the (all from Cell Signaling Technology, Beverly, MA, USA), first ORF was deleted by the use of site-directed and, as secondary antibodies, with peroxidase-conju- mutagenesis as follows—first, PCR was performed gated AffiniPure goat anti-rabbit IgG (H and L; Dianova, using the following primers: 50-CTCTAAGCCCACCT Hamburg, Germany), followed by enhanced chemilumi- GCTTCTTGCTGGAGGT-30 and 50-GGCTTACCTTCGAA nescence (ECL) detection (Amersham Pharmacia Bio- TTCACGGACTCTGGTCTGCTG-30 (amplification of tech, Freiburg, Germany). For the detection of IFN-l sIFN-lR1 with modified 30 end), 50-GAATATAACCCAA receptor components, we incubated blots with anti-IFN- GTCAATGGCTCTTTGTTGG-30 and 50-GGCTTACCTTCG lR1 pAbs (50 ng mlÀ1; Sigma-Aldrich), anti-IL-10R2 AATGGAATTTCCACACATCTCTC-30 (amplification of pAbs (600 ng mlÀ1; R&D Systems) and anti-GAPDH IL-22BP with modified 30 end). PCR conditions were as mAb (200 ng mlÀ1; clone 6C5; Millipore), followed by follows: initial denaturation at 94 1C for 2 min, followed by peroxidase-conjugated AffiniPure goat anti-rabbit IgG a two-step cycling consisting of 10 cycles with 30 s at 94 1C, (H and L), rabbit anti-goat IgG (H and L) and goat anti- 30 s at 60 1Cand30sat681C, and a further 30 cycles with mouse IgG (H and L) (all from Dianova) incubation, 30 s at 94 1C, 30 s at 65 1Cand30sat681C. The obtained respectively, and final ECL detection. For the detection of PCR products and complete vectors were cut with AaRI V5-tagged sIFN-lR1 and IL-22BP, we used cell lysates and BstBI (sIFN-lR1) or BstXI and BstBI (IL-22BP), and (generated as described in Wolk et al.50) and culture vector backbone fragments were ligated to digested PCR supernatants, native or treated with PNGase F (New products using T4 DNA-Ligase (New England Biolabs, England Biolabs) to remove N-linked carbohydrates, for Frankfurt am Main, Germany). Final plasmids were immunoblotting (30 mg total protein containing cell lysate propagated in E. coli (One Shot MAX Efficiency DH5a-T1; and 20 ml culture supernatant per lane for untreated Invitrogen) and reisolated (EndoFree Plasmid Maxi Kit; samples; 40 mg total protein containing cell lysate and Qiagen, Hilden, Germany). Successful mutagenesis was 15 ml culture supernatant per lane for deglycosylated

Genes and Immunity Target cells of type III interferons K Witte et al 712 samples). Blotted samples were then incubated with anti- and PCR was performed. First amplification was V5 mAb (Invitrogen), followed by goat anti-mouse IgG designed to cover sequences spanning exons V–VI, (H and L; Dianova) incubation and ECL detection. As a whereby the following primers and conditions were positive control, 10 ng per lane of the 53 kDa recombinant used: 50-CAGCCAGTCCAGATCACTCTCC-30 and 50- positope protein (Invitrogen) was used. TCCTCAATTTCTGATTCCCTCG-30 (Metabion, Martins- ried, Germany); initial denaturation for 3 min at 94 1C, Flow cytometric analysis followed by 35 cycles with 30 s at 94 1C, 25 s at 62 1C and

For the detection of intracellular phospho-STAT1 and 45 s at 72 1C; 1.5 and 1.7 mM MgCl2. A second amplifica- phospho-STAT3 in PBMCs, we fixed cells for 10 min (BD tion was designed to cover sequences spanning exons Cytofix; BD Biosciences), centrifugated at 300 g for 5 min II–VII using the following primers and conditions: and then permeabilized for 30 min on ice (BD Perm 50-CAGAATGTGACGCTGCTCTC-30 and 50-GGACATT Buffer III, BD Biosciences), followed by two washing GGGGCATTACATG-30; initial denaturizing for 3 min at steps (phosphate-buffered saline with 2% fetal bovine 94 1C, followed by 40 cycles at 94 1C for 30 s, 62 1C for

serum and 2 mM NaN3) and subsequent incubation with 25 s, 72 1C for 60 s, followed by 3 min final extension at

R-phycoerythrin (PE)-labeled phospho-STAT1 (clone 4a) 72 1C, using 1.5 mM MgCl2. Amplifications were run in a and phospho-STAT3 (clone 4) for 30 min in the dark at GeneAmp PCR system 9700 (Applied Biosystems). PCR room temperature. In all analyses of PBMCs, cells were products were separated on 1% agarose gel in Tris- stained in parallel with the following mAbs to allow acetate-EDTA buffer, purified from agarose gel pieces electronic gating and separate evaluation of subpopula- using the Easy Pure DNA purification kit (Biozym tions: fluorescein isothiocyanate-labeled HLA-DR, Diagnostik, Hess. Oldendorf, Germany) and checked DQ, DP (clone Tu¨ 39), allophycocyanin-labeled CD3 directly (first PCR) or after reamplification (second PCR) (clone SK7; BD Biosciences), PE-cyanin 5-labeled CD14 by sequencing using the ABI Prism BigDye Terminator (clone RMO52), fluorescein isothiocyanate-labeled CD16 cycle sequencing kit and the ABI 310 capillary (clone 3G8), CD56 (clone NCAM16.2) and CD57 (clone electrophoresis genetic analyzer (both from Applied NC1) (Beckmann Coulter, Krefeld, Germany). Biosystems). Cell-surface expression of MHC I was detected using PE-labeled anti-HLA-A, B, C mAb (clone G46-2.6) and Binding studies PE-labeled IgG1/k isotype mAb (clone MOPC-21) (BD For the investigation of cytokine-soluble cytokine recep- Biosciences). Specific expression was calculated as the tor interactions, we performed kinetic binding experi- difference between the mean fluorescence intensity (MFI; ments using surface plasmon resonance technology as geomean) obtained with the specific and isotype mAb. described previously.43 All measurements were per- All analyses were performed by means of a FACSCalibur formed using a Biacore X instrument, associated Biacore instrument and CellQuest software (BD Bioscience). X 2.2 control software, CM5 sensor chips and HBS-EP as running buffer (all from Biacore, Freiburg, Germany). In qPCR brief, IFN-l1 was immobilized on the sensor chip Panels of mostly pooled total RNA from human tissues (Biacore) surface of the first flow cell, whereas IL-22, were obtained from Clontech Laboratories (Heidelberg, immobilized on the second flow cell, served as a control Germany). The number of donors that each specific RNA protein. Ligand immobilization was achieved by coating was derived from was as follows: adrenal gland, 6; bone 25 mgmlÀ1 cytokine in 10 mM sodium acetate buffer (pH marrow, 69; brain, 2; fetal brain, 24; cerebellum, 1; colon, 4.5) for 16 (IFN-l1) or 20 min (IL-22) by an amine 2; heart, 1; kidney, 6; liver, 1; fetal liver, 63; lungs, 2; coupling chemistry standard procedure (Biacore) at a mammary gland, 8; pancreas, 9; placenta, 7; prostate, 19; flow rate of 5 ml minÀ1. The achieved ligand density salivary gland, 43; skeletal muscle, 12; skin, 6; small after immobilization corresponded to 2200 RU for intestine, 2; spinal cord, 13; spleen, 6; stomach, 2; testis, IFN-l1 and 1550 RU for IL-22. For binding analyses 45; thymus, 12; thyroid gland, 53; trachea, 84; uterus, 3. regarding sIFN-lR1 and IL-22BP purified from respec- Isolation of total RNA from isolated and/or cultured tively transfected HepG2 cell supernatant, different cells was performed using the Invisorb RNA kit II concentrations of sIFN-lR1 (10, 20, 40, 80, 160 mgmlÀ1) (Invitek, Berlin, Germany). Messenger RNA was reverse- and IL-22BP (10, 20, 40, 80 mgmlÀ1) were run over the transcribed and quantitative PCR analysis was per- sensor chip surface of both flow cells at 5 ml minÀ1 for formed in triplicate assays using the ABI Prism 7700 10 min. The receptor dose-dependent increase in RU as a Sequence Detection System (Applied Biosystems) as measure of the density of ship-bound receptor was described previously.51,52 For the detection of total assessed. To measure the affinity between sIFN-lR1 and IFN-lR1 and single splice variants, systems purchased IFN-l1, sIFN-lR1-Fc (second flow cell) and dimeric IL- from Applied Biosystems, together with the matching 22BP-Fc (first (reference) flow cell) (both from R&D system for the housekeeping gene hypoxanthine Systems) were coated on the sensor chip surface by the phosphoribosyl-transferase 1 (HPRT), were used. The standard amine coupling chemistry procedure (Biacore). analysis of HPRT was included to normalize IFN-lR1 The achieved ligand density after immobilization corre- expression, which was carried out by calculating IFN- sponded to 3050 RU for sIFN-lR1-Fc and 3300 RU for lR1 expression as the value 2 to the power of the sIL-22BP-Fc. After running different concentrations of difference between the threshold cycles for amplification IFN-l1 (100, 200, 400, 800, 1200, 1600, 2000 and of HPRT and IFN-lR1. 2400 ng mlÀ1)at50ml minÀ1 for 2 min over the sensor chip surface, steady-state affinity was calculated with Conventional PCR, gel electrophoresis and sequencing BIAevaluation software (Biacore, Freiburg, Germany) For investigating the mRNA expression of IFN-lR1 by performing global fitting. All cytokines (carrier free) variants, cDNA was synthesized as described above were purchased from R&D Systems.

Genes and Immunity Target cells of type III interferons K Witte et al 713 Statistical analyses signals in intestinal epithelial cells and murine CMV infection Differences between experimental treatment groups increases colonic IL-28A expression. Am J Physiol Gastrointest obtained with a minimum of five experiments were Liver Physiol 2005; 289: G960–G968. tested by the Wilcoxon matched-pairs signed-rank test 13 Mordstein M, Kochs G, Dumoutier L, Renauld JC, Paludan SR, (two-tailed) using SPSS 14.0 software (SPSS Inc., Chicago, Klucher K et al. Interferon-lambda contributes to innate IL, USA). An error probability of Po0.05 was considered immunity of mice against influenza A virus but not against significant. hepatotropic viruses. PLoS Pathog 2008; 4: e1000151. 14 Robek MD, Boyd BS, Chisari FV. Lambda interferon inhibits hepatitis B and C virus replication. J Virol 2005; 79: 3851–3854. Conflict of interest 15 Lasfar A, Lewis-Antes A, Smirnov SV, Anantha S, The authors declare no conflict of interest. Abushahba W, Tian B et al. Characterization of the mouse IFN-lambda ligand-receptor system: IFN-lambdas exhibit antitumor activity against B16 melanoma. Cancer Res 2006; 66: 4468–4477. Acknowledgements 16 Sato A, Ohtsuki M, Hata M, Kobayashi E, Murakami T. Antitumor activity of IFN-lambda in murine tumor models. We acknowledge B Ketel and A Buss for excellent J Immunol 2006; 176: 7686–7694. technical assistance and E Wallace for accurately proof- 17 Wolk K, Witte K, Witte E, Proesch S, Schulze-Tanzil G, reading the paper. This study was supported by Nasilowska K et al. Maturing dendritic cells are an important research Grant WO 1567/1-1 (to Kerstin Wolk) from the source of IL-29 and IL-20 that may cooperatively increase the innate immunity of keratinocytes. J Leukoc Biol 2008; 83: Deutsche Forschungsgemeinschaft (German Research 1181–1193. Foundation). 18 Maher SG, Sheikh F, Scarzello AJ, Romero-Weaver AL, Baker DP, Donnelly RP et al. IFN-alpha and IFN-lambda differ in their antiproliferative effects and duration of JAK/STAT References signaling activity. Cancer Biol Ther 2008; 7: nihpa47781. 19 Meager A, Visvalingam K, Dilger P, Bryan D, Wadhwa M. 1 Kotenko SV, Gallagher G, Baurin VV, Lewis-Antes A, Shen M, Biological activity of interleukins-28 and -29: comparison with Shah NK et al. IFN-lambdas mediate antiviral protection type I interferons. Cytokine 2005; 31: 109–118. through a distinct class II cytokine receptor complex. Nat 20 Ank N, Iversen MB, Bartholdy C, Staeheli P, Hartmann R, Immunol 2003; 4: 69–77. Jensen UB et al. An important role for type III interferon (IFN- 2 Sheppard P, Kindsvogel W, Xu W, Henderson K, Schlutsmeyer lambda/IL-28) in TLR-induced antiviral activity. J Immunol S, Whitmore TE et al. IL-28, IL-29 and their class II cytokine 2008; 180: 2474–2485. receptor IL-28R. Nat Immunol 2003; 4: 63–68. 21 Baize S, Pannetier D, Faure C, Marianneau P, Marendat I, 3 Pestka S, Krause CD, Sarkar D, Walter MR, Shi Y, Fisher PB. Georges-Courbot MC et al. Role of interferons in the control of Interleukin-10 and related cytokines and receptors. Annu Rev Lassa virus replication in human dendritic cells and macro- Immunol 2004; 22: 929–979. phages. Microbes Infect 2006; 8: 1194–1202. 4 Wolk K, Sabat R. Interleukin-22: a novel T- and NK-cell 22 Sommereyns C, Paul S, Staeheli P, Michiels T. IFN-lambda derived cytokine that regulates the biology of tissue cells. (IFN-lambda) is expressed in a tissue-dependent fashion and Cytokine Growth Factor Rev 2006; 17: 367–380. primarily acts on epithelial cells in vivo. PLoS Pathog 2008; 4: 5 Sabat R, Wallace E, Endesfelder S, Wolk K. IL-19 and IL-20: e1000017. two novel cytokines with importance in inflammatory 23 Wolk K, Kunz S, Witte E, Friedrich M, Asadullah K, Sabat R. diseases. Expert Opin Ther Targets 2007; 11: 601–612. IL-22 increases the innate immunity of tissues. Immunity 2004; 6 Doyle SE, Schreckhise H, Khuu-Duong K, Henderson K, 21: 241–254. Rosler R, Storey H et al. Interleukin-29 uses a type 1 interferon- 24 Dumoutier L, Lejeune D, Hor S, Fickenscher H, Renauld JC. like program to promote antiviral responses in human Cloning of a new type II cytokine receptor activating signal hepatocytes. Hepatology 2006; 44: 896–906. transducer and activator of transcription (STAT)1, STAT2 and 7 Dumoutier L, Tounsi A, Michiels T, Sommereyns C, Kotenko STAT3. Biochem J 2003; 370 (Part 2): 391–396. SV, Renauld JC. Role of the interleukin (IL)-28 receptor 25 Novick D, Cohen B, Rubinstein M. The human interferon tyrosine residues for antiviral and antiproliferative activity alpha/beta receptor: characterization and molecular cloning. of IL-29/interferon-lambda 1: similarities with type I inter- Cell 1994; 77: 391–400. feron signaling. J Biol Chem 2004; 279: 32269–32274. 26 Kirkwood JM, Tarhini AA, Panelli MC, Moschos SJ, Zarour 8 Gad HH, Dellgren C, Hamming OJ, Vends S, Paludan SR, HM, Butterfield LH et al. Next generation of immunotherapy Hartmann R. Interferon-\{lambda\} is functionally an inter- for melanoma. J Clin Oncol 2008; 26: 3445–3455. feron but structurally related to the interleukin-10 family. J Biol 27 Robak T. Current treatment options in hairy cell leukemia Chem 2009; 284: 20869–20875. and hairy cell leukemia variant. Cancer Treat Rev 2006; 32: 9 Zhou Z, Hamming OJ, Ank N, Paludan SR, Nielsen AL, 365–376. Hartmann R. Type III interferon (IFN) induces a type I IFN- 28 Sievert W. Management issues in chronic viral hepatitis: like response in a restricted subset of cells through signaling hepatitis C. J Gastroenterol Hepatol 2002; 17: 415–422. pathways involving both the Jak–STAT pathway and the 29 Weinstock-Guttman B, Ramanathan M, Zivadinov R. Inter- mitogen-activated protein kinases. J Virol 2007; 81: 7749–7758. feron-beta treatment for relapsing multiple sclerosis. Expert 10 Uze G, Monneron D. IL-28 and IL-29: newcomers to the Opin Biol Ther 2008; 8: 1435–1447. interferon family. Biochimie 2007; 89: 729–734. 30 Naka K, Takemoto K, Abe K, Dansako H, Ikeda M, 11 Coccia EM, Severa M, Giacomini E, Monneron D, Remoli ME, Shimotohno K et al. Interferon resistance of hepatitis C virus Julkunen I et al. Viral infection and Toll-like receptor agonists replicon-harbouring cells is caused by functional disruption induce a differential expression of type I and lambda of type I interferon receptors. J Gen Virol 2005; 86(Part 10): interferons in human plasmacytoid and monocyte-derived 2787–2792. dendritic cells. Eur J Immunol 2004; 34: 796–805. 31 Dellgren C, Gad HH, Hamming OJ, Melchjorsen J, Hartmann 12 Brand S, Beigel F, Olszak T, Zitzmann K, Eichhorst ST, Otte JM R. Human interferon-lambda3 is a potent member of the type et al. IL-28A and IL-29 mediate antiproliferative and antiviral III interferon family. Genes Immun 2009; 10: 125–131.

Genes and Immunity Target cells of type III interferons K Witte et al 714 32 Jordan WJ, Eskdale J, Boniotto M, Rodia M, Kellner D, 42 Wolk K, Kunz S, Asadullah K, Sabat R. Cutting edge: immune Gallagher G. Modulation of the human cytokine response by cells as sources and targets of the IL-10 family members? interferon lambda-1 (IFN-lambda1/IL-29). Genes Immun 2007; J Immunol 2002; 168: 5397–5402. 8: 13–20. 43 Wolk K, Witte E, Hoffmann U, Doecke WD, Endesfelder S, 33 Pekarek V, Srinivas S, Eskdale J, Gallagher G. Interferon Asadullah K et al. IL-22 induces lipopolysaccharide-binding lambda-1 (IFN-lambda1/IL-29) induces ELR(À) CXC chemo- protein in hepatocytes: a potential systemic role of IL-22 in kine mRNA in human peripheral blood mononuclear cells, in Crohn’s disease. J Immunol 2007; 178: 5973–5981. an IFN-gamma-independent manner. Genes Immun 2007; 8: 44 Jaitin DA, Roisman LC, Jaks E, Gavutis M, Piehler J, Van der 177–180. Heyden J et al. Inquiring into the differential action of 34 Mennechet FJ, Uze G. Interferon-lambda-treated dendritic interferons (IFNs): an IFN-alpha2 mutant with enhanced cells specifically induce proliferation of FOXP3-expressing affinity to IFNAR1 is functionally similar to IFN-beta. Mol suppressor T cells. Blood 2006; 107: 4417–4423. Cell Biol 2006; 26: 1888–1897. 35 Chi B, Dickensheets HL, Spann KM, Alston MA, Luongo C, 45 Jaks E, Gavutis M, Uze G, Martal J, Piehler J. Differential Dumoutier L et al. Alpha and lambda interferon together receptor subunit affinities of type I interferons govern mediate suppression of CD4T cells induced by respiratory differential signal activation. J Mol Biol 2007; 366: 525–539. syncytial virus. J Virol 2006; 80: 5032–5040. 46 Kalie E, Jaitin DA, Podoplelova Y, Piehler J, Schreiber G. The 36 Jordan WJ, Eskdale J, Srinivas S, Pekarek V, Kelner D, Rodia M stability of the ternary interferon-receptor complex rather than et al. Human interferon lambda-1 (IFN-lambda1/IL-29) the affinity to the individual subunits dictates differential modulates the Th1/Th2 response. Genes Immun 2007; 8: biological activities. J Biol Chem 2008; 283: 32925–32936. 254–261. 47 Lamken P, Lata S, Gavutis M, Piehler J. Ligand-induced 37 Serra C, Biolchini A, Mei A, Kotenko S, Dolei A. Type III and I assembling of the type I interferon receptor on supported lipid interferons increase HIV uptake and replication in human cells bilayers. J Mol Biol 2004; 341: 303–318. that overexpress CD4, CCR5, and CXCR4. AIDS Res Hum 48 Slutzki M, Jaitin DA, Yehezkel TB, Schreiber G. Variations in Retroviruses 2008; 24: 173–180. the unstructured C-terminal tail of interferons contribute to 38 Siebler J, Wirtz S, Weigmann B, Atreya I, Schmitt E, Kreft A differential receptor binding and biological activity. J Mol Biol et al. IL-28A is a key regulator of T-cell-mediated liver injury 2006; 360: 1019–1030. via the T-box transcription factor T-bet. Gastroenterology 2007; 49 Wolk K, Docke WD, von Baehr V, Volk HD, Sabat R. Impaired 132: 358–371. antigen presentation by human monocytes during endotoxin 39 Srinivas S, Dai J, Eskdale J, Gallagher GE, Megjugorac NJ, tolerance. Blood 2000; 96: 218–223. Gallagher G. Interferon-lambda1 (interleukin-29) 50 Wolk K, Kunz S, Crompton NE, Volk HD, Sabat R. Multiple preferentially down-regulates interleukin-13 over other T mechanisms of reduced major histocompatibility complex helper type 2 cytokine responses in vitro. Immunology 2008; class II expression in endotoxin tolerance. J Biol Chem 2003; 125: 492–502. 278: 18030–18036. 40 Dai J, Megjugorac NJ, Gallagher GE, Yu RY, Gallagher G. IFN- 51 Wolk K, Grutz G, Witte K, Volk HD, Sabat R. The expression of lambda1 (IL-29) inhibits GATA3 expression and suppresses legumain, an asparaginyl endopeptidase that controls antigen Th2 responses in human naive and memory T cells. Blood 2009; processing, is reduced in endotoxin-tolerant monocytes. Genes 113: 5829–5838. Immun 2005; 6: 452–456. 41 Moore KW, de Waal Malefyt R, Coffman RL, O0Garra A. 52 Wolk K, Witte E, Reineke U, Witte K, Friedrich M, Sterry W Interleukin-10 and the interleukin-10 receptor. Annu Rev et al. Is there an interaction between interleukin-10 and Immunol 2001; 19: 683–765. interleukin-22? Genes Immun 2005; 6: 8–18.

Genes and Immunity