Macrophages and Dendritic Cells as Actors in the Immune Reaction of Classical Hodgkin Lymphoma

Background The inflammatory infiltrate plays a pivotal role in classical Hodgkin lymphoma (cHL). Here, we focussed on the role of macrophages (MΦ) and dendritic cells (DC). Methods MΦ and DC infiltration was investigated in 106 cHL specimens using immunohistochemistry and cytokine expression was analyzed in a subset by real-time PCR. Human peripheral blood-derived monocytes, DC, MΦ stimulated with GM-CSF (MΦGM-CSF, pro-inflammatory MΦ-1-model) or M-CSF (MΦM-CSF, immunomodulatory MΦ-2-model) were incubated with cHL cell line (L1236, HDLM2) supernatants (SN). DC maturation or MΦ polarization were investigated by flow cytometry. Furthermore, the impact of DC or MΦ on cHL cell proliferation was analyzed by BrdU/CFSE assay. Results In cHL tissues mature myeloid (m)DC and MΦ predominated. High numbers of CD83+ mDC and low numbers of CD163+ MΦ were associated with improved disease specific survival. In numerous cHL specimens increased levels of both pro- and anti-inflammatory cytokines and of IL13 and GM-CSF were observed compared to reactive lymphadenopathies. Maturation of DC and induction and maintenance of an immunomodulatory MΦ phenotype were promoted by SN derived from cHL cell lines. TNFα neutralization in SN resulted in a significant inhibition of mDC maturation. DC and pro-inflammatory MΦ inhibited the proliferation of cHL cells. Conclusion Adopting an immunomodulatory phenotype is a potential mechanism for how MΦ promote immune evasion in cHL. Mature DC, in contrast, might participate in antitumoral immunity.

In classical Hodgkin lymphoma (cHL) few malignant Hodgkin-Reed-Sternberg (HRS) cells are embedded in a prominent inflammatory infiltrate [6,7]. The occurrence of DC in cHL has been reported and data on a potential influence on the prognosis of cHL patients were ambiguous [8][9][10]. Studies in breast cancer implicated that the tumor modulates DC maturation [11]. An inhibitory effect on DC maturation by tumoral production of mediators like IL10, TGFb and M-CSF has been proposed [1]. Moreover, in diverse tumors a beneficial effect of DC on clinical outcome was demonstrated [12][13][14].
Immune evasion might also be achieved by modulating the polarization of intratumoral macrophages (MW), being numerous in HL [15]. The tumor might promote their polarization towards an anti-inflammatory phenotype characterized by low cytotoxicity and low expression of inflammatory cytokines [16][17][18][19]. TAM and neoplastic cells produce diverse factors including IL10, TGFb, TNFa, PGE2 and IL1 known to promote tumor progression, induce regulatory T cells, suppress cytotoxic T and NK cells and influence DC maturation and MW polarization [18,[20][21][22][23][24]. In most studied malignancies the excessive presence of TAM was a poor prognosticator [4,20]. A tumor-promoting role of TAM in cHL first suggested decades ago was confirmed by gene expression profiling and immunohistochemical studies recently [15,[25][26][27][28][29][30]. TAM have been reported to show an intermediate phenotype that can exert pro-inflammatory/antitumoral functions under normoxic conditions [16,31] besides anti-inflammatory/ protumoral functions, which might explain that other groups found no association of MW numbers and survival in cHL [32][33][34].
In the present study we aimed to investigate the role of mDC and MW in cHL cases by analyzing their distribution and influence on outcome as well as the intratumoral cytokine profile. Furthermore, in a cell culture system the mutual effect of cHL cell lines on monocytes, monocyte-derived (mo)DC and MW and vice versa was investigated.

cHL tissue specimens
Formalin-fixed paraffin-embedded (FFPE) specimens of 106 cHL cases and of 10 reactive lymphadenopathies were retrieved from the Archives of the Institute of Pathology, University Hospital, Erlangen, Germany. Most cases have been included in a previous study [35]. An overview of the clinical data is provided in Table 1.
The use of FFPE material from the Archive of the Institute of Pathology was approved by the Ethics Committee of the Friedrich-Alexander-University of Erlangen-Nuremberg (ethic@zuv.uni-erlangen.de) on 24 th January 2005, waiving the need for retrospective consent for the use of archived rest material.
To assess the EBV state of CHL cases EBER-in situ hybridization (ISH) or latent membrane protein 1 (LMP1) specific immunohistochemistry was performed as described before [35].

Quantitative evaluation of histologic sections
A computer assisted hot spot analysis was performed on whole block sections. Images were acquired with a standard light microscope and a CCD camera or a Mirax Midi system (Zeiss, Jena, Germany) for digital slide scanning and Mirax viewer software (Zeiss). Positive cells were manually selected by a pathologist. Numbers of positive cells per mm 2 in the area with the densest infiltration were evaluated using the image analysis software ''Count'' (Biomas, Erlangen, Germany). In CD83 stains positive cells showing the morphology of neoplastic HRS cells were excluded from the analysis. For evaluation of CD123 only cells with plasmacytoid morphology were counted.

FFPE real-time PCR of cHL biopsies
From 10 mm FFPE sections total RNA isolation and DNAse digestion were performed, using a standardized fully automated isolation method based on germanium-coated magnetic beads (XTRAKT RNA kits, STRATIFYER Molecular Pathology GmbH, Cologne, Germany) in combination with the liquid handling robot XTRAKT SL (STRATIFYER) as previously described [36] for the analysis of TNF, IL10, TGFb1 and CALM2 (reference gene). For the analysis of IL4, IL13, GM-CSF, M-CSF and CALM2 (reference gene) RNA was isolated using Maxwell 16 LEV RNA FFPE Kit, according to the manufacturer's instructions. A probe of L1236-RNA was used as positive control and for comparison of the PCR-runs. Standard and no-template controls were included. DDC T qRT-PCR with probes for CALM2 (reference gene), TNF, IL10 and TGFb1 as well as for IL4, IL13, GM-CSF and M-CSF was performed with 40 cycles of nucleic acid amplification using QuantiFast Probe Assays for one-step qRT-PCR (Qiagen, Hilden, Germany). Assays were analyzed by the Applied Biosystems 7500 Fast Real-Time PCR System and the 7500 Software Version 2.0.6 (Applied Biosystems, Darmstadt, Germany).

Cell lines and isolation of moDC and MW
The cHL cell lines L1236 and HDLM2 were obtained in 2009 and 2002 from ''Deutsche Sammlung von Mikroorganismen und Zellkulturen'', Braunschweig, Germany. The cells showed heterogeneity in cell size and typical morphology including mono-, bi-, and also multi-nucleated cells. HDLM2 were suspension cells and L1236 semi-adherent as stated by the ''Deutsche Sammlung von Mikroorganismen und Zellkulturen'' and showed typical growth characteristics. Cells were negative for mycoplasma at the onset of the study and were re-tested negative at the end of the study. L1236 and HDLM2 and blood cells were maintained at 37˚C and 5% CO 2 in RPMI-1640 (Sigma-Aldrich) supplemented with 10% fetal bovine serum, 50 U/ml penicillin and 50 mg/ml streptomycin (Invitrogen, Darmstadt, Germany). PBMCs were isolated by Ficoll-Paque (GE Healthcare, Uppsala, Sweden) gradient centrifugation of peripheral blood from 40 healthy volunteers (mean age 35 years). Monocytes were isolated by plastic adherence. For generation of immature moDC, adherent cells after 2 h at 37˚C were incubated with 10 ng/ml rhIL4 and 25 ng/ml rhGM-CSF (CellGenix, Freiburg, Germany) for 6 days with cytokine renewal at day 2 and 5. To generate MW, adherent cells were incubated for 5 days with either 50 ng/ml rhGM-CSF (MW GM-CSF , a model of MW-1) or 50 ng/ml rhM-CSF (MW M-CSF , a model of MW-2). All other cytokines were also obtained from R&D Systems, Minneapolis, USA. moDC and MW purity and differentiation was confirmed by FACS analysis (see Figure S1). Of all volunteers written informed consent was obtained and the procedure was approved by the Ethics Committee of the Friedrich-Alexander-University of Erlangen-Nuremberg (Re.-No. 3832).

Statistical analyses
Continuous variables with normal distribution and homogenous variances were analyzed by unpaired or paired t-test. Mann-Whitney U-tests or Wilcoxon signed rank test were used for data with skewed distributions. Disease-specific survival (DSS) of cHL was defined from the receipt of tissue until the last follow-up or death caused by cHL. DSS rates of cHL patients were calculated according to Kaplan-Meier. DSS curves were compared with log-rank test. In all tests p,0.05 was accepted as statistically significant. Statistical analyses were performed using SPSS for Windows software (version 19.0 SPSS, IBM, Munich, Germany) or GraphPad Prism 5 for Windows software (version 5.02, GraphPad software Inc., San Diego, CA, USA).

Distribution and prognostic impact of MW and DC in cHL
CD1a and CD83 were used as markers of immature and mature mDC as described elsewhere [1] (Figure 1 A and B). In cHL the majority of mDC could be identified as mature CD83+ mDC and a significantly minor portion as CD1a+ immature mDCs (

Cytokine profile in cHL specimens
For the analyses of RNA isolated from cHL specimens, cases were selected with numbers per mm 2 of CD83+ or CD163+ cells below (low) or above (high) the 25 th or 75 th percentile, respectively. TNF, IL10, TGFb1 were investigated, which were reported in cHL tissue and can influence DC maturation/development and MW polarization [1,7,23,24]. Additionally, GM-CSF, M-CSF, IL4 and IL13 were analyzed, which might also partake in the polarization of MW [17,37]. Because of the degradation of RNA in FFPE material specifically designed taqman probes for sequences ,100 bp were applied. Results were normalized to RNA isolated from L1236 cells and were compared to reactive lymphadenopathies. Cases expected to have an inferior outcome, namely CD83 low and CD163 high, showed a significantly higher expression of IL10 and TGFb1 compared to the other cHL cases (Figure 3). A significantly induced expression of TNF, IL10, TGFb1, GM-CSF and IL13 was detected in cHL compared to rLA (Figures 3 to 5). The only cytokine which was found predominantly in rLA compared to cHL was IL4 ( Figure 5).

Influence of cytokines and cHL cell line SN on the maturation of moDC and polarization of peripheral blood monocytes and MW
Cytokines known to influence DC maturation and to be expressed in cHL [1,23] were selected to validate the experimental system. As anticipated, TNFa induced the activation marker CD40 and the maturation marker CD83. IL10 counteracted moDC development as shown by the upregulation of the monocyte marker CD14. In addition, moDC maturation was blocked by TGFb downregulating CD83 ( Figure S2 A). To reflect at least in part the diversity of immune mediators found in cHL, moDC were treated with a cytokine cocktail (TNFa/IL10/TGFb) or with SN of cHL cell lines L1236 and HDLM2. All treatments induced significant moDC activation and/or maturation similar to TNFa alone (  neutralizing TNFa antibodies a significant reduction of CD40 and CD83 upregulation was observed (Figure 6 B). SN generated from B-cells from PBMC as a control showed no effect on moDC when compared to medium controls ( Figure  S2 A).
MW GM-CSF (MW -1 model) expressed high levels of HLA-DR and lacked CD163 expression whereas MW M-CSF (MW-2 model) presented with high CD163 and low HLA-DR expression, as expected (Figure 7). To address the question whether MW polarization could be modulated by cHL-conditioned medium, monocytes, MW GM-CSF and MW M-CSF were incubated with L1236-SN and HDLM2-SN. Changes in the expression of CD68, a pan MW marker, the markers scavenger receptor CD163 and mannose receptor CD206, which are particularly found on immunomodulatory MW and of HLA-DR predominantly expressed by proinflammatory MW were analyzed (Figure 7). In MW M-CSF and monocytes a consistent induction of CD163 and CD206 upon treatment with L1236-SN and HDLM2-SN was seen, whereas the SN had no effect on CD163 expression in MW GM-CSF . CD206 was slightly down-regulated by L1236-SN in MW GM-CSF but not by HDLM2-SN.

Modulation of the proliferative activity of cHL cells by moDC and MW
To analyze the effect of moDC and MW on tumor cell proliferation, CFSE stained L1236 were treated with SN of moDC, MW GM-CSF or MW M-CSF and CFSE signal was measured as a reciprocal indicator of cell proliferation. A significant inhibition of L1236 growth was found after incubation with moDC-SN or MW GM-CSF -SN (Figure 8 A). MW M-CSF -SN, on the contrary, did not influence L1236 growth. Similar results were obtained measuring BrdU incorporation into DNA of L1236, when treated with MW GM-CSF -SN or MW M-CSF -SN (Figure 8 B). Co-culture experiments were performed to determine by BrdU assays whether cell-cell contact enhanced the effect of MW on L1236 proliferation. BrdU uptake by L1236 was significantly lower when co-cultured with MW GM-CSF compared to MW M-CSF similar to the findings after SN treatment (Figure 8 B and C). Since co-cultures were kept in 50% MW-conditioned medium containing GM-or M-CSF, we excluded an effect of these cytokines on L1236 growth (Figure S3 A). Proliferation of MW in co-culture was assessed by Ki67-CD68 double stainings. A weak proliferative activity was observed with no significant difference between MW GM-CSF and MW M-CSF , indicating that the difference in proliferation observed in co-culture can be assigned to L1236 (Figure S3 B).

Discussion
In the analysis of the inflammatory infiltrate of diverse tumors close attention was paid to the role of DC and MW in recent years. The present study focusses on their  . Expression of TNF, IL10 and TGFb1 in cHL cases with high or low numbers of mature mDC and CD163+ MW. RNA was isolated from FFPE cHL sections and subjected to qRT-PCR. The expression of TNF, IL10 and TGFb1 was determined. cHL cases were grouped into subgroups with low or high numbers of CD83+ or CD163+ infiltrating cells using the 25. or 75. percentile as cut-off, respectively and compared to a control group of benign reactive lymphadenopathy (rLA). All values are shown as fold change compared to the measured cytokine expression in the L1236 cell line. Mann-Whitney-U-tests were conducted with *P,0.05, **P,0.01 and ***P,0.001 comparing either all cHL cases with rLA or cHL cases with presumably adverse prognosis (CD83 low and CD163 high) with those with an expected better prognosis (CD83 high and CD163 low). Single bars represent the mean + SD of triplicates of single analyzed cases, each.
doi:10.1371/journal.pone.0114345.g003 Figure 4. Expression of GM-CSF and M-CSF in cHL cases with high or low numbers of mature mDC and CD163+ MW. RNA was isolated from FFPE cHL sections and subjected to qRT-PCR. The expression of GM-CSF and M-CSF was determined. cHL cases were grouped into subgroups with low or high numbers of CD83+ or CD163+ infiltrating cells using the 25. or 75. percentile as cut-off, respectively and compared to a control group of benign reactive lymphadenopathy (rLA). All values are shown as fold change compared to the measured cytokine expression in the L1236 cell line. Mann-Whitney-U-tests were conducted with *P,0.05, **P,0.01 and ***P,0.001 comparing either all cHL cases with rLA or cHL cases with presumably adverse prognosis (CD83 low and CD163 high) with those with an expected better prognosis (CD83 high and CD163 low). Single bars represent the mean + SD of triplicates of single analyzed cases, each. clinical outcome [8]. The adverse prognostic impact of MW found in multiple malignancies, however, has been confirmed for cHL [15,[25][26][27][28][29][30].
In our study mature mDC outnumbered immature mDC, which indicates that HRS cells do not abolish mDC maturation as a mean of immune evasion as was suggested for breast cancer [1]. A possible explanation might be that protumoral effects of factors promoting mDC maturation like TNFa, which is also angiogenic [38], might exceed the benefits of inhibiting mDC maturation. pDC were the most frequent DC subtype in our study. MW were by far the most numerous population observed. The fact that the number of CD163+ cells approximated the total number of CD68+ MW might indicate that a very relevant subgroup of TAM Figure 5. Expression of IL4 and IL13 in cHL cases with high or low numbers of mature mDC and CD163+ MW. RNA was isolated from FFPE cHL sections and subjected to qRT-PCR. The expression of IL4 and IL13 was determined. cHL cases were grouped into subgroups with low or high numbers of CD83+ or CD163+ infiltrating cells using the 25. or 75. percentile as cut-off, respectively and compared to a control group of benign reactive lymphadenopathy (rLA). All values are shown as fold change compared to the measured cytokine expression in the L1236 cell line. Mann-Whitney-U-tests were conducted with *P,0.05, **P,0.01 and ***P,0.001 comparing either all cHL cases with rLA or cHL cases with presumably adverse prognosis (CD83 low and CD163 high) with those with an expected better prognosis (CD83 high and CD163 low). Single bars represent the mean + SD of triplicates of single analyzed cases, each.
doi:10.1371/journal.pone.0114345.g005 Figure 6. Characterization of the maturation and activation of moDC. (A) moDC were treated for 24 h with TNFa, IL10 and TGFb (each 10 ng/ml), 50% L1236 or HDLM2 supernatants (SN) or with (B) 50% L1236 supernatants and either an IgG1 isotype control, an neutralizing TNFa antibody or a therapeutically applied TNFa antibody (infliximab). Supernatants were generated at a concentration of 1610 6 cells/ml medium after has some characteristics of immunomodulatory MW, as CD163 in model systems is particularly strongly expressed on MW-2.
As reported for breast cancer [39,40], no influence of CD1a+ immature mDC on DSS was found. Previous reports on the prognostic impact of CD1a+ cells in cHL were controversial, showing no association with outcome [9] or a poorer outcome in patients without CD1a+ cells in the tumor infiltrate [10]. In the present study, numerous CD83+ mature mDC were associated with improved outcome in cHL, as found in a variety of malignancies [1]. This could support the idea that mature DC promote antitumoral immunity by direct tumortoxicity or T cell priming. CD83 can also be expressed on pDC [41]. However, CD123+ pDC were also analyzed and had no influence on cHL outcome, as reported earlier [42]. Therefore, it appears improbable that the prognostic effect of CD83+ cells has to be attributed to an effect of pDC. As some morphological overlap with HRS cells may bias the analysis of CD83+ cells, this finding will need further confirmation in the future.
In therapy-refractory cHL a TAM-specific gene signature [15] and an association of high numbers of CD68+ MW with shortened progression-free survival have been reported [15]. Comparably other studies revealed that either CD163+ and/or CD68+ MW significantly correlated with inferior outcome [27][28][29][30]. Others, however, found no association of MW numbers and survival in cHL [32][33][34]. Here, we confirm a highly significant association of high numbers of CD163+ MW with inferior DSS, whereas CD68+ MW had no impact on outcome. CD163+ and CD68+ MW were more frequent in EBV+ compared to EBV-cHL, confirming previous findings [8,27,28,30,33]. This, however, was not associated with worse DSS of EBV+ cases (cut-off median: p50.06). Some explanations for the predominance of MW in EBV+ cHL were discussed. In pediatric cHL it was proposed that an antiviral immune reaction might result in an increased recruitment of both cytotoxic T cells and MW [8]. This would go in line with the report of a Th1 and antiviral gene profile in EBV+ cHL [42]. Moreover, monocytes were shown to induce proliferation and LMP1 expression in cells of EBV-associated nasal natural killer/T cell lymphoma [43] and a correlation of serum CD163 and plasma EBV-DNA in cHL was reported [44]. These findings might indicate a mutually inductive interaction of the virus and monocytes/MW. The exact mechanism, leading to an accumulation of MW in EBV+ cHL is, however, so far not understood. Next we investigated the cytokine profile of the cHL microenvironment. All investigated cytokines were found at least in a subset of cHL cases. A frequent expression of IL13 and M-CSF in cHL cases has been described, whereas IL4 was only detected in small amounts, going in line with our findings [45]. Comparing cHL cases to reactive lymphadenopathies, we found 48 h incubation. The expression of the DC activation marker CD40, the monocyte marker CD14 and the DC maturation marker CD83 was determined by flow cytometry. Mean fluorescence intensity (MFI) is indicated on the bar charts. Paired student's t-tests or Wilcoxon signed-rank tests were conducted with *P,0.05, **P,0.01 and ***P,0.001. doi:10.1371/journal.pone.0114345.g006 significantly higher amounts of TNF, IL10, TGFb1, GM-CSF and IL13 in cHL cases. Especially when comparing presumably adverse prognostic constellations in cHL (CD83 low or CD163 high) with cHL cases with an expected better prognosis (CD83 high or CD163 low) anti-inflammatory (IL10, TGFb1) cytokines were significantly upregulated in the adverse group. The upregulation of immunoregulatory cytokines IL-10 and TGFb1 in cHL and especially in the prognostically adverse group goes well in line with immune evasion. At first glance, in contrast, it appears paradoxical that TNF and GM-CSF are excessively produced in cHL. However, persistent inflammation has been connected to tumorigenesis [46,47]. Therefore, both, pro-and anti-inflammatory cytokines, might promote cHL growth. The presence of IL13 and M-CSF in cHL cases might be responsible for  the induction of MW-2-like CD163+ tumor-associated macrophages in cHL [17,37], whereas IL4 was detected infrequently and, if at all, seems to play a minor role.
Going in line with the dominance of mature mDC in cHL specimens, SN of cHL cell lines promoted the maturation and/or activation of moDC in our study. This effect was stronger for L1236 cells producing higher TNFa concentrations than HDLM2. When combining IL10 and TGFb that hamper DC maturation and maturation-inducing TNFa [23], all known to be expressed in cHL [7], the effect of TNFa on activation and maturation predominated. Moreover, neutralization of TNFa in L1236 SN decreased the observed effects on the activation and maturation of moDC, underlining the importance of TNFa.
For the investigation of MW, we used a published model representing two extreme polarization states of pro-inflammatory MW-1 and immunoregulatory MW-2 obtained by stimulation with M-CSF and GM-CSF [37], which we called MW GM-CSF or MW M-CSF depending on the used stimulus. One has, however, to keep in mind that depending on the stimulus used, many different activation states can be achieved far beyond the MW-1 and MW-2 model [48]. So obviously this model cannot fully reflect the intermediate phenotype described for TAM [16,31]. However, it can add to our understanding of possible interactions of polarized, circulating MW entering the tumor milieu with tumor cells. Incubating unstimulated peripheral blood derived monocytes with cHL SN alone sufficed to induce a MW-2-like phenotype with up-regulation of CD163 and CD206. Moreover, treatment with cHL cell SN conserved and enhanced the MW-2 phenotype in MW M-CSF , as also demonstrated by the upregulation of CD163 known to promote an anti-inflammatory microenvironment [49] and of CD206, which acts pro-angiogenic [50] and thus might promote tumor growth. cHL cell SN were, however, not able to convert MW-1 into MW-2 as a mean to debilitate potentially tumor-toxic effector cells. Soluble mediators which could be responsible for the promotion of an MW-2 phenotype in pre-differentiated MW and monocytes are IL4, IL13 and M-CSF [17]. Whereas IL4 was reported to be lacking or only very mildly expressed in L1236 [51] and HDLM2 [52], IL13 was shown to be expressed in both cell lines [45,53,54]. Therefore, IL13 is a good candidate for the MW-2-inducing effect of cHL cell SN. Moreover, the expression of M-CSF was described in cHL cells [55,56] and we could confirm this finding on RNA-level in L1236 (data not shown), so that M-CSF might also participate. The in vitro finding, that MW-2 markers were enhanced by cHL cell SN, together with the predominance and prognostic role of CD163+ cell in cHL specimens, might therefore indicate that immunoregulatory CD163+ MW might play an important role in the tumor site and are actively promoted and maintained by the tumor cells. The excess of pro-as well as anti-inflammatory cytokines in the tumor tissue, however, indicates, that not a typical MW-2 milieu is found in the tumor, going in line with the idea that TAM cannot be assigned to one of the extreme polarization patterns of MW-1 or MW-2.
When investigating the influence of moDC and MW on the proliferation of the L1236 cHL cell line, we found no effect of MW M-CSF on L1236 growth, whereas both moDC and MW GM-CSF inhibited cHL cell proliferation. This effect was not depending on cell-cell contact. Hence, soluble factors appeared sufficient to mediate some inhibitory properties. Going in line with our observations, Dufresne et al reported that MW-1 inhibited the proliferation of human bladder carcinoma cells [57]. GM-CSF stimulated monocytes, our model of MW-1 [58,59], and immature DC (here moDC) [60,61] were both shown to exert cytotoxic effects on tumor cell lines. It was demonstrated that both cell types can act via soluble mediators, although cell membrane-bound mediators appear to be the major players [59,60,62], and some functional role of secreted TNF was implicated for both [59,61,62]. Whereas in one study, however, TNFalpha, TRAIL and FasL were not up-regulated in GM-CSF-stimulated human MW [63], in another the cytotoxic effect of SN of GM-CSF stimulated monocytes was attenuated by administration of TNF-neutralizing antibodies [59]. Moreover, GM-CSF was shown to augment the release of H 2 O 2 , another potentially cytotoxic agent, from mouse macrophages [64]. However, as M-CSF had the same effect [64], it appears improbable that this could explain the difference in cytotoxicity observed in the present study between MW-1 and MW-2. Possibly other so far not characterized substances might also be operational in the antitumoral effect observed. In contrast to our findings in human bladder cancer MW-2 not only tolerated but even induced tumor cell growth [57] going in line with the observation that TAM density in breast cancer correlates positively with proliferative tumor activity [65].
In conclusion, mature mDC, which associated with improved outcome in cHL patients, might partake in the antitumoral immunity either by antigen presentation or by direct cytotoxicity. CD163+ MW, in contrast, were associated with inferior survival and might actively be induced and maintained by the tumor cells. However, no typical immunosuppressive profile but rather an excess of antiand pro-inflammatory cytokines was observed in cHL patients with adverse prognostic markers. Figure S1. Generation of moDC, pro-inflammatory MW GM-CSF and antiinflammatory MW M-CSF . (A) Human peripheral blood monocytes were differentiated into moDC, MW GM-CSF or MW M-CSF in the presence of IL4 + GM-CSF (6 days), GM-CSF (5 days) or M-CSF (5 days), respectively. Depicted are monocytes one day after isolation (left), moDC with branching projections, MW GM-CSF with a fried-egg morphology and MW M-CSF with a spindle-shaped morphology. All images were acquired with an Olympus IX81 (1006 magnification, 10x/0.30 Ph1 objective) using cellSens dimension software. (B) Purity analysis of moDC, MW GM-CSF and MW M-CSF . moDC purity was shown by FACS analysis of CD11c, CD14, CD3 and CD20 (BD Pharmingen) expression, being 98%, 12%, 0% and ,1%, respectively. MW purity was proven by FACS analysis of CD68, CD11b, CD3 and CD20 (BD Pharmingen). MW-1 were positive in 95%, 92%, ,1% and ,1%, MW-2 in 98%, 86%, ,1% and ,1%, respectively. Data are indicated as mean with SEM of 4 independent experiments. (C) Analysis of cytokine secretion. For the determination of IL10 and IL12 secretion DuoSet ELISAs (R&D Systems) were applied according to the manufacturer's instruction. For generation of supernatants (SN), cells were seeded at a concentration of 1610 6 /ml+/2100 ng/ml LPS (Sigma-Aldrich) and incubated for 24 h. Absorbance was measured at 450 nm and 540 nm as reference using Synergy 2 (BioTek, Bad Friedrichshall, Germany). Data are shown as mean with SEM of 3-4 independent experiments each. doi:10.1371/journal.pone.0114345.s001 (PDF) Figure S2. Characterization of the maturation and activation profile of moDC and cytokine profile of cHL cell lines L1236 and HDLM2. (A) moDC were treated for 24 h with 10 ng/ml TNFa, IL10 or TGFb (each 10 ng/ml) or with 50% B-cell supernatants. B-cells were isolated from human blood PBMC by MACS separation (positive selection) using CD19 MicroBeads, FcR Blocking reagent and MS columns for an OctoMACS separator (MACS Miltenyi Biotec, Bergisch Gladbach, Germany). Supernatants were generated at a concentration of 1610 6 cells/ml medium after 48 h incubation. The expression of the DC activation marker CD40, the monocyte marker CD14 and the DC maturation marker CD83 was determined by flow cytometry. As anticipated, TNFa induced the activation marker CD40 and the maturation marker CD83, IL10 counteracted moDC development as shown by the up-regulation of the monocyte marker CD14, and TGFb blocked moDC maturation by down-regulation of CD83. SN generated from B-cells from PBMC showed no effect on moDC when compared to medium controls Mean fluorescence intensity (MFI) is indicated on the bar charts. Isotype controls are depicted as dotted lines. Paired student's t-tests or Wilcoxon signed-rank tests were conducted with *P,0.05, **P,0.01 and ***P,0.001. (B) Cytokine profile of cHL cell lines L1236 and HDLM2. The expression of TNFa, TGFb and IL10 was analyzed by ELISA (R&D Systems). L1236 showed higher TNFa expression than HDLM2. Data are indicated as mean with SEM of at least 3 independent experiments. doi:10.1371/journal.pone.0114345.s002 (PDF)