Delineation of a Novel Dendritic-Like Subset in Human Spleen

RESEARCH ARTICLE

Delineation of a novel dendritic-like subset in human spleen

Sawang Petvises1, Dipti Talaulikar2,3,4 and Helen C O’Neill1

Dendritic cells (DCs) and monocyte subpopulations present in the human spleen were analyzed by flow cytometry in an attempt to identify the presence of a novel dendritic-like cell subset described previously in mice and named L-DCs. In this study, an equivalent of this novel murine subset was characterized in the human spleen, thus increasing our knowledge of the antigen-presenting cell types present in the human spleen. Human L-DCs were identified as a hCD11c1hCD11b1HLA-DR2hCD861 subset in the spleen, along with the previously described subsets of hCD1c1 DCs, hCD1231 plasmacytoid DCs (pDCs), hCD161 DCs and hCD1411 DCs. Three subsets of monocytes were also characterized. DC and monocyte subsets in human spleen had phenotypes similar to those of subsets in human blood. In line with murine studies, the presence of L-DC progenitors within the spleen was also investigated. When human splenocytes depleted of T and B cells were cocultured with the murine stromal line 5G3, hematopoiesis ensued and hCD11c1HLA-DR1 and hCD11c1HLA-DR2 cells were produced. The latter resemble L-DCs, which are also produced in murine spleen cocultures. Both subsets expressed hCD80 and hCD86, which identifies them as antigen-presenting cells, particularly DCs, and were highly endocytic. It is noteworthy that murine splenic stroma can serve as a support matrix for human hematopoiesis and DC production. These results support the hypothesis that 5G3 must express both cell-associated and soluble factors that can signal hematopoiesis in human and murine progenitors. Cellular & Molecular Immunology (2016) 13 , 443–450;doi:10.1038/cmi.2015.16;published online 20 April 2015

Keywords: dendritic cells; myelopoiesis; spleen; stroma

INTRODUCTION subsets with similar functions. The splenic microenvironment The spleen is a secondary lymphoid organ that is essential in the also supports development of dendritic and myeloid cells immune response against blood-borne antigens. Hematopoiesis from bone marrow-derived progenitors that enter the spleen in the spleen is mainly restricted to the production of erythro- through blood. cytes, although hematopoiesis can be triggered at times of stress Dendritic cell (DC) subsets in the murine spleen have been or upon ablation of bone marrow.1 A small number of HSCs broadly classified as conventional DCs (cDCs) and plasmacy- can migrate through the blood and lymphatic system to enter toid DCs (pDCs), and the cDC population can be further the spleen and other tissues.2,3 However, it is not yet known divided into two subsets: CD81 cDCs and CD82 cDCs. whether hematopoietic stem cells in the steady-state spleen can CD82 cDCs are capable of inducing a CD41 T-cell response self-renew and differentiate into a specific lineage of hemato- with production of inflammatory cytokines,5 while the CD81 poietic cells. cDC subset is superior in cross-presenting antigen to CD81 T Human and murine spleens are very similar, but minor cells.6 pDCs are the major producers of type I interferon.7 structural differences in the organization of the white pulp Previously, this lab identified a novel splenic dendritic-like cell are evident. The human spleen differs in that it contains a type called L-DCs in murine spleen with high potential to perifollicular region external to the marginal zone in the red induce CD81 T-cell responses but with no capacity to induce pulp, which is absent in murine spleen.4 In general, human and CD41 T-cell responses.8,9 Murine L-DCs are distinguishable by murine spleens both contain equivalent lymphoid and myeloid very high endocytic capacity and a lack of MHC-II expression.

1Research School of Biology, The Australian National University, Canberra, Australia; 2John Curtin School of Medical Research, The Australian National University, Canberra, Australia; 3ANU Medical School, College of Medicine Biology and Environment, The Australian National University, Canberra, Australia and 4Department of Haematology, Canberra Hospital, Canberra, Australia Correspondence: Professor HC O’Neill, Research School of Biology, The Australian National University, Canberra, ACT 0200, Australia. E-mail: [email protected] Received: 18 December 2014; Revised: 1 February 2015; Accepted: 1 February 2015 Human splenic dendritic cells S Petvises et al 444

Recent studies have shown that these same cells can be gener- Preparation of human splenocytes ated through in vitro hematopoiesis when murine spleen- or Fresh and/or cryopreserved human spleen samples were bone marrow-derived hematopoietic stem/progenitor cells are obtained from the ACT Haematology Research Tissue Bank cocultured with the splenic stromal cell line 5G3.10–14 The L- (Canberra Hospital, Canberra, Australia) with informed con- DCs produced are functionally distinct in their ability to acti- sent from patients undergoing splenectomies for suspected or vate CD81 but not CD41 T cells. Currently, L-DCs are thought known immune thrombocytopenia and/or lymphoma. Spleens to reflect a tissue-specific APC derived from progenitors endo- selected for this study had no histological evidence of lym- genous to spleen.15 The currently available evidence supports phoma. Spleen tissues were kept aseptically on ice and pro- the possible role of L-DCs in antigen presentation of blood- cessed within 12 h of removal. The tissues were chopped into borne antigens for CD81 T-cell activation. small pieces of 0.5–1.0 cm. Cell suspensions were obtained by DC development in the human spleen has not been as exten- forcing the tissue through a fine mesh sieve into DMEM with sively studied as it has in the murine spleen. In humans, DC washing by centrifugation at 300g at 4uC for 5 min. For the lysis subsets have been investigated mainly in blood due to greater of red blood cells, cells were resuspended in lysis buffer 1 1 accessibility of cells. Human blood DCs are HLA-DR CD11c (140 mM NH4Cl, 17 mM Tris base, pH 7.5) at room temper- cells, and four subsets have been differentiated according to the ature for 5 min; then cells were washed twice with DMEM by expression of hCD16, hCD1b/c, hCD304 and CD141.16–18 The centrifugation at 300g at 4uC for 5 min and resuspended at the hCD161 DC subset has been identified in both spleen and desired concentrations. blood.19 These cells resemble murine monocyte-derived DCs and are identifiable as hCD11c1hCD11b1HLA-DR1hCD161 Preparation of T and B cell-depleted spleen cells.19,20 The hCD1c1 DCs, delineated as hCD11c1hCD11b1 Splenocytes were depleted of T and B cells using MACS meth- HLA-DR1hCD1c1 cells, resemble murine myeloid DCs.21 These odology (Miltenyi Biotec: North Ryde, NSW, Australia). cells are the homolog of murine CD11b1CD41 DCs.22,23 The Biotinylated antibodies specific for hCD19 (HIB18; Biolegend, pDC subset comprises cells with a hCD11c1hCD11bloHLA- San Gabriel, CA, USA) and hCD3 (OKT3; Biolegend) were 7 DR1hCD1231 phenotype, which are the homolog of cells added to 10 cells and incubated on ice for 20 min. The cell found in murine spleen and blood.24,25 The hCD1411 DC, suspension was then washed twice with labeling buffer (degassed identified as Lin2hCD11c1HLA-DR1hCD1411 cells, are the PBS/pH 7.2, 0.5% bovine serum albumin and 2 mM EDTA) by murine homologue of CD1411 DCs.22,26 centrifugation at 300g at 4uC for 5 min. The supernatant was Information about DC subsets in the human spleen has discarded, and the cells were resuspended in 20 mlofMACSanti- been limited by tissue availability. Recently, Mittag and biotin microbeads. The cells were further incubated on ice for colleagues20 identified the same four subsets of DC in human 20 min and then washed twice by centrifugation. The cells were spleen and showed that they resemble subsets that were pre- resuspended in 500 ml labeling buffer and transferred to a MACS viously identified in human peripheral blood. However, the MS column placed in a SuperMACS II Separator (Miltenyi presence of a human equivalent of the L-DC subset described Biotec). The column was washed three times with 3 ml of label- in murine spleen has not yet been investigated. Here, human ing buffer, and the flow-through cells were collected and washed. . spleens were analyzed for DC and monocyte subsets using This procedure routinely gave 95% purity. antibody staining and flow cytometric analysis. Human Antibody staining spleens were also tested for the presence of L-DC progenitors For antibody staining, cells were washed and resuspended in by coculturing splenocytes with the murine 5G3 splenic stro- fluorescence-activated cell sorting (FACS) buffer (DMEM, mal line which supports the hematopoiesis of murine sple- 0.1% sodium azide, 1% fetal calf serum) at 13105–53105 nocytes. Cells produced in cocultures were phenotypically cells/100 ml. A 100 ml volume of cells was plated into a flexible and functionally characterized, and compared with murine 96-well microtiter plate and centrifuged at 300g at 4uC for L-DCs. 5 min. The cells were resuspended in 25 ml of FACS buffer containing 40 mg/ml of Fc block specific for FccII/IIIR MATERIALS AND METHODS (eBioscience, San Diego, CA, USA) for 15mins on ice before Splenic stromal cultures washing by centrifugation and discarding the supernatant. A The 5G3 splenic stromal line supports hematopoiesis in cocul- 25 ml aliquot of diluted primary antibody was added to the cells, tured bone marrow and splenocytes.27–29 The 5G3 stromal cells which were then incubated in the dark on ice for 25 min. The were maintained by scraping up adherent cells to passage every cells were washed twice with 150 ml of FACS buffer by centrifu- 3–4 days. Stromal cells were cultured at 37uCin5%CO2 in gation. When required, a 25 ml aliquot of diluted streptavidin Dulbecco’s modified Eagle’s medium (DMEM) supplemented conjugate was added to the cells, followed by incubation and with 4 mg/ml glucose, 6 mg/ml folic acid, 36 mg/ml L-aspara- washing as above. The cells were resuspended in 70 ml of FACS gine, 116 mg/ml L-asparagine hydrochloric acid, to which was buffer for flow cytometric analysis. Dead cell exclusion was added 10% fetal calf serum, 531024 M 2-mercaptoethanol, achieved by the addition of a final concentration of 1 mg/ml 10 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid, propidium iodide (PI) to the cells prior to flow cytometric 100 U/ml penicillin and 100 mg/ml streptomycin (sDMEM). analysis.

Cellular & Molecular Immunology Human splenic dendritic cells S Petvises et al 445

For multicolor staining, cocktails of up to five different prim- RESULTS ary antibodies were added in the first incubation step. Isotype The human spleen contains cells resembling murine L-DCs control antibodies and ‘fluorescence minus one controls’ were Human spleens were assessed for the presence of DC subsets used where appropriate to determine the background binding of using antibody staining and flow cytometric analysis. The ini- antibodies and to set gates. Antibodies and isotype controls were tial analysis was performed on T and B cell-depleted spleno- purchased from Biolegend. Primary antibodies specific for cytes. Gating on the basis of HLA-DR expression and CD11c or hCD11c (clone 3.9), HLA-DR (L243), hCD16 (3G8), hCD1c CD11b revealed five subsets that were distinguishable by fur- (L161), hCD123 (6H6), hCD80 (2D10), hCD86 (IT2.2), ther staining as DCs or dendritic-like L-DCs. Three subsets hCD141 (M80) and hCD14 (HCD14). To stain lineage (Lin)1 were found within the hCD11c1HLA-DR1 population, cells, a cocktail of biotinylated antibodies specific for human including hCD161 DCs, hCD1c1 DCs and hCD1231 pDCs, hematopoietic lineage cells was added to cell suspensions in line with previous studies20 (Figure 1a). Further analysis (10 mlper107 cells) and incubated for 10 min on ice. These were revealed that splenic pDCs expressed both hCD86 and specific forhCD3 (OKT3), hCD14 (HCD14), hCD19 (HIB18) hCD80, confirming their mature pDC status (Figure 1a). 1 and hCD56 (HCD56), which detect mature T cells, B cells, hCD141 DCs were delineated after lineage (Lin) cocktail anti- 1 natural killer cells, monocytes/macrophages and granulocytes. bodies were used to gate out mature Lin cells. Some gated T cells, B cells and myeloid cells expressed low levels of CD11c, so 1 Flow cytometry the selected CD11c population in Figure 1c was smaller than Flow cytometry was performed on an LSRII FACS machine that in Figure 1a. Further staining was used to detect HLA-DR, hCD11c and hCD141 (Figure 1c). A fifth population of DCs (Becton Dickinson: Franklin Lakes, NJ, USA). Voltages, para- 1 2 meters and event counts were set and determined using BD was identified within the hCD11c HLA-DR subset. Staining for hCD11b, hCD80 and hCD86 revealed a subset of FACSDIVA software (Becton Dickinson). Single-color controls 1 2 1 2 1 were used to set compensation. FlowJo software (FlowJo, hCD11c HLA-DR hCD11b hCD80 hCD86 cells. These were named human L-DCs because of their similarity to the Ashland, OR, USA) was used to analyze the data. In general, 9,12 cells were gated on the basis of forward scatter and side scatter. murine L-DC subset described previously (Figure 1b). Live cells were then gated out by the absence of PI staining. Analysis of monocyte subsets was also performed by flow Isotype-matched and ‘fluorescence minus one’ controls were cytometry. Initially, hCD14 and HLA-DR staining and gating used to set gates and to distinguish specific antibody binding. were used to identify the monocytic lineage. Given that human monocytes are HLA-DR1, the cells were initially divided by hCD14 and hCD16 expression to delineate three monocyte Endocytosis assay subpopulations (Figure 1d). Classical monocytes were iden- The capacity of cells to endocytose antigen was assessed by in tified as hCD1411hCD162 cells, intermediate monocytes as vitro measurement of endocytosis. The cells were washed twice hCD1411hCD1611 cells, and non-classical monocytes as with sDMEM and placed on ice for 10 min. FITC-conjugated hCD141hCD1611 cells. ovalbumin (OVA-FITC) (Molecular Probes, Eugene, OR, The proportion of each DC and monocyte subset was cal- USA) was added to the cells at a concentration of 2 mg/ml, culated with respect to the total number of nucleated cells in and the samples incubated at 37 uC for 45 min. Endocytosis the spleen using three different spleen samples. L-DCs were was halted by the addition of 100 ml of chilled PBS with 0.1% present in approximately equivalent numbers to hCD161 NaN . The cells were washed three times by centrifugation at 1 3 DCs, pDCs and hCD1 DCs; the size of these subsets were 300g at 4 C for 5 min. The supernatant was discarded, and the u not significantly different from one another (Wilcoxon Rank cells resuspended in FACS buffer for flow cytometric analysis. 1 Sum test; P # 0.05) (Figure 2). The hCD141 DC subset was the smallest population in the spleen and was present in Establishment of stromal cocultures significantly lower numbers than the other DC subsets but Initially, 5G3 stromal cells were grown to 80%–90% confluency. was not significantly different in number to any of the mono- Co-cultures established by plating T and B cell-depleted sple- cyte subsets (Figure 2). Each of the monocyte subsets was nocytes onto a 5G3 monolayer to study in vitro hematopoiesis found to be present in similar numbers with no significant and DC development. Cocultures were maintained at 37uC, 5% difference between them, although there were significantly CO2 and 95% humidity. Non-adherent cells were collected by fewer of them compared with the L-DC, hCD161 DC, pDC gently shaking the flask, followed by removal and replacement and hCD11 DC subsets (Wilcoxon Rank Sum test; P # 0.05) of the media at 3–4 day intervals. The cells were then analyzed (Figure 2). for cell surface marker expression using antibody staining and flow cytometric analysis at 14, 21 and 28 days. Murine splenic stroma supports development of human splenic DCs Statistical analysis Cocultures were established by plating T and B cell-depleted Data are presented as the mean6standard error (s.e.) for sam- human splenocytes on top of murine 5G3 stromal cells, and ple size n53. Given the small sample sizes, the Wilcoxon Rank maintaining these cultures for 28 days. This experiment assumes Sum Test was used to determine significance (P # 0.05). that murine and human splenic stromal cells are similar and

Cellular & Molecular Immunology Human splenic dendritic cells S Petvises et al 446

a 105 93 23.7 5.9 20.4 4 10 80.1 CD16+ DC 103 SSC SSC hCD16 102 hCD11c 0 0102 103 104 105 FSC HLA-DR hCD11b PI 8.3

b CD1c+ DC 0.1

25.6 hCD1c L-DC hCD11b hCD86 hCD80 0.2 97.7 0.2 92.5 hCD11b hCD11b 14.9 pDC hCD86 hCD80 hCD123 0 2.1 0 7.3 hCD11b HLA-DR HLA-DR c d

8.5 17.2 11.7 7.0 CD141+ DC Lin 6.9 hCD16 hCD141 5.6 HLA-DR hCD11c HLA-DR hCD14 hCD14 CD14++ CD16– Classical monocytes CD14++ CD16++ Intermediate monocytes CD14+ CD16++ Non-classical monocytes

Figure 1 Delineation of DC and monocyte subsets in human spleens. Human splenocytes were enriched by depletion of T and B cells and stained with antibodies to distinguish subsets by flow cytometry. PI staining was used to discriminate dead cells. Gates were set on bivariate plots using isotype control antibodies. Numbers on the gates reflect the percentage of positive cells. (a) HLA-DR and hCD11c staining was used to identify HLA-DR1hCD11c1 DCs, which were further divided into three distinct subsets with hCD11b, hCD11c, hCD16 and hCD1c staining: hCD161 DCs, hCD11 DCs and plasmacytoid (p) DCs. Expression of hCD80, hCD86 and hCD123 was used to identify pDCs as mature hCD1231hCD11c1hCD11b2hCD801hCD861 cells. (b) The HLA-DR2hCD11c1 subset was positive for hCDllb and hCD86 but not hCD80; thus, it resembled murine L-DCs. (c) In two separate experiments, a lineage (Lin) cocktail of antibodies (hCD3, hCD14, hCD19 and hCD56) was used to exclude Lin1 cells and to gate on hCD11c1 cells. CD1411 DCs were further identified as HLA-DR1.(d) Staining for HLA-DR was used to delineate monocytes. The three commonly known subsets of classical, intermediate and non-classical monocytes were then distinguishable on the basis of differences in hCD14 and hCD16 expression. Results from one spleen are reflective of three individual spleens. DC, dendritic cell; PI, propidium iodide.

that they can act as a niche or cell signaling matrix for in vitro Murine 5G3 stroma can therefore support the development hematopoiesis. Previous studies have demonstrated the ability of HLA-DR1 cells resembling DCs from T and B cell-depleted of the murine 5G3 splenic stroma to support myelopoiesis from human spleens over 28 days of coculture. Given that an hematopoietic stem/progenitor cells present in both murine extended period is necessary, it is thought that these cells arise bone marrow and spleen.12,14,15,30,31 Cell growth was moni- by hematopoiesis from progenitors present in the enriched tored, and non-adherent cells were collected and flow cytome- spleen cell population. The proportion of hCD11chiHLA-DR1 trically assessed at 7-day intervals. Live cells were gated, and cells produced increased to 30% by 28 days, reflecting the ability HLA-DR versus hCD11c staining was used to delineate any of 5G3 stroma to support human DC development. On the basis dendritic-like cells produced in the cocultures. of high CD11c and MHC-II expression, these cells resemble Three subsets of cells were identified from these cultures cDC-like cells, previously reported as a transient early popu- after 21 days: hCD11chiHLA-DR1, hCD11cloHLA-DR1 and lation produced in murine splenic cocultures.10,11 These cells hCD11cloHLA-DR2cells (Figure 3a). All subsets expressed disappear after 21–28 days of coculture. The hCD11cloHLA- hCD11b, hCD80 and hCD86, indicating production of mature DR2 subset was consistently produced and represented a maxi- dendritic-like cells in these cocultures. The cells produced in mum of 17% of cells over time. The expression of hCD80 and these cocultures were large, as evidenced by their high FSC hCD86 on these cells was higher than on the hCD11c1HLA- profile. Over time, no hCD1231 lymphoid cells were observed DR1 subset of cDC-like cells (Figure 3b). Phenotypically, in any cocultures. this hCD11cloHLA-DR2 subset resembles L-DCs produced in

Cellular & Molecular Immunology Human splenic dendritic cells S Petvises et al 447

Human splenic DC and monocytes DISCUSSION 1.5 Previous studies from this lab characterized L-DCs in murine

1.2 spleens as a novel APC subset that is highly endocytic and capable of cross-presenting antigens to CD81 T cells.9 In an 0.9 attempt to identify a human counterpart cell type, spleens collected from splenectomized patients were analyzed by antibody 0.6 staining and flow cytometry. Our study initially identified 1 1 2 % nucleated cells human L-DCs in the spleen as hCD11b hCD11c HLA-DR 0.3 cells which express the hCD86 costimulatory molecule but not 0 hCD80. The expression of only hCD86 on the cell surface of ex vivo human L-DCs reflects the phenotype of the murine equi- DC DC DC L-DC + PDC + + valent cell type9, and may indicate its status as an immature CD16 CD1c CD141 monocytes monocytes– monocytes ++ ++ dendritic-like cell type.

CD16 DCs in humans have been studied mainly in peripheral blood, + CD16 CD16 ++ ++ with four subsets identifiable amongst the Lin2hCD11c1HLA- CD14 CD14 CD14 DR1 DC population.32 Of these, hCD1c1 DCs and hCD1411 DCs represent myeloid DC subsets, while hCD161 DCs are Figure 2 Representation of DC and monocyte subsets in human considered monocyte-derived DCs.17,33 pDCs, the major produ- spleen. Splenocytes were analyzed by antibody staining and flow cyto- cers of type I interferon, are identified by the expression of metry as described in Figure 1. Cell frequency was calculated in terms 33 of the number of each cell type in relation to the total number of hCD123, hCD303 and hCD304. This study has identified 1 1 1 1 nucleated cells. Graphs show the means6s.e. for triplicate experi- hCD1c DCs, hCD16 DCs, hCD141 DCs and hCD123 ments, which involved the analysis of three individual spleen samples. pDCs in human spleens, which is consistent with previous find- Statistical significance was assessed using the Wilcoxon Rank Sum ings and resembles the subsets found in blood.20 The expression test (P # 0.05). No significant differences were detected between the of hCD86 and hCD80 by hCD1231 pDCs probably reflects the three monocyte subsets or between the three DC subsets of L-DCs, mature state of these cells in the human spleen. hCD161 DCs and pDCs, although each of the monocyte subsets was present in significantly smaller numbers than the three dominant DC Human monocytes have been characterized as cells that subsets (P values not shown). The hCD1411 DC subset was present express HLA-DR, a property that distinguishes them from mur- 2 34,35 in significantly lower numbers than the other four DC subsets. DC, ine monocytes which are MHC-II . Heterogeneity was dendritic cell. observed amongst monocytes in human spleens on the basis of their hCD14 and hCD16 expression. Three subpopulations of monocytes were identified as classical monocytes (hCD1411 murine cocultures, although in this human/mouse system, the hCD162), intermediate monocytes (hCD1411hCD1611)and 1 1 hCD11c HLA-DR cDC-like cells predominate over the L-DC non-classical monocytes (hCD141hCD1611). The phenotype 1 subset. The hCD11cloHLA-DR subset represents the majority of DC and monocyte subsets in human blood and human of cells produced in the cocultures, and the proportion of these spleen appear to be very similar, probably because of the cells approached 50% at day 21 (Figure 3b). This hCD11cloHLA- physiological influx and efflux of the same cells between blood 1 DR subset is unique to human cocultures and was not and spleen.36,37 observed in the murine system. Previously, murine splenic stroma was demonstrated to support in vitro hematopoiesisby its ability to produce L-DCs from Human DCs produced in cocultures are endocytic hematopoietic stem cells, as well as cDC-like cells. L-DCs Co-cultures of human spleen cells and 5G3 murine stromal reflect a novel type of APC that is distinguishable by stroma- cells were established for DC production. Cells collected at dependent development, lack of MHC-II expression, and high day 14 were incubated with OVA-FITC to measure endocyto- endocytic activity, which is consistent with their ability to effi- sis. The cells were then stained with HLA-DR and hCD11c to ciently cross-present antigens to CD8 T lymphocytes.10,11,38,39 delineate the three DC subsets described in Figure 3a and to The in vivo equivalent of L-DCs was later characterized in compare their endocytic potential. The hCD11chiHLA-DR1 mouse spleens and shown to have similar phenotype and func- cDC-like population, the hCD11cloHLA-DR1 novel subset tion.9 In cocultures of human splenocytes with the 5G3 murine and the hCD11cloHLA-DR2 L-DC-like subset produced in stromal line, several subsets of DC-like cells are produced, and cocultures were all capable of endocytosis (Figure 4). A subset they differ in their expression of HLA-DR and hCD11c. Three of hCD11c2HLA-DR2 cells was unable to take up OVA anti- subsets were apparent: the hCD11cloHLA-DR1 and hCD11chi gen and thus may represent myeloid cells or precursors which HLA-DR1 subsets of cDC-like cells, which are the majority of are known to be produced transiently in murine cocultures but cells present in the cocultures, and a minor hCD11c1 which disappear after approximately 21 days. While this popu- HLA-DR2 subset, which resembles murine L-DCs. All DC sub- lation is shown in Figure 4, it is not apparent in Figure 3, which sets produced appear to be activated in the in vitro cultures, as depicts cells from 28 day cocultures. indicated by their expression of both hCD80 and hCD86.17,33

Cellular & Molecular Immunology Human splenic dendritic cells S Petvises et al 448

a Human spleen co-culture over 5G3 Live PI– cells 5 5 41.0 57.8 5 33.2 66.8 5 0 0.6 5 10 21 10 10 10 10 104 16.9 104 104 104 104 103 103 103 103 103 SSC hCD86 48 hCD80 hCD11c 2 2 2 2 2 10 10 10 hCD123 10 10 0 0 0.6 0.6 0 000 44.7 54.7 0 0102 103 104 105 0102 103 104 105 0102 103 104 105 0102 103 104 105 0 50K 100K 150K 200K 250K HLA-DR hCD11b hCD11b hCD11b FSC

105 29.9 65.6 105 32.0 63.6 105 0 0.9 105 104 104 104 104

103 103 103 103 SSC

hCD86 2 hCD80 2 2 2 10 10 hCD123 10 10 0 2.5 2.0 0 2.9 1.5 0 32.4 66.7 0 0102 103 104 105 0102 103 104 105 0102 103 104 105 0 50K 100K 150K 200K 250K hCD11b hCD11b hCD11b FSC

105 24.8 67.4 105 25.1 70.0 105 0 0.4 105 104 104 104 104

103 103 103 103 SSC

hCD86 2 hCD80 2 2 2 10 10 hCD123 10 10 0 5.0 2.8 0 2.6 2.3 0 31.0 68.6 0 0102 103 104 105 0102 103 104 105 0102 103 104 105 0 50K 100K 150K 200K 250K hCD11b hCD11b hCD11b FSC

b Cell production

lo – 75 hCD11c HLA-DR + + + lo – + 60 hCD11c HLA-DR (hCD86 ) 100 hCD11c HLA-DR (hCD86 ) hCD11cloHLA-DR+ hCD11c+HLA-DR+(hCD80+) hCD11cloHLA-DR–(hCD80+) hCD11chiHLA-DR+ 45 75 50

30 50 % cells % cells 25 15 25

0 0 0 14 21 28 14 21 28 14 21 28 Days Days

Figure 3 Production of DCs in cocultures of human spleens with murine stroma. Human splenocytes were isolated by red blood cell lysis and T and B cell depletion. These cells were plated on a mono layer of the 5G3 murine splenic stromal line and maintained for 28 days. Over time, non- adherent cells were collected from the cultures and stained with antibodies for HLA-DR, hCD11c, hCD11b, hCD80, hCD86 and hCD123 to analyze the types of cells produced. PI staining was used to discriminate viable (PI2) cells. Gates were set on bivariate plots using isotype control antibody staining. Control staining is also shown as an overlay. (a) Cells were analyzed at 21 days by gating on hCD11c1 cells and their HLA-DR1 and HLA- DR2 subsets, which were further analyzed for the expression of DC markers. (b) The proportions of cell subsets produced over time were calculated. Graphs represent the mean6s.e. for triplicate cocultures. DC, dendritic cell; PI, propidium iodide.

Each of these subsets is endocytic and actively takes up FITC- produced in the murine system. While the subsets were present OVA antigen, which demonstrates their capacity to present in different relative proportions, they resemble the cDC-like antigen to T cells. The low yield of cells from human cocultures and L-DC-like cells produced in murine cocultures in terms of compared with murine stroma has precluded an analysis of the surface marker phenotype and endocytic capacity. In line with antigen-presenting capacity of these cells in T cell activation the murine cells, the L-DC-like subset in humans is a CD11clo assays. In contrast, cocultures of murine splenocytes with the subset, which may include both the HLA-DR2 and the novel murine 5G3 stromal line produce higher yields of cells. The sub- HLA-DR1 subsets, each of which has endocytic capacity. An L- set of hCD11cloHLA-DR2 cells appears most similar to L-DCs, DC equivalent cell in humans may therefore express HLA-DR. with the high endocytic capacity observed in the equivalent mur- A precedent for this hypothesis is exemplified by human mono- ine subset. cytes, which are HLA-DR1 cells, while similar monocyte sub- It is noteworthy that murine splenic stroma can suffice as a sets in mouse are MHC-II2. Another explanation is that the support matrix for hematopoiesis of human splenic hemato- hCD11cloHLA-DR2 population becomes activated due to in poietic progenitors. This is consistent with species conser- vitro culture and upregulates expression of MHC-II. This would vation of stromal cells, the molecules they produce, and the be consistent with the finding that only a hCD11cloHLA-DR2 signaling pathways that support hematopoiesis. It is also population was detectable amongst ex vivo-isolated splenocytes. important to note the high level of similarity between DC sub- Further assessment of the capacity of these cells to activate sets produced in the hybrid human/murine system and those CD41 and CD81 T cells and to cross-present antigen will be

Cellular & Molecular Immunology Human splenic dendritic cells S Petvises et al 449

Human spleen/5G3 endocytosis HLA-DR+hCD11c+ to that in mice in terms of function and phenotype verifies the 100 murine findings and supports the existence of this novel subset. 80 Its function in immunity is therefore of interest in humans, particularly during disease. 60

% of Max ACKNOWLEDGEMENTS 20 This work was supported by grants from the Australian Research Council (#DP13010703) and the National Health and Medical 0 0 102 103 104 105 Research Foundation of Australia (project grant # 585443). OVA FITC

PI– HLA-DR+hCD11clo 1 Brendolan A, Rosado MM, Carsetti R, Selleri L, Dear TN. Development 100 5 10 5.6 and function of the mammalian spleen. BioEssays 2007; 29: 166– 80 177. 104 2 Wolber FM, Leonard E, Michael S, Orschell-Traycoff CM, Yoder MC, 10.8 8.0 60 Srour EF. Roles of spleen and liver in development of the murine 103 hematopoietic system. Exp Hematol 2002; 30: 1010–1019. hCD11c 40 % of Max 3 Dor FJ, Ramirez ML, Parmar K, Altman EL, Huang CA, Down JD et al. 102 63.3 20 Primitive hematopoietic cell populations reside in the spleen: studies 0 in the pig, baboon, and human. Exp Hematol 2006; 34: 1573–1582. 0 4 Mebius RE, Kraal G. Structure and function of the spleen. Nat Rev 0 102 103 104 105 0 102 103 104 105 Immunol 2005; 5: 606–616. HLA-DR OVA FITC 5 de Smedt T, Pajak B, Muraille E, Lespagnard L, Heinen E, de Baetselier P et al. Regulation of dendritic cell numbers and HLA-DR–hCD11clo maturation by lipopolysaccharide in vivo. JExpMed1996; 184: 100 1413–1424. 6 Shortman K, Heath WR. The CD81 dendritic cell subset. Immunol Rev 80 2010; 234: 18–31. 60 7 Shortman K, Sathe P, Vremec D, Naik S, O’Keeffe M. Plasmacytoid dendritic cell development. Adv Immunol 2013; 120: 105–126. 40 8 Tan JK, Periasamy P, O’Neill HC. Delineation of precursors in murine % of Max spleen that develop in contact with splenic endothelium to give novel 20 dendritic-like cells. Blood 2010; 115: 3678–3685. 0 9 Tan JK, Quah BJ, Griffiths KL, Periasamy P, Hey YY, O’Neill HC. 0 102 103 104 105 Identification of a novel antigen cross-presenting cell type in OVA FITC spleen. J Cell Mol Med 2011; 15: 1189–1199. 10 Periasamy P, O’Neill HC. Stroma-dependent development of two HLA-DR–hCD11c– dendritic-like cell types with distinct antigen presenting capability. 100 Exp Hematol 2013; 41: 281–292. 11 Periasamy P, Petvises S, O’Neill HC. Development of two distinct 80 dendritic-like APCs in the context of splenic stroma. Front Immunol 60 2013; 4: 73. 12 Periasamy P, Tan JK, Griffiths KL, O’Neill HC. Splenic stromal niches 40 support hematopoiesis of dendritic-like cells from precursors in bone % of Max marrow and spleen. Exp Hematol 2009; 37: 1060–1071. 20 13 O’Neill HC, Griffiths KL, Periasamy P, Hinton RA, Petvises S, Hey YY 0 et al. Spleen stroma maintains progenitors and supports long-term 0 102 103 104 105 hematopoiesis. Curr Stem Cell Res Ther 2014; 9: 354–363. OVA FITC 14 Petvises S, O’Neill HC. Characterisation of dendritic cells arising from progenitors endogenous to murine spleen. PLoS One 2014; 9: Figure 4 Human DCs produced in cocultures are endocytic. Human e88311. DCs were produced in murine stromal cocultures over 14 days as 15 Hinton RA, O’Neill HC. Technical advance: in vitro production of described in Figure 3. They were incubated with 10 mg/ml OVA-FITC distinct dendritic-like antigen-presenting cells from self-renewing hematopoietic stem cells. J Leuk Biol 2012; 91: 341–346. at 37uC for 45 min prior to measurement of antigen uptake by flow 16 Dzionek A, Fuchs A, Schmidt P, Cremer S, Zysk M, Miltenyi S et al. cytometry. Staining for HLA-DR and hCD11c was used to delineate 1 1 1 lo 2 lo BDCA-2, BDCA-3, and BDCA-4: three markers for distinct subsets of HLA-DR hCD11c ,HLA-DR hCD11c ,HLA-DR hCD11c and dendritic cells in human peripheral blood. J Immunol 2000; 165: 2 2 HLA-DR hCD11c subsets. Endocytic activity was indicated by greater 6037–6046. uptake of OVA-FITC in cells treated at 37uC, compared with control cells 17 MacDonald KP, Munster DJ, Clark GJ, Dzionek A, Schmitz J, Hart DN. treated with OVA-FITC at 0uC. DC, dendritic cell. Characterization of human blood dendritic cell subsets. Blood 2002; 100: 4512–4520. performed once human splenic stromal cells are isolated, as 18 Palucka K, Banchereau J. Human dendritic cell subsets in these may yield higher numbers of cells in vitro and may facil- vaccination. Curr Opin Immunol 2013; 25: 396–402. 19 Grage-Griebenow E, Zawatzky R, Kahlert H, Brade L, Flad H, Ernst M. itate the further identification of an L-DC equivalent subset. Identification of a novel dendritic cell-like subset of CD641/CD161 Demonstration of a subset in the human spleen with similarities blood monocytes. Eur J Immunol 2001; 31: 48–56.

Cellular & Molecular Immunology Human splenic dendritic cells S Petvises et al 450

20 Mittag D, Proietto AI, Loudovaris T, Mannering SI, Vremec D, 30 Periasamy P, Tan JK, O’Neill HC. Novel splenic antigen-presenting Shortman K et al. Human dendritic cell subsets from spleen and cells derive from a Lin-c-kitlo progenitor. J Leuk Biol 2013; 93: 63– blood are similar in phenotype and function but modified by donor 69. health status. J Immunol 2011; 186: 6207–6217. 31 Tan JK, O’Neill HC. Haematopoietic stem cells in spleen have distinct 21 Shortman K, Liu YJ. Mouse and human dendritic cell subtypes. Nat differentiative potential for antigen presenting cells. J Cell Mol Med Rev Immunol 2002; 2: 151–161. 2010; 14: 2144–2150. 22 Haniffa M, Shin A, Bigley V, McGovern N, Teo P, See P et al. Human 32 Ziegler-Heitbrock L. The CD141 CD161 blood monocytes: their role tissues contain CD141hi cross-presenting dendritic cells with in infection and inflammation. J Leuk Biol 2007; 81: 584–592. functional homology to mouse CD1031 nonlymphoid dendritic 33 Kassianos AJ, Jongbloed SL, Hart DN, Radford KJ. Isolation of human cells. Immunity 2012; 37: 60–73. blood DC subtypes. Methods Mol Biol 2010; 595: 45–54. 23 Schlitzer A, Ginhoux F. Organization of the mouse and human DC 34 Geissmann F, Manz MG, Jung S, Sieweke MH, Merad M, Ley K. network. Curr Opin Immunol 2014; 26: 90–99. Development of monocytes, macrophages, and dendritic cells. 24 Onai N, Kurabayashi K, Hosoi-Amaike M, Toyama-Sorimachi N, Science 2010; 327: 656–661. Matsushima K, Inaba K et al. A clonogenic progenitor with 35 Sunderko¨tter C, Nilolic T, Dillon MJ, van Rooijen N, Stehling M, prominent plasmacytoid dendritic cell developmental potential. Drevets DA et al. Subpopulations of mouse blood monocytes differ Immunity 2013; 38: 943–957. in maturation stage and inflammatory response. J Immunol 2004; 25 Sathe P, Vremec D, Wu L, Corcoran L, Shortman K. Convergent 172: 4410–4417. differentiation: myeloid and lymphoid pathways to murine 36 Smeekens SP, van de Veerdonk FL, Joosten LA, Jacobs L, Jansen T, plasmacytoid dendritic cells. Blood 2013; 121: 11–19. Williams DL et al. The classical CD1411 CD162 monocytes, but not 26 Jongbloed SL, Kassianos AJ, McDonald KJ, Clark GJ, Ju X, Angel CE the patrolling CD141 CD161 monocytes, promote Th17 responses to et al. Human CD1411 (BDCA-3)1 dendritic cells (DCs) represent a Candida albicans. Eur J Immunol 2011; 41: 2915–2924. unique myeloid DC subset that cross-presents necrotic cell antigens. 37 Wong KL, Tai JJ, Wong WC, Han H, Sem X, Yeap WH et al. Gene J Exp Med 2010; 207: 1247–1260. expression profiling reveals the defining features of the classical, 27 Despars G, O’Neill HC. Heterogeneity amongst splenic stromal cell intermediate, and nonclassical human monocyte subsets. Blood lines which support dendritic cell hematopoiesis. In Vitro Cell Dev 2011; 118: e16–e31. Biol Anim 2006; 42: 208–215. 38 O’Neill HC, Wilson HL, Quah B, Abbey JL, Despars G, Ni K. Dendritic 28 Ni K, O’Neill HC. Long-term stromal cultures produce dendritic-like cell development in long-term spleen stromal cultures. Stem Cells cells. Br J Haematol 1997; 97: 710–725. 2004; 22: 475–486. 29 Ni K, O’Neill HC. Hemopoiesis in long-term stroma-dependent cultures 39 Quah B, Ni K, O’Neill HC. In vitro hematopoiesis produces a distinct from lymphoid tissue: production of cells with myeloid/dendritic class of immature dendritic cells from spleen progenitors with limited characteristics. In Vitro Cell Dev Biol Anim 1998; 34: 298–307. T cell stimulation capacity. Int Immunol 2004; 16: 567–577.

Cellular & Molecular Immunology