Distinct dendritic cell subsets exist as rare populations in lymphoid organs, and therefore are challenging to isolate in sufficient numbers and purity for immunological experiments. Here we describe a high efficiency, high yield method for isolation of all of the currently known major subsets of mouse splenic dendritic cells.
Dendritic cells (DCs) are professional antigen-presenting cells primarily responsible for acquiring, processing and presenting antigens on antigen presenting molecules to initiate T-cell-mediated immunity. Dendritic cells can be separated into several phenotypically and functionally heterogeneous subsets. Three important subsets of splenic dendritic cells are plasmacytoid, CD8αPos and CD8αNeg cells. The plasmacytoid DCs are natural producers of type I interferon and are important for anti-viral T cell immunity. The CD8αNeg DC subset is specialized for MHC class II antigen presentation and is centrally involved in priming CD4 T cells. The CD8αPos DCs are primarily responsible for cross-presentation of exogenous antigens and CD8 T cell priming. The CD8αPos DCs have been demonstrated to be most efficient at the presentation of glycolipid antigens by CD1d molecules to a specialized T cell population known as invariant natural killer T (iNKT) cells. Administration of Flt-3 ligand increases the frequency of migration of dendritic cell progenitors from bone marrow, ultimately resulting in expansion of dendritic cells in peripheral lymphoid organs in murine models. We have adapted this model to purify large numbers of functional dendritic cells for use in cell transfer experiments to compare in vivo proficiency of different DC subsets.
Dendritic cells (DC) were discovered almost forty years ago as the "large stellate (Greek dendron) cell" found in lymphoid organs1. Many studies have shown that DCs are the only antigen presenting cells that can effectively stimulate naïve T cells2. A major function of these cells is the uptake and presentation of antigens and their efficient processing and loading of these onto antigen presenting molecules. In mouse spleen, DCs can be separated into plasmacytoid and conventional subsets. The plasmacytoid DCs are characterized by low expression of CD11c and high levels of B220 and Gr-1. They are also positive for the surface marker mPDCA1 and secrete type I interferon in response to toll like receptor 9 (TLR9) ligands. The conventional DCs are high for CD11c and MHC class II expression. They can be split into three distinct subsets based on the surface expression of phenotypic markers such as CD4, CD8α, DEC205, CD11b and dendritic cell inhibitory receptor 2 (DCIR2, recognized by the 33D1 antibody) proteins 3,4. The CD8αPos DCs are also known as cDC1, are positive for DEC205, but negative for myeloid markers such as CD11b and 33D1. The CD8αNeg DCs, also called cDC2, are positive for 33D1, CD11b and CD4 but lack DEC205. The double negative subset (i.e., negative for both CD4 and CD8α) is relatively rare, and is negative for DEC205 and 33D1. It is the least characterized subset and may be a less differentiated form of CD8αNeg DC.
Phenotypic differences in the various DC subsets also extend to their in vivo functions. The CD8αNeg DCs are highly phagocytic and are thought to present exogenous antigen mainly via MHC class II to CD4 T cells 3. In contrast, the CD8αPos DCs are specialized for presentation of soluble protein antigen on MHC class I in a mechanism called cross-presentation. The outcome of cross-presentation depends on the activation status of these DCs5, and can either lead to expansion of cytotoxic T cells (CTLs) or development of regulatory T cells 62,7. Targeting of antigen to CD8αPos DCs using anti-DEC205-antibody-mediated delivery results largely in the deletion of T cells8, whereas presentation of antigens derived from infected apoptotic cells induces a strong CTL response 9.
In addition to recognition of peptide antigens, the mammalian immune system has evolved to recognize lipid and glycolipid antigens. These antigens are presented by CD1 molecules, which are MHC class I-like cell surface proteins that exist in multiple related forms in various mammals. In mice, a single type of highly conserved CD1 molecule called CD1d is responsible for presentation of glycolipid antigens 10. The major population of T cells that recognize CD1d/glycolipid complexes is called invariant NKT cells (iNKT cells). These cells express a semi-invariant T cell receptor (TCR) composed of an invariant TCRα chain that is paired with TCRβ chains that have limited diversity 11. Unlike conventional T cells that need to proliferate and differentiate to become activated effector T cells, iNKT cells exist as an effector population and start responding rapidly after glycolipid administration 12. Identification of physiologically relevant lipid antigen presenting cells is an active area of research, and several distinct cell types such as B cells, macrophages and DCs have been suggested to perform this function. However, it was demonstrated that the CD8αPos subset of DCs is the primary cell mediating uptake and presentation of lipid antigens to mouse iNKT cells 13 and glycolipid mediated cross-priming of CD8 T cells 14.
To compare the efficiency of antigen presentation by different antigen presenting cells, a straightforward approach is to transfer different types of purified APCs pulsed with equivalent amounts of antigen into naïve hosts. Cell transfer experiments of this type are often performed for immunological studies. However, performing transfer studies with ex vivo antigen treated DCs is challenging, since these cells exist as rare populations in lymphoid organs where they constitute less than 2% of total cells15. It is therefore necessary to enhance the development of these cells in donor animals to increase the efficiency of isolation protocols.
It is known that the common lymphoid and common myeloid progenitors, which are required for generation of pDC, CD8Pos and CD8Neg DC subsets, express fms-related receptor tyrosine kinase 3 (Flt-3). Upon in vivo Flt-3 ligand (Flt-3L) administration, emigration of Flt-3 expressing progenitor cells from bone marrow is increased, resulting in the increased seeding of peripheral lymphoid organs and the expansion of their mature DC progeny16. Expression of Flt-3 is lost during commitment to the B, T or NK cell differentiation pathways. Therefore, only minimal alterations are observed in these cells upon Flt-3L administration. Similar expansion in DC populations is observed in mice bearing tumors generated by implantation of a B16-melanoma cell line secreting murine Flt-3L, which provides a simple and economical method for providing sustained systemic levels of Flt-3L17,18. Using this approach, we have developed a protocol based on the implantation of B16-melanoma cells secreting Flt-3L to stimulate the expansion of all normal DC subsets in mouse spleen, thus greatly increasing the yields of these cells that can be isolated for subsequent experiments. We consistently find that within 10 – 14 days following subcutaneous implantation of the Flt-3 secreting tumor, mice develop splenomegaly with marked enrichment of DCs to constitute 40 – 60% of total spleen cells. From these spleens, different DC subsets can be isolated with high purity using standard commercially available cell purification kits that employ subset-specific phenotypic markers.
Animal experiments are done in accordance with approved guidelines from the institutional animal care and usage committee (IACUC). All procedures requiring sterility are performed in a biosafety cabinet.
1. Implantation of B16.Flt3L Melanoma in Mice
2. Preparation of Splenic Single Cell Suspension from Tumor Bearing Mice
3. Purification of Plasmacytoid DCs, CD8αPos DCs and CD8αNeg DCs from Splenic Single Cell Suspension
Note: Isolation of the splenic DC subsets from single cell suspension is a multi-step procedure as illustrated in Figure 1. Perform all the steps using pre-cooled buffer and an ice water bath to maintain temperature at 4 – 8 °C. All reagent volumes in the following section of the protocol are calculated for an initial input of 200 µl of splenic cell suspension at 108 cells/ml). This can be scaled up or down based on your need by changing the volume of cell suspension (always at a density of 1 X 108 cells/ml) and other reagents proportionately in the initial step (3.1.1 and 3.2.1). Adjust reagent volumes in all later steps accordingly. For example, if starting with 100 µl of splenic cell suspension at 1 x 108/ml, use half the stated reagent volumes in all steps.
4. Pulsing APCs with Antigens and Cell Transfer
The outcome of this procedure relies on the expansion of DC subsets by murine Flt-3L expressed by the implanted melanoma cells. The B16.Flt3L tumor was derived from a C57BL/6 mouse, and should be implanted into animals with that strain background in order to avoid failure of the tumor to become established due to rejection. In some cases, it may be desirable to use genetically modified mice to derive DCs with known defects in signaling pathways or receptors of interest. It is important to keep in mind that the tumor may develop with different kinetics in such animals, so it is important to monitor tumor growth based on appearance and palpation at the inoculation site. In our experience, once the tumor is palpable, the B16.Flt3L tumor bearing mice consistently show a marked increase in the splenic cellularity that correlates with expansion of all DC subsets (Figure 2A). We typically see a 2- to 3-fold increase in the number of total splenocytes obtained from B16.Flt3L melanoma-bearing mice compared to naïve animals. The increase in the number of DC in part contributes to the apparent reduction in the frequency of B cells, while there is little to no change in the absolute numbers of B cells. Interestingly, while a large expansion (~15- to 20-fold increase in the absolute numbers) is observed for conventional DCs in these mice, there is only a modest increase (about 4-fold) observed in the plasmacytoid DC subset.
The gating strategy to identify different DC subsets is shown in Figure 2A and is based on the expression of B220, CD11b, CD11c and CD8α as subset-specific phenotypic markers (Figure 2B). We also investigated DEC205, CD11b and mPDCA1 expression on these subsets as shown in Figure 2B. This shows the expected pattern, with mPDCA only on the pDC subset, whereas DEC205 is expressed highly on CD8αPos DCs and CD11b is detected most prominently on CD8αNeg DCs. We also analyzed the alterations in cell frequencies for each of the subsets purified at each step of the protocol described here (summarized in Figure 2C, with detailed examples of the flow cytometric data in Figure 3). The isolation of the splenic DC subsets described in this method is a multi-step procedure (Figure 1). Purification of plasmacytoid DCs relies on negative selection to enrich these cells after depletion of all undesired populations like B, T, NK cells and conventional DCs. Purification of the CD8αPos DCs, and CD8αNeg DCs is performed by positive selection of these cells after the depletion of T, B and NK cells. The high purity of each of the DC subsets obtained using this procedure (typically ~95%) is also illustrated in Figure 3.
Depending on the experiments to be performed, these cells can be pulsed with a desired antigen and transferred into histocompatible murine hosts for immunological studies. In addition, these cells can also be used for in vivo studies of stimulatory or regulatory activities of physiological DC subsets. The cell yields for the purified DC subsets per 2 x 107 total splenocytes are approximately one million pDCs, 2 million CD8αPos DCs and 4 million CD8αNeg DCs.
Figure 1. Isolation of DCs. A schematic illustrating the various steps of the protocol for isolation of different subsets of DCs. Please click here to view a larger version of this figure.
Figure 2. Analysis of Splenic DC Subsets in Mice with B16.Flt3L Tumors. A) Gating strategy is illustrated to identify the cell populations corresponding to pDCs (R1), CD8αPos DCs (R5) and CD8αNeg DCs (R6). B) Expression of mPDCA1, 33D1, DEC205 and CD11b on different subsets. The cells contained in gate R4 (a subset of lymphocytes; negative for B220, CD11c, CD11b) are used as a control population lacking the expression of these markers. C) Relative enrichment of DC and lymphocyte subsets isolated from the spleens of mice at day 14 after inoculation of B16.Flt3L melanoma cells is shown in the left plot as black bars, while the frequency of these cell populations in naïve mice is shown as white bars. The right plot displays the same data as cell numbers. Please click here to view a larger version of this figure.
Figure 3. Analysis of Splenic DC Subsets During Enrichment Protocol. A) Enrichment of plasmacytoid DCs (B220High and CD11cLow, R1) from 10% in the starting population of splenocytes from B16.Flt3L tumor bearing mice to ~95% purity in column flow through 1. Note that retentate cells eluted from this column are depleted of this population. B) Enrichment of conventional DCs (in CD11cHigh , mixture of CD8α positive and negative) from ~33% in the spleens of B16.Flt3L melanoma-bearing mice on day 10 after tumor implantation to ~78% in the flow through 2 fraction. In the column retentate 2 plots, the CD11c positive populations appear to be partially depleted, corresponding to the removal of substantial numbers of DCs during negative selection to deplete T, B and NK cells. Further fractionation of the flow through 2 suspension using positive selection for anti-CD8 binding yields >95% purity of CD8αPos DCs. The flow through from this was then subjected to positive selection for binding of anti-CD11c, yielding a population of >95% pure CD8αNeg DCs. Please click here to view a larger version of this figure.
Dendritic cells are accepted to be the major professional antigen presenting cells involved in the priming of T cell responses. Their main function is to survey the tissue microenvironment by taking up and processing antigens for presentation to T cells. In order to study the function of specific DC subsets, these need to be isolated in sufficient numbers using an approach that maintains their normal phenotype and functions. Most protocols rely on the isolation of specific DC subset from naïve mice. Since these cells are rare, the isolation procedures are not efficient and yield low numbers of cells. For example, one spleen will yield only about 0.5 - 1 million CD8pos DCs, while most experiments performed with transfer studies use at least this number of cells per animal. Thus, a major limitation of the existing methods is the inability to isolate sufficient cell numbers without using very large numbers of donor mice and substantial volumes of reagents. Our method substantially overcomes these problems by using mice implanted with a Flt-3L secreting tumor to greatly expand the DCs, allowing purification of much greater numbers of the three well-recognized DC subset using a series of immunomagnetic purification steps. The protocol described here therefore provides a major improvement over existing methods in terms of yields by at least several fold.
A few caveats exist for the isolation protocol described here. The major limitation is that the procedure used for isolating these cells may inadvertently alter their phenotype. DCs are exquisitely sensitive to endotoxin contamination and mechanical stimuli that may lead to changes in their phenotype and function. Another limitation of this technique is the reliance on Flt-3L secretion to expand DC subsets. This results in the lack of expansion in tissue resident DC subsets that do not express the Flt-3 receptor.
The protocol described here relies on cell enrichment using commercially available reagents. The kit for isolation of cell purification comes with a protocol that contains the recommended volumes of reagents to be used. The cell composition of tumor bearing mice is dramatically different from that of naïve mice (Figure 2C), and therefore we have modified the amounts for reagents used in our technique in order to achieve optimal cell purity. The DCs are adherent cells and may attach to the column matrix non-specifically. Equilibration of the column bed with buffer containing EDTA minimizes cell adherence, and the use of serial passage over two consecutive columns improves the cell purity further.
DCs are highly adherent cells and in vivo are generally closely associated with the extracellular tissue matrix. In order to increase the yield, it is very important to digest the splenic tissue with collagenase and DNase for sufficient time to extract a maximal number of these cells. It is also important to be sure that the splenic tissue fragments are completely submerged in the collagenase/DNase I solution while undergoing digestion. Another critical feature of DC isolation is to maintain strict sterility and endotoxin-free conditions during the isolation procedure, since these cells are highly endotoxin responsive. Use freshly prepared reagents from endotoxin-free stocks, and filter sterilize all reagents before use. All cell isolation buffers and media should be supplemented with FCS that is certified to be free of detectable endotoxin. These cells are also highly sensitive to mechanical stimuli and repeated mechanical stimulation may alter the maturation state of these cells. Although some mechanical force is needed to force the tissue through a cell strainer after collagenase digestion, this should be done as gently as possible. In addition, vigorous pipetting of the cells should be avoided as much as possible, and wide bore pipette tips should be used to minimize shear forces.
The adoptive transfer model is useful for investigation of antigen presentation and T cell priming by specific subsets of DCs. However, it has been demonstrated in other studies that transfer of antigen may occur from antigen pulsed DCs to endogenous DCs. Thus, the adaptive immune response instigated by antigen pulsed cells may reflect presentation by endogenous DCs presenting the "captured" antigen and not that of transferred APC. This kind of transfer is exponentially increased if the isolated DC preparations contain dead or dying cells. It is therefore important to carry out the DC isolation and transfer protocol as rapidly as possible, using sterile conditions with pre-cooled buffers to enhance cell survival. We chose a relatively short antigen pulsing period of 4 hr for this reason, with incubation in low binding cell culture plates to minimize the cell attachment during this step. In addition, extended culture condition may alter the maturation status of cultured DCs, and may influence the outcome of adoptive transfer experiments 22.
In summary, dendritic cell biology is a very active area of research, and development of a reliable method to generate DCs that faithfully reflect their in vivo phenotypic and functional divergence provides an immense opportunity to study these cells. This becomes especially important for studies involving chemicals that may modulate DC functions but cannot be administered systemically into animal models. Treating the isolated DC subsets of interest ex vivo with these chemicals is a practical and feasible approach that can help elucidate the mechanistic details of antigen presentation pathways.
The authors have nothing to disclose.
This work was supported by NIH/NIAID grant AI45889 to S.A.P. Flow cytometry studies were carried out using FACS core facilities supported by the Einstein Cancer Center (NIH/NCI CA013330) and Center for AIDS Research (NIH/NIAID AI51519).
0.05% Trypsin-EDTA | Life Technologies, Gibco | 25300-054 | |
Isoflurane | Sigma-Aldrich | CDS019936-250MG | |
Collagenase D | Roche Diagnostics | 11088858001 | |
DNase I (dry powder) | QIAGEN | 79254 | |
200 proof ethanol | Thermo Fisher Scientific | 9-6705-004-220 | Used to prepare 70% ethanol |
RBC lysis buffer | Sigma-Aldrich | R7757 | |
RPMI-1640 medium with L-glutamine | Life Technologies, Gibco | 11875-119 | |
DMEM medium with L-glutamine | Life Technologies, Gibco | 11995-073 | |
200 mM L-glutamine | Life Technologies | 25030081 | |
MEM non-essential amino acids | Life Technologies, Gibco | 11140-050 | |
MEM essential amino acids | Life Technologies | 11130-051 | |
β-mercaptoethanol | Life Technologies, Invitrogen | 21985-023 | |
Sodium pyruvate | Life Technologies | 11360-070 | |
HEPES | Life Technologies, Invitrogen | 15630 | |
Phosphate buffered saline (PBS Ca2+ and Mg2+ free pH 7.2) | Life Technologies, Invitrogen | 20012-050 | |
Dulbecco’s PBS (DPBS with Ca2+ and Mg2+) | Life Technologies, Gibco | 14040-182 | |
0.5 M Ethylenediaminetetraacetate (EDTA) solution | Life Technologies | 15575-020 | |
Bovine serum albumin (BSA) | Sigma-Aldrich | A2153 | |
Fetal calf serum | Atlanta Biologicals | S11050 | |
Penicillin/streptomycin | Life Technologies, Invitrogen | 15140-163 | |
Trypan blue (dry powder) | Sigma-Aldrich | T6146-5G | Prepare 0.08% with PBS |
Plasmacytoid Dendritic Cell Isolation Kit II, mouse | Miltenyi Biotech | 130-092-786 | |
Magnetic beads conjugated with anti-mouse CD11c | Miltenyi Biotech | 130-152-001 | |
CD8αPos mouse DC isolation kit | Miltenyi Biotech | 130-091-169 | |
Fc-gamma receptor blocking antibody (Clone 2.4G2) | BD Biosciences | 553141 | |
Anti-mouse CD11c-FITC | BD Biosciences | 553801 | |
Anti-mouse CD8α-PerCP | BD Biosciences | 553036 | |
Anti-mouse B220-PE | BD Biosciences | 553090 | |
1 ml syringes | BD | 26048 | |
23 G1 needle | BD | 305145 | |
100 mm Petri dishes | Thermo Fisher Scientific | 875712 | |
Surgical instruments | Kent Scientific | INSMOUSEKIT | |
Cell strainer (70 µm | BD | 352350 | |
Large Petri plates | Thermo Fisher Scientific | FB0875712 | |
Vacuum filtration system (500 ml(0.22 um | Corning | 431097 | |
LS columns | Miltenyi | 130-042-401 | |
Magnetic stand MACS separator | Miltenyi Biotec | 130-042-302 | |
Wide-bore 200 μl pipette tips | PerkinElmer | 111623 | |
Corning ultra-low attachment 96-well plates | Corning | CLS3474-24EA | |
6-8 week old female C57BL/6 mice | Jackson Research Laboratories, Jax | ||
Murine B16.Flt3L melanoma cell line | described by Mach et al. ,2000 | ||
Serum free DMEM and RPMI media | |||
500 ml DMEM with L-glutamine, or 500 ml RPMI-1640 with L-glutamine 5 ml MEM nonessential amino acids (100x, 10 mM) 5 ml HEPES Buffer (1 M) 5 ml L-glutamine (200 mM) 0.5 ml 2-mercaptoethanol (5.5 x 10-2 M) |
Mix all the ingredients in a biosafety cabinet with either DMEM or RPMI media depending on your need. Filter sterilize the media by passing through a 0.22 µm vacuum filtration system. | ||
Complete RPMI and DMEM media | Add 50 ml of heat inactivated fetal calf serum to serum free RPMI or DMEM media to obtain complete media. | ||
MACS buffer: | Add 2 ml of 0.5 M EDTA and 10 ml of heat inactivated fetal calf serum to 500 ml of Phosphate-buffered saline (PBS, Ca2+ and Mg2+ Free, pH 7.2) and 2 ug/ml 2.4G2 (Fc-Block antibody). | ||
FACS staining buffer | Dissolve sodium azide to 0.05% (0.25 g per 500 ml) in MACS buffer to obtain FACS staining buffer | ||
10x Collagenase D and DNase 1 stock solution Dissolve 1 g of collagenase D and 0.2 ml of DNase 1 stock (1 mg/ml, 100x) in 20 ml of PBS containing Ca++ and Mg++ to obtain a solution of approximately 1,000- |
2,000 Units of collagenase activity per ml This 10x stock solution can be prepared ahead of time and stored at -20 °C for several weeks. Dilute 1 ml of this with 9 ml of serum free RPMI immediately before use |