Killer cell lectin-like receptor G1
- Treg cell:
Regulatory T cell
CD4+CD25+ regulatory T cells (Treg cells) control immune responsiveness to a large variety of antigens. The isolation and therapeutic manipulation of Treg cells requires the use of reliable surface receptors that are selectively up-regulated in Treg cells. On the basis of global gene expression studies, we identified neuropilin-1 (Nrp1) as a specific surface marker for CD4+CD25+ Treg cells. Nrp1, a receptor involved in axon guidance, angiogenesis, and the activation of T cells, is constitutively expressed on the surface of CD4+CD25+ Treg cells independently of their activation status. In contrast, Nrp1 expression is down-regulated in naive CD4+CD25– T cells after TCR stimulation. Furthermore, CD4+Nrp1high T cells express high levels of Foxp3 and suppress CD4+CD25– T cells. Thus, Nrp1 constitutes a useful surface marker to distinguish Treg cells from both naive and recently activated CD4+CD25+ non-regulatory T cells.
Strong evidence is emerging that immune responsiveness to self- and allo-antigens can be controlled by naturally occurring CD4+CD25+ regulatory T cells (Treg cells) 1, 2. The key role of Treg cells in controlling self-reactive T cells has been established for CD4+CD25+ T cells in transfer models, where they were shown to suppress multi-organ-specific autoimmune disease induced either by depletion of Treg cells, reconstitution of T cell deficient mice with CD4+CD25– Tcells or thymectomy at day 3 after birth 3. Treg cells are anergic under certain culture conditions and exhibit in vitro and in vivo suppressor activity. The mechanisms of suppression include currently unknown cell surface molecules or cytokines operating most likely in a cell contact dependent manner. Although Treg cells are able to produce IL-10 and TGF-β, both cytokines are not necessarily required for the regulatory function of the cells in some systems 4–7. Recently, the transcription factor Foxp3 has been identified as being essential for the development and function of regulatory CD4+CD25+ Treg cells 8–10. Although it remains unclear which genes are induced by Foxp3 and mediate the regulatory activity, ectopic expression of Foxp3 in non-regulatory CD4+CD25– T cells generates Treg cells that are similar to naturally occurring CD4+CD25+ Treg cells; they are anergic and able to suppress T cell proliferation, and function both in vitro and in vivo8–10.
Characterization and manipulation of Treg cell function in vivo would be greatly facilitated by the identification of a specific cell surface receptor that should optimally be highly expressed exclusively in Treg cells independently of their activation status. Treg cells exhibit constitutive expression of cell surface proteins such as CD25, CTLA-4 and GITR 5, 11–14 that are commonly used to identify Treg cells. However, these receptors are highly induced upon TCR ligation of non-suppressive T cells thereby making it difficult to distinguish Treg from conventional CD4+ T cells activated in response to antigen; therefore these markers can not be considered as specific for Treg cells.
On the basis of global gene expression profiling, we identified neuropilin-1 (Nrp1), a semaphorin III receptor originally described to be involved in axon guidance and more recently as essential component of the immunological synapse in humans 15, 16, as a potent surface marker for Treg cells. In contrast to other genes up-regulated in Treg cells but also expressed following activation of naive T cells, Nrp1 expression is significantly down-regulated after T cell activation. Moreover, Nrp1 expression on Treg cells correlates with the expression of Foxp3 and in vitro suppressor function, thus representing a suitable new surface marker for the identification of Treg cells and their discrimination from recently activated T cells.
2 Results and discussion
2.1 Gene expression profiling reveals Nrp1 as marker for Treg cells
To identify a reliable marker that distinguishes Treg cells from non-regulatory naive T cells and activated CD4+ T cells, and to characterize these cells in more detail at the molecular level, we compared the gene expression profiles of naturally occurring polyclonal CD4+CD25+ Treg cells with their naive CD4+CD25– counterparts. In addition, monoclonal CD4+CD25+ Treg cells and naive CD4+CD25+ T cells of known antigen specificity as well as CD4+ T cells recently activated with their specific antigen were included in our analysis. RNA was isolated from the different cell pools and analyzed by microarrays (Affymetrix MOE430A).
Among the 22400 genes and expressed sequence tags (EST) analyzed, we identified a set of regulated genes that belong to a wide variety of gene classes encoding molecules involved in cell cycle, apoptosis and survival, signaling, general metabolism, transcription and translation factors as well as surface receptors (data not shown). Co-regulated genes were combined in clusters. By this approach we intended to identify those genes that are highly expressed on monoclonal and polyclonal Treg cells without being up-regulated upon T cell activation. As the transcription factor Foxp3 is well established as being a key regulator necessary for the function of Treg cells, we further focused on the analysis of genes that follow the same expression profile as Foxp3. Fig. 1 shows the cluster containing Foxp3 and in addition 86 additional genes that are over-expressed in Treg cells but exhibit low expression levels in naive T cells and are down-regulated in activated T cells (Fig. 1, and Supporting Information at http://www.wiley-vch.de/contents/ jc_2040/2004/24799_s.html). Among these genes, cell surface receptors are of special interest for the isolation, characterization and manipulation of Treg cells.
Two surface molecules identified by this approach were considered as candidate markers for Treg cells — Nrp1 and the killer cell lectin-like receptor G1 (KLRG1) (Fig. 1B). KLRG1 is an inhibitory C-type lectin previously described to be expressed on NK cells and activated CD8+ T cells 17. This molecule was highly expressed in CD4+CD25+ Treg cells, but was down-regulated 16 h as well as 4 days after T cell activation (Fig. 1B and data not shown). However, FACS analysis revealed that only a minor subpopulation (about 3%) of Treg cells expressed KLRG1 on the cell surface. Since depletion of KLRG1+ T cells from CD4+CD25+ Treg cells did not change their in vitro suppressive capacity (data not shown), KLRG1 cannot be considered as a marker for Treg cells and therefore was not considered further.
Nrp1 is a receptor for class III semaphorin subfamily and the heparin-binding forms of vascular endothelial growth factor (VEGF), known to be involved in axon guidance and angiogenesis 15, 16, and was more recently described as mediating homophilic interactions between DC and T cells in humans 18. Being co-regulated with Foxp3, Nrp1 represents a candidate marker for Treg cells, as it exhibits a high expression level in both polyclonal and monoclonal Treg cells but not in antigen-stimulated non-regulatory T cells (Fig. 1B). Moreover, Nrp1 expression was even significantly down-regulated following antigen-specific activation of naive T cells.
Extending our analysis to activated polyclonal CD4+CD25– T cells revealed that Nrp1 expression is down-regulated 16 h or 4 d after anti-CD3 + anti-CD28 stimulation (Fig. 2A), and even lost at day 11 post-stimulation (Fig. 2B). Since the expression of a suitable marker for Treg cells should be independent of activation, we examined the kinetics of Nrp1 expression in CD4+CD25+ Treg cells after in vitro stimulation. As shown in Fig. 2C, the mRNA level of Nrp1 in CD4+CD25+ Treg cells remains constantly high after in vitro stimulation with anti-CD3 + anti-CD28. The same results were obtained when stimulating the cells in the absence of CD28 or with the addition of high amounts of IL-2 to the cultures (data not shown).
Finally, Nrp1 expression on CD4+CD25+ Treg cells was confirmed by antibody surface staining and FACS analysis. To prove the specificity of the polyclonal anti-Nrp1 antibody, control staining on the hemagglutinin (HA)-specific T cell hybridoma 16.2.11, which was negative for Nrp1 in RT-PCR, was performed. Infection of 16.2.11 cells with an Nrp1-expressing retrovirus, but not with a control virus expressing GFP alone, leads to specific surface staining for Nrp1 as documented in Fig. 3A. Antibody staining of CD4+ T cells revealed that only CD4+CD25+ Treg cells express elevated levels of Nrp1 at their surface, whereas Nrp1 surface expression of CD4+CD25– T cells did not exceed background staining (Fig. 3B).
2.2 Nrp1 expression correlates with in vitro suppressor function
It has been reported that blocking of Nrp1 inhibits primary activation of human T cells 18. Therefore, we analyzed the influence of antibodies to Nrp1 on the stimulation of murine naive CD4+ T cells. Thus, sorted CD4+CD25– T cells were stimulated in vitro in the presence or absence of the anti-Nrp1 antibody H-286. In contrast to results obtained by Tordjman et al. in the human system 18, we found that anti-Nrp1 treatment strongly interferes with primary activation of murine CD4+ T cells (Fig. 4A). However, preincubation of CD4+CD25+ Treg cells with anti-Nrp1 did not affect suppressor function of CD4+CD25+ Treg cells (data not shown).
To address the question of whether Nrp1 expression on CD4+ T cells correlates with their regulatory phenotype, CD4+Nrp1high and CD4+Nrp1low T cells were isolated by cell sorting and tested for their capacity to inhibit proliferation of naive CD4+CD25– T cells. CD4+CD25+ Treg cells were included as an internal control. As shown in Fig. 4B, only CD4+Nrp1high T cells suppress proliferation of naive CD4+CD25– T cells, whereas CD4+Nrp1low cells lack the capacity to inhibit the growth of activated CD4+CD25– T cells. Thus, we could demonstrate that Nrp1 expression on CD4+ T cells correlates with in vitro suppressive function.
2.3 Linked expression of Nrp1 and Foxp3
Treg cells have been recently shown to specifically express the forkhead transcription factor Foxp3, which represents a critical regulator for the development and function of these cells 8–10. Since Nrp1 and Foxp3 follow the same expression pattern in regulatory, naive and activated CD4+ T cells (Fig. 1) and CD4+Nrp1high T cells have regulatory properties, we addressed the question of whether the level of Nrp1 expression in Treg cells correlates with elevated levels of Foxp3 mRNA. Analysing sorted Nrp1high and Nrp1low T cells for the expression of Foxp3 by RT-PCR revealed that Foxp3 message correlates well with Nrp1 expression in the CD4+Nrp1high T cells (Fig. 5A). Also, the expression of other Treg cell markers such as PD-1, αEβ7 (CD103) and CTLA-4 was found to be increased in CD4+Nrp1high cells (data not shown).
As CD4+Nrp1high Treg cells exhibit high levels of Foxp3 mRNA, we next investigated whether ectopic expression of Foxp3 in non-regulatory CD4+CD25– T cells would induce Nrp1 expression in these cells. To this end, naive CD4+CD25– T cells were infected with a Foxp3-encoding retrovirus and analyzed for the expression of Nrp1. It has been described before that Foxp3 expression confers regulatory function to the infected T cells 10. This was confirmed in our hands (data not shown). Moreover, RT-PCR analysis (Fig. 5B) indeed revealed elevated Nrp1 levels in Foxp3-transduced T cells. Thus, Foxp3 expression is accompanied with suppressor function as well as up-regulation of Nrp1 expression in CD4+ T cells, further corroborating Nrp1 as selective marker for Treg cells.
Taken together, the aim of this study was the identification of a surface marker on Treg cells, because the currently known cell surface receptors for these cells such as the B7-family members CTLA-4, PD1, and ICOS, as well as the members of the TNF receptor superfamily GITR (TNFRSF18) or the αEβ7 integrin 11–14, 19, 20, are also up-regulated in activated non-regulatory CD4+ T cells. On the basis of gene expression profiling, we identified genes that are co-regulated with Foxp3 — known to be a key regulator for Treg cells. Our major finding was that Nrp1 is over-expressed in the subpopulation of CD4+CD25+ Treg cells independently of their activation status. Moreover, in conventional T cells Nrp1 is expressed at a significantly lower level and is further down-regulated upon T cell activation. Our data further suggest that the expression of the Nrp1 gene is regulated by the transcription factor Foxp3.
Nrp1 was initially described as semaphorin III and VEGF165 receptor, being essential for axonal guidance and vascularization, respectively 15. Recently, Nrp1 has been also identified on hematopoietic cells and shown to be involved in the formation of the immunological synapse, as it co-localizes with CD3 at the naive T-cell–DC interface in human T cells 16. Also, some Nrp1 antibodies inhibit DC-induced proliferation of human resting T cells 18. The latter result was elaborated in murine polyclonal CD4+CD25– T cells stimulated with anti-CD3 antibodies in spite of the fact that these cells express only tiny amounts of Nrp1 (Fig. 4). One might speculate that because of the much higher expression in polyclonal CD4+CD25+ Treg cells (Fig. 2 and 3), Nrp1 is involved in the functional control of these cells. However, further studies are required to support or refute such a hypothesis.
In the nervous system, Nrp1 is able to either repel or attract different axons and thus Nrp1 might have distinct roles in the immune system as well. Interestingly, agrin, a glycoprotein important in the formation of neuronal synapses, has also been found on T lymphocytes, consistent with some molecular similarities of the neurological and immunological synapses 16. Although the detailed immunological functions of Nrp1 in immune regulation remain currently unclear, we conclude that Nrp1 represents an additional surface marker to distinguish Treg cells from naive and activated T cells which in concert with other Treg cell markers will greatly facilitate the characterization and manipulation of Treg cell function in vivo.
3 Materials and methods
3.1 Mice and cell lines
BALB/c mice were obtained from Harlan (Borchen, Germany). TCR-HA transgenic mice, expressing a TCRα β specific for peptide 110–120 from influenza HA presented by I-Ed, have been described previously 21. Mice aged 6–16 weeks were used for experiments and were maintained under specific pathogen free conditions. The HA-specific and MHC class II-restricted T cell hybridoma 16.2.11 was kindly provided by W. Gerhard (Wistar Institute, Philadelphia, PA).
3.2 Antibodies and flow cytometry
The monoclonal antibody 6.5 (anti-TCR-HA) was purified from a hybridoma supernatant and used in FITC-labeled form. Anti-CD4 (GK1.5) and anti-CD25 (PC61), from BD Biosciences (San Jose, CA), were used as biotin, FITC or PE conjugates. Anti-Nrp1 (H-286, rabbit polyclonal IgG) and the corresponding isotype-control antibody (normal rabbit IgG, control Ig) were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). PE–streptavidin, cychrome–streptavidin and FITC-labeled goat anti-rabbit-Ig (mouse, rat and human absorbed), from BD Biosciences, were used as secondary reagents. Flow cytometric analyses were done on a FACSCalibur with the CellQuest software (BD Biosciences). Cells were sorted with a MoFlow cell sorter (Cytomation, Fort Collins, CO).
3.3 Cell isolation
For cell sorting experiments, red blood cell-depleted splenocytes from TCR-HA and BALB/c mice were labeled with anti-CD4, anti-TCR-HA (6.5) and anti-CD25 in variable combinations. For separation of cells expressing high levels of Nrp1, splenocytes were stained with anti-Nrp1, and goat anti-rabbit-Ig as the secondary antibody. Labeled cells were separated with a MoFlow and purity was >97%. Isolated cells were either used for RNA purification or for T cell proliferation experiments.
3.4 Proliferation assay
Sorted naive CD4+CD25– T cells, CD4+CD25+ Treg cells, CD4+Nrp1high cells, or CD4+Nrp1low cells (5×104), alone or in addition to CD4+CD25– T cells at a ratio of 1:1, were stimulated with 1 μg/ml soluble anti-CD3 (2c11) without the addition of APC. Alternatively, 10 μg/ml anti-Nrp1 or the corresponding isotype-control antibody was added together with the anti-CD3 antibody. Background proliferation was quantified by [3H]thymidine incorporation in cells incubated without anti-CD3. Proliferation assays were done for 72 h in 96-well flat-bottom plates in a final volume of 200 μl IMDM medium that contained 10% fetal calf serum. 1 μCi/well [3H]thymidine was added for the last 6 h of the experiment and thymidine incorporation was measured by scintillation counting.
3.5 In vitro T cell activation
Splenic CD4+CD25– T cells or CD4+CD25+ T cells from BALB/c mice were sorted and cultured in 12-well plates in the presence of 0.75 μg/ml anti-CD3 (platebound) and 1 μg/ml anti-CD28 (soluble). At different times after stimulation, cells were recovered for RNA preparation and FACS analysis. For antigen-specific stimulation, red blood cell-depleted spleen cells from TCR-HA mice were stimulated with 10 μg/ml of the MHC class II-restricted HA peptide HA110–120. After 16 h of culture, cells were harvested and labeled with anti-CD4, 6.5 (anti-TCR-HA) and anti-CD25. CD4+6.5+CD25+ T cells, representing in vitro-activated T cells, were sorted and used for RNA preparation.
3.6 Retroviral infection and analysis of transduced CD4+CD25+ T cells
Foxp3 cDNA (amino acids 1–429) was amplified by RT-PCR from sorted CD4+CD25+ cells using specific primers (5′-CTG-ACA-AGA-ACC-CAA-TGC-CCA-ACC-3′ and 5′-CTC-CCC-CGC-CCA-CCT-TTT-CT-3′). After cloning into pCR2.1 TOPO (Invitrogen, Karlsruhe, Germany), the Foxp3 cDNA was inserted into an MCSV-based retroviral vector encoding eGFP under control of an IRES site. Nrp1 cDNA (amino acids 1–924) was isolated from pB II SK-m.nrp1 EP (kindly provided by H. Fujisawa, Tokio, Japan) and cloned into the MCSV-based retroviral vector described above. The GPE86+ ecotropic packaging cell line was transfected as described previously 10. Retrovirus-containing culture supernatant was collected after 24 h and passed through 0.45 μm filters. Sorted CD4+CD25– splenocytes were activated as described above (see Sect. 3.5) and, after 48 h of culture, cells were incubated with virus-containing supernatant supplementedwith 20 mM Hepes, 8 μg/ml polybrene and 100 U/ml recombinant IL-2 (not in the case of 16.2.11) following centrifugation for 60 min at 5000×g. Twenty-four hours post-infection, 50% of the culturesupernatant was removed and fresh medium was added. GFP+ cells were sorted after 7–10 days.
3.7 Real-time RT-PCR
Total RNA was prepared from sorted T cells using the TriFast FL reagent from PeqLab (PeqLab Biotechnology, Erlangen, Germany). The cDNA was synthesized using oligo-dT primers and MLV RT polymerase (Invitrogen), following the manufacturer's recommendations. Alternatively, RNA was isolated with the RNeasy kit (Qiagen, Hilden, Germany) following cDNA synthesis using the Superscript II Reverse Transcriptase (Invitrogen). Quantitative real-time RT-PCR was performed in an ABI PRISM cycler (Applied Biosystems) using a SYBR Green PCR kit from Applied Biosystems or Stratagene and specific primers optimized to amplify 90–250 bp fragments from the different genes analyzed. A threshold was set in the linear part of the amplification curve, and the number of cycles needed to reach itwas calculated for every gene. Relative mRNA levels were determined by using included standard curves for each individual gene and further normalization to RPS9. Melting curves established the purity of the amplified band. Primer sequences are: Nrp1 (5′-GCC-TGC-TTT-CTT-CTC-TTG-GTT-TCA-3′, 5′-GCT-CTG-GGC-ACT-GGG-CTA-CA-3′); Foxp3 (5′-CTG-GCG-AAG-GGC-TCG-GTA-GTC-CT-3′,5′-CTC-CCA-GAG-CCC-ATG-GCA-GAA-GT-3′); and RPS9 (5′-CTG-GAC-GAG-GGC-AAG-ATG-AAG-C-3′, 5′-TGA-CGT-TGG-CGG-ATG-AGC-ACA-3′).
3.8 DNA microarray hybridization and analysis
The quality and integrity of the total RNA isolated from 5×104 cells was controlled by running all samples on an AgilentTechnologies 2100 Bioanalyzer (Agilent Technologies, Waldbronn, Germany). For RNA-amplification, the first round was done according to Affymetrix without biotinylated nucleotides using the Promega P1300 RiboMax Kit (Promega, Mannheim, Germany) for T7 amplification. For the second round of amplification the precipitated and cleaned aRNA was converted to cDNA using random hexamers (Pharmacia, Freiburg, Germany). Second-strand synthesis and probe amplification was done like in the first round, except for an incubation with RNAse H before first-strand synthesis to digest the aRNA and for the use of the T7T23V oligo for initiation of the synthesis of the second strand. The concentration of biotin-labeled cRNA was determined by UV absorbance. In all cases, 12.5 μg of each biotinylated cRNA preparation was fragmented and placed in a hybridization cocktail containing four biotinylated hybridization controls (BioB, BioC, BioD, and Cre) as recommended by the manufacturer. Samples were hybridized to an identical lot of Affymetrix MOE430A for 16 h. After hybridization the GeneChips were washed, stained with SA–PE and read using an Affymetrix GeneChip fluidic station and scanner.
3.9 Data analysis
Analysis of microarray data was performed using the Affymetrix Microarray Suite 5.0, Affymetrix MicroDB 3.0 and Affymetrix Data Mining Tool 3.0. For normalization, all array experiments were scaled to a target intensity of 150, otherwise using the default values of the Microarray suite. Filtering of the results was done as follows: genes are considered as strongly regulated when their fold change is greater than or equal to 2, or less than or equal to –2, the statistical parameter for a significant change is less than 0.01 [change p-value for changes called increased (I)] or greater than 0.99 [change p-value for changes called decreased (D)]. Additionally, the signal difference of a certain gene should be greater than 100. Genes are considered as weakly regulated when their fold change is greater than or equal to 1.5, or less than or equal to –1.5, the statistical parameter for a significant change is less than 0.001 or greater than 0.999 and the signal difference of a certain gene should be greater than 40.
We thank Patricia Gatzlaff, Tanja Toepfer and Silvia Prettin for excellent technical assistance as well as Lothar Gröbe for cell sorting. We thank H. Fujisawa for kindly providing us the plasmid pB II SK-m.nrp1 EP encoding the Nrp1 cDNA. This work was supported by the Deutsche Forschungsgemeinschaft, the Deutsche Krebshilfe and the VolkswagenStiftung.