The cellular prion protein is preferentially expressed by CD4⁺ CD25⁺ Foxp3⁺ regulatory T cells

Abstract

Post-translational modification of the cellular prion protein (PrP^C) is intimately associated with the pathogenesis of prion disease, yet the normal function of the protein remains unclear. PrP^C is expressed in lymphoid cells and is known to be a T-cell activation antigen. Further, transcription profiling studies of regulatory T cells have shown preferential overexpression of PrP^C, suggesting a possible role in regulatory function. We report that both the expression of PrP message and cell surface PrP^C levels are increased in murine CD4⁺ CD25⁺ regulatory T cells compared with CD4⁺ CD25⁻ cells. However, PrP^0/0 mice do not show altered regulatory T-cell numbers or forkhead box P3 (Foxp3) expression levels, or impaired regulatory T-cell function in vitro. Nevertheless, the preferential expression of surface PrP^C by regulatory T cells raises the possibility that therapeutic ligation of PrP^C might alter immune regulation.

Introduction

The cellular isoform of prion protein (PrP^C) is central to the pathogenesis of prion disease. However, its physiological function has not been fully elucidated. The mature human PrP^C species consists of 231 amino acids with an unstructured N-terminus and a structured C-terminus folded into three α-helices and a two-strand β-sheet. After addition of a glycosyl-phosphatidylinositol (GPI) anchor and up to two carbohydrate moieties, PrP^C is predominantly trafficked to the cell surface membrane where it enters lipid raft domains in both neurons and lymphocytes.¹ More than 10 candidate PrP^C ligands have been proposed (reviewed in ²), from which it can be inferred that either the principal receptor for PrP^C is undefined, or PrP^C is shared by a variety of cellular pathways with multiple binding partners. An important consideration is whether PrP^C has a single function in all tissues in which it is expressed, or has multiple tissue-specific roles. While there is some evidence that the role of PrP^C in neurons relates to neurotransmission,^3,4 the protein must have additional functions as PrP^C is expressed in many non-excitable cells, including glia and various lymphoid and non-lymphoid organs.^5–8

Within the immune system, PrP^C has been detected on T and B lymphocytes, natural killer (NK) cells, monocytes, dendritic cells and follicular dendritic cells.^9–14 Surface expression of PrP^C on human T lymphocytes is high and is further up-regulated upon T-cell activation and memory differentiation, strongly suggestive of a functional role in T-cell biology.^10,11,15 Indeed, many attempts have been made to define the role of PrP^C in the immune system, particularly through comparison of wild-type and knockout mice. From these studies a number of rather subtle effects have been documented, without any evidence for gross immune dysregulation.^15–21

In recent years there has been considerable interest in characterizing the key molecules and cellular interactions underlying regulatory T-cell (Treg) function.^22,23 Gene expression microarray analysis has in some, although not all, studies suggested that PrP may be up-regulated in naturally occurring Tregs.^24–27 However, PrP expression in Tregs has not been directly characterized. We therefore analysed PrP expression in CD4⁺ CD25⁺ forkhead box P3⁺ (Foxp3⁺) Tregs. In parallel we also examined Treg development, numbers and function in PrP-deficient mice.

Materials and methods

Mice

Adult wild-type FVB/N and C57/BL6 mice were used where described. Prnp^0/0 mice were originally obtained on a C57BL/6 × Sv129 background as described previously.¹⁶ The line used here was derived from the original Zurich I mouse but crossed onto the FVB/N background for 10 generations. All experiments on PrP^0/0 mice used age-matched wild-type FVB/N mice as controls. All mice were housed in accordance with institutional and UK Home Office requirements at the CBS, Hammersmith Hospital, Imperial College, London.

RNA and cDNA preparation

Mice were killed and spleens, thymi and lymph nodes (including the axillary and inguinal, termed ‘peripheral’, and mesenteric nodes) were aseptically dissected. Single-cell suspensions were made and RNA prepared by the acid phenol method using Trizol (Invitrogen, Paisley, UK). The RNA concentration and nucleic acid/protein ratio were determined using a Nanodrop spectrophotometer (Nanodrop, Wilmington, DE). RNA was diluted to 60–200 ng/ml. cDNA was prepared using SuperScipt III RNase HRT (Invitrogen) according to the manufacturer’s instructions and stored at −20° until further use.

Quantitative real-time polymerase chain reaction

Quantitative analysis of prion protein (Prnp), Foxp3 and 18S transcript levels was performed using TaqMan ‘assays on demand’, which include polymerase chain reaction (PCR) primers with TaqMan MGB probes [carboxyfluorescein (FAM) dye-labelled] (Applied Biosystems, Foster City, CA). Values for each transcript were obtained by running triplicate monoplex PCR reactions for each sample, containing cDNA equivalent to 200 ng RNA. Reactions were performed in a 20-μl volume on a Mx3000P real-time PCR thermocycler (Stratagene, La Jolla, CA), programmed at 50° for 2 min, followed by 10 min at 95° and then 50 cycles of 15 seconds at 95° and 1 min at 60°. Standard curves were constructed by running PCR reactions on a series of fivefold serial dilutions of cDNA. The PCR amplification signal was expressed as a ΔR value, reflecting the intensity of reporter dye emission for each sample at the end of each cycle minus the baseline signal during the initial PCR cycles. Efficiencies of the PCR reactions for Prnp, Foxp3 and 18S were compared by plotting Δ threshold cycle number (ΔC_T) (where ΔC_T is the C_T value for the Prnp or Foxp3 assay at a specific dilution minus the C_T value for the 18S assay at the same dilution) against log(dilution). Efficiencies of the Prnp and 18S standard curves were 99·7% and 98·0% respectively, while the Foxp3 PCR had an efficiency of 76·6%. Mean C_T values for Prnp and 18S were obtained for each experimental sample. The abundance of Prnp transcript in each sample was then calculated relative to 18S quantity using the ΔΔC_T method. For each sample, Prnp transcription was first normalized to its 18S value using the equation:

One sample (S₀) was then used as a ‘baseline’ value. Prnp transcription in all other experimental samples (S_X) was then calculated relative to this value using the equation:

Because the efficiencies of the Foxp3 and 18S real-time reactions were divergent, the ΔΔC_T method could not be used to calculate relative Foxp3 abundance. Instead, relative transcript quantities for each sample were calculated from C_T values using the standard curve equation for each primer. Foxp3 expression in each sample was then normalized to 18S expression, allowing direct comparison between samples.

Antibodies and flow cytometry

Anti-PrP mouse immunoglobulin G1 (IgG1) monoclonal antibody (mAb) ICSM18 (D-Gen Ltd, London, UK) was fluorescein isothiocyanate (FITC)-conjugated using the FluoroTag-FITC kit (Sigma, St Louis, MO). FITC-conjugated mouse IgG1 (eBioscience, San Diego, CA) was used as a control mAb for ICSM18-FITC. Other fluorochrome-conjugated antibodies (and isotype controls) were purchased from eBioscience as follows: FITC-conjugated: anti-mouse CD4; phycoerythrin (PE)-conjugated: anti-mouse CD4, anti-mouse CD25 and anti-mouse Foxp3; PECy5-conjugated: anti-mouse CD4 and anti-mouse CD8; and allophycocyanin (APC)-conjugated: anti-mouse CD25 and anti-mouse Foxp3. Antibodies against cell surface antigens were added at saturating concentrations to cells suspended in RPMI with 1% volume/volume (v/v) fetal calf serum (FCS). For Foxp3 detection, cells were then treated with Fix/Perm solution (eBioscience) and stored overnight at 4°. After washing, anti-Foxp3 mAb or isotype control was added to each well and cells further incubated at 4° for 30 min. Cells were washed and analysed immediately by flow cytometry on single or double-laser FacsCalibur machines (BD, Oxford, UK) and data were analysed using CellQuest software (BD). For analysis of PrP expression, the geometric mean values for PrP and an isotype control (measuring auto-fluorescence and non-specific binding) were obtained. PrP expression was defined as the difference between these two values (Δ geometric mean).

Sorting of lymphocytes into CD25⁺ and CD25⁻ CD4⁺ fractions

Sorting was provided as a service by the MRC Clinical Sciences Centre flow cytometry core (Imperial College, London, UK). Briefly, after staining with anti-CD4-PECy5 and anti-CD25-PE mAbs, cells were re-suspended in fluorescence-activated cell sorting (FACS) buffer [Ca²⁺/Mg²⁺-free phosphate-buffered saline (PBS) supplemented with 1 mm ethylenediaminetetraacetic acid (EDTA), 25 mm HEPES and 1% v/v FCS] at 1 × 10⁷ cells/ml. Using flow cytometry, lymphocytes were detected by forward (FSC) and side scatter (SSC) characteristics and CD4⁺ cells identified by fluorescence in the FL3 channel. Gates were drawn around CD25⁺ and CD25⁻ fractions, distinguished by high or low fluorescence in the FL2 channel, respectively. These were then collected simultaneously into separate tubes for subsequent RNA extraction.

Functional studies of PrP^+/+ and PrP^0/0 FVB/N murine Tregs

Single-cell splenocyte suspensions were made as above from 6- to 12-week-old PrP^+/+ and PrP^0/0 FVB/N mice. CD4⁺ cells were negatively isolated using Dynabeads (Invitrogen) according to the manufacturer’s instructions. These were further fractionated into CD25⁺ and CD25⁻ populations by incubation with biotinylated anti-CD25 followed by Streptavidin MicroBeads (Miltenyi Biotech, Bergisch Gladbach, Germany). CD4⁺ CD25⁺ T cells were then positively selected from CD4⁺ CD25⁻ cells on MiniMACS magnetic columns (Miltenyi Biotech). The purity of retrieved cells was determined by flow cytometry to be > 84% for CD4⁺ CD25⁻ cells and > 92% for CD4⁺ CD25⁺ cells. There was no significant difference in the purity of PrP^+/+ cells compared with PrP^0/0 cells. Purified CD4⁺ CD25⁺ T cells (1 × 10⁵ cells/well) were cultured, in 96-well round-bottomed plates, either alone or with CD25⁻ T cells at indicated ratios (between 1 : 16 and 1 : 1 CD25⁺ to CD25⁻ cells) in the presence of Epoxy DynaBeads (1 bead/5 cells; Invitrogen) coated with anti-CD3 and anti-CD28 mAb. Plates were incubated at 37° with 5% CO₂. After 3 days, proliferation was measured by tritiated thymidine incorporation during the final 16 hr of culture.

Statistical analysis

Statistical analysis was performed using GraphPad InStat (GraphPad Software, San Diego, CA). Data sets were analysed for statistical significance using the t-test. Error bars represent standard deviations from the mean.

Results

CD4⁺ CD25⁺ Foxp3⁺ Tregs express higher levels of surface PrP^C than conventional T cells

Flow cytometry was used to identify Tregs from wild-type FVB/N mice. After gating on CD4⁺ cells, we observed higher surface PrP^C expression on Foxp3⁺ compared with Foxp3⁻ T cells (Fig. 1a and b). Approximately 19% of the Foxp3⁺ T cells were considered PrP^C positive, compared with < 1% of Foxp3⁻ T cells. Successive gating on CD25 and Foxp3 negative and positive populations demonstrated that approximately 7% of CD4⁺ splenocytes expressed both surface CD25 and intracellular Foxp3 (data not shown). Histograms illustrating PrP^C expression confirmed that non-regulatory (conventional) FVB/N CD4⁺ splenocytes were essentially PrP^C negative (Fig. 1d), while Tregs represented a distinct population with intermediate PrP^C expression (Fig. 1c). Calculating the delta geometric mean for PrP^C expression, we observed approximately 10-fold higher intensity of surface PrP^C expression by CD4⁺ CD25⁺ Foxp3⁺ Tregs than conventional CD4⁺ T cells (Fig. 1e). Similarly, in C57BL/6 mice we observed that approximately 18% of CD4⁺ CD25⁺ splenocytes were PrP^C positive compared with 6% of CD4⁺ CD25⁻ cells (data not shown). Delta geometric mean PrP^C expression was approximately 3·5 times higher in CD4⁺ CD25⁺ T cells compared with CD4⁺ CD25⁻ T cells (data not shown).

Figure 1.

Correlation of cellular prion protein (PrP^C) with CD25 and forkhead box P3 (Foxp3) in murine CD4⁺ T cells. PrP^C expression, measured by flow cytometry, was compared between FVB/N CD4⁺ CD25⁻ Foxp3⁻ and CD4⁺ CD25⁺ Foxp3⁺ T cells. (a) A representative dot plot is shown, acquired by gating on the CD4⁺ population, demonstrating higher surface PrP^C expression by Foxp3⁺ compared with Foxp3⁻ cells. (b) Control plot in which anti-PrP monoclonal antibody (mAb) was replaced by immunoglobulin G1 (IgG1) mAb without surface antigen specificity. (c) Histogram demonstrating surface PrP^C expression by CD4⁺ CD25⁺ Foxp3⁺ T cells; light grey curve, PrP^C; filled black curve, isotype control. (d) Histogram demonstrating surface PrP^C expression by CD4⁺ CD25⁻ Foxp3⁻ T cells; grey curve, PrP^C; filled black curve, isotype control. (e) Cells with a regulatory T cell (Treg) phenotype demonstrated significantly higher surface PrP^C expression (t-test).

PrP is transcriptionally up-regulated in Tregs

To confirm these findings at the transcriptional level, we used flow cytometry to separate C57BL/6 CD4⁺ T cells into CD25⁺ and CD25⁻ fractions. We then determined Foxp3 and Prnp mRNA expression in these two populations by real-time PCR. The effective separation of Tregs from other T cells using this method was validated by the finding that the CD4⁺ CD25⁺ fraction contained ∼100 times more Foxp3 mRNA than the CD4⁺ CD25⁻ population, suggesting that we had discriminated CD4⁺ Tregs from T effector cells (Fig. 2a). Further, Prnp expression was ∼4·5-fold higher in CD4⁺ CD25⁺ cells than in CD4⁺ CD25⁻ cells (Fig. 2b). We therefore conclude that Prnp is preferentially expressed by Tregs.

Figure 2.

Expression of prion protein (Prnp) and forkhead box P3 (Foxp3) by murine CD4⁺ splenocytes. Foxp3 (a) and Prnp (b) transcription was analysed by real-time reverse transcriptase–polymerase chain reaction (RT-PCR) in CD25⁻ and CD25⁺ CD4⁺ splenocytes from C57BL/6 mice. The CD4⁺ CD25⁺ fraction contained significantly more Foxp3 and Prnp mRNA than the CD4⁺ CD25⁻ population (t-test after logarithmic transformation).

PrP^0/0 mice have normal numbers of Tregs and Foxp3 expression levels

These data raised the question of Treg ontogeny in the absence of PrP. Using flow cytometry, we determined the percentage of CD4⁺ CD25⁺ Foxp3⁺ cells in thymus, spleen, and mesenteric and peripheral (non-mesenteric) lymph nodes from PrP^+/+ and PrP^0/0 mice. We found no deficit in Treg number in PrP^−/− mice, indicating that these cells do not require PrP for their thymic development or maintenance in the periphery (Fig. 3a). The level of Foxp3 expression was also no different between PrP^+/+ and PrP^0/0 mice (Fig. 3b).

Figure 3.

Regulatory T cell numbers and forkhead box P3 (Foxp3) expression in prion protein (PrP)^+/+ and PrP^0/0 mice. (a) The percentage of CD4⁺ cells co-expressing CD25 and Foxp3 in the thymus and peripheral lymphoid tissues does not differ between PrP^+/+ and PrP^0/0 FVB/N mice. (b) Foxp3 expression levels determined by flow cytometry in CD4⁺ CD25⁺ Foxp3⁺ regulatory T cells (Tregs) in the thymus and peripheral lymphoid tissues do not differ between PrP^+/+ and PrP^0/0 FVB/N mice.

PrP^0/0 regulatory T cells have intact suppressor function

Finally, we determined whether Treg function would be affected by embryonic deletion of PrP. CD4⁺ lymphocytes were purified from spleens of PrP^+/+ and PrP^0/0 mice and further split into CD25⁺ and CD25⁻ fractions. Wild-type CD4⁺ CD25⁻ cells were stimulated with anti-CD3- and anti-CD28-coated beads in the presence of an increasing number of syngeneic or congeneic CD4⁺ CD25⁺ cells. As a control, CD4⁺ CD25⁻ cells were cultured alone to assess the maximum possible proliferative response; mean counts were 97 004 in the lanes to which PrP^+/+ Tregs were subsequently added and 106 506 in the lanes where PrP^0/0 Tregs were added. Addition of Tregs resulted in almost complete suppression of proliferation at high ratios of Tregs:effectors (Fig. 4). We found that the ability of PrP^0/0 Tregs to suppress proliferation of wild-type non-regulatory cells was intact (Fig. 4). Cultures of PrP^0/0 CD4⁺ CD25⁻ T cells with wild-type and PrP^0/0 Tregs gave similar results (data not shown).

Figure 4.

Functional assay of prion protein (PrP)^+/+ and PrP^0/0 regulatory T cell (Treg) suppression. Wild-type CD4⁺ CD25⁻ T cells were stimulated with anti-CD3 and anti-CD28 in the presence of either PrP^+/+ or PrP^0/0 Tregs at indicated ratios of CD25⁺ to CD25⁻ cells. The response is expressed as a percentage of the proliferation of CD4⁺ CD25⁻ cells stimulated in the absence of Tregs. PrP^0/0 Tregs have intact capacity to suppress proliferation of wild-type T cells.

Discussion

Gene expression microarray analysis has suggested that PrP is transcriptionally up-regulated in certain classes of Treg.^24,25 We therefore sought to characterize PrP expression in Tregs directly. Using CD25 and Foxp3 staining to identify naturally occurring Tregs, we demonstrated a 10-fold increase in surface PrP expression compared with non-regulatory (conventional) CD4⁺ T cells. Further, using real-time RT-PCR in cells split into CD4⁺ CD25⁺ and CD4⁺ CD25⁻ fractions, we demonstrated that cells bearing a regulatory phenotype have increased Prnp mRNA.

Although PrP^C is preferentially expressed by regulatory T cells, unlike Foxp3 it is not exclusively expressed by Tregs. PrP^C therefore has more in common with CD25, glucocorticoid-induced tumour necrosis factor receptor (GITR), CD40 and cytotoxic T-lymphocyte antigen 4 (CTLA-4), representing an inducible marker in activated conventional T cells but constitutively up-regulated in Tregs. Indeed, the findings we describe here are highly reminiscent of those with GITR, one of the defining markers of Tregs.²⁸ GITR^−/− Treg cells suppress responder T cells similarly to wild-type Tregs and the number of Treg cells is not significantly different between the strains.²⁹ However, exacerbation of inflammatory and autoimmune phenotypes has been reported as a consequence of administering anti-GITR antibodies in vivo. It will be important to investigate these effects in relation to PrP^C.

The effects of embryonic deletion of PrP on Treg development and number have not previously been explored. We found that PrP^0/0 mice have normal numbers of regulatory T cells in the thymus and secondary lymphoid organs. Further, Foxp3 expression level did not differ between PrP^+/+ and PrP^0/0 Tregs. Finally, we demonstrated that PrP^0/0 Tregs have intact suppressor function. In addition, we did not identify any features of spontaneous autoimmunity in knockout mice (data not shown).

To date, experiments on immune function in PrP null mice have produced conflicting results. Some groups have not been able to demonstrate any impact of PrP deletion on functional T-cell responses,^16,20 while others have reported reduced proliferative responses of PrP^0/0 T cells to some mitogens.^15,18,19 Aguzzi and co-workers reported normal CD8⁺ expansion and antibody production after vesicular stomatitis virus (VSV) and lymphocytic choriomeningitis virus (LCMV) infection in mice lacking PrP alone or PrP and its partial homologue Doppel,³⁰ while another group has reported modest changes in leucocyte infiltrate during zymosan-induced peritonitis.³¹ However, more robust infectious and immunological challenges to PrP^0/0 mice may expose more striking phenotypic alterations that might reflect altered T-cell regulatory or effector functions.

There is increasing evidence that Treg function is finely controlled and subject to modifying signals, such as toll-like receptor (TLR)-8 signalling.³² We have not demonstrated functional differences resulting from embryonic PrP deletion. However, PrP^C ligation may alter T-cell function through mechanisms not modelled by gene knockout techniques.²¹ Further studies will address whether this also applies to Treg function. The existence of functionally specialized T-cell populations that show enhanced PrP^C expression has implications for the systemic use of anti-PrP therapies; the effects of such agents on immune function will need to be carefully modelled prior to clinical application.

Conflict of interest statement

JC is a Director and JC and GSJ are shareholders and consultants of D-Gen Limited, an academic spin-out company working in the field of prion disease diagnosis, decontamination and therapeutics. D-Gen markets one of the routine antibodies (ICSM 18) used in this study.