Cell mediated immune responses against human prion protein

Abstract

Vaccination approaches that may provide protection against the abnormal form of prion protein (PrP^Sc) have recently focused on the ability of antibodies to prevent PrP^Sc propagation. Progress has been hampered due to the difficulty in generating antibody responses in wild type mice, which is believed to be a consequence of T cell tolerance to the normal form of prion protein (PrP^C). The problem of tolerance can be avoided using transgenic mice unable to express PrP^C. This study examines active PrP specific T cell responses that can be produced in PrP null (PrP 0/0) mice using simple peptide vaccination procedures. Spleenocytes recovered from vaccinated PrP 0/0 mice were tested in vitro for their specificity with T cell recognition demonstrated through a proliferative response to the peptide. Analysis of mRNA also indicates the stimulation of a heterogenous population of T cells with an increase in cytokines and cytotoxicity associated mRNA. Responsive T cells were expanded using a T cell cloning procedure and demonstrated an ability to recognize the mature human prion protein. These clones may potentially be used to negate the problem of T cell tolerance in wild type mice.

INTRODUCTION

Prion diseases are a group of neurodegenerative diseases including Bovine Spongiform Encephalopathy (BSE), scrapie, Creutzfeldt-Jakob disease, Kuru and fatal familial insomnia. Particularly alarming is the manifestation of variant CJD (vCJD) believed to be a consequence of exposure to BSE associated prions [1,2]. The long-term impact of this disease is difficult to assess due to its long incubation period, problematic diagnosis, possible transfusion transmission of PrP^Sc and the unknown quantities of BSE infected meat that have entered the human food chain [3–7]. PrP^C and PrP^Sc are poor immunogens, limiting the success of vaccination strategies that may provide some protective immunity against PrP^Sc. Focus has shifted to the generation of PrP 0/0 mice where the ablation of PrP^C prevents the development of T cell tolerance. Once vaccinated with PrP protein or DNA, PrP 0/0 mice are able to mount a humoral response, producing antibodies specific for PrP [8,9]. Anti-PrP antibodies have been developed for applications in a variety of fields including immunohistochemistry, Western blotting, flow cytometry, ELISA and PrP purification [10–13].

There is growing interest in potential vaccine approaches to provide protection against PrP^Sc, as recent studies have shown that anti-PrP antibodies may reduce prion propagation [14,15]. Although PrP^C or PrP^Sc themselves do not initiate significant immune responses in wild type mice, vaccination of peptides in conjunction with strong adjuvants can induce a significant anti-PrP response capable of inhibiting PrP^Sc propagation [16]. Unfortunately, such extreme vaccination strategies are likely to provide the required humoral response at the cost of generating problematic autoimmune reactions [17]. Self-tolerance has also been successfully bypassed through the generation of mice that transgenically express variable regions from the monoclonal antibody 6H4. These mice appear capable of inhibiting PrP^Sc pathogenesis, demonstrating that immunological inhibition of prion disease is feasible [15].

Another potential method of overcoming self-tolerance is to use PrP specific T cells derived from PrP 0/0 mice and vaccinate them into wild type (parental strain) mice. This study examines methods of generating PrP specific T cells in PrP 0/0 mice in addition to analysing their phenotype and specificity. Examination of their phenotype provides information as to their possible therapeutic use once transferred into wild type mice and their specificity will identify any possible risks of autoimmune complications.

To generate PrP specific T cells, we have used three basic approaches. Firstly we vaccinated PrP 0/0 mice with a synthetic peptide conjugated to keyhole limpet haemocyanin (KLH). The choice of peptide sequence was based on a predicted ability to bind to MHC molecules [18,19]. Analysis highlighted a good candidate sequence designated PrP^159−166 (see Fig. 1) which was predicted to bind well to MHC molecules for the PrP 0/0 mice H2 haplotype. The second approach involved using a DNA construct designed to express human PrP and the third approach used a combination of DNA and peptide vaccinations.

Fig. 1.

The location of the sequence chosen (NQVYYRPM) for this vaccination study in both (a) the normal cellular isoform of human prion protein (PrP^C) and (b) the disease associated form (PrP^Sc)

MATERIALS AND METHODS

Reagents

Dulbecco's modified eagle's medium (DMEM), fetal bovine serum (FCS), phorbol 12-myristate 13-acetate (PMA), psoralen, ionomycin, concanavalin A (ConA), 4′,6-Diamidino-2-phenyindole (DAPI) and Trizol reagent were purchased from Sigma Aldrich Company Ltd (Dorset UK). The synthetic peptide was generated and conjugated to KLH by Sigma Genosys (Cambridge, UK). Con A stimulated rat spleenocytes supernatant (CRSS), transfected A1A cell lines, pCIhPrP (plasmid expressing human PrP) and IL-2 were obtained from NIBSC stocks. The P9A8 mouse anti-PrP antibody was also generated in house and derived from PrP 0/0 mice vaccinated with DNA expressing PrP. Taq DNA polymerase was purchased from Promega (Southampton UK). 5× first strand buffer and M-MLV reverse transcriptase was purchased from Gibco-Invitrogen Paisley UK. Primers for RT-PCR were obtained from MWG-Biotech Ltd (Milton Keynes, UK). CD28 antimouse antibody was purchased from Pharmingen, Oxford UK. Triton X was purchased from BDH (Dorset UK). Vectorshield, biotinylated horse antimouse IgG and streptavidin R-phycoerythrin purchased from Vector Laboratories, Peterborough UK. Rat antimouse α-tubulin and goat antirat IgG FITC was purchased from Serotec (Oxford UK).

Vaccinations

The PrP 0/0 mice used in this study carry a null mutation and are derived from the 129/Ola strain [20]. PrP 0/0 mice were divided into groups of five. The first group was given four monthly intramuscular vaccinations with 100 µg pCIhPrP plasmid expressing the human prion gene. The second group was vaccinated twice subcutaneously every two weeks with 100 µg of PrP^159−166-KLH in PBS. The final group was given four 100 µg intramuscular vaccinations of pCIhPrP plasmid every month followed by two subcutaneous boosts with 100 µg of PrP^159−166. Control studies included five PrP 0/0 mice vaccinated subcutaneously twice with 100 µg KLH every two weeks, and five PrP 0/0 mice given four monthly 100 µg intramuscular vaccinations of the empty plasmid pCI.

DNA synthesis assay

Two weeks after the final vaccination, spleens were harvested and cell suspensions prepared by passage through a 70-µm nylon mesh. Cells were seeded at 2 × 10⁶ cells per ml in 96 well plates. Wells were treated with PrP^159−166 at a final concentration of 100 µg per ml. Controls included wells treated with ConA at 1 µg per ml, or ovalbumin at 100 µg per ml or left untreated. Ovalbumin was also used at 100 µg per ml to examine nonspecific proliferation. All treatments were carried out in triplicate. Proliferation of spleenocytes was analysed at day 3 by pulsing cells with 0·5 µ Ci per well of ³H-thymidine for 12 h prior to harvesting for cell associated ³H-thymidine incorporation using liquid scintillation counting. The stimulation index was calculated as mean counts per minute of treated wells/mean counts per minute of unstimulated wells.

Rt pcr

Functional analysis was carried out on spleenocytes from PrP^159−166-KLH vaccinated mice as these mice were to be used for subsequent generation of T cell lines and clones. Spleens from five naïve mice were also harvested to examine RNA expression prior to vaccination. Spleenocytes were seeded in 24 well plates, at a concentration of 2 × 10⁶ cells per ml and used at 1 ml per well. Wells were treated with PrP^159−166 at 100 µg per ml, ovalbumin at 100 µg per ml, or Con A at 1 µg per ml or left untreated. On day 3 total cellular RNA was extracted from cells using Trizol reagent. Total cellular RNA was resuspended in diethyl-pyrocarbonate (DEPC) treated H₂0. 1–3 µg of RNA was primed with oligo DT at 65°C for 10 min before being reverse transcription using an RT mix of 5 × first strand buffer, 1 µg BSA, 24 units of RNAGuard (RNase inhibitor), 5 mm of each dNTP and 2 units of reverse transcriptase at 37°C for 1 h. The cDNA products were then subjected to PCR using a range of primers (see Table 1). Reaction conditions were 30 cycles containing a denaturing step at 94°C for 30 s, an annealing step at 50°C for 30 s and an elongation step at 72°C for 90 s. PCR products were run on a 2% agarose gel, visualized with ethidium bromide and UV transillumination. The intensity of each band was determined by digital image analysis using Kodak 1D analysis software and indexed relative to β-actin.

Table 1.

A list of primers used for PCR of treated spleenocytes and clones.

Sense	Primer name	Sequence
Sense	Murine PrP*	5′ CAT TTT GGC AAC GAC TGG GAG GAC  3′
Antisense	Murine PrP*	3′ GAC TCC ATC AAA GGG ACC TGA AGC  5′
Sense	Human PrP*	5′ ATG GCG AAC CTT GGC TGC TGG AT 3′
Antisense	Human PrP*	3′ GAC CGT GCG CTG CTT GAT TGT GAT  5′
Sense	Actin	5′ GTG GGC CGC TCT AGG CAC CAA  3′
Antisense	Actin	3′ CTC TTT GAT GCT ACG CAC GAT TTC 5′
Sense	IL-4	5′ ATG GGT CTC AAC CCC CAG CTA GT 3′
Antisense	IL-4	3′ GCT CTT TAG GCT TTC CAG GAA GTG  5′
Sense	IFN-γ	5′ TGA ACG CTA CAC ACT GCA TCT TGG  3′
Antisense	IFN-γ	3′ CGA CTC CTT TTC CGC TTC CTG AG 5′
Sense	GRANZYME A	5′ CTC TGG TCC CCG GGG CCA TC 3′
Antisense	GRANZYME A	3′ TAT GTA GTG AGC CCC AAG AA 5′
Sense	PERFORIN	5′ TGA GGT AGG AGA CTG CCT GAA  3′
Antisense	PERFORIN	3′ ATA GCC TGT CTC AGA GCC TCC  5′
Sense	FAS LIGAND	5′ GCT GAG GAG GCG GGT T  3′
Antisense	FAS LIGAND	3′ ACT TGG TAT TCT GGG TCA GGG  5′

^*denotes PrP primers used for PCR of A1A transfected and untransfected cells

Expansion of responsive cells

Antigen reactive T cells were obtained from draining lymph nodes of mice vaccinated with PrP^159−166-KLH. Cells were seeded in 96 well plates at 5 × 10³ cells per well in 100 µl of DMEM. They were fed IL-2 at 10 IU per well with irradiated spleenocytes at 6 × 10⁴ cells per well and 10 µg of PrP^159−166 per well. After one week, wells were fed 100 µl of fresh medium containing 2% CRSS. At day 14, clones were derived using limiting dilution and treated with 10⁵ irradiated spleenocytes and 100 µg PrP^159−166 per ml. Clones were identified in plates with less than 10 positive wells and expanded using a feeder cocktail of IL-2 at 100 IU per ml, PMA at 0·2 µg per ml, ionomycin at 7·3 µg per ml and antimurine CD28 antibody at 0·2 µg per ml. Following three days of stimulation, IL-2 was added at 100 IU per ml. At day seven, cells were rested with 100 µl of fresh medium plus 2% CRSS per well. Cells were re-stimulated again at day 14 and continued through this cycle of stimulation and rest to provide sufficient cells for further analysis. Following every cycle of expansion, clones were tested for retention of specificity by incubating 5 × 10³ cells in the presence of 6 × 10⁴ feeder cells and 10 µg of PrP^159−166. ³H-Thymidine was measured at day 3 using the protocol previously described.

Characterization of A1A cells

Identified and expanded clones were tested for their response to mature human prion protein expressed in A1A cells. This cell line was derived from PrP 0/0 mice lung tissue and transfected in vitro with a pCImPrPEH plasmid or pCIhPrPEH plasmid constructed to express mouse and human PrP, respectively. Stably transfected cells were selected via plasmid expressed hygromyocin resistance using hygromyocin at 100 µg per ml. Transfection and expression was confirmed using RT-PCR as described previously with primers specific for mouse and human PrP. To examine the cellular expression of PrP by A1A cells, immunofluorescence microscopy was carried out on transfected and nontransfected A1A cells grown for 24 h in 6 well plates on poly l-lysine coated glass coverslips. Cells were washed three times in PBS followed by fixation in 4% formaldehyde for 10 min at room temperature and a further three washes in PBS. Cells were permeabilized for 5 min at 4°C using 0·5% Triton X to allow intracellular penetration of antibodies. Traces of detergent were removed with a further three washes of PBS and nonspecific binding sites blocked using 0·1% BSA in PBS for 10 min at room temperature. Cells were washed three times with PBS before being incubated with rat anti α-tubulin (1/100 dilution) and P9A8 (1/50 dilution) for 1 h at room temperature. Cells were then washed three times with PBS and incubated with biotinylated horse antimouse IgG and goat antirat IgG FITC for 30 min at room temperature. Following three washes in PBS the biotinylated antibody was conjugated to a fluorochrome using strepavidin R-phycoerythrin (1/50 dilution) for 20 min. Cell nuclei were labelled with DAPI and incubated at room temperature for 5 min Excess nuclear stain was removed using PBS and coverslips were mounted on glass slides with Vectorshield prior to viewing under a Nikon E400 elipse fluorescent microscope.

Clone specificity and phenotype

To determine the specificity of T cell clones, cells were seeded at a concentration of 1 × 10⁵ cells in 96 round bottom well plates at a ratio of 1 : 1 with irradiated spleenocytes. A1A and transfected A1A cells were incubated with psoralen at 10 µg per ml and u.v. irradiated for 30 min. Cells were then washed three times with PBS. Clones were treated with A1A, A1ApCIhPrPEH or A1ApCImPrPEH cells at 1 × 10⁵ cells per ml. Controls again included leaving wells unstimulated or ConA treatment at 1 µg per ml. All treatments were carried out in triplicate and DNA synthesis assayed as previously described.

RT-PCR was carried out to analyse clone phenotype. Clones which demonstrated proliferation to A1A cells expressing PrP were seeded in 24 well plates at a concentration of 1 × 10⁶ cells per well and treated with ConA at 1 µg per ml. Following 3 days incubation, RT-PCR was carried out and DNA primed for IL-4, IFN-γ, perforin, granzyme A and Fas-L using the protocol described earlier.

RESULTS

Proliferation of spleenocytes

Spleenocyte response was assessed two weeks after the last vaccination. Spleens from individual mice were treated separately to determine the efficacy of vaccination approaches. Spleenocytes from both PrP^159−166-KLH and pCIhPrP/PrP^159−166-KLH vaccinated mice demonstrated in vitro proliferation to PrP^159−166 at day 3 (Fig. 2). The extent of proliferation varied between individuals and not all mice responded to in vitro exposure to the peptide. Spleenocytes from pCIhPrP (DNA) only vaccinated mice demonstrated little or no proliferation when treated with PrP^159−166. No spleenocytes demonstrated a significant response to ovalbumin in any of the vaccination groups. Reponses to ConA varied widely, although generally ConA responses were substantially greater than those to the peptide. In control vaccinated and naïve mice no response to PrP^159−166 or ovalbumin was seen. Mice vaccinated with KLH demonstrated proliferation to KLH and ConA only.

Fig. 2.

(a) ³H-thymidine incorporation in spleenocytes from pCIhPrP, pCIhPrP/PrP^159−166-KLH and PrP^159−166-KLH vaccinated groups treated with peptide, ovalbumin, ConA or left untreated. Proliferation to the peptide was not evident in mice vaccinated with pCIhPrP only. Most individuals in groups vaccinated with pCIhPrP/PrP^159−166–KLH and PrP^159−166–KLH show proliferative response to the peptide. No individuals demonstrated proliferated in response to treatment with ovalbumin. Individuals in all vaccination groups show a variable response to ConA. (b) ³H-thymidine incorporation in spleenocytes from control groups including naïve mice, KLH vaccinated and pCI vaccinated mice. Spleenocytes in KLH vaccinated mice show proliferation to KLH and ConA only. pCI vaccinated mice and naïve mice show no response to KLH, PrP^159−166 or ovalbumin. ConA responses appear to be variable.

Profile of responsive spleenocytes

The phenotypic profile of naïve mice and mice vaccinated with PrP^159−166-KLH was analysed using RT-PCR. Spleens were kept in the order corresponding to those in the proliferation studies. RT-PCR was carried out on mRNA isolated from spleenocytes treated with PrP^159−166, ovalbumin, ConA or left blank. β-actin was used as a positive control. The relative levels of murine IL-4, IFN-γ, perforin, granzyme A and Fas-ligand were indexed to that of β-actin. The naïve mice demonstrated low background levels of IL-4, IFN-γ, granzyme A and Fas-L in wells treated with PrP^159−166, ovalbumin or left blank, although bands corresponding to perforin expression was evident in some of the treated groups. When naïve mice where treated with ConA they demonstrated an increase in the mRNA expression of IL-4, IFN-γ and granzyme A. The levels of Fas-L remained relatively low.

In KLH-PrP^159−166 vaccinated mice, four out the five mice showed a relative increase in IFN-γ mRNA in response to the peptide compared to that of the negative control (Fig. 3). However a similar but reduced response was observed in naïve mice. Levels of IL-4 mRNA appeared to be elevated in three of the five vaccinated mice exposed to the peptide. To examine potential cytotoxic T cell and natural killer cell activity, levels of perforin and granzyme A mRNA were assessed. Three of the five vaccinated mice demonstrated a relative increase in granzyme A mRNA expression in response to the peptide. Ovalbumin also appeared to generate an increase in granzyme A mRNA in three out of the five compared to the negative control. Expression of Fas-Ligand appeared relatively unchanged in PrP^159−166, untreated and ConA treated cells.

Fig. 3.

(a) RT-PCR results of spleenocytes from naïve mice treated with PrP^159−166, ovalbumin, ConA or left untreated. The order of individuals corresponds to those tested for proliferation in Fig. 2b. Graphs under images demonstrate the relative levels of IL-4, IFN-γ, Perforin, Granzyme A and Fas Ligand indexed to β-actin expression. Low levels of IL-4, IFN-γ, granzyme A and Fas-Ligand mRNA in all treatments except those treated with ConA. Some expression of Perforin was seen in treatments other than ConA. (b) RT-PCR results of spleenocytes from PrP^159−166-KLH vaccinated mice again treated with PrP^159−166, ovalbumin, ConA or left untreated. The order of individuals corresponds to those tested for proliferation in Fig. 2a. In some individuals, expression of IL-4, perforin and granzyme A appears higher in cells pulsed with PrP^159−166 compared to the negative control.

Characterization of transfected A1A cells

To confirm successful transfection, RT-PCR was carried out on mRNA isolated from both transfected and nontransfected A1A cells. As expected, results showed an absence of PrP expression in untransfected cells (Fig. 4a). A band corresponding to murine PrP was seen in pCImPrPEH transfected A1A cells and bands for human PrP in pCIhPrPEH transfected A1A cells confirming successful transfection and gene expression (Fig. 4a).

Fig. 4.

(a) PCR products from A1A cells derived from PrP 0/0 lung tissue. Lanes 1–3 are the untransfected cell lines. Lanes 4–6 are A1A cells transfected with the pCImPrP plasmid. Lanes 7–9 are A1A cells transfected with the pCIhPrP plasmid. Lanes 1,4 and 7 are β-actin cDNAs. Lanes 2,5 and 8 are bands corresponding to murine PrP primed RT-PCR and lanes 3,6 and 9 are bands corresponding to human primed PrP. (b) Image shows immunoflourescence of A1A cells (i) untransfected (ii) transfected with pCImPrP (iii) transfected with the pCIhPrP. Cells are adherent and spread on substratum. Blue is DAPI nuclear staining, green is α-tubulin staining red corresponds to PrP. Yellow is superimposition of red and green emission spectra. Positive staining for PrP is seen predominantly within the cell (arrow A). Some positive staining appears to be colocalized with α-tubulin (arrow B). Magnification × 1000

In culture, A1A cells are adherent cells with a spindle like morphology similar to that of fibroblasts and remain unchanged following transfection. Immunoflourescence demonstrated similar cellular distributions of PrP in both pCImPrPEH and pCIhPrPEH transfected cell lines. Staining for PrP appeared to be predominantly within intracellular compartments with some PrP colocalizing with α-tubulin (Fig. 4b). No PrP expression was seen in nontransfected A1A cells.

Specificity of cloned cells and phenotype

From T cell cloning procedures, five potential T cell clones were identified and further examined for proliferation with the transfected A1A cells. Following three days of stimulation, three out of the five clones demonstrated a high stimulation index in response to A1A cells expressing pCIhPrPEH (Fig. 5). No proliferation was seen to the A1A transfected with pCImPrPEH. All clones responded well to treatment with ConA. The three clones that proliferated when treated with pCIhPrP were phenotyped using RT-PCR. Two of the clones demonstrated expression of IFN-γ, perforin and granzyme A mRNA. The final responsive clone demonstrated expression of IL-4 mRNA.

Fig. 5.

(a) Stimulation Index of cloned cells derived from PrP^159–166-KLH vaccinated mice when coincubated with A1A, A1ApCIhPrP, A1ApCImPrP cells or ConA. Clones 2, 3 and 5 proliferate in the presence of A1A pCIhPrPEH transfected cells. (b) PCR products from clones 2, 3 and 5 which demonstrated proliferation to AIApCIhPrPEH cells. The image demonstrates bands corresponding for IFN-γ, granzyme A and perforin in clones 2 and 3 whilst clone 5 demonstrates bands corresponding to only IL-4.

DISCUSSION

Spleenocyte proliferation in response to PrP^159−166 in both PrP^159−166-KLH and pCIhPrP, PrP^159−166-KLH vaccinated mice clearly demonstrates that PrP 0/0 mice are capable of generating cell mediated immune response to PrP derived sequences. Although plasmids expressing the PrP are capable of inducing humoral responses, the lack of T cell proliferation in DNA only vaccinated mice indicates that the plasmid vaccination approach may be inappropriate for the generation of T cell responses against PrP^159−166. In KLH only vaccinated mice, no proliferation was shown to PrP^159−166 indicating that there are no cross-reacting epitopes between PrP^159−166 and KLH. Naïve mice demonstrated no response to PrP^159−166 showing that vaccination is essential to generate T cells specific for this eptiope. Some individual mice in this study showed abnormal mitogenic responses to ConA that may be a consequence of the absence of PrP^C in these mice and suggests a potential functional role of prion protein in T cell activation and proliferation [21]. An additional possibility for this variation may be due to the proliferation assay being carried out at day 3 which may be suboptimal for ConA treated spleenocytes.

PrP^159−166-KLH vaccinated mice were chosen for cloning and expansion since the lack of a proliferative response in pCIhPrP only vaccinated mice indicated it was not effective in priming T cell immunity to PrP^159−166. Analysis of mRNA expression patterns in spleenocytes from peptide vaccinated mice using RT-PCR were difficult to conclusively interpret. Although expression of some cytokine and cytotoxic mRNA appear elevated in response to the peptide, responses in naïve mice and in ovalbumin treated mice make it difficult to conclusively determine the phenotype of the responsive T cells. The ambiguity of these results demonstrate the potential benefit of cloning responsive T cells to more closely ascertain their phenotype.

Transfected A1A cells were used as a source of PrP as they are easy to maintain and quick to expand. Successfully transfected cells were selected through their resistance to hygromyocin. RT-PCR helped confirmed expression of human and mouse PrP in A1A transfected cells. Immunolocalization of PrP on the cells expressing both human and mouse PrP^C demonstrated similar distribution of PrP with the majority of expression within intracellular compartments. Some PrP also appeared to be colocalized with α-tubulin which may reflect transport of PrP within the cell.

Proliferation of T cell clones to the pCIhPrP transfected A1A cell lines demonstrated recognition of the T cells to mature human prion protein. Only three of the five clones evidenced proliferation which may be due to a loss of specificity during expansion. The timing of stimulation following rest may be a critical factor for these cells. The lack proliferation to A1A cells transfected with pCImPrP could be due to a one amino acid difference in the octamer peptide sequence between mouse and human, suggesting that this amino acid is a vital for TCR binding. Additionally, the lack of response to murine PrP in vitro also indicates that these T cells are unlikely to recognize murine PrP^C expression in wild type mice. RT-PCR on the three clones demonstrated two different phenotypes. Clone 1 and 2 expressing mRNA of IFN-γ, perforin and granzyme A consistent with the expression pattern of cytotoxic T cells. Clone 3 demonstrates IL-4 consistent with T helper type 2 cells.

In this study we have demonstrated a vaccination approach to generating T cells that recognize human PrP. If vaccinated into wild type mice, IL-4 expressing clones similar to clone 3 may provide a useful tool in providing the additional T cell help to raise antibodies against either murine or human PrP once these mice are subsequently vaccinated with human PrP. As some antibodies recognizing PrP can inhibit PrP^Sc propagation, this approach may have some therapeutic potential.